Artificial Intelligence in Biomedical 3D Printing: Mapping the Evidence

Abstract

This study provides an integrated synthesis of Artificial Intelligence (AI) applications in Biomedical 3D Printing, mapping the conceptual and structural evolution of this rapidly emerging field. The bibliometric analysis, based on 229 publications indexed in the Web of Science Core Collection (2018–2025) and visualised in CiteSpace, identifies three interconnected research domains: AI-driven design and process optimisation, data-assisted bioprinting for tissue engineering, and the development of smart and adaptive materials enabling 4D functionalities. The results highlight a clear progression from algorithmic control of additive manufacturing parameters toward predictive modelling, deep learning, and autonomous fabrication systems. Leading contributors include China, India, and the USA, while journals such as Applied Sciences, Polymers, and Advanced Materials act as major dissemination platforms. Emerging clusters around “4D printing”, “deep learning”, and “shape memory polymers” indicate a shift toward intelligent, sustainable, and personalised biomanufacturing. In addition, a qualitative synthesis of the most influential papers complements the bibliometric mapping, providing interpretative depth on the scientific core driving this interdisciplinary evolution. Overall, the study reveals the consolidation of a multidisciplinary research ecosystem in which computational intelligence and biomedical engineering converge to advance the next generation of adaptive medical fabrication technologies.

1. Introduction

Recent advances in machine learning, deep learning, and computer vision—core branches of artificial intelligence (AI)—are increasingly addressing critical challenges in bioprinting [1]. By analysing large datasets and inferring material properties, these tools automatically adjust design parameters, enhancing the precision, efficiency, and reproducibility of bioprinting processes [2]. AI also enables improving printing settings, monitoring the quality of printed tissue in real time, and creating scaffolds that resemble natural biological structures [3,4,5].

The combination of AI and biomedical three-dimensional (3D) printing represents a significant advancement in personalised healthcare, tissue engineering, and regenerative medicine. Beyond enhancing design control, this convergence makes it easier for medical research to move forward more quickly, improves patient outcomes, and might even enable the creation of fully functional artificial organs [6,7,8].

In recent years, the combination of biomedical 3D printing and AI has changed the way personalised medicine and healthcare work [9,10]. Bioprinting, or biomedical 3D printing, can produce complex biological tissues and structures that can be used to fabricate organs tailored to each patient. Nevertheless, the process requires meticulous control, optimisation, and predictive modelling, which can be difficult and time-consuming [11,12].

Additive manufacturing (AM), or 3D printing, has emerged as a versatile platform for fabricating biomedical devices, scaffolds, wearable technologies, and smart systems such as soft robotics, actuators, and flexible electronics [13,14,15,16]. It can be used in addition to traditional methods for making complex materials using layer-by-layer strategies.

There has been a worldwide shortage of transplantable organs for several years. The ethical issues surrounding animal experimentation, which continue to rank among the most significant concerns, emphasise the need for alternatives to in vivo testing. Two emerging technologies that have significantly changed people’s lives are AI and Three-dimensional printing. Three-dimensional Bioprinting, which extends 3D printing to enable the creation of complex structures with high deposition accuracy, is one of the most promising methods for biofabrication [17,18,19,20].

AI algorithms use generative design techniques to create and optimise 3D models. These algorithms create structures that are more efficient, stronger, and lighter than those produced by traditional design techniques by analysing constraints such as material type, weight, and strength [21].

Real-time 3D printing process monitoring is enabled by machine learning systems that can automatically adjust operating parameters and detect flaws such as warping or misaligned layers. These interventions increase product consistency, decrease material waste, and improve dimensional accuracy [22].

By forecasting the characteristics and performance of materials, AI makes it easier to find and assess new materials for additive manufacturing. Consequently, advanced materials for aerospace parts, biomedical implants, and sustainable manufacturing applications are developed more quickly through this process. Predictive models powered by AI evaluate sensor data from additive manufacturing machinery to estimate maintenance needs, reducing downtime and increasing equipment longevity [23,24].

AI-driven additive manufacturing enables the production of patient-specific products in the healthcare industry, such as tissue scaffolds, dental implants, and customised prosthetics. AI systems create highly accurate, individualised designs by processing medical imaging data, including computed tomography (CT) and magnetic resonance imaging (MRI) scans [25].

AI optimises printing settings to reduce production waste, energy use, and material consumption. This method promotes environmentally friendly additive manufacturing when combined with recyclable or biodegradable materials [26]. AI controls the placement, density, and viability of cells in biomedical 3D printing, guiding the creation of intricate biological tissues. These capabilities enable applications in organ regeneration, drug testing, and patient-specific medical care [27].

Overall, the integration of 3D printing and AI has redefined the boundaries of biomedical manufacturing, providing a foundation for precision medicine, automated fabrication, and sustainable innovation. This review maps the scientific evidence in this rapidly evolving interdisciplinary field.

The purpose of this research is to provide a comprehensive review, in the form of a review-of-reviews, for the applications of AI in biomedical 3D printing, intending to map the conceptual and structural development of this fast-developing and expanding subject.

The structure of this paper is as follows: In Section 2, the Data Collection, Search Strategy, and Bibliometric Analysis Tool are developed, while in Section 3, Results and Discussions, the Publication Trend, Growth Dynamics, Bibliometric Analysis Tool, the Global Research Landscape, Journal Sources, and Co-citation Analysis are performed. Section 3.4 adds a qualitative synthesis of the most influential studies, complementing the quantitative mapping with interpretative insights into the field’s conceptual evolution. Finally, Section 4 presents the Conclusions.

2. Materials and Methods

2.1. Data Collection and Search Strategy

This bibliometric study was conducted using the Web of Science (WoS) Core Collection database, chosen for its rigorous indexing standards and its comprehensive coverage of high-impact journals across engineering, biomedical, and computational sciences. The purpose of this stage was to ensure the inclusion of relevant and influential publications at the intersection of advanced manufacturing technologies, AI, and biomedical applications. While other major indexing platforms such as Scopus or PubMed were not included, the focus on WoS allowed for a consistent dataset suitable for bibliometric analysis using CiteSpace. Future studies could expand the scope by including additional databases to enhance coverage.

The search was carried out on 13 October 2025, covering the period from 2018 to 2025. This time interval was selected to capture both the emergence of AI applications in additive manufacturing and the most recent developments in biomedical 3D printing. The search query was carefully formulated using Boolean logic and topic search (TS) operators, integrating key concepts that dominate the scientific discourse within this interdisciplinary domain.

To build a representative dataset, five conceptual search sets were defined, as summarised in

Table 1. Bibliometric search strategy and conceptual sets used in data collection.

Search Set	Concept	Search Term (WoS Query)	Initial Count
#1	Artificial Intelligence	TS = (“Artificial Intelligence”)	281,644
#2	Additive Manufacturing	TS = (“3d printing” OR “additive manufacturing”)	161,274
#3	Biomedical Applications	TS = (“Biomedical”)	259,500
#4	Bioprinting	TS = (“Bioprinting”)	8863
#5	Smart Materials	TS = (“smart materials”)	49,178

. These included Artificial Intelligence, Additive Manufacturing/3D Printing, Biomedical Applications, Bioprinting, and Smart Materials.

The intermediate sets were combined using Boolean operators to identify overlapping research domains. First, #1 AND #2 returned studies integrating artificial intelligence with additive manufacturing (1744 records). In parallel, #3 OR #4 OR #5 grouped studies related to biomedical engineering, bioprinting, and smart materials (313,715 records). The final search (#6 AND #7) identified documents for both intersections, yielding 234 relevant publications.

Following this, several refinement stages were applied. The dataset was limited to English-language documents, and records not directly related to Biomedical 3D Printing were excluded. After this filtering, 231 papers remained; two were subsequently removed due to thematic inconsistency, resulting in a final dataset of 229 documents selected for in-depth analysis.

For each selected record, bibliographic metadata were extracted, including titles, authors, abstracts, keywords, publication years, sources, and citation counts. The analysis focused on document types representing substantive research—review articles, original articles, early access papers, and proceedings.

The predominance of review-type publications (over 60% of the dataset) indicates a maturing research field where theoretical frameworks and methodological consensus are emerging. This consolidation phase validates the need for a review-of-reviews approach to map accumulated knowledge and identify emerging directions comprehensively.

2.2. Bibliometric Analysis Tool

The bibliometric and network analysis was performed using CiteSpace (version 6.3.R1), an advanced visualisation software widely employed for mapping research evolution and identifying collaboration patterns in the scientific literature. CiteSpace was selected for its capability to detect citation bursts, compute betweenness centrality, and visualise co-citation clusters, allowing both structural and temporal analyses of a given knowledge domain.

The time slicing parameter was set to the full period covered by the dataset (2018–2025), enabling observation of the progressive evolution of this interdisciplinary research area. The Pathfinder and Pruning the Merged Network algorithms were applied to simplify network complexity by retaining only the most significant links and nodes. This process improved interpretability and visual clarity, ensuring that the resulting network reflected the field’s underlying conceptual structure.

To achieve robust thematic labelling, clusters were identified using three complementary algorithms: Log-Likelihood Ratio (LLR), Latent Semantic Indexing (LSI), and Mutual Information (MI). Each method provided a different perspective on cluster semantics, allowing for cross-validation of thematic relevance and stability.

For each network (references, journals, institutions, countries, and keywords), key indicators were computed:

• Citation counts, represent global influence and recognition.
• Betweenness centrality indicates structural importance as a bridge between distinct thematic areas.
• Citation burstness, reflecting sharp increases in citation frequency over short intervals and marking rapidly emerging research frontiers.
• Sigma value, integrating both burstness and centrality to highlight nodes with transformative potential.

Visualisation outputs were generated in both timeline view and cluster view, illustrating the chronological distribution of clusters and their persistence across years. These visual maps allowed the identification of major knowledge domains, emerging themes, and key turning points within the field of AI-driven biomedical additive manufacturing.

By combining quantitative network indicators with qualitative interpretation, the analytical framework established in this study provides a comprehensive overview of how AI has transformed biomedical 3D printing research—from isolated studies to an increasingly interconnected, data-driven discipline.

3. Results and Discussions

3.1. Publication Trend and Growth Dynamics

The annual evolution of publications (Figure 1) shows a clear, consistent upward trend in research output on AI and Biomedical Additive Manufacturing over the 2018–2025 period. The number of papers increased steadily from four publications per year during 2018–2020 to 14 in 2021, followed by a notable acceleration in 2022 (33) and 2023 (32), and a substantial rise in 2024 (44) and 2025 (94).

This quantitative growth demonstrates that the topic has gained sustained scientific attention, transitioning from an emerging area of interest to a more structured and established research field.

During the initial years (2018–2020), the relatively small number of studies indicates a formative phase in which the potential of AI-assisted 3D printing for biomedical use was still being explored. The limited data from this period suggest that most contributions were conceptual or experimental, focusing on algorithmic feasibility, process control, and basic design optimisation.

Starting in 2021, a continuous increase can be observed, marking the beginning of a phase of expansion and diversification. The visible rise in the number of articles coincides with the broader integration of AI methodologies—such as machine learning, computer vision, and data-driven modelling—into biomedical engineering and materials research. The emergence of such cross-domain studies reflects growing collaboration between computational scientists and biomedical researchers, a pattern that bibliometric visualisation tools such as CiteSpace also confirm in later sections of this analysis.

The period 2022–2024 is characterised by consistent annual growth and by the appearance of a higher share of review papers, indicating that researchers began to synthesise accumulated knowledge and identify research directions more systematically. This is a typical sign of field consolidation, as bibliometric maturity often corresponds with increased synthesis activity and thematic clustering.

The apparent peak in 2025, with 94 publications, further emphasises the field’s current vitality. Because indexing for 2025 is still ongoing, the number should be interpreted as partial rather than final; however, the high value already recorded reinforces the conclusion that AI-enhanced biomedical additive manufacturing has become a dynamic, rapidly expanding domain.

Overall, the evolution presented in Figure 1 reveals three distinguishable stages:

• Emergence (2018–2020)–sporadic and exploratory publications, mostly conceptual;
• Expansion (2021–2023)–steady growth and diversification of themes and techniques;
• Consolidation (2024–2025)–sustained output and increased analytical and review-type contributions.

This trajectory, combined with the cumulative pattern of citations observed in subsequent analyses, supports the interpretation that the integration of AI into biomedical additive manufacturing has evolved from a niche focus into a recognisable interdisciplinary research field, characterised by stable growth, international collaboration, and thematic maturity.

3.2. Global Research Landscape and Journal Sources

3.2.1. Geographic Distribution

The research effort is highly concentrated geographically, with a clear leadership structure formed by the People’s Republic of China, India, and the USA (

Table 2. Top contributing countries in AI-enhanced biomedical additive manufacturing research.

Country	Documents	Percentage (%)
Peoples R China	50	21.645%
India	46	19.913%
USA	44	19.048%
England	20	8.658%
Germany	14	6.061%

). Together, these three countries generate more than 60% of the total output identified in this study.

This indicates that most research on integrating AI and biomedical additive manufacturing is conducted in regions with established academic infrastructure and active technological development in both AI and 3D printing.

England (8.658%) and Germany (6.061%) follow as secondary contributors, maintaining consistent publication activity and collaborative links with top-producing countries. Their presence in the leading group reinforces the idea that this topic has attracted substantial attention across both Asian and Western research ecosystems.

3.2.2. Source Journals

The analysis of publication sources (

Table 3. Leading journals publishing research on AI-assisted biomedical additive manufacturing.

Journal Name	Documents	Percentage (%)
Applied Sciences Basel	7	3.030%
Biomimetics	5	2.165%
Polymers	5	2.165%
Advanced Materials	4	1.732%
Frontiers in Bioengineering and Biotechnology	4	1.732%

) confirms the interdisciplinary character of research combining AI and biomedical additive manufacturing. Relevant papers appear across journals covering materials science, biomedical engineering, computer science, and applied technology, showing that the topic bridges several scientific communities.

Applied Sciences (Basel) records the highest number of papers, followed by Polymers, Advanced Materials, Biomimetics, and Frontiers in Bioengineering and Biotechnology. These journals emphasise complementary aspects such as material optimisation, process control, and biocompatibility evaluation.

Overall, the distribution across specialised and multidisciplinary journals indicates that AI-assisted biomedical additive manufacturing has become a visible and expanding research field, supported by contributions from both engineering and biomedical perspectives.

3.3. Co-Citation Analysis

3.3.1. Co-Cited Reference Analysis

Co-cited reference analysis is a bibliometric method used to identify the intellectual structure of a research field. It examines how often two references are cited together in later publications, assuming that frequently co-cited papers share related concepts or themes.

Figure 2a presents the network of cited articles and their co-citation relationships, reflecting the intellectual connections within the research domain. Figure 2b displays the resulting clusters, each corresponding to a distinct research topic or field. The title assigned to each cluster encapsulates its principal theme, thereby providing insight into the major areas of scholarly focus.

The network shown in Figure 2b consists of 7 clusters, summarised in

Table 4. Summary of co-citation clusters derived from the reference network.

Cluster ID	Size	Silhouette	Label (LSI)	Label (LLR)	Label (MI)	Average Year
1	28	0.904	4D printing technologies	4D printing technologies (28.65, 1.0 × 10−4)	3D-printed microneedle feature (0.98)	2019
2	20	0.874	ai-driven 3D bioprinting	ai-driven 3D bioprinting (21.29, 1.0 × 10−4)	structure-function integrated tissue regeneration (0.63)	2021
3	18	0.915	deep learning-powered powder bed fusion	3D printing (21.16, 1.0 × 10−4)	learning approach (0.56)	2022
4	17	0.955	extrusion-based bioprinting system	3D extrusion bioprinting (30.09, 1.0 × 10−4)	applications potential (0.05)	2018
5	16	0.994	shape memory polymer review	shape memory polymer review (27.47, 1.0 × 10−4)	tissues engineering (0.69)	2020
6	13	0.918	4D printing	recent advancement (16.88, 1.0 × 10−4)	learning approach (0.58)	2022
7	12	0.885	4D bioprinting	4D bioprinting (25.98, 1.0 × 10−4)	learning approach (0.1)	2021

.

The largest cluster (#1) has 28 members and a silhouette value of 0.904. It is labelled as 4D printing technologies by both LLR and LSI, and as 3D-printed microneedle feature (0.98) by MI.

The major citing article of the cluster is: Pugliese, R (2022) “Artificial intelligence-empowered 3d and 4d printing technologies toward smarter biomedical materials and approaches” [16].

The most cited members within this cluster are Shahrubudin et al. (2019), published in Procedia Manufacturing [28], Ahmed et al. (2021) in Polymer [29], and Ngo et al. (2018) in Composites Part B: Engineering [30], having received 8, 7, and 6 citations, respectively.

The second largest cluster (#2) has 20 members and a silhouette value of 0.874. It is labelled as ai-driven 3D bioprinting by both LLR and LSI, and as structure-function integrated tissue regeneration (0.63) by MI.

The major citing article of the cluster is: Zhang, Z (2025) “Ai-driven 3d bioprinting for regenerative medicine: from bench to bedside” [31].

The most cited members in this cluster are Yu et al. (2020), published in International Journal of Bioprinting [32], Gu et al. (2020) in Asian Journal of Pharmaceutical Sciences [33], and Schwab et al. (2020) in Chemical Reviews [34], each receiving seven citations within this network.

The third largest cluster (#3) includes 18 members and has a silhouette value of 0.915. It is labelled as 3D printing by LLR, deep learning-powered powder bed fusion by LSI, and learning approach (0.56) by MI. The most influential citing article in this cluster is Lu (2024), titled “3D Printing of Biologics—What Has Been Accomplished to Date?” [35].

The most cited members in this cluster include Goh et al. (2021), published in Artificial Intelligence Review [36], with 17 citations; Equbal et al. (2025) in International Journal of Lightweight Materials and Manufacture [37], cited 5 times; and Zhou et al. (2024) in Sensors (Basel) 8 [38], with 4 citations.

The fourth largest cluster (#4) has 17 members and a silhouette value of 0.955. It is labelled as 3D extrusion bioprinting by LLR, extrusion-based bioprinting system by LSI, and application potential (0.05) by MI.

The major citing article of the cluster is: Zhang, YS (2021) “3d extrusion bioprinting” [39].

The most cited members within this cluster are Conev et al. (2020), published in Tissue Engineering Part A [40], with 10 citations; Gu et al. (2018) in Materials Horizons [41], cited 3 times; and Jin et al. (2020) in Advanced Intelligent Systems [42], also with 3 citations.

The fifth largest cluster (#5) has 16 members and a silhouette value of 0.994. It is labelled as a shape memory polymer review by both LLR and LSI, and as tissues engineering (0.69) by MI.

The major citing article associated with this cluster is Oladapo (2023), titled “Shape Memory Polymer Review for Flexible Artificial Intelligence Materials of Biomedical Applications”, published in Materials Chemistry and Physics [43].

The most cited members within this cluster are Zhu et al. (2021) in Nature Reviews Materials [44], with 17 citations; Wang et al. (2022) in Advanced Materials [45], cited 7 times; and Gungor-Ozkerim et al. (2018) in Biomaterials Science [46] with 3 citations.

The sixth largest cluster (#6) has 13 members and a silhouette value of 0.918. It is labelled as a recent advancement by LLR, 4D printing by LSI, and a learning approach (0.58) by MI.

The major citing article for this cluster is Azher (2025), titled “Revolutionising the Future of Smart Materials: A Review of 4D Printing, Design, Optimisation, and Machine Learning Integration”, published in Advanced Materials Technologies [47].

The most cited members in this cluster include Pugliese et al. (2022) in Polymers (Basel) [16] with 6 citations; Sajjad et al. (2024) in Advanced Industrial and Engineering Polymer Research [48] with 5 citations; and Chen et al. (2024) in Advanced Materials [49], also cited 5 times.

The seventh largest cluster (#7) has 12 members and a silhouette value of 0.885. It is labelled as 4D bioprinting by both LLR and LSI, and as a learning approach (0.1) by MI.

The major citing article in this cluster is Tamir (2024), titled “A Review of Advances in 3D and 4D Bioprinting: Toward the Mass Individualization Paradigm”, published in the Journal of Intelligent Manufacturing [50].

The most cited members of this cluster are Bozkurt et al. (2021) in Journal of Materials Research and Technology [51], Liu et al. (2021) in Bioactive Materials [52], and Oh et al. (2023) in International Journal of Bioprinting [53], each with 3 citations.

Across all clusters, the top-ranked item by citation count is Goh et al. (2021) [36] from Cluster #3, with 17 citations, followed by Zhu et al. (2021) [44] from Cluster #5, also with 17 citations. The third is Conev et al. (2020) [40] from Cluster #4, with 10 citations, while Shahrubudin et al. (2019) [28] from Cluster #1 ranks fourth with 8 citations.

The top-ranked items by bursts are shown in

Table 5. References exhibiting the strongest citation bursts.

Bursts	Node Name	DOI	Cluster ID
3.29	Ngo TD, 2018, COMPOS PART B-ENG, V143, P172	10.1016/j.compositesb.2018.02.012	1
2.76	Ligon SC, 2017, CHEM REV, V117, P10212	10.1021/acs.chemrev.7b00074	1
2.73	Kuang X, 2018, ACS APPL MATER INTER, V10, P7381	10.1021/acsami.7b18265	1
2.29	Jin ZQ, 2019, MANUF LETT, V22, P11	10.1016/j.mfglet.2019.09.005	1
2.23	Shahrubudin N, 2019, PROCEDIA MANUF, V35, P1286	10.1016/j.promfg.2019.06.089	1
2.18	Momeni F, 2017, MATER DESIGN, V122, P42	10.1016/j.matdes.2017.02.068	1
2.18	Aoyagi K, 2019, ADDIT MANUF, V27, P353	10.1016/j.addma.2019.03.013	1
1.92	Goh GD, 2021, ARTIF INTELL REV, V54, P63	10.1007/s10462-020-09876-9	3
1.83	Wan X, 2020, ADV SCI, V7, P0	10.1002/advs.202001000	5
1.83	Jin ZQ, 2020, ADV INTELL SYST-GER, V2, P0	10.1002/aisy.201900130	4

.

The connectivity structure of the co-citation network is reflected in

Table 6. Most interconnected references ranked by degree.

Degree	Node Name	DOI	Cluster ID
36	Cheng YH, 2017, ADV MATER, V29, P0	10.1002/adma.201703900	0
36	Arslan H, 2019, ADV SCI, V6, P0	10.1002/advs.201800703	0
36	Cheng Y, 2018, SENSOR ACTUAT B-CHEM, V255, P3117	10.1016/j.snb.2017.09.137	0
28	Conev A, 2020, TISSUE ENG PT A, V26, P1359	10.1089/ten.tea.2020.0191, 10.1089/ten.TEA.2020.0191	4
23	Goh GD, 2021, ARTIF INTELL REV, V54, P63	10.1007/s10462-020-09876-9	3
18	Jin ZQ, 2019, MANUF LETT, V22, P11	10.1016/j.mfglet.2019.09.005	1
17	Zhu ZJ, 2021, NAT REV MATER, V6, P27	10.1038/s41578-020-00235-2	5
14	Lee J, 2020, BIOFABRICATION, V12, P0	10.1088/1758-5090/ab8707	2
12	Chen BQ, 2022, ADV FUNCT MATER, V32, P0	10.1002/adfm.202201843	2
12	Blaeser A, 2016, ADV HEALTHC MATER, V5, P326	10.1002/adhm.201500677	4

, which lists the references with the highest degree values.

Table 7. References ranked by betweenness centrality.

Centrality	Node Name	DOI	Cluster ID
0.35	Conev A, 2020, TISSUE ENG PT A, V26, P1359	10.1089/ten.tea.2020.0191, 10.1089/ten.TEA.2020.0191	4
0.31	Goh GD, 2021, ARTIF INTELL REV, V54, P63	10.1007/s10462-020-09876-9	3
0.25	Zhu ZJ, 2021, NAT REV MATER, V6, P27	10.1038/s41578-020-00235-2	5
0.13	Jin ZQ, 2019, MANUF LETT, V22, P11	10.1016/j.mfglet.2019.09.005	1
0.08	Ngo TD, 2018, COMPOS PART B-ENG, V143, P172	10.1016/j.compositesb.2018.02.012	1
0.08	Bayraktar Ö, 2017, POLYM ADVAN TECHNOL, V28, P1044	10.1002/pat.3960	4
0.07	Shahrubudin N, 2019, PROCEDIA MANUF, V35, P1286	10.1016/j.promfg.2019.06.089	1
0.06	Wan X, 2020, ADV SCI, V7, P0	10.1002/advs.202001000	5
0.06	Kantaros A, 2024, APPL SCI-BASEL, V14, P0	10.3390/app14062550	6
0.05	Chen BQ, 2022, ADV FUNCT MATER, V32, P0	10.1002/adfm.202201843	2

summarises the references ranked by betweenness centrality, highlighting those acting as structural bridges within the co-citation network.

Table 8. References with the highest sigma index values.

Sigma	Node Name	DOI	Cluster ID
1.68	Goh GD, 2021, ARTIF INTELL REV, V54, P63	10.1007/s10462-020-09876-9	3
1.31	Jin ZQ, 2019, MANUF LETT, V22, P11	10.1016/j.mfglet.2019.09.005	1
1.29	Ngo TD, 2018, COMPOS PART B-ENG, V143, P172	10.1016/j.compositesb.2018.02.012	1
1.16	Shahrubudin N, 2019, PROCEDIA MANUF, V35, P1286	10.1016/j.promfg.2019.06.089	1
1.12	Wan X, 2020, ADV SCI, V7, P0	10.1002/advs.202001000	5
1.06	Aimar A, 2019, J HEALTHC ENG, V2019, P0	10.1155/2019/5340616	3
1.06	Hamel CM, 2019, SMART MATER STRUCT, V28, P0	10.1088/1361-665X/ab1439	1
1.06	Gu GX, 2018, MATER HORIZ, V5, P939	10.1039/c8mh00653a	4
1.06	Ligon SC, 2017, CHEM REV, V117, P10212	10.1021/acs.chemrev.7b00074	1
1.04	Momeni F, 2017, MATER DESIGN, V122, P42	10.1016/j.matdes.2017.02.068	1

presents the top ranked items by sigma value.

Overall, the results summarised in Table 5, Table 6, Table 7 and Table 8 show that the studies with the highest structural influence are those by Goh et al. (2021) [36], Zhu et al. (2021) [44] and Conev et al. (2020) [40], characterised by high centrality and sigma index values, indicating a bridging impact across different research domains. The distribution of these works reveals a progression from the technological foundation of 3D printing (2018–2019), to the integration of machine learning into manufacturing processes (2020–2022), culminating in the emergence of 4D bioprinting and adaptive materials (2023–2025).

The co-citation analysis outlines a coherent thematic transition—from geometric control in printing to biologically oriented intelligent manufacturing—reflecting the consolidation of an interdisciplinary field at the intersection of engineering, materials science, and biomedicine.

3.3.2. Journals Analysis

The analysis of co-cited journals provides insight into the disciplinary foundations and thematic convergence of research at the intersection of AI and Biomedical Additive Manufacturing. The journal co-citation network (Figure 3a) and its clustered structure (Figure 3b) reveal the multidimensional nature of this research landscape, where materials science, bioengineering, and computational intelligence interact to shape a unified knowledge domain.

Five main clusters were identified, as summarised in

Table 9. Summary of journal co-citation clusters.

Cluster ID	Size	Silhouette	Label (LSI)	Label (LLR)	Label (MI)	Average Year
0	62	0.851	additive manufacturing	3D-printed microrobot (34, 1.0 × 10−4)	drug delivery system (1.31)	2021
1	56	0.862	recent advance	3D bioprinting (41.58, 1.0 × 10−4)	bacterial infection (1.59)	2022
2	52	0.919	Future prospect	additive manufacturing (118.12, 1.0 × 10−4)	bacterial infection (2.67)	2022
3	13	0.961	emerging trend	functional design (42.48, 1.0 × 10−4)	bacterial infection (0.21)	2024
4	8	0.937	3D printing	biomedical application (34.64, 1.0 × 10−4)	bacterial infection (0.29)	2024

, each corresponding to a distinct area of scholarly influence.

The largest cluster (#0) comprises 62 members and exhibits a silhouette value of 0.851. It is labelled as 3D-printed microrobot by LLR, additive manufacturing by LSI, and drug delivery system (1.31) by MI. The major citing article in this cluster is Huo (2025), titled “Advancing Microneedle Technology for Multiple Distinct Target Organs Drug Delivery through 3D Printing: A Comprehensive Review”, published in Advanced Composites and Hybrid Materials [54]. The most cited sources within this cluster include Advanced Materials (118 citations), Scientific Reports (UK) (113 citations), and Advanced Functional Materials (105 citations).

The second largest cluster (#1) consists of 56 members and has a silhouette value of 0.862. It is labelled as 3D bioprinting by LLR, recent advance by LSI, and bacterial infection (1.59) by MI. The major citing article is Razzaq (2025), “Additive Manufacturing for Biomedical Bone Implants: Shaping the Future of Bones”, published in Materials Science & Engineering R: Reports [55].

The most-cited members in the second-largest cluster (#1) include Acta Biomaterialia (91 citations), Biomaterials (88 citations), and Biofabrication (75 citations).

The third largest cluster (#2) comprises 52 members and has a silhouette value of 0.919. The major citing article in this cluster is Azher (2025), titled “Revolutionising the Future of Smart Materials: A Review of 4D Printing, Design, Optimisation, and Machine Learning Integration”, published in Advanced Materials Technologies [47]. The most cited sources are Materials (111 citations), Additive Manufacturing (94 citations), and Materials & Design (94 citations).

The fourth largest cluster (#3) contains 13 members and exhibits a silhouette value of 0.961. It is labelled as functional design by LLR, emerging trend by LSI, and bacterial infection (0.21) by MI. The major citing article is Dananjaya (2025), “MXenes and Its Composite Structures: Synthesis, Properties, Applications, 3D/4D Printing, and Artificial Intelligence–Machine Learning Integration”, published in Progress in Materials Science [56].

The most-cited members of this cluster are Biosensors and Bioelectronics (27 citations), Analytical Chemistry (22 citations), and Biosensors (Basel) (18 citations).

The fifth largest cluster (#4) comprises 8 members and has a silhouette value of 0.937. It is labelled as biomedical application by LLR, 3D printing by LSI, and bacterial infection (0.29) by MI. The major citing article in this cluster is Dananjaya (2025), “MXenes and Its Composite Structures: Synthesis, Properties, Applications, 3D/4D Printing, and Artificial Intelligence–Machine Learning Integration”, published in Progress in Materials Science [56].

The most cited sources associated with this cluster include Nanomaterials (Basel) (31 citations), Progress in Additive Manufacturing (24 citations), and ACS Applied Polymer Materials (21 citations).

Across the network, the most influential journals by citation count include Advanced Materials, Scientific Reports (UK), Materials, Advanced Functional Materials, and Polymers (Basel) (

Table 10. Most frequently cited journals.

Citation Counts	Node Name	Cluster ID
118	Adv Mater	0
113	Sci Rep-UK	0
111	Materials	2
105	Adv Funct Mater	0
100	Polymers-Basel	0
97	Acs Appl Mater Inter	0
94	Mater Design	2
94	Addit Manuf	2
92	Nat Commun	0
91	Acta Biomater	1

).

Journals with strong citation bursts—such as Nature Materials, Rapid Prototyping Journal, and 3D Printing and Additive Manufacturing (

Table 11. Journals with the strongest citation bursts.

Bursts	Node Name	Cluster ID
7.20	Nat Mater	0
6.52	Rapid Prototyping J	2
6.22	3D Print Addit Manuf	2
5.66	Lab Chip	0
5.40	J Manuf Sci E-T Asme	2
4.96	P Natl Acad Sci Usa	0
4.81	Biomaterials	1
4.65	Mater Horiz	0
4.58	J Mater Sci	4
4.53	Angew Chem Int Edit	0

)—highlight moments of rapid growth and conceptual breakthrough, often coinciding with the integration of AI algorithms into design and printing workflows.

The top-ranked items by degree are presented in

Table 12. Most interconnected journals ranked by degree.

Degree	Node Name	Cluster ID
65	ADV FUNCT MATER	0
63	Acs Appl Mater Inter	0
52	Adv Healthc Mater	1
48	Biomaterials	1
47	Acs Nano	0
44	Adv Mater	0
42	Biofabrication	1
39	Acta Biomater	1
38	Sci Rep-Uk	0
37	Nat Commun	0

.

Moreover, sources such as Additive Manufacturing, Materials & Design, and Acta Materialia demonstrate both high centrality and sigma values (

Table 13. Journals ranked by betweenness centrality.

Centrality	Node Name	Cluster ID
0.15	Adv Funct Mater	0
0.11	Acs Appl Mater Inter	0
0.08	Mater Design	2
0.08	Addit Manuf	2
0.07	Rapid Prototyping J	2
0.07	Acta Mater	2
0.06	Acs Nano	0
0.06	Mat Sci Eng C-Mater	1
0.05	Adv Healthc Mater	1
0.05	Sensor Actuat B-Chem	3

and

Table 14. Journals with the highest sigma index values.

Sigma	Node Name	Cluster ID
1.59	Rapid Prototyping J	2
1.36	Addit Manuf	2
1.31	Mater Design	2
1.27	Acta Mater	2
1.21	Biomaterials	1
1.11	Angew Chem Int Edit	0
1.08	J Control Release	1
1.06	Chem Soc Rev	0
1.05	Sci Adv	0
1.05	Nat Mater	0

), confirming their role as structural bridges linking the materials science core with the biomedical and computational subfields. In this context, centrality reflects the structural importance of a journal as a connector between different thematic domains, while the sigma value combines both centrality and citation burstness, highlighting sources with transformative potential.

The journal co-citation landscape reflects the multidisciplinarity of AI-enhanced additive manufacturing. The coexistence of journals from engineering, materials, and biomedical domains signifies not fragmentation, but rather convergence—a gradual formation of a cohesive research ecosystem where design intelligence, material adaptability, and clinical relevance intersect.

3.3.3. Country Co-Occurrence Analysis

The analysis of country co-occurrence reveals the global structure of collaboration and knowledge production in the field of AI-assisted biomedical additive manufacturing. As illustrated in Figure 4, the network of national affiliations demonstrates a high degree of interconnectivity, with a few dominant hubs shaping the intellectual and technological landscape.

Five major clusters were identified (

Table 15. Summary of country co-occurrence clusters.

Cluster ID	Size	Silhouette	Label (LSI)	Label (LLR)	Label (MI)	Average Year
0	13	0.833	mechanical properties	mechanical properties (65.4, 1.0 × 10−4)	smart glove (1.62)	2023
1	12	0.63	3D printing	smart eyeglass frame (44.32, 1.0 × 10−4)	learning approach (1.3)	2022
2	12	0.711	4D printing	synthesis properties application (59.31, 1.0 × 10−4)	using blockchain (1.83)	2023
3	11	0.948	new insight	creating new opportunities (37.53, 1.0 × 10−4)	synergies advantage (0.38)	2022
4	6	0.78	using new advanced photopolymerisable resin	using new advanced photopolymerisable resin (27.24, 1.0 × 10−4)	ti-based alloys quality evaluation (0.03)	2022

), each reflecting different regional focuses and research synergies. Cluster labels were generated using three algorithms—LSI (Latent Semantic Indexing), LLR (Log-Likelihood Ratio), and MI (Mutual Information)—which automatically extract representative terms to characterise the thematic focus of each cluster.

The first cluster is led by the major citing article of Tuninetti (2025), titled “Biomimetic Lattice Structures Design and Manufacturing for High Stress, Deformation, and Energy Absorption Performance”, published in Biomimetics [57].

The most cited contributing countries in this cluster are People’s Republic of China (47 citations), Saudi Arabia (13 citations), and South Korea (10 citations).

The second largest cluster (#1) includes 12 members and has a silhouette value of 0.63. It is labelled as smart eyeglass frame by LLR, 3D printing by LSI, and learning approach (1.3) by MI. The major citing article is Rahman (2023), “Review of Intelligence for Additive and Subtractive Manufacturing: Current Status and Future Prospects”, published in Micromachines. The most cited countries represented in this cluster are the United States (43 citations), Canada (11 citations), and Singapore (10 citations).

The third largest cluster (#2) also comprises 12 members and presents a silhouette value of 0.711. It is labelled as synthesis properties application by LLR, 4D printing by LSI, and using blockchain (1.83) by MI. The major citing article is Dananjaya (2025), “MXenes and Its Composite Structures: Synthesis, Properties, Applications, 3D/4D Printing, and Artificial Intelligence–Machine Learning Integration”, published in Progress in Materials Science [56].

The most cited countries in this cluster are India (45 citations), England (20 citations), and Germany (13 citations).

The fourth largest cluster (#3) contains 11 members and has a silhouette value of 0.948. It is labelled as creating new opportunities by LLR, new insight by LSI, and synergies advantage (0.38) by MI. The major citing article is Hassoun (2023), “Creating New Opportunities for Sustainable Food Packaging through Dimensions of Industry 4.0: New Insights into the Food Waste Perspective”, published in Trends in Food Science & Technology [58].

The most cited members in this cluster are: 13 Italy, 9 Spain, and 8 Portugal.

The fifth largest cluster (#4) has 6 members and a silhouette value of 0.78. It is labelled as using new advanced photopolymerisable resin by both LLR and LSI, and as Ti-based alloys quality evaluation (0.03) by MI. The major citing article of the cluster is Baila (2024), “3D Printing of Personalised Stents Using New Advanced Photopolymerizable Resins and Ti-6Al-4V Alloy”, published in the Rapid Prototyping Journal [59].

Overall, the co-occurrence results confirm the leading roles of China, India, and the USA, which together account for over 60% of the dataset’s total research output (

Table 16. Most cited countries in AI-bioprinting research.

Citation Counts	Node Name	Cluster ID
47	Peoples R China	0
45	India	2
43	USA	1
20	England	2
13	Italy	3
13	Saudi Arabia	0
13	Germany	2
11	Canada	1
10	Singapore	1
10	South Korea	0

). Their dominance is complemented by emerging contributions from European and Middle Eastern countries, suggesting a gradual diversification of expertise and collaborative focus.

Nations such as Turkey, Italy, Singapore, and Iran exhibit strong citation bursts (

Table 17. Countries with the strongest citation bursts.

Bursts	Node Name	Cluster ID
2.00	Turkey	2
1.52	Singapore	1
1.28	Italy	3
0.93	Iran	1
0.73	Norway	4
0.67	Switzerland	1
0.00	Peoples R China	0
0.00	India	2
0.00	USA	1
0.00	England	2

), signalling growing influence and recent acceleration in scientific visibility.

From a network perspective, countries with both high degree and sigma values—such as the USA, India, and Turkey (

Table 18. Most interconnected countries ranked by degree.

Degree	Node Name	Cluster ID
25	USA	1
22	India	2
20	Peoples R China	0
20	Turkiye	3
18	Italy	3
17	Saudi Arabia	0
17	France	3
16	Portugal	3
15	England	2
14	Singapore	1

,

Table 19. Countries ranked by betweenness centrality.

Centrality	Node Name	Cluster ID
0.23	Peoples R China	0
0.21	USA	1
0.18	India	2
0.16	Saudi Arabia	0
0.16	Poland	4
0.13	England	2
0.10	Turkiye	3
0.06	Italy	3
0.06	Portugal	3
0.06	Germany	2

and

Table 20. Countries with highest sigma index values.

Sigma	Node Name	Cluster ID
1.11	Turkey	2
1.08	Italy	3
1.06	Iran	1
1.05	Singapore	1
1.00	Peoples R China	0
1.00	USA	1
1.00	India	2
1.00	Saudi Arabia	0
1.00	Poland	4
1.00	England	2

)—act as structural bridges connecting technologically advanced regions with emerging innovation hubs.

This structure reflects a globally distributed yet interdependent research ecosystem in which AI-enabled additive manufacturing for biomedical applications evolves through sustained international collaboration, shared technological platforms, and cross-disciplinary exchange.

3.3.4. Institutional Co-Occurrence Analysis

The institutional co-occurrence analysis highlights the organisational structure of global research collaboration within the field of AI-enhanced additive manufacturing for biomedical applications. As shown in Figure 5, the network consists of institutions that frequently co-appear as affiliations in the same publications, indicating active collaboration links and shared authorship across organisations.

Figure 5 shows that several institutions act as central nodes in the network. Among them, the Indian Institute of Technology System (IIT System) and the National Institute of Technology System (NIT System) are prominent hubs, connected to multiple other institutions. This suggests that these institutions are consistently involved in collaborative work and contribute to multi-institution research teams. The Massachusetts Institute of Technology (MIT) is also positioned as a core node with multiple connections, indicating its role in linking different research groups.

Harvard University and its medical affiliates, together with Brigham and Women’s Hospital, form a closely connected subgroup, suggesting an active clinical and biomedical translation component within this research area. Nanyang Technological University is likewise part of the main network structure, appearing in connection with both engineering- and medically oriented institutions. The California State University System is also represented in the network, indicating participation from large university systems as well as specialised research centres.

Figure 6 shows the top 10 institutions exhibiting the strongest citation bursts, highlighting those with the most rapid increase in research impact over the analysed period. These institutions can be interpreted as emerging leaders in the field, either because of accelerated publication activity or a recent concentration of highly cited work.

In Figure 6, the Strength value reflects the intensity of the citation burst, with higher values indicating a sharper and more significant increase in research impact. The rightmost column illustrates the time span of each burst between 2020 and 2025, with red segments denoting the specific years of intensified citation activity.

The institutional co-occurrence pattern confirms that collaborations between technical universities, medical research affiliates, and large university systems drive the field.

3.3.5. Analysis of Keywords

The keyword co-occurrence analysis identifies the conceptual structure and emerging themes in the research field of AI-assisted additive manufacturing for biomedical applications. By mapping the relationships among recurring terms, it becomes possible to trace the evolution of dominant topics and research priorities.

The visualisation presented in Figure 7 reveals a dense and interconnected keyword network. The most prominent nodes correspond to keywords such as “3D printing”, “additive manufacturing”, “artificial intelligence”, “3D bioprinting”, “scaffolds”, “prediction”, and “mechanical property”. Their large node sizes and central positioning indicate high frequency and strong link strength, reflecting their pivotal role in connecting different research areas. The dense connections between these nodes illustrate the field’s interdisciplinary nature—combining engineering, materials science, and biomedical research within an AI-driven framework.

Keywords such as “artificial intelligence” and “prediction” are closely linked to “additive manufacturing”, showing the growing reliance on data-driven approaches and machine learning models for process optimisation. The association between “scaffolds” and “bioprinting” underscores the biomedical focus of many studies, where 3D and 4D printed structures are developed for tissue regeneration and implant design.

The clustering analysis of keywords, presented in Figure 7b, identifies seven major thematic clusters labelled from #0 to #7, indicated in

Table 21. Summary of major keyword co-occurrence clusters.

Cluster ID	Size	Silhouette	Label (LSI)	Label (LLR)	Label (MI)	Average Year
0	28	0.789	3D bioprinting	3D bioprinting (74.36, 1.0 × 10−4)	3D-manufactured monolithic stage (0.64)	2023
1	21	0.666	additive manufacturing	artificial intelligence technologies (43.93, 1.0 × 10−4)	redefining medical application (0.84)	2022
2	21	0.73	4D printing technologies	additive manufacturing (60.7, 1.0 × 10−4)	3D-manufactured monolithic stage (2.73)	2022
3	13	0.752	drug delivery	additive manufacturing (51.28, 1.0 × 10−4)	smart additive manufacturing (0.7)	2022
5	6	1	3D printing	annealed 3D-printed pla-date pit (39.97, 1.0 × 10−4)	3D-manufactured monolithic stage (0.56)	2023
6	6	0.879	pilot suit engineering	pilot suit engineering (33.59, 1.0 × 10−4)	molecular analysis (0.05)	2024
7	5	0.938	filler	polymer composite (22.06, 1.0 × 10−4)	additive manufacturing (0.06)	2024

.

The major citing article of the largest cluster is Yadav (2025), “Pioneering 3D and 4D Bioprinting Strategies for Advanced Wound Management: From Design to Healing”, published in Small [60].

The most cited members in this cluster are 23 scaffolds, 12 tissue engineering, and 12 3D bioprinting.

The second largest cluster (#1) has 21 members and a silhouette value of 0.666. The major citing article of the cluster is Omairi (2021), “Towards Machine Learning for Error Compensation in Additive Manufacturing”, published in Applied Sciences (Basel) [61].

The most cited members in this cluster are 19 in technology, 12 in mechanical properties, and 11 in deep learning.

The third largest cluster (#2) has 21 members and a silhouette value of 0.73. The major citing article is Prasittisopin (2024), “How 3D Printing Technology Makes Cities Smarter: A Review, Thematic Analysis, and Perspectives”, published in Smart Cities [62].

The most cited members in this cluster are 72 artificial intelligence, 55 additive manufacturing, and 40 design.

The fourth largest cluster (#3) has 13 members and a silhouette value of 0.752. The major citing article is Khalili (2023), “Advanced Therapy Medicinal Products for Age-Related Macular Degeneration: Scaffold Fabrication and Delivery Methods”, published in Pharmaceuticals [63].

The most cited members in this cluster are 25 3D, 16 4D printing, and 12 system.

The fifth largest cluster (#5) has 6 members and a silhouette value of 1.0. The major citing article is Fouly (2023), “Investigating the Mechanical Properties of Annealed 3D-Printed PLA-Date Pits Composite”, published in Polymers [64].

The most cited members in this cluster are 52 in 3D printing, 5 in surgery, and 2 in rehabilitation medicine.

The sixth largest cluster (#6) has 6 members and a silhouette value of 0.879. The major citing article is Lye (2025), “Microfluidic-Enabled Nanomedicine: A Comprehensive Review of Recent Advances and Translational Potential”, published in Microfluidics and Nanofluidics [65].

The most cited members in this cluster are 6 devices, 4 composites, and 3 sensors.

The seventh largest cluster (#7) has 5 members and a silhouette value of 0.938. The major citing article is Arora (2024), “A Comprehensive Review on Fillers and Mechanical Properties of 3D-Printed Polymer Composites”, published in Materials Today Communications [66].

Overall, the keyword analysis shows that the field has evolved from fundamental studies on additive manufacturing and mechanical properties toward advanced applications in bioprinting, drug delivery, and 4D adaptive systems. The presence of artificial intelligence and prediction as central connecting terms reflects the integration of computational methods into material design and fabrication optimisation. This transition indicates a clear shift from process-oriented research to intelligent, application-driven innovation in biomedical additive manufacturing.

Table 22. Most frequently cited keywords.

Citation Counts	Node Name	Cluster ID
72	artificial intelligence	2
55	additive manufacturing	2
52	3D printing	5
40	design	2
30	machine learning	2
27	fabrication	2
25	3D	3
23	scaffolds	0
19	technology	1
16	4D printing	3

lists the most frequent keywords, confirming that “artificial intelligence”, “additive manufacturing”, “3D printing” and “design” dominate the research landscape.

To capture emerging trends,

Table 23. Keywords with the strongest citation bursts.

Bursts	Node Name	Cluster ID
3.01	deep learning	1
2.88	technology	1
2.18	biomedical applications	1
1.76	shape memory	3
0.88	mechanical property	1
0.78	system	3
0.62	prediction	1
0.00	artificial intelligence	2
0.00	additive manufacturing	2
0.00	3D printing	5

lists the keywords with the strongest citation bursts. The leading burst terms—“deep learning” (3.01), “technology” (2.88), and “biomedical applications” (2.18)—are all located within Cluster #1, underscoring the rapid convergence between computational intelligence and biomedical fabrication.

Table 24. Most interconnected keywords ranked by degree.

Degree	Node Name	Cluster ID
27	scaffolds	0
26	additive manufacturing	2
23	3D bioprinting	0
22	artificial intelligence	2
21	3D printing	5
20	mechanical property	1
20	prediction	1
20	3D	3
17	system	3
17	fabrication	2

, which ranks keywords by degree, highlights “scaffolds” (27), “additive manufacturing” (26), “3D bioprinting” (23), and “artificial intelligence” (22) as the most interconnected nodes in the dataset. These terms appear in clusters #0, #2, and #5, confirming that the intersection of biofabrication, AI, and additive manufacturing forms the conceptual core of the field.

Table 25. Keywords ranked by betweenness centrality.

Centrality	Node Name	Cluster ID
0.20	scaffolds	0
0.19	additive manufacturing	2
0.18	3D printing	5
0.15	artificial intelligence	2
0.14	mechanical property	1
0.13	prediction	1
0.11	3D bioprinting	0
0.10	3D	3
0.10	devices	6
0.09	fabrication	2

lists the top keywords by centrality, a measure of their bridging capacity within the network. Here again, “scaffolds” (0.20) and “additive manufacturing” (0.19) display the highest values, reinforcing their function as connectors between technological and biomedical research clusters. “3D printing” (0.18) and “artificial intelligence” (0.15) maintain strong intermediary positions, linking algorithmic approaches to practical fabrication.

Table 26. Keywords with the highest sigma index values.

Sigma	Node Name	Cluster ID
1.20	technology	1
1.15	deep learning	1
1.12	mechanical property	1
1.08	prediction	1
1.04	system	3
1.04	biomedical applications	1
1.01	shape memory	3
1.00	scaffolds	0
1.00	additive manufacturing	2
1.00	3D printing	5

ranks keywords by their sigma value, which combines centrality and burst strength to identify transformative concepts. The leading positions are held by “technology” (1.20), “deep learning” (1.15), and “mechanical property” (1.12), indicating that the most significant recent breakthroughs occur where AI-based modelling intersects with material performance analysis. “Prediction” (1.08) and “biomedical applications” (1.04) also show high sigma values, underscoring the accelerating integration of AI techniques in biomedical engineering.

The temporal evolution of research themes is depicted in Figure 8, which displays the timeline view of the keyword clusters identified in the co-occurrence network.

Each horizontal line corresponds to a thematic cluster, and the nodes along it represent the keywords that are active in a specific year. The colour gradient—from blue (earlier years) to red (recent years)—illustrates the field’s chronological progression.

Cluster #0 (3D bioprinting) remains the most significant and most persistent throughout the entire 2020–2025 period. Its evolution—from early keywords such as cells, hydrogels, and collagen to later terms like bioinks, constructs, and wound healing—shows the steady advancement from material and structural research toward functional tissue engineering and clinical translation.

Cluster #1 (additive manufacturing) spans a wide temporal range, connecting early methodological concepts (prediction, mechanical property, design) with emerging computational approaches (deep learning, neural network, powder bed fusion). This indicates a gradual integration of AI into process optimisation and material analysis.

Cluster #2 (4D printing technologies) appears as a coherent sequence of terms such as shape memory, smart materials, and industry 5.0, representing the shift toward adaptive and time-dependent materials and manufacturing systems.

Cluster #3 (drug delivery) emerges more prominently after 2022, linking biomaterials and cell research with controlled-release systems and personalised medical applications.

Clusters #5 to #7 (3D printing, pilot suit engineering, and filler) show more specialised or peripheral developments. They focus on mechanical design, rehabilitation devices, and composite materials, extending the impact of additive manufacturing beyond biomedical contexts.

Overall, Figure 8 reveals a clear chronological trajectory: early research (2020–2021) concentrated on materials and structural parameters; mid-period studies (2022–2023) emphasised AI-driven modelling and process optimisation; while recent publications (2024–2025) highlight bioprinting, regenerative medicine, and adaptive manufacturing systems. This timeline visualisation confirms the ongoing convergence of artificial intelligence, materials innovation, and biomedical engineering into an integrated, evolving research domain.

3.4. Qualitative Synthesis of Core Scientific Contributions

To deepen the thematic interpretation of the results, a qualitative synthesis was conducted of the 25 most-cited papers from the corpus of 229 studies.

Table 27. Summary of the 25 most influential papers integrating ai and biomedical 3D printing.

N0	Authors	Research Area	Main Focus	Key Findings	Times Cited
1	Goh et al. [36]	Computer Science	Examines the integration of machine learning techniques in additive manufacturing (AM) to improve design, process optimisation, in situ monitoring, and cybersecurity within smart manufacturing systems.	Identifies artificial neural networks (ANN) and convolutional neural networks (CNN) as the most effective algorithms for process optimisation and anomaly detection in AM. Highlights the importance of standardised data formats, high-resolution sensors, and advanced data-acquisition techniques to enable predictive modelling, real-time monitoring, and the creation of digital twins for future manufacturing and biomedical applications.	511
2	Zhang et al. [39]	Science and Technology	Examines extrusion-based three-dimensional (3D) bioprinting methodologies, emphasising the selection of hardware, software, and bioinks for fabricating complex and functional tissue constructs.	Identifies recent advancements that enhance the precision and versatility of 3D extrusion bioprinting, highlighting its growing role in tissue engineering and translational biomedical applications, and outlining future directions for improving biofabrication performance and reproducibility.	289
3	Zhu et al. [44]	Science and Technology	Explores the convergence of artificial intelligence, functional materials, and advanced 3D printing to enable the direct fabrication of wearable and implantable biomedical devices on complex target surfaces.	Demonstrates that AI-driven control systems enhance precision and adaptability in both ex situ and in situ 3D printing, supporting autonomous medical manufacturing, real-time environmental sensing, and the development of personalised therapeutic solutions.	268
4	Seoane-Viaño et al. [67]	Pharmacology and Pharmacy	Reviews the progress of 3D printing technologies in pharmaceutical development, emphasising their potential to create personalised drug products and accelerate clinical translation.	Highlights advances in producing flexible-dose medicines tailored to individual patient needs and outlines current challenges that hinder clinical adoption, proposing strategies to bridge the gap between experimental pharmaceutical 3D printing and its integration into routine medical practice.	206
5	Meng et al. [68]	Orthopaedics	Analyses the current state of 3D printing technologies for producing customised metal implants in orthopaedic surgery, focusing on their design, materials, and clinical applicability.	Highlights that 3D-printed metal implants enhance structural precision, biocompatibility, and osseointegration while reducing surgery time and costs. Notes challenges related to software, materials, and regulation, and envisions future integration with AI, 4D printing, and big data to achieve fully personalised orthopaedic solutions.	147
6	Dabbagh et al. [69]	Science and Technology	Reviews recent advances in 3D printing technologies for the fabrication of microrobots, emphasising their biomedical and environmental applications and the potential for intelligent actuation and control.	Describes how 3D-printed microrobots can perform targeted tasks such as drug delivery, minimally invasive surgery, and imaging. Identifies emerging trends in integrating smart materials, artificial intelligence, and physical intelligence to enhance autonomy and biocompatibility, while outlining current challenges in clinical translation and regulatory standardisation.	123
7	Mamo et al. [8]	Engineering; Materials Science	Provides a comprehensive overview of additive manufacturing technologies and their applications in the medical field, discussing process types, biomedical materials, and technological challenges.	Emphasises the potential of additive manufacturing to deliver customised, cost-effective medical devices, such as implants, prosthetics, and surgical guides. Identifies material limitations, regulatory barriers, and the need for standardisation, while highlighting the growing role of AI and the Internet of Medical Things (IoMT) in enhancing safety, efficiency, and personalisation in biomedical manufacturing.	103
8	Guo et al. [70]	Materials Science; Metallurgy and Metallurgical Engineering	Reviews state-of-the-art multi-material 3D printing technologies for titanium-based biomaterials, focusing on their structural design, mechanical optimisation, and biomedical applicability.	Highlights how 3D-printed titanium alloys enhance osseointegration, mechanical tunability, and biocompatibility in bone tissue engineering. Identifies challenges in balancing strength and elasticity and emphasises the potential of artificial intelligence and machine learning to optimise process parameters and design efficiency.	94
9	Dananjaya et al. [71]	Materials Science	Discusses the synthesis, properties, and applications of graphene quantum dots (GQDs) in polymer composites, emphasising their integration into 3D printing and the role of machine learning in materials optimisation.	Demonstrates that GQD-based composites improve electrical, mechanical, and thermal performance, expanding their use in sensors, energy devices, and biomedicine. Identifies how AI enables predictive modelling and accelerated material design for next-generation nanocomposites.	94
10	Ikram et al. [72]	Materials Science; Polymer Science	Reviews the development of functional polymer nanocomposites for additive and 4D printing of smart biomedical devices responsive to environmental stimuli.	Shows that additive manufacturing enables the creation of adaptive, cost-efficient, and durable structures for biomedical applications. Notes ongoing research into novel materials and intelligent manufacturing methods to enhance flexibility, stability, and patient-specific device functionality.	92
11	Muhindo et al. [73]	Pharmacology and Pharmacy	Reviews the expanding role of 3D and 4D printing technologies in pharmaceutical manufacturing, focusing on personalised drug delivery systems and production automation.	Highlights the potential of integrating AI, machine learning, and the Internet of Things to streamline biopharmaceutical manufacturing, reduce human intervention, and enable intelligent fabrication of personalised medicines.	80
12	Pugliese et al. [16]	Polymer Science	Analyses advances in 4D printing for biomedical applications, emphasising smart, stimuli-responsive materials capable of dynamic transformation and in situ adaptation.	Describes how AI accelerates material design, enhances printing precision, and supports real-time control for surgical and biomedical uses, while outlining the emerging concept of 5D printing as AI-driven intelligent fabrication.	80
13	Basu et al. [74]	Engineering; Materials Science	Introduces the concept of “biomaterialomics”, integrating AI, data science, and additive manufacturing to accelerate the development of next-generation biomaterials and implants.	Proposes a data-driven framework that connects processing, structure, and properties of biomaterials, enabling predictive modelling, patient-specific implant design, and digital twin simulations for biomedical innovation.	79
14	Velásquez-García et al. [75]	Engineering	Reviews additive manufacturing technologies for biomedical applications, particularly metal 3D printing, focusing on material diversity, process efficiency, and biocompatible design.	Identifies opportunities to enhance additive manufacturing through AI–based optimisation, multi-material printing, and the creation of advanced load-bearing and biodegradable implants for clinical use.	76
15	Wang, ZJ et al. [76]	Chemistry; Science and Technology	Reviews the use of bioprinting for developing smart wound dressings and biomedical devices, focusing on polymer-based inks, fabrication methods, and functional design strategies.	Highlights seven major modification approaches—ranging from chemical synthesis to multi-material and in situ bioprinting—that enhance healing efficacy and device performance, while discussing future directions involving 4D and AI-assisted bioprinting.	68
16	Dong et al. [77]	Chemistry; Materials Science; Physics; Polymer Science	Surveys the fabrication methods, driving mechanisms, and emerging applications of soft robotics designed to mimic biological systems and enhance interaction in complex environments.	Identifies advances in flexible materials, structural design, and control strategies that enable multi-degree actuation and adaptability in medical and industrial contexts, emphasising AI integration for next-generation functional robotics.	64
17	Carou-Senra et al. [78]	Chemistry; Science and Technology	Examines inkjet printing as an additive manufacturing technique for personalised drug delivery and biomedical applications, focusing on pharma-ink formulation and precise material deposition.	Demonstrates how the combination of tailored substrates, intelligent hardware, and AI-based control enhances drug printing accuracy and therapeutic efficiency, positioning inkjet printing as a key enabler of patient-specific medicine manufacturing.	64
18	Sarabi et al. [79]	Biochemistry and Molecular Biology	Reviews the biomedical applications of optical fibres made from soft, biocompatible, and biodegradable materials, with emphasis on their use in implants, sensors, and therapeutic devices.	Highlights the role of 3D printing in fabricating flexible optical systems and discusses how AI enhances data analysis and predictive modelling for next-generation optical diagnostics and treatment technologies.	60
19	Rahman et al. [80]	Chemistry; Science and Technology	Provides a comprehensive analysis of the integration of AI across additive, subtractive, and hybrid manufacturing processes across various industries, including the medical and aerospace sectors.	Identifies that AI-driven monitoring and defect detection improve manufacturing precision and process automation. Notes challenges in data quality, real-time optimisation, and model generalisation, emphasising the need for advanced connectivity and sensing systems.	56
20	Park et al. [81]	Cell Biology; Biotechnology and Applied Microbiology	Summarises technological progress in tissue engineering and regenerative medicine (TERM) driven by AI and big data analytics, focusing on innovation trends and convergence research.	Highlights three major advancements in 2017: biomaterial development with 3D cell printing, exosome-based regenerative therapies, and 3D culture systems. Suggests that AI and data integration will accelerate clinical translation in TERM.	53
21	Ameta et al. [82]	Science and Technology; Energy and Fuels; Materials Science	Reviews the evolution of sustainable 3D and 4D printing technologies and their expanding role in industrial and biomedical manufacturing.	Demonstrates that 3D/4D printing improves customisation, energy efficiency, and sustainability across sectors, particularly in medical applications. Foresees deeper integration with AI to enable adaptive smart materials and environmentally responsible production systems.	49
22	Zhang et al. [31]	Engineering; Materials Science	Proposes a structured methodology for integrating AI into 3D bioprinting within the Quality by Design (QbD) framework to enhance clinical translation.	Identifies AI’s role in improving bioink formulation, structural modelling, and real-time process control, emphasising data-driven design and sensing as key enablers for scalable, patient-specific bioprinting applications.	45
23	Oladapo et al. [43]	Materials Science	Analyses the evolution of flexible 4D printing technologies and their use in developing adaptive, self-healing polymer composites for biomedical applications.	Highlights the potential of AI-enhanced materials (AIM) to replicate dynamic tissue behaviour and improve actuation control. Notes that the field remains nascent, with challenges in demonstrating practical biomedical integration and large-scale applicability.	43
24	Boban et al. [83]	Automation and Control Systems; Engineering	Reviews post-processing methods used to improve surface quality and dimensional accuracy in metal additive manufacturing (MAM), focusing on their relevance to biomedical and industrial components.	Summarises laser, chemical, and hybrid finishing techniques that enhance mechanical and functional performance. Emphasises automation through AI and machine learning to optimise surface modification and process efficiency in metallic AM parts.	43
25	Tian et al. [84]	Automation and Control Systems; Computer Science; Robotics	Explores the integration of data-driven and machine learning tools into additive manufacturing for optimising material and structural design.	Shows that AI-driven topology optimisation and tool-path planning improve material efficiency and product performance. Suggests future frameworks combining digital manufacturing and intelligent data analytics to advance Industry 4.0-driven innovation.	39

provides an overview of these works, highlighting each work’s main research focus and key findings. This synthesis helps contextualise bibliometric patterns within the field’s conceptual landscape, revealing how additive manufacturing, AI, and biomedical engineering converge to shape current research priorities.

The synthesis of the 25 most-cited studies provides a consolidated view of the scientific directions that define the intersection of AI and Biomedical Additive Manufacturing. The reviewed works converge toward three dominant areas: (1) optimisation of 3D and 4D printing processes through AI-based modelling and control (Goh et al. [36], Rahman et al. [80], Zhang et al. [31]); (2) development of smart, stimuli-responsive, and biocompatible materials supporting functional and adaptive biomedical structures (Pugliese et al. [16], Oladapo et al. [43], Ikram et al. [72]); and (3) applications focused on personalised medicine, ranging from tissue engineering and regenerative therapies to customised implants and drug delivery systems (Seoane-Viaño et al. [67], Meng et al. [68], Muhindo et al. [73]).

Across this corpus, AI consistently appears as an enabling technology that improves precision, repeatability, and data integration in fabrication workflows. Machine learning techniques, especially artificial neural networks and convolutional neural networks, are the most frequently employed to enhance process monitoring, error detection, and prediction of mechanical or biological performance (Goh et al. [36], Omairi et al. [61]). At the same time, studies addressing material aspects emphasise the convergence between functional polymers, composites, and metallic alloys, where AI contributes to design optimisation and multi-material coordination (Guo et al. [70], Dananjaya et al. [71]). The pharmaceutical and biomedical applications (Seoane-Viaño et al. [67], Muhindo et al. [73], Carou-Senra et al. [78]) demonstrate a shift toward automated and personalised fabrication of drug delivery systems. At the same time, orthopaedic and tissue engineering research (Meng et al. [68], Park et al. [81]) highlights additive manufacturing as a route to patient-specific, clinically applicable solutions.

Although the predominance of review-type publications reflects a consolidating research field, it also reveals a limited volume of primary experimental studies. This pattern indicates important opportunities for future development. Strengthening translational research will be essential for connecting algorithmic advances with clinically validated outcomes, particularly in tissue engineering, implantables, and drug delivery systems. At the same time, progress in explainable and robust AI frameworks is needed to ensure transparency, reproducibility, and regulatory acceptance in biomedical fabrication workflows. Finally, increasing emphasis on clinical relevance, high-throughput experimentation, and scalable manufacturing strategies can accelerate the transition from conceptual demonstrations to deployable biomedical solutions. Together, these directions outline a pathway for evolving the field beyond synthesis toward measurable, practice-oriented impact.

Overall, the analysis confirms that the most influential publications form a coherent narrative linking process intelligence, material innovation, and biomedical translation. These studies collectively illustrate the conceptual consolidation of AI-assisted additive manufacturing, outlining its methodological maturity and its orientation toward clinically meaningful applications.

A further dimension that emerges from the reviewed literature, yet remains insufficiently articulated, concerns the regulatory challenges associated with the clinical translation of AI-enabled biomedical 3D printing. Multiple influential studies highlight that despite rapid technological progress, printed constructs, and medical devices must comply with complex, evolving regulatory frameworks that vary across jurisdictions [67,68,69]. These challenges include the lack of harmonised standards for bioink characterisation, mechanical reliability, sterility assurance, and long-term biocompatibility, as well as uncertainties regarding how AI-driven decision systems should be validated and audited within existing medical device regulations. In the case of personalised implants and patient-specific pharmaceutical dosage forms, authors such as Seoane-Viaño et al. [67] emphasise the difficulty of defining batch consistency and quality control when each printed unit differs from the previous one. Similarly, Meng et al. [68] and Dabbagh et al. [69] point out that the integration of smart materials, embedded sensors, and adaptive algorithms complicates certification pathways that were originally designed for static, conventionally manufactured devices. The need for standardised testing protocols, transparency in machine-learning models, and updated guidance for in situ or point-of-care fabrication remains a major barrier to widespread clinical adoption. As the scientific ecosystem increasingly moves toward autonomous, data-driven fabrication, addressing these regulatory bottlenecks becomes essential for ensuring safety, reproducibility, and public trust in next-generation biomanufacturing technologies.

3.5. Study Limitations

This study has some limitations that should be acknowledged. First, the bibliometric dataset relies exclusively on the Web of Science Core Collection, which may exclude relevant contributions indexed in other major scientific databases. Second, the analysis covers the period 2018–2025, meaning that earlier foundational work and very recent publications indexed after the search date are not represented. Third, the predominance of review-type papers in the dataset may bias thematic interpretations toward conceptual consolidation rather than experimental advances. Fourth, the clustering and mapping outcomes depend on CiteSpace parameters, meaning that different pruning strategies or time slicing could yield slightly different network structures. Finally, the qualitative synthesis focuses only on the 25 most-cited papers, which provides depth but does not fully capture the diversity of less-cited but innovative research emerging at the fringes of the field. These limitations should be considered when interpreting the study’s findings and their generalisability.

4. Conclusions

The integration of AI into Biomedical 3D Printing represents a transformative step toward intelligent, adaptive, and personalised fabrication systems. Through a bibliometric analysis of 229 publications from 2018 to 2025, this study reveals a consistent, accelerating growth in research interest at the intersection of computational modelling, materials science, and biomedical engineering.

The evolution of the field follows a clear trajectory from early algorithmic and process-control studies to the current phase of deep learning–assisted bioprinting and 4D adaptive materials. Core research clusters—focused on AI-driven design optimisation, scaffold engineering, and shape-memory polymers—reflect a shift from static material fabrication to dynamic, self-responsive biomedical systems. Countries such as China, India, and the United States have emerged as the most active contributors, supported by globally recognised journals including Applied Sciences, Polymers, and Advanced Materials.

The results highlight an ongoing convergence of intelligent manufacturing and regenerative medicine, supported by advances in prediction, monitoring, and materials adaptability. Future research will likely focus on developing explainable AI frameworks, enhancing real-time control of bioprinting processes, and ensuring scalability, reproducibility, and ethical integration of AI technologies within clinical applications.

In addition, the qualitative synthesis presented in Section 3.4 enriches these findings by revealing how the most cited studies collectively shaped the field’s evolution—from algorithmic design optimisation to predictive, data-driven biomanufacturing. This interpretative perspective complements the quantitative mapping and reinforces the understanding of AI-assisted additive manufacturing as a conceptually mature and scientifically coherent domain.

Overall, this study establishes the conceptual and structural foundation for understanding AI-driven biomedical additive manufacturing as a coherent, data-centric, and evolving research ecosystem.