Convergence of Integrated Sensing and Communication (ISAC) and Digital-Twin Technologies in Healthcare Systems: A Comprehensive Review

Abstract

Modern healthcare systems are under growing strain from aging populations, urbanization, and rising chronic disease burdens, creating an urgent need for real-time monitoring and informed decision-making. This survey examines how the convergence of Integrated Sensing and Communication (ISAC) and digital-twin technologies can meet that need by analyzing how ISAC unifies sensing and communication to gather and transmit data with high timeliness and reliability and how digital-twin platforms use these streams to maintain continuously updated virtual replicas of patients, devices, and care environments. Our synthesis compares ISAC frequency options across sub-6 GHz, millimeter-wave, and terahertz bandswith respect to resolution, penetration depth, exposure compliance, maturity, and cost, and it discusses joint waveform design and emerging 6G architectures. It also presents reference architecture patterns that connect heterogeneous clinical sensors to ISAC links, data ingestion, semantic interoperability pipelines using Fast Healthcare Interoperability Resources (FHIR) and IEEE 11073, and digital-twin synchronization, and it catalogs clinical and operational applications, together with validation and integration requirements. We conduct a targeted scoping review of peer-reviewed literature indexed in major scholarly databases between January 2015 and July 2025, with inclusion restricted to English-language, peer-reviewed studies already cited by this survey, and we apply a transparent screening and data extraction procedure to support reproducibility. The survey further reviews clinical opportunities enabled by data-synchronized twins, including personalized therapy planning, proactive early-warning systems, and virtual intervention testing, while outlining the technical, clinical, and organizational hurdles that must be addressed. Finally, we examine workflow adaptation; governance and ethics; provider training; and outcome measurement frameworks such as length of stay, complication rates, and patient satisfaction, and we conclude that by highlighting both the integration challenges and the operational upside, this survey offers a foundation for the development of safe, ethical, and scalable data-driven healthcare models.

1. Introduction

Health systems face concrete, measurable pressures after COVID-19: a sustained workforce gap, rising multi-morbidity in aging populations, and persistent backlogs in chronic–acute care. These pressures expose a specific technical bottleneck: current remote and in-facility monitoring pipelines still deliver fragmented, low-context, and not-always-timely data to decision makers, limiting prevention and early intervention [1,2,3].

Fragmentation typically arises from the siloing of electronic health records (EHRs) and device-specific applications that hinder interoperability, whereas timeliness is hindered by batch uploads and delayed integration of bedside/IoT streams.

These gaps are consistently documented in multi-institution audits and interoperability initiatives, such as those based on Fast Healthcare Interoperability Resources (FHIR) standards [4,5,6].

What ISAC and digital twins are—briefly. Integrated Sensing and Communication (ISAC) reuses the same radio front end and waveforms to sense and connect, enabling co-designed latency, reliability, and resolution [7]. A digital twin (DT) is a continuously updated computational replica of a patient, device, or unit that turns telemetry into forecasts and recommendations [8]. Complementary advances in mobile computing likewise demonstrate DT-driven closed-loop control in operational settings. For example, ELITE learns routing policies in a digital twin and enacts them in a software-ized vehicular network [9], a “learn-in-twin, act-in-world” pattern directly relevant to clinical operations and hospital-flow digital twins.

Why combine them? The rationale for coupling ISAC and DT rests on three mechanism-level complementarities:

(i) Data freshness: Joint sensing-and-link budgets keep twin inputs current enough for deterioration forecasting (minutes to seconds rather than hours). (ii) Context richness: ISAC adds contact-free motion, respiration, localization, and ambient dynamics to clinical streams, reducing false alarms in twins that otherwise see only vital-sign snapshots or EHR entries. (iii) Closed-loop operation: Twin outputs (risk and recommended action) can reconfigure ISAC beams, bandwidth, or sampling to focus on the most informative signals and support timely interventions.

As summarized in Figure 1, patient and environmental signals (vital signs, infusion/IoT devices, and contactless RF sensing) are carried over a unified ISAC link for timely transmission. Edge/cloud processing fuses the streams; a digital twin ingests them to infer risk and update model states. The resulting clinical insights return to the bedside and adapt ISAC parameters (e.g., beam direction, bandwidth, and sampling), reinforcing the three motivations mentioned above—freshness, context, and closed-loop operation.

At present, peer-reviewed real-world evidence that an end-to-end digital twin–ISAC stack improves patient outcomes (e.g., mortality, unplanned transfer to the intensive care unit (ICU), or 30-day readmission) or key performance indicators (KPIs) of hospital workflow (e.g., emergency department (ED) length of stay and bed turnover) remains very limited.

Nevertheless, adjacent components provide accumulating evidence in wards and emergency departments—such as contactless respiratory monitoring accuracy, real-time location system (RTLS) and indoor positioning system (IPS) process metrics, and hospital-flow digital twins demonstrated in simulation or pilots. For a consolidated summary of quantitative indicators and the separation of gray-literature cases, see Section 6.1 (Clinical Validation Status of DT–ISAC) and Section 6.2 (Recommended Evaluation Protocol for DT–ISAC in Hospitals).

Concrete payoffs: Coupled ISAC–DT stacks target well-scoped use cases: (a) early warning for COPD (Chronic Obstructive Pulmonary Disease) or heart-failure exacerbations using contact-free respiration and activity plus medication/encounter history, (b) fall risk and wandering detection with centimeter-level indoor localization feeding ward-level twins, and (c) hospital-flow twins that ingest asset and staff mobility to optimize bed turnover and escalation pathways. Each case relies on the trio mentioned above—freshness, context, and a closed loop—rather than on generic “data-driven” claims.

Scope and contributions of this survey: We focus on ISAC–DT convergence in healthcare and (i) compare ISAC frequency options (sub-6 GHz, mmWave, and THz) in terms of resolution, penetration depth, exposure compliance, and cost; (ii) present reference architectures that connect heterogeneous clinical sensors to ISAC links, data ingestion, semantic interoperability pipelines (e.g., FHIR/IEEE 11073), and twin synchronization; (iii) catalog clinical and operational applications, together with validation and integration requirements; and (iv) synthesize cross-cutting gaps in standardization, security/assurance, and clinical evidence generation.

Paper organization: Section 2 outlines the search strategy and selection criteria that underpin the review. Section 3 provides background and a literature review, including an introduction to the enabling-stack perspective. Section 4 develops the technical analysis of ISAC technology and frequency bands for healthcare. Section 5 reviews digital-twin technology in healthcare, including patient/device/facility twins and synchronization/closed-loop workflows. Section 6 analyzes ISAC–DT integration challenges and opportunities, with detailed subsections on clinical validation signals (Section 6.1) and a recommended evaluation protocol for hospital deployment (Section 6.2). Finally, Section 7 concludes this paper.

2. Review Methods: Identification, Selection, and Quantitative Synthesis

2.1. Search Strategy and Selection Criteria

This review follows a transparent, reproducible search strategy appropriate for a targeted scoping review of ISAC and digital-twin (DT) applications in healthcare.

Databases and sources: We searched IEEE Xplore, PubMed/MEDLINE, ACM Digital Library, and Scopus for scholarly literature. To cover standards and regulatory/industry guidance, we consulted 3GPP, IEEE (including standards), HL7 (FHIR), IETF, NIST, and relevant regulator (e.g., FDA/HHS) repositories and technical reports.

Time window: Publications from January 2015 through July 2025 were considered; earlier foundational or regulatory works were selectively included when essential for propagation fundamentals or compliance context.

Queries and keywords: We combined terms for joint sensing/communication and for digital twins with clinical contexts. Representative Boolean strings are presented below.

• ("integrated sensing and communication"
  OR ISAC OR "joint sensing and communication")
  AND (health* OR clinic* OR hospital)
• ("digital twin" OR "physiological twin")
  AND (patient OR device OR ward OR hospital)
• (mmWave OR "millimeter wave" OR THz OR "sub-6")
  AND (vital OR respiration OR localization OR radar)
  AND (health* OR clinic*)
• FHIR OR "HL7 FHIR" AND (interoperability OR EHR)
• (security OR safety OR regulatory OR HIPAA OR FDA)
  AND (ISAC OR "digital twin") AND health*

Inclusion criteria: (1) English-language publications; (2) peer-reviewed articles, full conference papers, or authoritative standards/technical reports from SDOs or regulators (3GPP, IEEE, HL7, NIST, or IETF/FDA/HHS); (3) direct relevance to healthcare use of ISAC or DTs or their integration; (4) reporting of methods, evidence, or design trade-offs.

Exclusion criteria: (1) non-archival abstracts/posters; (2) non-English publications; (3) duplicates.

Screening and selection: Two reviewers independently screened titles/abstracts, followed by full-text eligibility review; disagreements were resolved by discussion. We logged decisions and maintained the final bibliography in LATEX/BibTEX.

Data extraction and synthesis: For included items, we captured the sensing modality/band, DT modeling, DT synchronization (update timing/policies and temporal/spatial alignment), interoperability (FHIR/EHR), validation endpoints, and security/safety/regulatory notes.

2.2. Quantitative Synthesis

We conducted a quantitative synthesis limited to the peer-reviewed articles already cited by this survey; gray literature (e.g., vendor reports and preprints) was excluded. Each item was labeled by domain (healthcare vs. non-healthcare) and by focus (DT only, ISAC only, ISAC–DT integration, or other/background). In this synthesis, Other includes sensing-only and communication-only papers.

Table 1. Domain × focus distribution (peer-reviewed citations only; $n=218$).

Domain	DT	ISAC	ISAC–DT	Other	Total
Healthcare	24	9	4	38	75
Non-healthcare	21	57	6	59	143
Total	45	66	10	97	218

shows that roughly one-third of the peer-reviewed corpus falls under healthcare (75/218, 34.4%), while ISAC–DT integration remains rare—only 10 papers (4.6%) overall, of which 4 are in healthcare.

• ISAC only (66) is broken down into healthcare (9) vs. non-healthcare (57); DT-only (45) breaks down as healthcare (24) vs. non-healthcare (21).
• The scarcity of healthcare ISAC–DT work confirms the early stage of convergence and motivates and guides evaluation elsewhere in this paper.

Table 2. ISAC-only papers by domain and subcategory ($n=66$).

Domain	RF/Radar Sensing	Comm. Tech	General ISAC	Total
Healthcare	8	0	1	9
Non-healthcare	14	24	19	57
Total	22	24	20	66

indicates that healthcare ISAC research is overwhelmingly RF/radar-centric (8/9), with no communications-centric ISAC in healthcare, pointing to a clinical emphasis on contact-free biosignal/positioning rather than network-side ISAC design.

Table 3. Distribution by bibliographic database (selected venues).

Database	DT	ISAC	ISAC–DT	Other	Total
Elsevier	16	3	1	12	32
IEEE	4	24	3	6	37
IEEE Xplore	2	26	2	7	37
Springer	11	2	0	8	21
PubMed	9	0	0	3	12
ACM	0	6	0	7	13
MDPI	3	5	1	8	17

reveals a clear venue split. The IEEE family concentrates on ISAC-only work—75.8% (50/66 when combining IEEE journals and IEEE Xplore)—and serves as a hub for 6G/mmWave/THz standards and RF sensing fundamentals, whereas DT work is concentrated in the Elsevier/Springer venues (27/45, 60.0%) and clinical DT appears in PubMed; notably, no ISAC items are indexed in PubMed in our corpus.

Table 4. Yearly distribution (2020–2025; $n=218$).

Year	DT	ISAC	ISAC–DT	Other	Total
2020	3	1	0	2	6
2021	5	2	0	3	10
2022	7	10	1	6	24
2023	8	15	2	9	34
2024	9	20	3	12	44
2025	13	18	4	21	56
Total	45	66	10	53	218

shows a clear upward trajectory across all focus areas. DT-only publications rise steadily, reaching a new high in 2025 (13 papers). ISAC-only work also grows, peaking in 2024, with 20 papers, and maintaining a strong presence in 2025 (18 papers).

ISAC–DT integration, while still rare, exhibits gradual uptake: from 0 before 2022 to 2 papers in 2023, 3 in 2024, and 4 in 2025.

Finally, the Other category (sensing-only and communication-only papers) increases markedly in 2025 (21 papers), suggesting broader diversification of research themes alongside the maturing DT and ISAC lines.

3. Background and Literature Review

This section provides a comprehensive overview of the foundational concepts, technological developments, and current research landscape related to Integrated Sensing and Communication (ISAC) technology, digital-twin technology, and the challenges facing modern healthcare systems. By establishing this background, we aim to contextualize the subsequent discussions on the convergence of these technologies and their potential impact on healthcare delivery.

3.1. ISAC Technology Overview

3.1.1. Definition and Evolution of ISAC

Integrated Sensing and Communication (ISAC) represents a paradigm shift from traditional wireless systems where sensing and communication functions operate independently. ISAC systems unify these functions, sharing hardware, software, and spectral resources to achieve greater efficiency, reduced complexity, and enhanced performance [7]. This integration enables simultaneous sensing of the environment and transmission of information using the same signal waveforms and processing infrastructure.

The evolution of ISAC technology can be traced through several developmental stages:

1.

Early dual-function radar–communication systems (1960s–1990s): Initial efforts focused on military applications where radar systems were modified to carry communication signals, primarily through simple modulation techniques [7,10].

2.

Cognitive radio and spectrum sharing (2000s): The development of cognitive radio technologies laid the groundwork for dynamic spectrum access and sharing between sensing and communication functions [11].

3.

Joint radar–communication systems (2010s): Research began exploring truly integrated approaches where waveform design, signal processing, and resource allocation were optimized for both functions simultaneously [7].

4.

Modern ISAC systems (2020s): Current ISAC technologies leverage advanced signal processing; multiple-input, multiple-output (MIMO) techniques; and artificial intelligence to achieve unprecedented levels of integration and performance [7].

The emergence of ISAC has been accelerated by several technological trends, including the development of software-defined radio platforms, advances in signal processing algorithms, and the increasing demand for spectrum-efficient wireless systems. The upcoming 6G wireless networks are expected to fully embrace ISAC as a core technology, enabling new applications across various domains, including healthcare [7].

3.1.2. Integration Principles of Sensing and Communication

ISAC systems integrate sensing and communication in four tightly coupled layers. First, in the hardware layer, a single set of antenna arrays, RF chains, and base-band processors is reused for both tasks, which cuts cost and power while enabling agile beam steering with massive-MIMO or phased-array front ends [7].

Second, in the waveform layer, engineers co-design dual-functional signals whose ambiguity-function sharpness meets radar-resolution demands, while their spectral efficiency meets data-rate targets; the literature distinguishes communication-centric, sensing-centric, and fully joint designs, and recent robust formulations keep both qualities stable under channel uncertainty [7,12].

Third, in the resource-allocation layer, time, frequency, spatial beams, and power are adaptively split between the two services so that latency and detection probability trade-offs remain near-optimal, even as traffic or scene dynamics change [13].

Finally, in the signal-processing layer, a single received echo is decomposed into user data and sensing parameters by sparsity-aware, compressed sensing, and machine learning methods, as surveyed in [14].

3.1.3. Role of ISAC in 6G Networks

Integrated Sensing and Communication (ISAC) is widely recognized in the peer-review literature as a cornerstone of 6G wireless networks that will extend radio functionality far beyond data transport [7,15].

In a fully fledged 6G environment, ISAC is expected to enable the following:

(i)

Ubiquitous sensing, whereby dense, distributed transceivers perceive the radio environment and provide rich context awareness [15];

(ii)

Centimeter-level localization for users and smart objects by exploiting wide-bandwidth and joint radar–communication signal processing [16];

(iii)

Gesture and activity recognition through contact-free Wi-Fi or mmWave sensing that classifies human motion patterns in real time [17];

(iv)

High-resolution imaging and mapping that turns the network into a distributed radar/camera array [7];

(v)

Non-invasive health monitoring via metrics such as heart beat, respiration, or blood-pressure tracking via RF reflections [18].

By fusing these sensing services with the ultra-reliable, low-latency links of 6G, the radio network becomes a cyber–physical fabric capable of maintaining live digital twins of people and spaces, an approach that aligns directly with emerging healthcare twin platforms [19].

3.2. Digital-Twin Technology Overview

3.2.1. Definition and Concept

A digital twin is a high-fidelity, virtual representation of a physical entity, process, or system that remains bidirectionally coupled to its real-world counterpart throughout the entire life cycle [20,21]. In contrast to one-off simulations, a twin ingests real-time data and can, via feedback control, influence the physical asset.

The concept is commonly described by six characteristics. Fidelity refers to the geometric, physical, and behavioral accuracy of the model [22]. Connectivity denotes the continuous, low-latency data exchange that keeps virtual and physical states synchronized, while Historicity emphasizes the systematic storage of time-stamped data for trend analysis and prediction [23]. Analytics embeds model-based and data-driven reasoning that turns raw telemetry into actionable insight [23]. Interaction highlights user-oriented interfaces—ranging from dashboards to immersive XR—that allow humans to query or steer the twin [24]. Finally, Evolution captures the ability to evolve and adapt as the physical entity changes over time [25].

Digital-twin principles scale from component-level replicas (e.g., a heart chamber) to system-level twins of entire hospitals, thanks to modular abstraction and hierarchical data management [22]. Such multi-scale fidelity is particularly valuable in healthcare, where physiological, device, and facility data must co-evolve; recent surveys foresee digital-twin platforms enabling personalized therapies and hospital-wide optimization [19,26,27].

3.2.2. Current Applications in Healthcare

Digital-twin technology has already demonstrated significant potential across various healthcare domains:

Patient-specific modeling: GPU-accelerated cardiac twins can reproduce multi-physics heart dynamics and forecast individual responses to interventions, allowing cardiologists to test therapy strategies virtually [28]. In respiratory medicine, digital twins of chronic lung-disease patients help clinicians tailor ventilator settings and anticipate progression [29].

Hospital operations: Facility-level twins integrate EHR, asset tracking, and environmental data to improve the 30-second response rate (QoS₃₀) by 17–21%p [30,31].

Medical-device management: A systematic review shows that digital twin-based predictive maintenance frameworks cut unplanned downtime and extend device life by coupling real-time telemetry with physics models [32]. Implantable devices such as pacemakers are beginning to adopt similar virtual monitoring strategies to fine-tune parameters non-invasively [27].

Clinical trials and drug development: In silico cohorts driven by patient digital twins can identify likely responders, shorten trial duration, and lower cost; recent work also outlines concrete regulatory pathways and real-world case studies spanning oncology and cardiometabolic disease [33,34,35].

Public health management: Population-scale twins that couple epidemiological and capacity data are being explored to predict outbreak trajectories and optimize intervention logistics; a recent framework highlights data-sharing requirements for pandemic-level deployment [36].

These examples illustrate how healthcare twins move from proof of concept to operational impact, spanning bedside decision support to health-system resilience [19,26,37].

3.2.3. Key Enabling Technologies

Figure 2 summarizes the enabling layers discussed in this section. The basic stack comprises the following:

(i)

Sensors/IoMT (ECG, radar, UWB tags, wearables, and infusion pumps) that feed raw and derived observations;

(ii)

The ISAC link (PHY/MAC) spanning sub-6 GHz, mmWave, UWB, and THz bands with URLLC scheduling;

(iii)

A data pipeline for ingestion, QA/QC, logging, and semantic mapping to FHIR, HL7, and IEEE 11073;

(iv)

Twin engines providing physiology-based modeling, inference/analytics, and simulation;

(v)

Clinical applications that realize monitoring, CDS, and personalization.

Cross-cutting layers—interoperability, deployment, and assurance—appear alongside the basic stack to highlight system integration concerns that apply across layers.

The implementation of digital-twin technology in healthcare relies on several enabling technologies that map directly to Figure 2:

Internet of Things (IoT) and Internet of Medical Things (IoMT) (i): Dense networks of wearables, implantables, and smart devices stream real-time biosignals and context data that feed patient digital twins [8].

Cloud and edge computing (ii–iii): Hybrid cloud–edge offloading frameworks lower latency for bedside analytics and allow for scalable twin synchronization [38].

Artificial intelligence and machine learning (iv): Supervised and self-supervised models learn physiological patterns from twin data to forecast deterioration and recommend interventions [39].

Advanced visualization (v): AR/VR rendering enables interactive 3D exploration of cardiac or pulmonary twins, improving surgical planning and situational awareness [40].

Simulation and modeling (iv): High-fidelity organ twins—such as GPU-accelerated heart models [28] and multi-scale lung twins [29]—merge physics with data-driven layers to reproduce patient-specific dynamics.

Interoperability standards (iii): FHIR-based APIs expose granular EHR data as resources that can be queried or updated by twin engines, ensuring semantic consistency across devices and apps [5].

3.3. Healthcare System Challenges

3.3.1. Aging Population and Chronic Disease Management

The global demographic shift toward aging populations presents significant challenges for healthcare systems worldwide. According to the World Health Organization, the proportion of people aged 60 and older will nearly double from 12% to 22% between 2015 and 2050, rising from 900 million to 2 billion people in this age group [41,42].

This demographic transition has profound implications for healthcare:

Increased prevalence of chronic conditions: Older populations experience higher rates of chronic diseases such as cardiovascular disease, diabetes, dementia, and arthritis. These conditions typically require continuous monitoring and management rather than episodic care [43,44].

Multi-morbidity: The co-occurrence of multiple chronic conditions becomes increasingly common with age, complicating treatment approaches and requiring coordinated care across multiple specialties [45].

Long-term care needs: Many older adults need assistance with activities of daily living, creating demand for both institutional and home-based long-term care services [46].

Healthcare workforce challenges: The aging of the healthcare workforce itself, combined with increasing demand for services, creates staffing challenges that necessitate more-efficient care delivery models and new technologies [2].

Traditional healthcare models—designed primarily for acute, episodic care—are poorly suited to address these pressures. Continuous monitoring, preventive interventions, and tightly coordinated care are increasingly recognized as essential to manage chronic conditions while maximizing patients’ quality of life and independence.

3.3.2. Data Integration and Analysis Difficulties

Modern healthcare generates enormous volumes of data from diverse sources, including electronic health records (EHRs), medical imaging, laboratory tests, wearable devices, and environmental sensors. However, several challenges impede the effective integration and analysis of this data:

Data Silos: Healthcare data often exists in isolated systems with limited interoperability, preventing the formation of comprehensive patient records and system-wide analytics [47]. These silos result from a combination of technical, organizational, and regulatory factors, including proprietary systems, departmental boundaries, and privacy concerns.

Data Heterogeneity: Healthcare data spans multiple modalities, formats, and structures—from structured clinical measurements to unstructured clinical notes and complex imaging data [48]. This heterogeneity complicates data integration and requires sophisticated approaches for harmonization and analysis.

Data Quality Issues: Healthcare data frequently suffers from quality problems, including missing values, inconsistencies, errors, and biases [49]. These issues can undermine the reliability of analyses and limit the utility of the data for decision support.

Real-time Processing Challenges: Many healthcare applications require real-time or near-real-time data processing, particularly for the monitoring of critically ill patients or detecting emergent conditions [50]. Traditional data processing pipelines often introduce unacceptable latencies for these time-sensitive applications.

Privacy and Security Concerns: Healthcare data is subject to strict privacy regulations and security requirements, which can create barriers to data sharing and integration [51]. Balancing data accessibility for legitimate clinical and research purposes with privacy protection remains a significant challenge.

Addressing these data integration and analysis difficulties is essential for realizing the potential of data-driven healthcare approaches, including those enabled by ISAC and digital-twin technologies. Advanced data integration frameworks, interoperability standards, and privacy-preserving analytics techniques are being developed to overcome these challenges and enable more comprehensive, timely, and actionable healthcare insights.

3.3.3. Remote Monitoring and Telehealth Limitations

The COVID-19 pandemic accelerated the adoption of telehealth and remote monitoring solutions, demonstrating their potential to extend healthcare access and enable continuous care outside traditional clinical settings. However, several limitations constrain the effectiveness of current approaches:

Sensing Limitations: Current remote monitoring technologies often provide limited physiological data compared to in-person clinical assessments. Many important clinical parameters cannot be reliably measured with existing consumer-grade devices, creating gaps in remote patient monitoring [52].

Connectivity: These programs rely on consistent Internet connectivity, which may be unavailable or unreliable in rural areas, developing regions, or for socioeconomically disadvantaged populations. These gaps can exacerbate healthcare disparities [53].

Data Integration: Collected streams often reside separately from clinical records, complicating integration into workflows and decision-making processes and limiting their utility for comprehensive care management [54].

Contextual Awareness: Current approaches often lack environmental and activity context, which is crucial for interpreting physiological data and making appropriate clinical decisions [55].

Patient Engagement: Sustained participation with such technologies remains difficult; many patients discontinue devices over time or experience alert fatigue, which limits long-term program effectiveness [56].

The integration of ISAC and digital-twin technologies offers potential solutions to many of these limitations by enabling more comprehensive sensing, efficient data transmission, contextual awareness, and intuitive visualization of patient status. These technologies could transform remote monitoring from a limited supplement to in-person care to a robust approach for continuous, context-aware health management.

3.4. Convergence Opportunities

The convergence of ISAC and digital-twin technologies presents unique opportunities to address healthcare challenges through several mechanisms.

3.4.1. Enhanced Data Collection and Integration

ISAC technology can significantly enhance the data collection capabilities that underpin digital twins as follows:

Multimodal Sensing: ISAC systems can simultaneously monitor multiple physiological parameters, environmental conditions, and behavioral patterns, providing a more comprehensive data foundation for digital twins [57].

Continuous Monitoring: The energy efficiency and compact form factors of recent ISAC hardware prototypes enable longer duration telemetry with less patient burden, supporting the continuous updating of digital twins [58].

Contextual Awareness: ISAC’s environmental sensing capability can supply high-fidelity context (e.g., ambient vibration, motion, or environmental dynamics) that enriches physiological data interpretation within digital-twin models [59].

Efficient Data Transmission: The converged PHY/MAC design of ISAC networks reduces latency and improves link reliability, ensuring timely data delivery to digital-twin engines [60].

3.4.2. Real-Time Feedback and Intervention

The bidirectional nature of both ISAC and digital-twin technologies enables closed-loop systems for healthcare management:

Predictive Alerts: Digital twins analyze streaming ISAC data to forecast adverse events or disease exacerbations before they occur, enabling preventive interventions [19].

Adaptive Monitoring: Based on digital-twin predictions and risk assessments, ISAC systems can dynamically adjust their sensing parameters to focus on the most relevant physiological signals or environmental factors [61].

Personalized Guidance: Digital twins can generate personalized recommendations for patients based on their current status and predicted trajectories, which can be communicated through ISAC-enabled devices [62].

Remote Intervention: In some cases, therapeutic interventions can be delivered remotely through ISAC-enabled devices based on digital-twin assessments, such as medication reminders, stimulation-therapy adjustments, or environmental modifications [19].

3.4.3. System-Level Optimization

Beyond individual patient care, the convergence of ISAC and digital-twin technologies enables system-level optimization of healthcare delivery across several dimensions.

First, digital twins of hospitals and healthcare networks—powered by ISAC-generated data on patient flow, equipment usage, and staff mobility—can support dynamic resource allocation and scheduling optimization [63,64]. These capabilities are particularly valuable in maximizing efficiency in high-demand environments such as emergency departments or intensive care units, where real-time insights into operational conditions can drive better decision-making.

Moreover, system-level digital twins allow healthcare planners to simulate future patient populations and care demands, enabling predictive capacity planning. This approach helps healthcare systems anticipate bottlenecks, allocate staff and beds appropriately, and plan infrastructure investments strategically [63].

In the context of public health emergencies—such as pandemics or natural disasters—the distributed sensing capabilities of ISAC, combined with digital-twin models, can provide real-time situational awareness across facilities and regions. These systems enable coordinated emergency responses by visualizing resource status, patient surges, and intervention effects in a shared operational picture [65,66].

Additionally, the continuous collection and analysis of operational workflow data through ISAC enable digital twins to identify inefficiencies and deviations from best practices. These insights support continuous improvement initiatives aimed at streamlining healthcare processes and reducing unnecessary costs [67].

Altogether, the integration of ISAC and digital-twin technologies offers a scalable and data-driven approach to optimizing healthcare systems at the operational and strategic levels, complementing the benefits of personalized care with system-wide intelligence.

Because limitations and research needs are distributed across the literature review, the ISAC (spectrum/hardware) analyses, and the digital-twin integration analyses,

Table 5. Cross-cutting research gaps synthesized from the literature review and from analyses across the ISAC and digital-twin sections.

Area	Observed Limitation in Prior Work	Research Need/Direction	Where Covered
External validity	Robustness and reliability across varied clinical conditions not established Limited evidence on stability over time in monitoring Lack of standardized validation/benchmarking methodology	Standardized validation and benchmarking Longitudinal evaluation for stability Alignment with clinically relevant outcomes	Section 4.5.2
DT synchronization	ISAC sensing/link cadences not aligned with DT update policies Ingestion/QA pipelines often omit provenance and synchronized timestamps Semantic alignment across EHR/device and ISAC streams is fragile	Align cadences and record provenance for synchronized updates QA tool chains for replay/trust in timestamps/identifiers Robust semantic mapping for heterogeneous sources	Section 6 Section 4.5.3
Latency-aware inference	End-to-end delay (sensing / link / edge–cloud / actuation) not modeled Pipelines do not balance model fidelity with latency/throughput constraints Limited real-time/closed-loop clinical evidence	Placement under latency budgets; throughput-aware pipelines Closed-loop evaluation with ISAC-in-the-loop constraints Reporting of detection/decision/action timelines	Section 6 Section 6.3
Model calibration & VVUQ	Uncertainty propagation under link/sensor variability under-specified Personalization remains brittle; high-dimensionalparameter spaces	VVUQ routines that condition on ISAC telemetry quality/cadence Robust calibration with external validation across sites/subgroups	Section 6.6
Semantic interoperability	Mapping ISAC features to FHIR-compatible resources incomplete Inconsistent semantics/lineage across EHR and device streams	FHIR-aligned schemata/APIs with validation Timestamp/identifier policies for synchronization and replay	Section 4.5.3
Security, privacy, and safety	ISAC-specific threat models under-specified (exposure/power limits and leakage surfaces) On-link protections and privacy constraints not co-designed with sensing Band-dependent exposure limits (SAR for Sub-6; power density for mmWave/THz)	Threat models and requirements for ISAC waveforms/links Privacy-by-design and on-link protections integrated with DT cadence Design within SAR/PD limits; power/beam control policies	Section 4.5.1 Section 4.6.5 Section 4.6.4
Spectrum and hardware limits	Resolution–penetration–cost trade-offs vary by band Hardware maturity constraints (Sub-6/mmWave/THz) Wearable/implantable power budgets and miniaturization constraints	Evidence-driven band selection Multimodal sensing across bands Power-aware designs (duty cycling and burst scheduling) and safe in situ power/beam strategies	Section 4.3.3 Section 4.1.3 Section 4.2.3 Section 4.5.1
Organizational adoption	Workflow integration and change management needs Training/governance not fully specified Outcome measurement frameworks underdeveloped	Co-design with clinical workflows Training and governance programs Deployment guidance and outcome evaluation	Section 6.5
Equity and infrastructure	Performance variation across facilities and populations Unreliable electricity/broadband in LMICs; capex/opex of dense mmWave small cells	Prospective studies in constrained settings Cost/benefit and accessibility metrics	Section 4.5.4

Abbreviations: DT, Digital Twin; ISAC, Integrated Sensing and Communication; EHR, Electronic Health Record; FHIR, Fast Healthcare Interoperability Resources; SAR, Specific Absorption Rate; PD, Power Density; mmWave, millimeter wave (30–300 GHz); THz, terahertz (0.1–10 THz); LMICs, Low- and Middle-Income Countries; QA, Quality Assurance; VVUQ, Verification, Validation, and Uncertainty Quantification.

aggregates these cross-cutting items and provides pointers to each item’s primary discussion.

Specifically, it covers the following: External validity and evaluation (Section 4.5.2); DT synchronization—aligning ISAC sensing/link cadences with DT update policies, recording provenance and synchronized timestamps, and robust semantic mapping (Section 4.5.3 and Section 6); latency-aware inference—placement under latency budgets, throughput-aware pipelines, and closed-loop timelines (Section 6 and Section 6.3); semantic interoperability (FHIR-aligned schemata/APIs and synchronization/replay policies, Section 4.5.3); model calibration and VVUQ conditioned on ISAC telemetry quality with external validation (Section 6.6); security, privacy, and safety—ISAC-specific threat models, on-link protections, and band-dependent exposure limits (Section 4.5.1, Section 4.6.4 and Section 4.6.5); spectrum a hardware limits with safe power/beam strategies (Section 4.1.3, Section 4.2.3, Section 4.3.3 and Section 4.5.1); Organizational adoption (Section 6.5); and equity and infrastructure in resource-constrained settings (Section 4.5.4).

4. Technical Analysis of ISAC Technology and Frequency Bands

This section provides a detailed technical analysis of Integrated Sensing and Communication (ISAC) technology with a specific focus on the various frequency bands employed in healthcare applications. We examine the characteristics, advantages, limitations, and specific applications of millimeter-wave (mmWave), terahertz (THz), and other relevant frequency bands in the context of healthcare monitoring and digital twin integration.

Table 6. Comparative summary of ISAC frequency bands for healthcare applications. TRL = technology readiness level (EU Horizon 2020 scale [69]).

Frequency Band	Typical Sensing Resolution	Penetration Through Tissue/Obstacles	Usable Bandwidth	Approx. Hardware Cost (2025)	Safety Metric and Compliance	TRL
mmWave (30–300 GHz)	Medium (centimeter level)	Moderate (LOS preferred)	2–4 GHz per channel	Moderate (phased-array RFICs)	Power density ≤ 10 W m−2	6–7
THz (0.1–10 THz)	High (sub-millimeter)	Low (surface only) < 1 mm in tissue	10–100 GHz per channel	High (lab prototypes)	Power density ≤ 1 W m−2	3–4
Sub-6 GHz	Low (decimeter level)	High (good wall and tissue penetration)	≤200 MHz per channel	Low (commodity chipsets)	SAR-based, within ICNIRP limits	8–9

offers an at-a-glance comparison of the three principal ISAC frequency bands—sub-6 GHz, mmWave (30–300 GHz), and THz (0.1–10 THz)—against seven practical criteria (sensing resolution, tissue penetration, usable bandwidth, approximate 2025 hardware cost, regulatory ICNIRP (International Commission on Non-Ionizing Radiation Protection) safety limits [68], and technology readiness level [69]), thereby enabling clinicians and engineers to select the most suitable band for a given healthcare scenario.

4.1. Millimeter-Wave (mmWave) Band

4.1.1. Characteristics and Properties

Millimeter wave (mmWave) refers to the portion of the electromagnetic spectrum spanning frequencies from 30 to 300 GHz, with corresponding wavelengths between 1 and 10 mm [70]. This spectrum offers a set of technical properties that make it especially attractive for Integrated Sensing and Communication (ISAC) in healthcare environments.

One of the primary advantages of the mmWave band lies in its abundant bandwidth. Compared with sub-6 GHz bands, mmWave channels can provide several-hundred-megahertz to multi-gigahertz bandwidths, enabling both gigabit-class data transmission and fine-grained sensing [70]. These characteristics are critical for healthcare applications that require simultaneous real-time monitoring and high-data-rate communication, such as remote diagnostics or immersive telepresence.

The short wavelength of mmWave signals also contributes to fine spatial resolution. In radar-based sensing, range resolution improves with increased bandwidth, while angular resolution benefits from the shorter wavelengths associated with mmWave. Recent studies have achieved centimeter-level localization and sub-degree angular differentiation for patient positioning and respiratory motion tracking [71,72].

Another distinctive aspect of mmWave is its reliance on beamforming and directional communication. Owing to high free-space path loss, mmWave systems employ large antenna arrays to generate narrow, high-gain beams that both compensate for attenuation and enhance spatial selectivity—thereby reducing multi-user interference and enabling concurrent communication–sensing in dense environments [73].

However, mmWave signals also present propagation challenges. They suffer from increased atmospheric absorption—particularly around the 60 GHz oxygen absorption peak—and have a limited ability to penetrate obstacles such as walls or the human body. Instead, signals tend to reflect and scatter, producing complex multipath environments that must be addressed in system design [74].

Finally, the realization of mmWave systems demands advanced hardware and signal-processing solutions. High-frequency RF front ends, compact phased arrays, and ultra-fast ADC/DACs are essential to support precise beamforming and channel estimation. Recent SiGe BiCMOS transceiver prototypes for biomedical imaging illustrate the required design trade-offs between bandwidth, energy efficiency, and physical size—particularly important for wearable or implantable medical devices [75].

Altogether, mmWave presents a compelling option for ISAC-based healthcare applications, offering high-resolution, bandwidth-rich communication and spatial precision—provided that its limitations in propagation and hardware complexity are properly managed.

4.1.2. Applications in Healthcare ISAC (mmWave)

Millimeter-wave (mmWave) technology, owing to its unique characteristics, is driving several key applications in healthcare-focused Integrated Sensing and Communication (ISAC) systems.

One prominent application area is vital-sign monitoring. mmWave radar systems leverage their sensitivity to detect the minute chest-wall displacements associated with respiration and heartbeat. Studies at 60 GHz and 77 GHz have shown that contact-free systems can estimate breathing and pulse with accuracy comparable to that of clinical contact sensors—even when subjects wear loose clothing or are in motion [71,76].

Beyond vital signs, the fine spatial resolution of mmWave enables detailed tracking of human movement for gait analysis and fall detection. Continuous assessment of step symmetry, stride length, and fall events is invaluable for elderly or mobility-impaired patients and can be streamed into digital-twin models for risk prediction [77,78].

Sleep monitoring also benefits from mmWave sensing. By capturing respiration, gross body movement, and even heartbeat during sleep, mmWave devices deliver polysomnography-comparable metrics without burdening patients with electrodes or belts [71,79].

In facility management, mmWave systems enable precise indoor localization and tracking. Narrow beams and sub-degree angular resolution allow for centimeter-level positioning of patients, staff, and equipment, feeding operational digital twins that optimize workflows and resource allocation [80,81].

Finally, the abundant bandwidth in mmWave bands supports rapid transmission of high-volume medical data—e.g., real-time sensor feeds or raw MRI/CT image streams—so that digital-twin models remain synchronized with the physical world [82].

Taken together, these applications illustrate mmWave’s crucial role in advancing healthcare ISAC systems toward more responsive, efficient, and patient-centered care delivery.

4.1.3. Limitations and Challenges

While millimeter-wave (mmWave) technology offers promising capabilities for healthcare ISAC applications, it also presents several notable challenges that must be addressed for effective deployment.

First, the limited penetration ability of mmWave signals poses a fundamental constraint. Due to their short wavelengths, mmWave signals struggle to penetrate solid materials such as walls, furniture, and even the human body. As a result, sensing and communication are often restricted to line-of-sight scenarios or must rely on reflected paths, degrading reliability and coverage in complex hospital environments [83,84].

Moreover, mmWave frequencies experience significant free-space path loss and are highly sensitive to blockage by objects or human bodies. This leads to substantial coverage challenges—particularly in dynamic indoor settings where signal paths frequently change. To achieve comprehensive sensing and communication coverage, dense deployments of access points or sensing nodes are often required, increasing system complexity and cost [85,86].

Hardware complexity further exacerbates these challenges. mmWave systems demand phased-array antennas, high-speed data converters, and precise RF front ends. Such tight manufacturing tolerances and advanced beamforming capabilities raise system cost, limiting scalability in resource-constrained healthcare environments [87].

In addition, the high-frequency operation and intensive signal processing associated with mmWave systems incur elevated power consumption—problematic for battery-powered medical wearables, where energy efficiency is critical [88].

Finally, the directional nature of mmWave communication necessitates sophisticated beam-forming and beam-tracking algorithms to maintain reliable links in dynamic environments, adding computational burden for real-time healthcare sensing and communication [14].

Overall, while mmWave technology holds transformative potential for healthcare ISAC, mitigating its physical, architectural, and operational limitations is essential for successful clinical integration.

4.2. Terahertz (THz) Band

4.2.1. Characteristics and Properties

The terahertz (THz) band, spanning frequencies from 0.1 to 10 THz, bridges the gap between microwave and infrared regions of the electromagnetic spectrum. This frequency range offers several unique characteristics pertinent to healthcare ISAC applications.

One of the most notable features of the THz band is its ultra-wide bandwidth. This vast bandwidth enables ultra-high-speed communication, potentially reaching terabits per second, and allows for extremely fine sensing resolution. Such capabilities are crucial for applications requiring high data rates and precise measurements [89].

THz radiation interacts with molecular rotational and vibrational states, producing distinctive spectral signatures for various biological materials. This property facilitates material identification and compositional analysis through spectroscopic techniques, making THz spectroscopy a valuable tool in biomedical diagnostics [90,91].

In terms of penetration properties, THz waves can traverse certain non-metallic materials, such as clothing, paper, and plastics, while being strongly absorbed by water and reflected by metals. This selective penetration enables unique sensing applications, such as non-invasive imaging, while maintaining certain privacy and safety advantages [92].

Propagation characteristics of THz waves include extreme free-space path loss; strong atmospheric absorption, particularly due to water vapor; significant scattering from surface roughness comparable to the wavelength; and limited diffraction around obstacles due to quasi-optical behavior. These factors necessitate careful consideration in system design and deployment [93].

Hardware implementation in the THz band presents significant challenges. The limited availability of efficient sources and sensitive detectors; difficulties in signal generation, modulation, and detection; and precision requirements for components and alignment contribute to the complexity. Additionally, thermal noise and stability issues pose further obstacles to the development of reliable THz systems [94].

Despite these challenges, ongoing research and technological advancements continue to address these limitations, paving the way for the integration of THz technology into healthcare ISAC applications.

4.2.2. Applications in Healthcare ISAC (THz)

The terahertz (THz) frequency band, spanning from 0.1 to 10 THz, offers unique properties that enable a range of specialized applications in healthcare Integrated Sensing and Communication (ISAC) systems.

One significant application is non-invasive biomedical imaging. THz imaging delivers high-resolution, non-ionizing visualization of surface and subsurface structures, proving useful for skin cancer detection, burn injury assessment, dental imaging, and wound-healing monitoring [95,96].

Sensitivity to water-content variations allows for detailed imaging without risk of ionizing radiation. THz spectroscopy enables molecular sensing by exploiting characteristic rotational–vibrational fingerprints, supporting applications such as glucose monitoring, hydration assessment, and biomarker detection in breath or bio-fluids [97,98].

The extremely short wavelength of THz waves supports ultra-high-resolution vital-sign monitoring. Systems have resolved micrometer-scale chest motions and pulse waveforms, offering fine-grained physiological data for advanced digital-twin models [99,100].

For secure communication, the high directionality and limited propagation range of THz links inherently reduce eavesdropping risk; experimental studies demonstrate physical-layer secrecy capacities exceeding lower frequency systems [101,102].

Finally, the massive bandwidth available in the THz band enables ultra-high-speed data transfer. A recent 0.22 THz link achieved 84 Gb/s over 1.26 km, illustrating the feasibility of real-time updates for large digital-twin datasets and uncompressed 8K medical video [103].

Collectively, these applications underscore THz technology’s transformative potential for healthcare ISAC—advancing diagnostics; patient monitoring; and secure, low-latency data exchange.

4.2.3. Limitations and Challenges

Despite the promising applications of terahertz (THz) technology in healthcare ISAC systems, several significant challenges hinder its practical implementation.

Firstly, THz system technology remains less mature compared to established lower frequency systems. Challenges persist in efficient signal generation, sensitive detection, and stable processing, limiting the immediate deployability of THz solutions in clinical environments [89].

Secondly, THz signals experience extreme path loss and atmospheric absorption, particularly from water vapor molecules, severely restricting their effective propagation range. Indoor THz links are typically limited to under 10 meters without sophisticated line-of-sight engineering [93].

Moreover, the strong absorption of THz radiation by water confines penetration into biological tissues to shallow depths—often less than one millimeter. This restricts sensing to surface or near-surface structures, limiting applications that require imaging of deeper tissues [92].

Cost and system complexity present further hurdles. THz devices require specialized materials, precision fabrication, and highly sensitive components, resulting in significantly higher costs than mmWave systems [94].

Finally, the THz band currently lacks comprehensive standardization and established regulatory frameworks tailored for healthcare applications. This regulatory uncertainty inhibits investment and slows the transition of THz technology from research laboratories to clinical practice [104].

Overcoming these challenges is critical to realizing the full potential of THz-enabled healthcare ISAC systems.

4.3. Sub-6 GHz Band

4.3.1. Characteristics and Properties

The sub-6 GHz frequency band, encompassing frequencies below 6 GHz, plays a pivotal role in wireless communications, including cellular networks, Wi-Fi, and IoT protocols. Its propagation characteristics are particularly advantageous for healthcare ISAC applications.

Firstly, sub-6 GHz signals exhibit favorable propagation properties—lower free-space path loss than higher frequency bands, better penetration through walls and human tissue, and reduced sensitivity to blockage and atmospheric conditions. These features enable reliable non-line-of-sight (NLOS) communication, even in complex indoor environments such as hospitals [105,106].

Secondly, the superior coverage and range of sub-6 GHz frequencies allow for a broader area coverage with fewer access points, making deployments more cost-effective for large healthcare facilities [107,108,109].

Moreover, technologies operating in the sub-6 GHz band benefit from decades of development: widely available components, mature standards and protocols, efficient signal-processing techniques, and comprehensive regulatory frameworks [110,111,112].

However, the sub-6 GHz spectrum offers limited bandwidth relative to mmWave or THz bands—typically tens to a few hundred MHz per channel—constraining achievable data rates and sensing resolution for bandwidth-hungry healthcare ISAC applications [85].

Finally, many sub-6 GHz bands are heavily utilized, leading to interference and spectrum availability challenges. The growing number of users and services vying for limited resources demands sophisticated spectrum management strategies [113].

4.3.2. Applications in Healthcare ISAC (Sub-6 GHz)

The sub-6 GHz frequency band offers several advantages for healthcare ISAC applications, owing to its favorable propagation characteristics and established technological ecosystem.

Firstly, the superior penetration and range of sub-6 GHz signals facilitate whole-building monitoring in healthcare facilities. These signals can traverse walls and floors, enabling comprehensive tracking of patient and staff movements, equipment utilization, and environmental conditions, which are essential components of facility-wide digital twins [114].

Secondly, despite offering lower resolution than mmWave or THz systems, sub-6 GHz radar can effectively detect vital signs such as respiration and heart rate at distances of several meters, even through obstacles. This capability is valuable for general patient monitoring in hospital rooms or home environments [115].

Thirdly, sub-6 GHz sensing can identify human activities and movements based on characteristic signal patterns. This enables the monitoring of patient mobility, activities of daily living, and potential emergencies such as falls, enriching patient digital twins with behavioral and functional data [17].

Furthermore, the sub-6 GHz band provides reliable connectivity for medical devices, wearables, and environmental sensors that feed data to digital-twin systems. The established standards and protocols in this band facilitate interoperability and system integration [116].

Lastly, for remote patient monitoring applications, sub-6 GHz communications offer greater range and reliability than higher bands, enabling digital-twin updates, even in challenging environments or rural settings [117]. A consolidated view of these band-specific applications follows in

Table 7. Healthcare applications and representative studies in the ISAC frequency band.

Band (§ Link)	Specific Examples/Techniques	Key Characteristics
mmWave (30–300 GHz; see Section 4.1.2)
Vital-sign monitoring	Contact-free breathing and pulse estimation at 60/77 GHz	Accuracy comparable to that of clinical contact sensors (reported prototypes/studies)
Gait analysis and fall detection	Fine-grained human motion tracking; gait symmetry/stride measures	Useful for older adults or patients with limited mobility
Sleep monitoring	Respiration, body motion, and heart-related signals during sleep	PSG-comparable metrics without electrodes/belts
Indoor localization	Centimeter-level localization/tracking (patients/staff/equipment)	Feeds operational digital twins for workflow/resource optimization
High-volume data transfer	Streaming of high-volume medical data (e.g., real-time sensor feeds and raw MRI/CT image streams)	Abundant bandwidth for timely twin synchronization
THz (0.1–10 THz; see Section 4.2.2)
Biomedical imaging (non-invasive)	THz imaging for skin/thermal wound/dental/lesion assessment	Non-ionizing; sensitive to water-content changes
Spectroscopy for biosensing	Molecular/biomarker sensing; hydration assessment via THz spectroscopy	Distinct spectral signatures enable identification
Ultra-high-resolution vital-sign monitoring	Vital-sign sensing with very fine spatial/temporal granularity	Ultra-fine sensing resolution for physiology-aware DT inputs
Physical-layer security	Directionality and limited propagation range studied for PLS	Reflection-path risks noted; security aspects investigated
Ultra-high-speed data links	0.22 THz link, achieving 84 Gb/s over 1.26 km	Real-time large DT datasets/uncompressed video support
Sub-6 GHz (see Section 4.3.2)
Facility-wide monitoring	Tracking of movement/utilization/occupancy across wards/facilities	Facility-level digital-twin components
Vital-sign sensing	Respiration/heart-related sensing at several meters, even through obstacles	Room/home tolerance with obstacle penetration
Human activity recognition	ADL/mobility detection; fall alerts	Enriches patient digital twins with behavioral context
Medical device connectivity	Connectivity for wearables and environmental sensors	Standards/protocols facilitate interoperability and system integration
Remote patient monitoring	Reliable updates in challenging or rural environments	Greater range/reliability than higher bands

, which also provides § links to the mmWave, THz, and sub-6 subsections.

4.3.3. Limitations and Challenges

While the sub-6 GHz frequency band offers several advantages for healthcare ISAC applications, it also presents certain limitations and challenges that must be considered.

Firstly, the longer wavelength and limited bandwidth of sub-6 GHz signals restrict sensing resolution, making them less suitable for applications requiring fine detail or precise localization. This limitation affects the ability to detect subtle physiological changes or small movements, which are critical for certain medical diagnostics [118].

Secondly, the limited bandwidth constrains communication data rates, potentially creating bottlenecks for applications requiring transmission of large data volumes, such as high-resolution medical imaging. This constraint can hinder real-time data transmission and affect the performance of data-intensive healthcare applications [119].

Thirdly, the widespread use of sub-6 GHz bands increases the potential for interference, particularly in dense healthcare environments with numerous wireless devices. This interference can degrade both sensing and communication performance, leading to unreliable data and potential misdiagnoses [120].

Furthermore, many sub-6 GHz bands are subject to strict regulatory limitations on transmission power and bandwidth, constraining system design and performance. These regulations can limit the deployment flexibility and scalability of healthcare ISAC systems [104].

Lastly, the superior penetration of sub-6 GHz signals raises privacy concerns, as sensing may inadvertently capture information from unintended areas or subjects. Ensuring patient privacy and data security becomes more challenging, necessitating robust privacy-preserving mechanisms [121].

4.4. Multi-Band Integration

4.4.1. Complementary Characteristics

Integrating multiple frequency bands in healthcare ISAC applications leverages the unique advantages of each band to enhance overall system performance. Lower frequency bands, such as sub-6 GHz, offer extensive coverage and superior penetration capabilities, making them ideal for broad-area monitoring and communication reliability [7,122]. Conversely, higher frequency bands like mmWave and THz provide higher resolution and finer sensing detail [123], which are crucial for applications requiring precise localization and detailed physiological measurements.

This multi-band approach allows systems to dynamically select the most appropriate frequency band based on specific application requirements and environmental conditions. For instance, sub-6 GHz frequencies can be used for general patient monitoring throughout a facility, ensuring consistent connectivity, even in challenging environments, while mmWave or THz bands can be employed for high-resolution imaging in critical zones [7,124].

Moreover, different frequency bands interact distinctively with biological tissues and the surrounding environment, offering diverse sensing information. This sensing diversity enables a more comprehensive understanding of patient conditions and environmental factors, which significantly enhances the accuracy and reliability of digital-twin models in healthcare [125].

Additionally, employing multiple frequency bands enhances communication reliability. Systems can intelligently switch between bands to maintain optimal performance, using higher frequencies for high-speed data transmission and reverting to lower frequencies under challenging conditions [123,124].

Finally, intelligent allocation of communication tasks across various frequency bands optimizes spectral efficiency. High-bandwidth applications can be directed to mmWave or THz bands, while control signaling and lower bandwidth communications utilize sub-6 GHz frequencies, ensuring efficient use of available spectral resources [7].

4.4.2. Implementation Approaches

The integration of multiple frequency bands in healthcare ISAC systems can be achieved through several implementation approaches, each leveraging the unique advantages of different frequency bands to enhance system performance and reliability.

Firstly, deploying heterogeneous networks involves the coordination of complementary networks operating in different frequency bands. This approach allows for the optimization of overall performance by utilizing existing infrastructure while incorporating new capabilities. For instance, sub-6 GHz bands can ensure wide coverage and robust connectivity, while mmWave or THz bands provide high-resolution sensing in critical areas [7,126].

Secondly, the development of multi-band devices capable of operating across various frequency bands is crucial. These devices can intelligently select the appropriate band based on application requirements and environmental conditions. Modern software-defined radio techniques and reconfigurable architectures significantly enhance flexibility and adaptability [124,127].

Thirdly, hierarchical sensing architectures offer an efficient way to manage sensing and communication resources. Lower frequency systems provide wide-area awareness and trigger localized, high-resolution sensing through higher frequency systems when needed, optimizing both resource utilization and sensing fidelity [128,129].

Moreover, cross-band information fusion emerges as a powerful strategy to combine heterogeneous sensing data from multiple frequency bands. By integrating information from sub-6 GHz, mmWave, and THz bands, healthcare ISAC systems can achieve improved accuracy, robustness, and reliability, even in complex environments [127,130].

In healthcare deployments, hybrid operation across sub-6 GHz and higher bands has been shown to satisfy low-latency reliability requirements while reserving high-resolution capability via short, localized bursts; representative peer-reviewed reports include the following:

• In-hospital hotspot sensing/telepresence: A mixed-reality telesupervised ultrasound platform on a private 5G network installed with 4.7 GHz (coverage) and 28 GHz (hotspots) base stations with a local 5G core; measured throughput reached ≈ 1.45 Gbps down/147 Mbps up, and end-to-end latencies for concurrent streams were ∼49 ms (ultrasound image), ∼196 ms (HMD view), and ∼797 ms (360° camera), enabling real-time guidance [131].
• Air–ground continuity in emergency care: A helicopter EMS program used 2.6 + 4.9 GHz dual-frequency collaborative networking to maintain access from vertical take-off/landing to cruise, with network switching delays ≤ 30 ms, supporting continuous telemetry and consultation during flight [132].
• Regional emergency response: A mixed-frequency private 5G emergency system expanded the rescue radius from ∼5 km to ∼60 km and reduced cross-district response time from ∼60 min to $<20$ min, illustrating multi-band orchestration translating into operational KPIs [133].
• Private 5G SA in hospitals: A standalone private 5G build supported mobile ward rounds, remote ultrasound, and inter-hospital links with measured inter-site latency ∼14 ms and downlink/uplink rates of ∼790/91 Mbps, providing the substrate on which sub-6 coverage and high-band bursts can be scheduled per application phase [134].

Taken together, these deployments illustrate a pragmatic pattern for healthcare ISAC: anchor control and low-rate telemetry in the sub-6 GHz band for coverage and mobility and invoke higher band bursts for high-resolution imaging, telepresence, or dense sensing, thereby optimizing the latency–resolution trade-off in situ.

Looking ahead and consistent with these hybrid patterns, recent roadmaps for next-generation ISAC systems, including near-field ISAC, suggest that dynamic orchestration across bands will become even more crucial in healthcare scenarios demanding ultra-reliable and low-latency monitoring and intervention [135].

4.4.3. Challenges in Multi-Band Integration

While multi-band integration in healthcare ISAC systems offers significant advantages, it also introduces substantial technical and practical challenges that must be carefully addressed.

One major challenge lies in the hardware complexity. Supporting a wide range of frequency bands necessitates more sophisticated RF front ends, multiple antenna arrays, and broader bandwidth signal-processing capabilities, all of which increase system cost, size, and power consumption [136,137]. Particularly in healthcare environments, where device miniaturization and energy efficiency are critical, these hardware demands pose notable barriers.

Another important issue concerns the coordination overhead. Managing operations across multiple bands requires frequent signaling, control exchanges, and dynamic reconfiguration, which can introduce significant latency and overhead if not carefully optimized [136,138]. This is especially problematic for real-time healthcare applications, where low latency and high reliability are paramount.

Calibration and synchronization across different frequency bands represent additional technical hurdles. For coherent sensing and reliable cross-band data fusion, systems must maintain tight timing and phase alignment among bands, a task complicated by variations in propagation characteristics and hardware impairments [136]. Without precise synchronization, the benefits of multi-band integration could be undermined by errors and inconsistencies.

Finally, the lack of comprehensive standards for multi-band ISAC systems further complicates their deployment. In healthcare settings, where interoperability among heterogeneous devices and platforms is essential, the absence of unified protocols and certification frameworks creates integration risks and operational inefficiencies [4,139]. Thus, standardization efforts will play a crucial role in enabling widespread and seamless multi-band ISAC adoption in the healthcare sector.

4.5. Healthcare-Specific Considerations

The deployment of Integrated Sensing and Communication (ISAC) systems within healthcare environments necessitates careful attention to domain-specific safety, regulatory, clinical, and integration requirements.

4.5.1. Safety and Regulatory Aspects

First and foremost, ISAC systems must ensure compliance with electromagnetic exposure guidelines to safeguard patients and healthcare personnel. Depending on the operating frequency, different metrics apply; for example, the Specific Absorption Rate (SAR) is typically used for sub-6 GHz systems, while power-density limits are applied to higher frequency bands such as mmWave and THz [140]. These standards are critical for preventing adverse biological effects associated with wireless energy exposure.

Another vital consideration is medical-device compatibility. ISAC systems must not interfere with sensitive medical equipment, including pacemakers, infusion pumps, and patient monitors [141]. Careful frequency planning, adaptive power control, and comprehensive compatibility testing are necessary to mitigate potential risks.

In addition, medical sensing components of ISAC systems may fall under the regulatory definition of medical devices, thereby requiring approval from agencies such as the FDA in the United States or the EMA in Europe, depending on their intended functionality and claims [142]. Navigating these regulatory pathways is crucial for lawful deployment and clinical acceptance.

Finally, any ISAC system collecting patient data must comply with stringent healthcare privacy regulations, such as HIPAA in the U.S. and GDPR in the EU [143,144]. This compliance includes secure data handling, explicit consent management, and adherence to appropriate data-use limitations.

4.5.2. Clinical Validation Requirements

Beyond regulatory approval, the clinical adoption of ISAC technologies hinges on rigorous validation processes. Measurement accuracy must be demonstrated relative to established clinical standards across diverse patient demographics and environmental settings [145,146]. Simply achieving technical performance is insufficient; the clinical relevance and utility of ISAC-derived metrics must be confirmed through studies showing clear contributions to diagnosis, monitoring, or treatment decisions [147].

Furthermore, healthcare systems demand high robustness and reliability from any deployed technology. ISAC systems must characterize their performance limitations, error behaviors, and failure modes to ensure safe operation under varied clinical conditions [148].

For applications involving long-term patient monitoring, stability over time becomes paramount. Systems must demonstrate measurement consistency despite seasonal changes, device aging, or environmental variability, ensuring that long-term trends remain medically reliable [149].

4.5.3. Integration with Healthcare Information Systems

The true clinical value of ISAC systems will only be realized if they integrate seamlessly with existing healthcare information infrastructure. Integration with electronic health records (EHRs) is particularly important to ensure that ISAC-derived data contributes directly to clinical workflows and decision support [150].

Achieving this requires adherence to established interoperability standards such as HL7 FHIR, DICOM for imaging data, and IEEE 11073 for medical-device communications [151]. Compliance with these standards facilitates integration across heterogeneous systems and vendors, a necessity in complex healthcare environments.

Moreover, ISAC outputs must be incorporated into clinical workflows without introducing excessive cognitive or operational burden. Effective alerting mechanisms, intuitive data visualization, and seamless documentation processes are critical for clinician adoption [152].

Finally, healthcare ISAC systems must manage the vast data volumes they generate through scalable storage, efficient preprocessing, and responsible life-cycle management. Recent research on healthcare-oriented data-lake architectures demonstrates how distributed storage and metadata-driven governance can meet these requirements while supporting secondary uses such as research, quality improvement, and epidemiological surveillance [153].

4.5.4. Equity Perspective

Advanced ISAC links and data-hungry digital-twin engines risk widening the digital divide if the prerequisites for connectivity and compute are not met in low- and middle-income countries (LMICs). A 2024 World Health Organization policy analysis observes that LMICs still exploit barely 5% of their health data, citing chronic shortfalls in broadband, power, and cloud infrastructure that disproportionately burden already resource-constrained systems [154].

The World Economic Forum likewise identifies unreliable Internet access and electricity supply as the single most significant obstacle to scaling digital health solutions in these settings [155]. Even when 5G roll-outs are planned, mmWave ISAC coverage depends on dense small-cell grids that “drive up the cost of deploying mmWave networks at scale”, rendering such deployments impractical for rural or peri-urban clinics [156].

Bridging these socioeconomic and infrastructural gaps—through investment in affordable sub-6 GHz connectivity, shared cloud or edge resources, and open-source modeling platforms—will be essential to ensure that ISAC-enabled twins benefit global health rather than entrenching existing inequities.

4.6. Future Trends and Research Directions

The future of Integrated Sensing and Communication (ISAC) technology in healthcare is being shaped by rapid advancements in enabling technologies, the emergence of novel applications, and evolving efforts in standardization.

4.6.1. Technological Advancements

Several technological trends are poised to dramatically enhance the capabilities and accessibility of ISAC systems. Advances in semiconductor technologies—for example, silicon photonics, compound semiconductors, and integrated photonic platforms—promise to improve performance, energy efficiency, and cost effectiveness across all operating frequency bands [157].

Artificial intelligence and machine learning are increasingly being incorporated into ISAC signal-processing pipelines. AI-enhanced algorithms enable improved sensing accuracy, robust interference mitigation, and adaptive operation in complex, dynamic healthcare environments [158,159].

Moreover, innovations in energy harvesting and the development of ultra-low-power electronics are expected to support long-lifetime or even battery-free ISAC devices—crucial for continuous patient monitoring and minimally invasive sensing [160].

Finally, metamaterials and advanced antenna designs are anticipated to improve ISAC system performance. Such technologies enhance directivity, sensitivity, and miniaturization, enabling new deployment scenarios within constrained or sensitive healthcare environments [161].

4.6.2. Emerging Applications

The expanding technical capabilities of ISAC systems are enabling new healthcare applications. One promising direction is the development of implantable and ingestible ISAC devices that can monitor internal physiological parameters and wirelessly communicate data to external systems [162].

Furthermore, the use of advanced THz sensing holds the potential to enable non-invasive monitoring at the molecular or cellular level. Such capabilities would support early disease detection and precision medicine initiatives by providing detailed, real-time physiological information [163].

Another promising frontier lies in the development of closed-loop therapeutic systems. These systems could not only monitor a patient’s condition but also autonomously deliver therapeutic interventions based on real-time ISAC sensing data, creating highly responsive, personalized treatment models [164].

At the broader environmental level, pervasive ISAC sensing could contribute to the creation of ambient-intelligence systems in healthcare facilities. These systems would enhance patient safety, optimize care delivery, and anticipate patient needs without requiring direct user interaction [165].

4.6.3. Standardization Efforts

Standardization efforts will play a pivotal role in ensuring the interoperability, security, and clinical acceptance of healthcare ISAC systems. New standards specific to ISAC technologies are already emerging—for example, the draft IEEE P802.11bf amendment for Wi-Fi sensing—which provides a structured framework for interoperability and performance validation [166].

To realize full clinical integration, existing healthcare data standards must also evolve to accommodate ISAC-derived data streams. Research prototypes that augment HL7 FHIR resources for real-time physiological sensing illustrate a viable path forward [167].

Due to the pervasive and sensitive nature of ISAC sensing, the establishment of rigorous security and privacy standards is of critical importance. The ongoing IEEE P2933 initiative, addressing Trust, Identity, Privacy, Protection, Safety, and Security (TIPPSS) for clinical IoT devices, illustrates the potential trajectory of such standardization efforts [168].

Finally, standardized validation and benchmarking methodologies are needed for objective evaluation of ISAC technologies prior to regulatory approval and clinical adoption. Recent work on open benchmarks and evaluation metrics for joint communication–sensing prototypes provides an early foundation [169].

In summary, the convergence of technological advances, emerging applications, and robust standardization may catalyze the next wave of ISAC innovation in healthcare and substantially improve patient monitoring, diagnosis, and care delivery.

4.6.4. Standardization Hurdles for ISAC-Enabled Healthcare Devices

Successful clinical uptake of ISAC-enabled sensors will depend as much on mature, healthcare-oriented standards as on algorithmic progress. However, the present landscape shows sizable gaps. Vocabulary project IEEE P3384 is still at the draft-ballot stage and covers only generic terms, while no single specification harmonizes operation across sub-6 GHz, mmWave, and THz medical bands [7,136,170]. ETSI GR ISC 001 and 3GPP TR 22.837 catalog healthcare use cases but explicitly defer conformance profiles and test procedures to future work, leaving vendors without a certification target [114,171].

Regulatory alignment is likewise immature. The FDA’s 2023 final guidance on medical-device cybersecurity makes no reference to joint RF sensing, and recent radar-based vital-sign studies confirm that applicants must navigate an ad hoc predicate path [172,173].

Interoperability frameworks such as HL7 FHIR still lack reference mappings for high-rate ISAC telemetry, a challenge repeatedly noted in systematic reviews of FHIR deployments [174].

Security and privacy standardization also lags. IEEE P2933 (TIPPSS) [168] proposes a trust framework for clinical IoT, but its current draft omits ISAC-specific attack surfaces such as covert localization and waveform spoofing, while academic surveys on IoMT security continue to flag the absence of protocol-level counter-measures for joint sensing links.

Until these gaps are closed, commercial roll-out of ISAC-based healthcare devices is likely to remain fragmented and slow.

4.6.5. Security Issues and Mitigation Strategies in 6G ISAC Systems

The dual-functional nature of integrated sensing and communication (ISAC) in 6G exposes both communication data and high-resolution sensing returns to adversaries, creating attack surfaces that traditional link-layer encryption alone cannot fully mitigate [175]. In particular, illegitimate receivers can exploit the transmitted dual-use waveform to infer environmental or target information (“sensing eavesdropping”) while also attempting to decode confidential messages, which forces a careful balance between sensing performance and secrecy [176].

A primary risk is joint eavesdropping on the sensing and communication channels, where the same waveform that supports target estimation also leaks information to malicious observers [176]. Artificial noise (AN)-aided beamforming and secure precoding have therefore been formulated to maximize the legitimate radar/communication utility subject to explicit information-leakage constraints at sensing eavesdroppers, with simulations showing improved mutual information for the legitimate receiver while bounding the eavesdropper’s sensing capability [175,176].

Beyond AN, reconfigurable radio environments help reshape the wire-tap channel: phase-coupled intelligent omni-surfaces (IOSs, a generalization of RISs) can partition space into communication and sensing regions, form virtual line-of-sight links for targets, and simultaneously enhance physical-layer secrecy through joint optimization of communication/sensing beamformers and surface phases [177]. Recent studies also show that a malicious RISscan spoof echoes and bias delay/Doppler/angle estimates, underscoring the need for spoof-resilient signal designs and channel checks in practical ISAC deployments [178].

Although the highly directional nature of THz beams suggests inherent security, reflections and beam-misalignment events still open side lobes that adversaries can exploit. Measurements at 0.3–0.5 THz reveal that an eavesdropper placed within a first-order reflection path can recover up to 40% of transmitted symbols without being in the direct line of sight [101]. Proposed mitigations include random phase perturbation of pilot tones, rapid beam hopping combined with AN masking [179], and RIS-assisted null steering that adapts to specular reflectors in real time [180].

RIS/IOS-assisted PLS can reduce information leakage while preserving sensing fidelity via environment shaping, complementing AN and cooperative jamming [177,180]. In parallel, rate-splitting multiple access (RSMA) offers extra degrees of freedom for secure ISAC by flexibly splitting messages and managing interference; recent tutorials report that RSMA improves secrecy/throughput trade-offs and is a promising access strategy for ISAC [181,182].

The main challenge is joint optimization of sensing, communication, and security objectives under the latency and compute constraints typical of edge–cloud pipelines. Practical deployments will likely combine AN/RIS-based PLS with lightweight authentication and privacy-preserving processing at the edge; end-to-end evaluations that co-optimize secrecy metrics and sensing accuracy remain an open research need [175].

5. Digital-Twin Technology in Healthcare

This section explores the application of digital-twin technology in healthcare contexts, examining the fundamental concepts, implementation approaches, and current applications across different healthcare domains. We analyze how digital twins are transforming patient care, healthcare facility operations, and medical device management, with a particular focus on their potential integration with ISAC technology.

5.1. Patient-Centric Digital Twins

5.1.1. Conceptual Framework

Patient-centric digital twins are virtual models that replicate the physiological, behavioral, and contextual characteristics of individual patients, supporting personalized healthcare strategies [26,183,184]. These digital twins operate across multiple scales, from molecular and organ-specific representations to whole-body systems.

A critical element of this framework is physiological modeling, where mathematical and computational techniques simulate biological structures and functions. These range from biophysical simulations, such as cardiac electrophysiology models, to phenomenological models that capture observable relationships such as glucose–insulin dynamics [25,185].

An essential complement to modeling is the data integration layer, which collects and harmonizes heterogeneous patient data from electronic health records, imaging modalities, wearable devices, implantables, and environmental sensors [183,184]. Addressing challenges in data heterogeneity, semantic alignment, and temporal synchronization is critical for maintaining an accurate, dynamic digital twin.

Personalization mechanisms adapt generic models to reflect individual patient characteristics through parameter estimation, structural adaptation, and machine learning techniques [26,186]. This personalization ensures that predictions and simulations are clinically meaningful.

Furthermore, predictive analytics extend digital twins beyond descriptive models by forecasting future physiological states, disease trajectories, or treatment outcomes [25,187]. Visualization and interaction interfaces translate these complex insights into actionable clinical decision support, while feedback mechanisms allow digital twins to issue recommendations or even trigger automated interventions, creating a closed-loop system between the virtual and physical entities [26,188].

5.1.2. Implementation Approaches

Several approaches have emerged for the construction of patient-centric digital twins.

Physics-based modeling relies on fundamental biological principles to simulate physiological processes. For example, finite element methods are applied in cardiac mechanics, while computational fluid dynamics model vascular flow [25,185]. Although explanatory power is high, these methods require extensive patient-specific data and computational resources.

In contrast, statistical and machine learning-based methods model physiological behavior directly from empirical data without strict reliance on mechanistic understanding [183,186]. Regression techniques, neural networks, and time-series models are common but can struggle with extrapolation beyond trained conditions.

Hybrid modeling combines mechanistic and data-driven approaches to leverage the advantages of both paradigms [25,187]. For instance, a cardiac twin might use a physics-based structure while refining parameters through machine learning optimization.

Multi-scale integration frameworks further expand modeling fidelity by linking molecular, cellular, organ, and systemic models, although this increases the complexity of parameter tuning and system coupling [26,186].

Federated implementations distribute different components of the digital twin across multiple secure systems, preserving privacy by keeping sensitive data local while maintaining logical integration [188,189].

5.1.3. Current Applications and Case Studies

Digital twin technology has moved beyond proofs of concept and is now being evaluated in prospective or real-world studies across multiple specialties [190].

In cardiovascular medicine, digital twins that integrate cardiac imaging, electrophysiology, and biophysical simulation have been shown to personalize pacemaker programming and valve repair [191]. A review reported in [192] summarizes how such twins improve risk stratification, clinical-trial design, and therapy selection, while a machine learning study reported in [193] describes population-scale heart-twin cohorts for decision support.

In diabetes management, a twelve-month real-world study of the Twin Health platform that followed 1985 adults with type 2 diabetes reported a mean HbA_1c reduction of 1.8 % and a 4.8 kg weight loss while reducing medication burden [194].

The Cleveland Clinic is running a randomized trial of an AI-enabled whole-body digital-twin system that records up to 3000 variables per day and dynamically balances 87 essential nutrients [195]. For type 1 diabetes, an evaluative case study [196] concludes that twin-based closed-loop “artificial pancreas” controllers improve time in range and reduce hypoglycemic and hyperglycemic excursions.

In oncology, the FarrSight^®-Twin platform replicated phase-II/III trial outcomes and improved response-rate prediction in breast, pancreatic, and ovarian cancers [197]. A study [35] perspective details efforts to create immune-oncology twins that generate virtual patients for therapy optimization.

Neurological applications include a brain-twin framework that fuses neuro-imaging, genomics, and clinical outcomes to model tumor impact, as reported in [198], and a comprehensive review [199] that maps a translational pathway from in silico cortex models to personalized neuromodulation.

Orthopedic surgery is another emerging area: a study [200] shows that fracture-repair twins predict implant stress distribution and residual deformity to aid in surgeon planning, and a review reported in [201] summarizes musculoskeletal twins for gait analysis, rehabilitation, and implant design.

5.2. Healthcare Facility and System Digital Twins

Digital twins at the healthcare facility and system levels provide tools for simulating, optimizing, and monitoring the delivery of healthcare services across organizational scales. This section presents the conceptual framework, implementation approaches, and current applications of these digital twins, emphasizing their growing importance for healthcare resilience, efficiency, and patient safety.

5.2.1. Conceptual Framework

Healthcare facility and system digital twins construct dynamic virtual representations of physical environments, operational processes, and resource flows [26,183,187]. At the foundation lies spatial representation, encompassing three-dimensional modeling of healthcare infrastructure, from basic floor plans to detailed Building Information Models (BIMs) enriched with equipment and environmental parameters [22,184].

Resource management components model the availability, utilization, and allocation of critical assets such as staff, beds, medical equipment, and supplies. These modules enable dynamic optimization of resource deployment in response to fluctuating demands [25].

Environmental monitoring layers continuously gather data on temperature, humidity, air quality, and occupancy patterns, supporting infection control, patient comfort, and energy optimization strategies [26,184].

Operational intelligence modules transform raw data streams into actionable insights through performance dashboards, predictive analytics, and what-if scenario simulations, enhancing decision-making at both strategic and tactical levels [25,187].

Finally, robust integration interfaces connect the digital twin with electronic health records, hospital information systems, building management platforms, and medical device networks, ensuring alignment with the evolving physical environment [22,184].

5.2.2. Implementation Approaches

Implementing healthcare facility digital twins typically leverages multiple complementary modeling strategies.

Discrete Event Simulation (DES) models healthcare processes as sequences of discrete operations, ideal for analyzing patient flow, resource bottlenecks, and queuing dynamics in settings such as emergency departments and surgical suites [183,187].

Agent-Based Modeling (ABM) captures emergent behaviors by simulating interactions among autonomous agents such as patients, staff, and equipment. ABM is particularly valuable for the modeling of infection transmission, adaptive behavior, and complex human–system interactions [26,202].

System Dynamics (SD) models healthcare organizations as interconnected stocks, flows, and feedback loops evolving over time, supporting analysis of long-term trends in resource availability, workforce planning, and population health [203].

The integration of Building Information Modeling (BIM) couples physical infrastructure data with operational simulations, enabling facility design optimization, energy management, and maintenance scheduling [22,184].

Internet of Things (IoT) platforms provide real-time data streams from distributed sensors throughout healthcare environments, enabling digital twins to maintain continuous, accurate representations of physical systems [26,188].

Hybrid approaches increasingly combine multiple modeling techniques. For example, a hospital-wide digital twin may integrate DES for patient flow, ABM for infection modeling, SD for capacity planning, and BIM for infrastructure management to provide a comprehensive operational picture [184,187].

5.2.3. Current Applications and Case Studies

Healthcare facility and system digital twins are being deployed in a range of practical applications.

Hospital capacity management represents a major area of deployment. During the COVID-19 pandemic, digital twins were used to simulate surge scenarios, optimize ICU bed allocations, and manage emergency resource scaling, integrating epidemiological models with patient flow and facility constraints [26,187].

Emergency department optimization employs digital twins to model patient arrivals, triage processes, treatment flows, and discharge pathways. Real-time implementations allow for proactive congestion management, reduced waiting times, and optimized staff utilization [204,205].

Operating room management benefits from digital twins that simulate surgical scheduling, procedure durations, turnover times, and resource requirements. These models enable dynamic adjustments to operating-room utilization strategies, minimizing delays and maximizing throughput, even under uncertainty [25,206].

Infection control has emerged as a vital application, with digital twins modeling airflow patterns, contact networks, and facility layouts to predict and mitigate nosocomial transmission risks. Such models informed facility redesigns and operational protocols during the COVID-19 pandemic [26,207].

Energy management and environmental sustainability initiatives are supported by digital twins that optimize HVAC operations, lighting, and occupancy-driven energy usage, improving patient comfort while reducing operational costs [184,187].

Together, these applications demonstrate the transformative potential of healthcare facility and system digital twins when deployed with robust modeling, real-time data integration, and advanced predictive analytics.

5.3. Medical Device and Equipment Digital Twins

Digital twins of medical devices and healthcare equipment are transforming how critical technologies are monitored, maintained, and optimized throughout their life cycle. By creating dynamic virtual counterparts, these systems enhance device reliability, operational efficiency, and patient safety across healthcare environments [25,208,209].

5.3.1. Conceptual Framework

The conceptual foundation for medical device digital twins consists of several inter-related components designed to support continuous life-cycle management.

Device modeling serves as the core, employing computational representations of mechanical, electrical, thermal, and control systems [208,210]. These models enable simulations under various operational and environmental conditions, supporting failure prediction and performance optimization.

Performance monitoring layers integrate real-time operational data, tracking usage patterns, environmental conditions, and system-health indicators. Machine learning analytics detect degradation trends, anomalies, or early indicators of potential failures [211,212].

Predictive maintenance strategies leverage historical and real-time data to anticipate maintenance needs and schedule interventions proactively, minimizing unplanned downtime and extending equipment lifespans [211].

Configuration management modules maintain detailed records of hardware configurations, software versions, calibration statuses, and maintenance activities, ensuring regulatory compliance and aiding in fault diagnosis [184,213].

Usage analytics examine how devices are utilized within clinical workflows, identifying patterns that inform improvements in device design, user training, and deployment strategies [26,211].

Finally, digital thread integration ensures continuity across the device life cycle, linking design, manufacturing, clinical operation, and end-of-life management into a seamless feedback system for continuous improvement [25,213].

5.3.2. Implementation Approaches

Implementing medical device digital twins involves several complementary methodologies.

Physics-based modeling utilizes engineering principles and finite element methods to simulate device structure and behavior under different stressors [208,210]. These models offer explanatory power critical for fault diagnosis and optimization.

Data-driven modeling captures operational behavior patterns through statistical models and machine learning algorithms, enabling anomaly detection and predictive insights without detailed physical knowledge [211,212].

Edge computing platforms allow real-time data processing to be performed locally on the device, reducing latency and mitigating reliance on continuous network connectivity—particularly valuable for mobile and remote healthcare applications [184,214].

Cloud-based Digital-Twin-as-a-Service (DTaaS) solutions provide scalable infrastructure for storing, analyzing, and visualizing twin data while integrating with broader healthcare information systems [215].

Augmented-reality interfaces overlay maintenance instructions, diagnostic information, or operational alerts directly onto physical devices, boosting technician effectiveness [183].

Blockchain-secured twins safeguard data integrity and traceability—essential in regulated healthcare environments [188].

5.3.3. Current Applications and Case Studies

Several case studies illustrate the real-world impact of medical device digital twins in healthcare operations.

In medical imaging, digital twins for MRI, CT, and ultrasound equipment continuously monitor critical components such as cryogen levels, gradient systems, and RF amplifiers, predicting maintenance needs and ensuring imaging quality [25,26,216].

Infusion pumps are increasingly managed through digital twins that track medication delivery accuracy, monitor battery health, and analyze usage trends. These capabilities improve fleet management, maintenance scheduling, and medication safety [211,217].

Ventilator digital twins, particularly during the COVID-19 pandemic, enabled remote performance monitoring, optimization of gas delivery accuracy, and dynamic maintenance management across hospital networks [187,218].

In laboratory environments, analyzers and automation systems are digitally twinned to optimize reagent use, maintain analytical quality, and enhance laboratory workflow efficiency, supporting higher throughput and reliability [219].

Implantable devices such as pacemakers, defibrillators, and neurostimulators employ digital-twin models to monitor device–tissue interactions, predict battery depletion, and personalize therapy [26,220].

Taken together, these applications show a clear shift from reactive maintenance to proactive, predictive, and personalized management across patients, devices, and facilities.

To aid in navigation,

Table 8. Applications and case studies in healthcare DTs.

DT Scope	Specific Applications/Cases	Reported Evidence/Key Outcomes
Patient-centric digital twins—see Section 5.1.3
Cardiovascular (patient-level DT)	Digital twins that integrate cardiac imaging, electrophysiology, and biophysical simulation	Personalization of pacemaker programming and therapy selection; population-scale heart-twin cohorts discussed
Diabetes management	Real-world study of the Twin Health platform	Twelve-month real-world study reporting HbA1c reduction (mean decrease reported)
Oncology	FarrSight®-Twin-type oncology twins that generate virtual patients for therapy optimization	Replication of phase-II/III trial outcomes and improved therapy optimization outlined
Neurology	Brain-twin framework fusing neuro-imaging, genetics, and neural models	Pathway from in silico cortex models to personalized neuromodulation
Orthopedics	Musculoskeletal/orthopedic twins for gait analysis, rehabilitation, and implant design	Case-study evidence showing applicability to surgeon planning/implant design
Healthcare facility/system digital twins—see Section 5.2.3
Emergency department flow	Twins model patient arrivals, triage processes, treatment flows, and discharge pathways	Surge-scenario simulation; optimized staff utilization reported
Operating room management	Twins simulate surgical scheduling, procedure durations, turnover times, and resource needs	Robust scheduling and throughput under uncertainty described
Infection control	Twins model airflow patterns, contact networks, and facility layouts	Prediction/mitigation of nosocomial transmission discussed
Energy/sustainability	Twins optimize HVAC operations, lighting, and related building systems	Energy-efficiency gains and operational-cost impacts outlined
Medical device and equipment digital twins—see Section 5.3.3
Imaging/ultrasound systems	Digital twins monitor critical components (e.g., transducers)	Predictive maintenance/reliability improvements described
Infusion/delivery devices	Twins track medication delivery accuracy and battery/pressure health	Remote performance monitoring and anomaly detection reported
Ventilation/gas delivery	Twins support optimization of gas delivery	Remote monitoring and performance optimization described (including pandemic-period operations)
Clinical lab/analyzers	Laboratory automation systems digitally twinned	Reagent-use optimization and maintenance of analytical quality discussed
Implantable/therapeutic devices	Pacemakers, defibrillators, and neurostimulators with DT-enabled monitoring	Shift from reactive to proactive maintenance/therapy management noted

compiles the patient-, device-, and facility-level digital-twin use cases and provides § links to the corresponding “Current Applications and Case Studies” blocks.

6. ISAC–DT Integration Challenges and Opportunities

The integration of digital-twin technology into healthcare systems presents both formidable challenges and promising opportunities. This section discusses technical, clinical, and organizational aspects that influence the success of digital twin adoption.

6.1. Clinical Validation Status of DT–ISAC

At present, peer-reviewed real-world evidence that an end-to-end digital-twin–ISAC stack improves patient outcomes (e.g., mortality, unplanned transfer to the intensive care unit (ICU), or 30-day readmission) or hospital workflow key performance indicators (KPIs) (e.g., emergency department (ED) length of stay or bed turnover) remains very limited; most published studies report simulation outcomes or process-only metrics rather than pre/post or controlled clinical endpoints [63].

Nevertheless, adjacent components of the stack have accumulating clinical evidence: (i) contactless sensing accuracy in wards (e.g., radar respiratory rate) [221], (ii) RTLS/IPS-derived process measures in live ED operations [222], and (iii) hospital-flow digital twins demonstrating reduced waiting time/LOS in simulation or pilots [63].

Table 9. Evidence relevant to clinical validation of DT–ISAC: numeric endpoints from peer-reviewed studies.

Use Case and Setting	Sample/Period and Design	Key Numeric Findings
Contactless respiratory rate (UWB radar) Setting: University hospital emergency ward (adults at rest)	Sample/period: $n=32$ Duration: median of 42 min Design: Method comparison vs. reference (Nox T3s)	Bias of $0.0$–$0.1$ breaths/min 95% limits of agreement $\approx [-1.1,1.2]$ breaths/min Trend concordance of $96\%$ No missed/false clinical alarms No gap $>5$ min [221]
ED staff indoor positioning (UWB IPS) Setting: French Level-1 Emergency Department	Sample/period: 46 days Tags: 27 Design: Observational tracking + ML classifier	Doctors’ care-related time share of 26–$39\%$ Triage/ICU nurses $\sim 50\%$ Nurses’ walking distance rises with occupancy Job-category classifier accuracy of $96\%$ [222]
Continuous monitoring and outcomes (capnography/oximetry) Setting: Med-surg wards (PCA opioids)	Period: 2012–2015 Design: Before/after (capnography rollout)	Opioid-induced respiratory rescue incidence $0.4\%\to 0.2\%$ Transfers to higher level of care reduced by $79\%$ (7.6→1.6 per month) [223]
Continuous pulse-ox surveillance and rescue/ICU transfer Setting: Postoperative units	Period: 11 mo before/10 mo after Design: Before/after (surveillance with pager alerts)	Rescue events $3.4\to 1.2$ per 1000 discharges ICU transfers $5.6\to 2.9$ per 1000 patient-days (Comparison units: no change) [224]
DT for flow (simulation/pilots) Setting: Hospital ED/wards (various)	Design: Simulation/pilot reports	DT models validated against site data Scenarios report reduced waiting time and LOS (Simulation evidence; limited real-world endpoints) [63]

Abbreviations: ED, Emergency Department; ICU, Intensive Care Unit; UWB, Ultra-Wideband; IPS, Indoor Positioning System; LOS, Length of Stay.

summarizes numeric indicators from representative studies, and Box 1 separates vendor case reports.

Box 1. Vendor case reports (gray literature)

GE HealthCare Command Center: capacity/surge digital twin; qualitative claims on utilization and planning (Link: https://www.gehealthcare.com/insights/article/making-informed-capacity-decisions-with-digital-twin-technology, accessed on 9 September 2025). AnyLogic NHS hospital twin: whole-journey flow modeling; scenario projections on queues/LOS (Link: https://www.anylogic.com/resources/case-studies/hospital-digital-twin-to-improve-operations-and-enhance-patient-experience/, accessed on 9 September 2025).

6.2. Recommended Evaluation Protocol for DT–ISAC in Hospitals

To close this gap, we outline evaluation protocols and a minimum reporting checklist tailored to DT–ISAC deployments in hospitals. When feasible, cluster randomized or stepped-wedge cluster designs provide robust causal identification and accommodate phased roll-outs [225,226].

Otherwise, interrupted time series (ITS) with seasonality controls and segmented regression offer a strong quasi-experimental alternative for service-wide changes [227,228]. For AI-enabled decision support, reporting should follow CONSORT-AI/SPIRIT-AI extensions, and quality-improvement or service evaluations should adopt SQUIRE 2.0 [229,230,231]. Where patient-level risk models or digital-twin surrogates are evaluated, TRIPOD and PROBAST help ensure transparent reporting and risk-of-bias assessment [232,233].

Co-primary endpoints should capture both patient outcomes and workflow KPIs, with secondary metrics for sensing fidelity, twin calibration, closed-loop timeliness, safety, and cost to serve:

• Design: baseline ≥ 6–12 months; phased roll-out; risk adjustment; seasonality [227,228].
• ISAC fidelity: MAE/LoA vs. reference; valid coverage (% time usable); alert yield (PPV, alarms/bed-day).
• DT fidelity: fit to arrivals/occupancy/LOS (MAPE); recalibration cadence; external validation across sites [232,233].
• Closed loop: time to detection/decision/action; execution %; override rates.
• KPIs: ED/ward LOS (median/IQR and upper tails), boarding hours, rescue/ICU transfer per 1000 patient-days, code events, staff walking distance/time share.

For ease of implementation and peer review,

Table 10. Minimum reporting checklist for DT–ISAC clinical validation.

Item	Requirement
Design	Pre-intervention baseline duration; roll-out schedule (cluster/stepped wedge) Risk-adjustment covariates/methods; seasonality controls; ITS parameters (if applicable)
Population	Inclusion/exclusion criteria; case-mix descriptors Sites and periods
Intervention	DT–ISAC architecture and deployment mode (silent vs. live) Sensing/communication budgets; edge–cloud placement
Comparator	Standard care or historical controls Randomization/sequence rules
Outcomes	Clinical and workflow endpoints; precise definitions and observation windows
Metrics	AUROC/AUPRC; calibration; PPV@alert $\Delta$KPI with confidence intervals
Operational	Alert rate per bed day; acknowledgment (ack) latency Human-in-the-loop protocol and over-ride policy
Generalization	External validation across sites/subgroups Drift monitoring plan
Safety	RF exposure; alarm fatigue Fail-safe behavior; over-ride rules
Governance	Data protection and audit Change management; regulatory pathway
Economics	Cost-to-serve; incremental staffing ROI sensitivity analysis

Abbreviations: AUROC, area under the receiver operating characteristic curve; AUPRC, area under the precision–recall curve; PPV, positive predictive value; ITS, interrupted time series; KPI, key performance indicator; CI, confidence interval; ROI, return on investment; RF, radio frequency; ack, acknowledgment.

consolidates the same minimum reporting items into a compact checklist aligned with the headings above (Design, ISAC fidelity, DT fidelity, Closed loop, and KPIs). Each row specifies what must be reported (definitions, units, and time windows where relevant) so that studies can be reproduced and compared across sites. Abbreviations used in the table are listed directly beneath it for clarity.

6.3. Technical Integration Challenges

Integrating digital twins with ISAC-derived telemetry and links introduces challenges that are not purely “DT” problems but arise at the joint sensing–communication–computation boundary. Below, we frame each issue as a coupling between ISAC signal/link properties and twin pipelines, with pointers to the relevant technical trade-offs elsewhere in the paper.

• Data integration and freshness (ISAC → DT ingestion).
  • Healthcare digital twins must harmonize heterogeneous sources—electronic health records (EHRs), medical devices, imaging, and wearables—and contact-free ISAC streams (e.g., respiration/motion and indoor localization). Variations in formats, sampling cadences, semantics, and quality require ingestion, semantic mapping, and QA pipelines that preserve data freshness for deterioration forecasting [25,26,184,185,234,235]. In practice, this means aligning ISAC sensing/link cadences with DT update policies and recording provenance so that twin states reflect current bedside context (see interoperability mechanisms in Section 4.5.3 and band-driven sensing constraints in Section 4.1.3, Section 4.2.3 and Section 4.3.3).
• Compute and placement under end-to-end latency budgets.
  • Detailed physiological modeling (multi-scale, fluid-dynamics-based, or neuro-cardiac) is computationally intensive. When the twin closes the loop on ISAC (e.g., adapting beam direction, bandwidth, or sampling), end-to-end delay spans sensing, the wireless link, edge/cloud compute, and actuation. Balancing model fidelity with latency/throughput is therefore an integration constraint, not only a modeling choice [208,210,214,236]. Section 4.1.3, Section 4.2.3 and Section 4.3.3 detail band-level trade-offs that drive feasible sampling/update rates, while this section specifies how placement across edge–cloud systems affects closed-loop timeliness.
• Uncertainty and credibility with link-/sensor-induced variability.
  • Twin predictions inherit uncertainty from physiologic variability, measurement noise, packet loss/reordering, and model approximations [187,237]. With ISAC in the loop, changes in link quality or sensing geometry can shift observation models and calibration, so uncertainty must be quantified and propagated with awareness of sensing/link conditions and communicated to clinicians to avoid overconfidence. This motivates VVUQ routines that condition on ISAC telemetry quality and cadence (see calibration and validation considerations in Section 6.6 and Section 4.5.2).
• Interoperability and provenance for synchronized updates.
  • Despite the adoption of Fast Healthcare Interoperability Resources (FHIR), achieving consistent semantics and lineage across EHR/device data and ISAC streams remains difficult [184,235]. ISAC features (e.g., contact-free vital signs and localization) must be mapped to FHIR-compatible resources with timestamps/identifiers that the twin can trust for synchronization and replay. Practical pathways and QA tool chains are discussed in Section 4.5.3.
• Privacy and security with sensing–communication co-design.
  • Digital twins carry sensitive patient data and, thus, require cryptographic protections, fine-grained access control, and secure architectures compliant with HIPAA/GDPR [26,188]. ISAC-specific considerations (e.g., exposure/power limits, side-channel/leakage surfaces, and on-link protections) must be addressed alongside application-layer measure (see safety/regulatory and security discussions in Section 4.5.1 and Section 4.6.5). Integrating these constraints with the twin’s update cadence helps ensure that privacy and safety are maintained without sacrificing timeliness.

Taken together, these challenges are integration constraints at the ISAC–DT boundary: data freshness and semantic alignment (ingestion), latency-/throughput-aware placement (closed loop), credibility under variable sensing/link quality (VVUQ), interoperable synchronization with provenance (FHIR/EHR), and privacy/safety/security that respects band and power limits (Section 4.1.3, Section 4.2.3 and Section 4.3.3, Section 4.5.1, Section 4.5.3 and Section 4.6.5).

6.4. Clinical Integration Opportunities

Despite technical challenges, pairing digital twins with ISAC sensing and communication enables opportunities that are actionable in clinical workflows [26,238].

• Clinical decision support with ISAC-informed context.
  • Patient-specific twins that simulate interventions and predict outcomes become more actionable when continuously fed by ISAC telemetry (e.g., contact-free respiration/motion and indoor localization), which keeps model states current; recommended actions can, in turn, adapt beam direction, bandwidth, or sampling in a closed loop [214,239]. Latency/throughput considerations for such loops are discussed in Section 6.3, with band-driven sensing constraints presented in Section 4.1.3, Section 4.2.3 and Section 4.3.3.
• Continuous monitoring and early warning.
  • ISAC-enabled contact-free streams can reveal subtle physiologic changes before overt signs, enabling proactive interventions [25,187]. Operating points depend on sensing band and link budget (mmWave/THz for resolution; sub-6 GHz for coverage/penetration) and on update policies that preserve data freshness for twin ingestion (see Section 4.1.3, Section 4.2.3 and Section 4.3.3 and Section 6.3).
• Personalized treatment planning with richer observability.
  • ISAC-derived context (mobility, respiration patterns, and environmental dynamics) improves parameterization and boundary conditions for individualized models, supporting tailored therapy while controlling risk [185,208]. Credibility hinges on calibration and validation procedures that account for sensing/link variability (see Section 4.5.2 and Section 6.6).
• In silico testing and closed-loop execution.
  • Virtual testing of interventions prior to clinical execution enhances safety and optimizes strategies, particularly in high-risk settings [210,240]. When deployed with ISAC, recommended actions can be executed with closed-loop adaptations of the sensing/communication stack, subject to safety/regulatory and security constraints (see Section 4.5.1 and Section 4.6.5).
• Remote and distributed care.
  • For home and rural scenarios, sub-6 GHz links improve range and reliability so that twins remain synchronized under patchy connectivity, while secure pipelines protect patient data [26,188]. Interoperability and provenance for mixed EHR/device/ISAC streams are covered in Section 4.5.3, and end-to-end timeliness considerations are discussed in Section 6.3.

6.5. Organizational and Workflow Integration

Beyond technical and clinical considerations, organizational factors determine whether ISAC-enabled digital twins move from prototypes into routine care [184,241].

• Workflow integration and alert routing.
  • Twin outputs must be embedded into existing pathways (rounds, escalation, and discharge) without undue burden and with clear ownership for closed-loop actions that may reconfigure ISAC sensing/communication (e.g., beam direction, bandwidth, and sampling) [187,242]. Practical workflow fit depends on sensing-band and placement choices for contact-free monitoring and localization (see band trade-offs in Section 4.1.3, Section 4.2.3 and Section 4.3.3 and latency- and throughput-aware pipelines in Section 6.3).
• Training and competency.
  • Clinicians and biomedical engineers require training to interpret ISAC-derived telemetry (contact-free respiration/motion and indoor localization), understand twin assumptions and uncertainty, and incorporate outputs into decision-making [25,26]. Competency also includes provenance-aware data handling and interoperability (FHIR/EHR) so that updates ingested by the twin are trusted and timely (see Section 4.5.2, Section 4.5.3 and Section 6.6).
• Governance, safety, and security.
  • Effective governance balances clinical leadership and technical expertise to manage data use, system integrity, and ethics [208,243,244]. For ISAC–DT, this explicitly includes RF exposure/power limits and spectrum coexistence in clinical areas, along with privacy and security controls aligned to HIPAA/GDPR. Related considerations are detailed in Section 4.5.1 and Section 4.6.5.
• Change management and human in the loop.
  • Introducing ISAC–DT loops requires strategies that address resistance, build trust, and clarify the boundary between automation and clinician oversight [185,210]. Alarm policies, escalation paths, and fallback behaviors should be co-designed so that latency budgets and alert routing remain compatible with staffing patterns (see integration aspects in Section 6.3).
• Outcome measurement and continuous improvement.
  • Organizations should track clinical, operational, and economic endpoints to assess the impact of ISAC–DT deployments, including patient outcomes (length of stay, complications, and satisfaction), operational flow (throughput, and time to intervention), and technical indicators that reflect ISAC coupling (update freshness, end-to-end latency, and alert precision/recall) [26,63]. Evaluation frameworks and validation considerations are discussed in Section 4.5.2.

6.6. Model Calibration Challenges for Physiological Digital Twins

Model calibration remains one of the most critical barriers to realizing truly personalized and clinically actionable physiological digital twins. Establishing credibility requires rigorous verification, validation, and uncertainty quantification (VVUQ), yet high-dimensional parameter spaces and sparse, noisy patient data make formal validation and error propagation extremely challenging.

These difficulties are amplified in cardiac electrophysiology, where ill-posed inverse problems must infer tissue conductivities or ion-channel parameters from indirect observations and where computational cost grows rapidly with model fidelity [190,191,245,246,247].

• Credibility and VVUQ anchored in clinical use.
  • Calibration must produce models whose predictive claims align with clinical endpoints and uncertainty bounds that clinicians can act upon. This requires VVUQ routines tailored to the target pathway (e.g., early warning and therapy planning) and reported in a reproducible manner [245,246]. Guidance on validation endpoints appears in Section 4.5.2.
• Online calibration under ISAC timing and freshness.
  • With ISAC-derived telemetry (e.g., radar-based respiration/motion and indoor localization) feeding the twin, calibration and re-calibration must respect end-to-end latency budgets and update policies so that twin states remain current for decision support. The sensing band and link budget govern feasible sampling and assimilation rates (Section 4.1.3, Section 4.2.3 and Section 4.3.3); integration pipelines and placement choices for timely updates are discussed in Section 6.3.
• Data assimilation and synchronization across scales.
  • Continuous multimodal streams—radar-derived vital signs, electrocardiography, and imaging—must be assimilated with consistent timestamps and provenance. Temporal misalignment, jitter, and dropouts can destabilize filters used in calibration and degrade predictive accuracy [7,23,136,190,248]. Provenance capture and semantic mapping for synchronized updates are covered in Section 4.5.3.
• Personalization under uncertainty.
  • Patient-specific parameters should reflect genetic variation, lifestyle factors, and heterogeneous disease trajectories, yet limited longitudinal data increase estimation uncertainty. Methods that propagate uncertainty from sensors and links through calibrated parameters to outputs help avoid overconfidence [23,245,247,249]. Reporting practices for uncertainty and calibration quality are summarized in Section 4.5.2.
• Resource-aware model fidelity and placement.
  • Physiological models (multi-scale or fluid dynamics-based) are computationally demanding; fidelity must be balanced against latency/throughput constraints of the ISAC link and the edge–cloud compute path [208,210,214,236]. Practical operating points that keep closed-loop operation feasible are linked to band-dependent constraints in Section 4.1.3, Section 4.2.3 and Section 4.3.3 and integration considerations in Section 6.3.
• Privacy, safety, and governance implications for calibration data.
  • Calibration often requires access to sensitive raw signals and intermediate states. Pipelines must comply with HIPAA/GDPR while respecting RF exposure/power limits and spectrum coexistence in clinical areas [26,188]. Related safety/regulatory and security considerations appear in Section 4.5.1 and Section 4.6.5.
• From prototypes to reproducible evidence.
  • Collectively, these hurdles motivate physics-informed machine learning, Bayesian data-assimilation pipelines, open benchmark datasets, and harmonized VVUQ protocols tailored to physiological twins so that prototypes can mature into reliable clinical decision-support systems [245,248]. Alignment with the ISAC sensing/communication cadence and provenance standards (Section 4.5.3) helps ensure that calibration results are both credible and operationally timely.

7. Conclusions

This survey examined the convergence of Integrated Sensing and Communication (ISAC) and digital-twin (DT) systems as a pathway to timelier, more context-rich, and closed-loop healthcare. We reviewed enabling technologies and spectrum trade-offs, implementation approaches, and application domains for ISAC and detailed how DTs operate across patients, devices, and facilities. In doing so, we emphasized the mechanism-level complementarities of freshness, context, and feedback that underlie an ISAC–DT closed loop.

Methodologically, our review followed a transparent process documented in Section 2 and a quantitative synthesis constrained to peer-reviewed citations already included in this survey; gray literature was excluded. These constraints clarify scope and support reproducibility while keeping the evidentiary base focused on archival scholarship.

Across sub-6 GHz, mmWave, and THz bands, we compared band-level trade-offs for healthcare ISAC and outlined reference architecture patterns that connect heterogeneous clinical sensors to ISAC links, data ingestion, semantic interoperability pipelines (e.g., FHIR/IEEE 11073), and twin synchronization. The synthesis highlights cross-cutting integration challenges—synchronization under latency budgets, interoperability and semantics, VVUQ and assurance, security/privacy/safety, and standardization gaps—that must be addressed for reliable clinical translation. Notably, peer-reviewed real-world evidence for end-to-end ISAC–DT stacks in healthcare remains limited; however, adjacent components show promising signals. To bridge this gap, we consolidated practical clinical validation indicators (Section 6.1) and proposed an evaluation protocol for hospital deployment (Section 6.2), linking technical metrics to operational and clinical endpoints. These contributions provide a concrete path for prospective studies and staged adoption.

In summary, ISAC–DT convergence offers a means to move from reactive to proactive and individualized care, but its impact will depend on harmonized semantics and reporting, latency-aware pipelines, rigorous validation and uncertainty quantification, and security- and privacy-preserving designs. We anticipate that multi-site prospective studies guided by the above protocol will help establish external validity and accelerate safe, equitable deployment in diverse healthcare settings.