Diagnostics

27 pages, 4440 KB

Open AccessArticle

Optimization-Driven Hybrid Machine Learning Framework for Brain Tumor Classification in MRI with Metaheuristic Feature Selection

by Yasin Özkan, Yusuf Bahri Özçelik and Aytaç Altan

Diagnostics 2026, 16(5), 819; https://doi.org/10.3390/diagnostics16050819 - 9 Mar 2026

Cited by 2 | Viewed by 716

Abstract

Background/Objectives: Brain tumors are among the most severe neurological disorders, and their variability in size, morphology, and anatomical location complicates early and accurate diagnosis. Although magnetic resonance imaging (MRI) is the most reliable non-invasive modality for tumor detection, manual interpretation remains time-consuming, subjective, [...] Read more.

Background/Objectives: Brain tumors are among the most severe neurological disorders, and their variability in size, morphology, and anatomical location complicates early and accurate diagnosis. Although magnetic resonance imaging (MRI) is the most reliable non-invasive modality for tumor detection, manual interpretation remains time-consuming, subjective, and susceptible to human error. This study aims to develop an optimization-driven hybrid machine learning framework for accurate and computationally efficient automatic brain tumor classification. Methods: The dataset includes 834 MRI images (583-training, 123-validation, 128-independent test). Because YOLOv11 detects tumor and non-tumor regions separately, the sample size doubled during region-based analysis, and all subsequent stages were conducted at the regions of interest (ROI) level. On the independent test set, YOLOv11 achieved 98.87% mAP@50, 98.54% precision, and 98.21% recall. The proposed framework combines automated tumor localization with image standardization using Gaussian noise reduction and bilinear interpolation. From the processed MR images, 39 entropy-based features were extracted. To enhance diagnostic performance and eliminate redundant information, the superb fairy-wren optimization algorithm (SFOA) was applied for feature selection and compared with particle swarm optimization (PSO), Harris hawk optimization (HHO), and puma optimization (PO). Final classification was primarily performed using k-nearest neighbors (kNN), while support vector machines (SVM) were used for comparative evaluation. Results: SFOA reduced the feature dimensionality from 39 to 5 features while achieving 99.20% classification accuracy on the independent test set. In comparison, PSO selected 10 features, HHO selected 6 features and PO selected 10 features, all achieving 98.45% accuracy. The best performance obtained with SVM was 98.45% accuracy (HHO-SVM), which remained lower than the 99.20% achieved by the proposed SFOA-kNN model. Conclusions: The results indicate that combining entropy-based feature extraction with SFOA-driven feature selection and kNN classification significantly enhances diagnostic accuracy while reducing computational complexity, highlighting the strong potential of the proposed framework for integration into computer-aided diagnosis systems to support clinical decision-making. Full article

(This article belongs to the Special Issue 3rd Edition: AI/ML-Based Medical Image Processing and Analysis)

► Show Figures

Figure 1

13 pages, 856 KB

Open AccessArticle

Sun-Exposed vs. Non-Sun-Exposed Areas: Epidemiology and Pathogenesis of Non-Metastatic Merkel Cell Carcinoma

by Alexander Yakobson, Ronen Brenner, Hanna T. Frumin Edri, Anna Ievko, Sofiia Turaieva, Tanzilya Tairov, Ilia Berezhnov, Shlomit Fenig, Eyal Fenig, Tomer Ziv-Baran, Sabri El-Saied and Walid Shalata

Diagnostics 2026, 16(5), 818; https://doi.org/10.3390/diagnostics16050818 - 9 Mar 2026

Cited by 1 | Viewed by 515

Abstract

Background: Merkel cell carcinoma (MCC) is a rare and aggressive cutaneous neuroendocrine malignancy. The prognostic impact of sun exposure at the primary tumor site in localized and locally advanced MCC remains incompletely defined. We aimed to compare clinicopathologic characteristics and survival outcomes between [...] Read more.

Background: Merkel cell carcinoma (MCC) is a rare and aggressive cutaneous neuroendocrine malignancy. The prognostic impact of sun exposure at the primary tumor site in localized and locally advanced MCC remains incompletely defined. We aimed to compare clinicopathologic characteristics and survival outcomes between sun-exposed and non-sun-exposed MCC in a large, multi-center Israeli cohort. Methods: We retrospectively identified 249 patients diagnosed with localized or locally advanced MCC between January 1985 and December 2020. Of these, 225 patients met eligibility criteria and were included in the analysis: 142 with sun-exposed primary tumors (cohort A) and 83 with non-sun-exposed tumors (cohort B). Baseline characteristics included age, sex, tumor size, lymph node (LN) involvement at diagnosis, disease-free survival (DFS), and overall survival (OS). Results: Median age at diagnosis was similar between cohorts (~73 years), with a male predominance in both groups. LN involvement was significantly more frequent in non-sun-exposed tumors compared with sun-exposed tumors (57.0% vs. 30.0%, p < 0.001), while tumor size distribution did not differ significantly. Median DFS was numerically longer in sun-exposed patients (58.0 vs. 47.8 months, p ≈ 0.18), whereas median OS favored non-sun-exposed patients (89.7 vs. 79.7 months, p ≈ 0.21), though neither difference reached statistical significance overall. Females demonstrated longer DFS and OS than males across both cohorts. Among LN-negative patients, non-sun-exposed tumors were associated with significantly improved OS (105.9 vs. 91.4 months, p ≈ 0.03), particularly in males. Primary tumor size further stratified outcomes: non-sun-exposed patients had significantly superior OS for tumors <2 cm and both improved DFS and OS for tumors ≥2 cm. Conclusions: In this large real-world MCC cohort, sun exposure status was associated with distinct patterns of nodal involvement and survival in clinically relevant subgroups. Non-sun-exposed MCC demonstrated favorable survival outcomes, particularly in LN-negative disease and across tumor size categories, suggesting underlying biological differences that merit further investigation. Full article

(This article belongs to the Special Issue Advances in Diagnosis and Management of Skin Diseases)

► Show Figures

Figure 1

12 pages, 5200 KB

Open AccessArticle

Comparison of Manual, Semi-Automatic, and Automatic CT-Based Methods for Liver Volume Segmentation

by Berna Dogan, Sadik Bugrahan Simsek, Sefa Sonmez, Merve Nur Ozgen Sonmez, Omur Dasci and Zafer Ozmen

Diagnostics 2026, 16(5), 817; https://doi.org/10.3390/diagnostics16050817 - 9 Mar 2026

Viewed by 619

Abstract

Background/Objectives: To evaluate whether semi-automatic and automatic CT-based liver segmentation methods can provide clinically acceptable volumetric agreement compared with manual segmentation while improving processing efficiency in routine practice. Methods: CT images from 86 individuals were retrospectively analyzed. Liver volumes were calculated [...] Read more.

Background/Objectives: To evaluate whether semi-automatic and automatic CT-based liver segmentation methods can provide clinically acceptable volumetric agreement compared with manual segmentation while improving processing efficiency in routine practice. Methods: CT images from 86 individuals were retrospectively analyzed. Liver volumes were calculated using manual segmentation, RVX Semi-Automatic, RVX Deep Learning, and TotalSegmentator. Differences among methods were assessed using repeated-measures ANOVA. Agreement with manual segmentation was evaluated using a Bland–Altman analysis, while the Dice Similarity Coefficient (DICE) and Hausdorff Distance (HD) quantified spatial overlap and boundary deviation, respectively. Processing times were recorded. Results: Mean liver volumes were 1503.9 ± 356.0 cm³ (manual), 1512.6 ± 373.6 cm³ (RVX Semi-Automatic), 1549.8 ± 367.9 cm³ (RVX Deep Learning), and 1518.3 ± 365.8 cm³ (TotalSegmentator). RVX Deep Learning produced significantly higher volumes compared with manual segmentation (p < 0.001), whereas RVX Semi-Automatic and TotalSegmentator showed no significant differences (p > 0.05). DICE values were 0.911 ± 0.032, 0.946 ± 0.018, and 0.938 ± 0.021 for RVX Semi-Automatic, RVX Deep Learning, and TotalSegmentator, respectively. HD values were highest for the deep learning-based method. Processing times were shortest for RVX Deep Learning and longest for manual segmentation. Conclusions: Semi-automatic and automatic liver segmentation methods substantially reduce processing time while maintaining clinically acceptable volumetric agreement. Among the evaluated approaches, TotalSegmentator showed the closest agreement with manual segmentation, supporting its use in routine CT-based liver volumetry. Deep learning-based segmentation, although faster, tended to overestimate volume, potentially limiting its use in applications requiring high volumetric precision. Full article

(This article belongs to the Special Issue Recent Advances in Abdominal Imaging)

► Show Figures

Figure 1

12 pages, 10516 KB

Open AccessArticle

Automated Assessment of Ki-67 Labeling Index Using Cell-Level Detection and Classification in Whole-Slide Images

by Masayuki Tsuneki, Meng Li and Fahdi Kanavati

Diagnostics 2026, 16(5), 816; https://doi.org/10.3390/diagnostics16050816 - 9 Mar 2026

Viewed by 1023

Abstract

Background: The Ki-67 labeling index (LI) is a widely used marker of tumour proliferation, yet its manual assessment is time-consuming and subject to substantial inter-observer variability. Automated methods may improve reproducibility, but their clinical relevance depends on achieving performance comparable to expert [...] Read more.

Background: The Ki-67 labeling index (LI) is a widely used marker of tumour proliferation, yet its manual assessment is time-consuming and subject to substantial inter-observer variability. Automated methods may improve reproducibility, but their clinical relevance depends on achieving performance comparable to expert pathologists. Method: We evaluated an artificial intelligence (AI)-based, cell-level system for automated Ki-67 LI assessment that detects and classifies individual tumour cell nuclei as Ki-67-positive or -negative. After nuclear detection using a pre-existing cell detection model, a lightweight convolutional neural network classifier operating on nucleus-centred patches was trained, and then applied to cases independently assessed by three pathologists. Agreement between AI-derived and human Ki-67 LI values was compared directly with inter-pathologist agreement across a range of proliferation levels. Results: The AI-based cell classification achieved 98% AUC on a test set consisting of 71K positive and 170K negative image patches centred on nuclei. On the automated Ki-67 LI assessment, the AI system demonstrated concordance with expert pathologists comparable to human inter-observer variability. Conclusions: These results support the potential of cell-level automated Ki-67 assessment as a reproducible decision-support tool for routine histopathological practice. Full article

(This article belongs to the Special Issue Artificial Intelligence in Pathological Image Analysis—2nd Edition)

► Show Figures

Figure 1

19 pages, 3435 KB

Open AccessArticle

Glaucoma Classification Using a NFNet-Based Deep Learning Model with a Customized Hybrid Attention Mechanism

by Sandeep Angara, Loc Tran and Jongwoo Kim

Diagnostics 2026, 16(5), 815; https://doi.org/10.3390/diagnostics16050815 - 9 Mar 2026

Cited by 1 | Viewed by 626

Abstract

Background/Objectives: Glaucoma is a leading cause of irreversible blindness worldwide, making accurate and efficient detection methods essential. One primary concern with glaucoma is that it often presents no early symptoms. Vision loss typically begins at the periphery and progresses unnoticed until it significantly [...] Read more.

Background/Objectives: Glaucoma is a leading cause of irreversible blindness worldwide, making accurate and efficient detection methods essential. One primary concern with glaucoma is that it often presents no early symptoms. Vision loss typically begins at the periphery and progresses unnoticed until it significantly affects central vision. Due to this gradual and usually silent progression, early detection through regular eye exams is vital for preventing permanent vision loss. Methods: In this study, we propose a hybrid attention mechanism that recalibrates feature maps from the feature extractor for glaucoma detection. We explored normalization-free ResNet (NF-ResNet) architectures to evaluate the proposed attention mechanism, specifically NF-ResNet-26, NF-ResNet-50, and NF-ResNet-101, in comparison to baseline state-of-the-art ResNet variants. Our approach was evaluated on three publicly available glaucoma datasets, LAG, EyePACS, and BrG, to differentiate between normal and glaucomatous from fundus images. Results: The experimental results demonstrate that our proposed hybrid attention module, combined with normalization-free architectures, significantly enhances performance compared to state-of-the-art ResNet variants. The proposed attention model based on the normalization-free ResNet-50 achieved an accuracy of 0.9394 on the LAG dataset, 0.9117 on the EyePACS dataset, and 0.9020 on the BrG dataset. When evaluated on the combined dataset, the model achieved an accuracy of 0.9193, sensitivity of 0.9182, and specificity of 0.9202. Conclusions: The results from these representative datasets for glaucoma detection highlight the exceptional performance of our attention module, establishing it as a highly competitive classification model in the field of glaucoma detection. Full article

(This article belongs to the Section Machine Learning and Artificial Intelligence in Diagnostics)

► Show Figures

Figure 1

21 pages, 1380 KB

Open AccessArticle

Association Between Serum Testosterone Levels and Coronary Artery Stenosis: A Cross-Sectional Study in Central European Population

by Pavol Fülöp, Zuzana Pella, Tibor Porubän, Peter Hreško, František Pavol Zajac, Mariana Dvorožňáková, Štefan Tóth and Dominik Pella

Diagnostics 2026, 16(5), 814; https://doi.org/10.3390/diagnostics16050814 - 9 Mar 2026

Viewed by 583

Abstract

Background: The relationship between testosterone and coronary artery disease (CAD) remains a subject of debate. Most studies suggest an inverse association—lower testosterone, higher risk. However, data from Central European populations undergoing coronary angiography are limited. Objectives: To investigate the association between [...] Read more.

Background: The relationship between testosterone and coronary artery disease (CAD) remains a subject of debate. Most studies suggest an inverse association—lower testosterone, higher risk. However, data from Central European populations undergoing coronary angiography are limited. Objectives: To investigate the association between serum testosterone levels and angiographically confirmed coronary artery stenosis in a Slovak population. Methods: This cross-sectional study included 129 consecutive stable patients (84 men, 45 women; mean age 64.3 ± 9.7 years) undergoing elective coronary angiography for suspected stable coronary artery disease. Significant coronary stenosis was defined as ≥50% luminal narrowing in any major epicardial vessel. Serum testosterone, lipid profile, and traditional risk factors were assessed. Univariate and multivariate logistic regression models were constructed to evaluate independent associations of coronary stenosis. Results: Coronary stenosis ≥ 50% was present in 74 patients (57.4%). Notably, patients with stenosis had significantly higher testosterone levels (6.62 ± 2.79 vs. 4.85 ± 3.50 ng/mL, p = 0.002). In univariate analysis, testosterone showed a significant association (OR 1.197 per ng/mL, OR 1.784 per SD, p = 0.003). In multivariate analysis adjusted for age, sex, diabetes mellitus, and LDL (low-density lipoprotein) cholesterol, testosterone remained independently associated (adjusted OR 2.043 per SD, 95% CI 1.221–3.420, p = 0.007), as did diabetes mellitus (OR 2.60, p = 0.032). Conclusions: Elevated serum testosterone is paradoxically associated with increased prevalence of coronary stenosis in our cohort. These findings from stable, chronic CAD patients may work fundamentally differently from what is observed in acute coronary syndromes, where stress-induced testosterone suppression may confound observed associations. Full article

(This article belongs to the Section Clinical Diagnosis and Prognosis)

► Show Figures

Figure 1

12 pages, 938 KB

Open AccessArticle

Assessment of Optimal Stent Implantation with the Use of Optical Coherence Tomography in Patients with Coronary Artery Disease

by Alexandros Kaperonis, Alexandru Scafa-Udriște, Cosmin Mihai, Vlad Bataila, Bogdan Marian Drăgoescu, Vlad Ploscaru, Diana Zamfir, Radu Popescu, Daniel Tonu and Lucian Calmac

Diagnostics 2026, 16(5), 813; https://doi.org/10.3390/diagnostics16050813 - 9 Mar 2026

Viewed by 697

Abstract

Background/Objective: Percutaneous coronary intervention (PCI) has a pivotal role in the treatment of coronary artery disease (CAD). Although PCI is generally guided only angiographically, advancements in intravascular imaging, particularly in optical coherence tomography (OCT), may offer significant advantages. OCT provides high-resolution cross-sectional [...] Read more.

Background/Objective: Percutaneous coronary intervention (PCI) has a pivotal role in the treatment of coronary artery disease (CAD). Although PCI is generally guided only angiographically, advancements in intravascular imaging, particularly in optical coherence tomography (OCT), may offer significant advantages. OCT provides high-resolution cross-sectional images that allow for a more detailed assessment of lesion characteristics and procedural outcomes, which are not fully available with angiography. These findings are associated with or predictive of major adverse cardiovascular events (MACE), encouraging the use of OCT in PCI procedures. This study sought to characterize the role of post-PCI OCT imaging in PCI optimization in patients with CAD. Methods: This retrospective study includes patients who underwent OCT-guided PCI. A total of 64 patients with various types of CAD were included. The primary endpoint was the identification of suboptimal stent implantation as evaluated with OCT after stent implantation, and the secondary endpoint was the assessment of the possibility to achieve optimal stent implantation after further OCT-guided optimization based on standard definitions of optimal PCI. Results: In total, 73 vessels were studied, 42.46% (31) had a stent expansion index (SEI) of < 80%, 31.51% (23) had an SEI between 80–90%, and 26.03% (19) had an SEI of more than 90%. Minimum stent area (MSA) of more than 4.5 mm² was found in 82.19% (60) of vessels, while 17.80% (13) had an MSA below the cut-off value. Suboptimal stent implantation was identified in 35.61% (26) of vessels, including underexpansion 9.58% (7), malapposition 15.06% (11), stent edge dissection 6.85% (5), plaque burden or lipid-rich pool in the stent edges 2.73% (2), and tissue protrusion 1.36% (1). Post-PCI OCT optimization resulted in significant improvements, with only 6.84% (5) of the vessels still not achieving all OCT criteria for optimal stent implantation. Conclusions: In patients with CAD, post-PCI OCT evaluation provided useful information, otherwise unavailable by angiography alone. We identified that 35.61% (26) of the targeted vessels, were suboptimally stented. OCT imaging was able to provide procedural and strategic guidance for optimization until the appropriate results, based on our criteria, were achieved in most of the lesions. Full article

(This article belongs to the Special Issue Multimodal Cardiac Imaging: Diagnostic and Prognostic Advances)

► Show Figures

Figure 1

11 pages, 574 KB

Open AccessArticle

Evaluation of the Impact of Different Skeletal Orthodontic Anomalies on Condylar Asymmetry Using Cone-Beam Computed Tomography

by Muhammet Bahattin Bingul, Seda Kotan, Saadet Cinarsoy Cigerim and Mevlude Yuce Polat

Diagnostics 2026, 16(5), 812; https://doi.org/10.3390/diagnostics16050812 - 9 Mar 2026

Viewed by 495

Abstract

Background/Objectives: This study aims to evaluate mandibular condylar asymmetry in individuals with different types of skeletal malocclusions using a three-dimensional imaging technique, and to determine the relationship between these anomalies and condylar asymmetry. Methods: The study included 100 individuals who visited [...] Read more.

Background/Objectives: This study aims to evaluate mandibular condylar asymmetry in individuals with different types of skeletal malocclusions using a three-dimensional imaging technique, and to determine the relationship between these anomalies and condylar asymmetry. Methods: The study included 100 individuals who visited the Department of Orthodontics Faculty of Dentistry between 2015 and 2020 and underwent Cone-Beam Computed Tomography (CBCT) imaging for various reasons. The evaluation of condylar asymmetry was performed using the Habets method, and measurements were carried out with the NemoCeph V.2017 software. Participants were categorized into skeletal Class I (2–4°), Class II (>4°), and Class III based on their ANB angles. For statistical analysis, frequency distribution, the Kruskal–Wallis H test, and Spearman’s correlation coefficient were used. Results: No statistically significant relationship was found between gender and skeletal classifications (p > 0.05). In terms of age, the mean age of individuals in the Class III group was significantly lower than that of those in the Class II group (p < 0.05). A weak positive correlation was observed between condylar and ramal indices in the overall sample (p = 0.029); however, this correlation was found to be moderate and statistically significant only within the Class III group (p = 0.002). Conclusions: The presence of a significant relationship between condylar and ramal asymmetries in individuals with Class III malocclusion indicates an increased risk of developing facial asymmetry if left untreated. These findings underscore the importance of skeletal malocclusions in influencing condylar morphology. Full article

(This article belongs to the Special Issue Diagnosis and Management in Oral and Maxillofacial Surgery)

► Show Figures

Figure 1

11 pages, 5084 KB

Open AccessArticle

AI-Assisted OCT Imaging for Core Needle Biopsy Guidance: The 1st in Humans Study

by Nicusor Iftimia, Poonam Yadav, Michael Primrose, Gopi Maguluri, Jack Jones, John Grimble and Rahul Anil Sheth

Diagnostics 2026, 16(5), 811; https://doi.org/10.3390/diagnostics16050811 - 9 Mar 2026

Viewed by 667

Abstract

Background: The heterogeneous nature of cancer with varying degrees of fat, necrosis, fibrosis, and varying degrees of tissue repair severely impacts the success of acquiring adequate tissue samples during percutaneous image-guided biopsy. Although ultrasound or CT fluoroscopy are used to identify tumor [...] Read more.

Background: The heterogeneous nature of cancer with varying degrees of fat, necrosis, fibrosis, and varying degrees of tissue repair severely impacts the success of acquiring adequate tissue samples during percutaneous image-guided biopsy. Although ultrasound or CT fluoroscopy are used to identify tumor location and thus to guide biopsy needle insertion, these technologies do not provide the necessary resolution to determine tissue composition and enable the selection of the most appropriate location for biopsy specimen extraction. As a result, biopsy must be repeated, leading to significant cost to the health care system. Methods: In this study, we introduce a combined optical imaging/artificial intelligence (OI/AI) methodology for the real-time assessment of tissue morphology at the tip of the biopsy needle, prior to the collection of a biopsy specimen. Addressing a significant clinical challenge, this approach aims to reduce the proportion of biopsy cores—currently as high as 40%—that yield low diagnostic value due to elevated adipose or low tumor content. Our methodology employs micron-scale optical coherence tomography (OCT) imaging to obtain detailed structural tissue information using a minimally invasive needle probe. The OCT images are automatically analyzed using a convolutional neural network (CNN)-driven AI software developed by our team. A U-net style architecture was used to segment regions of tumor from the OCT scans. U-Net is a specialized convolutional neural network (CNN) architecture designed for fast, precise image segmentation, which involves classifying each pixel in an image to outline objects. This streamlined approach shows promise to provide clinicians with real-time results, supporting more accurate and informed decisions regarding biopsy site selection. To evaluate this technology, we conducted a clinical study using a custom-made OCT imager and recorded OCT images from patients diagnosed with liver cancers. Expert OCT interpreters supplied annotated reference images that were used to train a custom AI algorithm. Results: OCT imaging with ~10 mm axial and 20 mm lateral resolution enabled the collection of high-quality images of the tissue. The AI analysis was performed offline. UNet achieved an AUC of ~0.877 on the validation dataset, indicating promising performance for the relatively small data set used to train the model. The AI model matched human interpretations approximately 90% of the time, highlighting its promise for making biopsy procedures both more accurate and more efficient. Conclusions: A novel OCT instrument and AI software were evaluated for assessing tissue composition at the tip of biopsy needle. The OCT instrument produced micron-scale resolution images of the tissue, enabling AI analysis and accurate real-time discrimination of tissue type. This preliminary study demonstrated the clinical potential of this technology for improving biopsy success. Full article

(This article belongs to the Special Issue Expanding Horizons: Optical Coherence Tomography in Cross-Specialty Diagnostic Practice)

► Show Figures

Figure 1

23 pages, 1189 KB

Open AccessArticle

Atherogenic Lipid Indices in Colorectal Cancer: Metabolic Associations and Survival Outcomes

by Răzvan Alexandru Marinescu, Daniela Marinescu, Lidia Boldeanu, Ana-Maria Ciurea, Marius Bică, Ștefan Pătrașcu, Victor Dan Eugen Strâmbu, Petru Adrian Radu, Petrica Popa, Mohamed-Zakaria Assani, Mihail Virgil Boldeanu and Valeriu Șurlin

Diagnostics 2026, 16(5), 810; https://doi.org/10.3390/diagnostics16050810 - 9 Mar 2026

Viewed by 583

Abstract

Background/Objectives: Type 2 diabetes mellitus (T2DM) and atherogenic dyslipidemia have been implicated in colorectal cancer (CRC) development, but their prognostic relevance after cancer diagnosis remains unclear. This study aimed to evaluate the association between T2DM, lipid-derived atherogenic indices, and survival outcomes in patients [...] Read more.

Background/Objectives: Type 2 diabetes mellitus (T2DM) and atherogenic dyslipidemia have been implicated in colorectal cancer (CRC) development, but their prognostic relevance after cancer diagnosis remains unclear. This study aimed to evaluate the association between T2DM, lipid-derived atherogenic indices, and survival outcomes in patients with CRC. Methods: We conducted a retrospective cohort study including 240 CRC patients, of whom 60 had coexisting T2DM. Overall survival (OS) and disease-free survival (DFS) were analyzed using the Kaplan–Meier (KM) method and log-rank tests. In the absence of recurrence-specific data, DFS was defined as time to death or last follow-up. Lipid-related indices, including the atherogenic index of plasma (AIP), atherogenic coefficient (AC), remnant cholesterol (RC), non-high-density lipoprotein cholesterol (non-HDL-C), triglyceride–glucose (TyG) index, and triglyceride-to-HDL cholesterol ratio (TG/HDL-C), were evaluated by tertiles in KM analyses. Multivariable Cox proportional hazards models were constructed to assess the independent prognostic value of AIP, AC, and RC (entered separately as a continuous variable standardized to 1 standard deviation), adjusted for age, sex, adjuvant chemotherapy, radiotherapy, and T2DM status. Sensitivity analyses were performed in stage III–IV patients. Results: During follow-up, 28 deaths occurred. OS did not differ significantly between CRC patients and those with CRC coexisting with T2DM (log-rank p-values = 0.220). DFS analyses showed no significant differences across tertiles of any lipid-related index (all log-rank p-values > 0.05), with overlapping survival curves and no consistent dose–response patterns. In adjusted Cox models, AIP (hazard ratio [HR] per 1 SD = 0.71, 95% CI 0.48–1.06), AC (HR = 0.72, 95% CI 0.44–1.20), and RC (HR = 0.66, 95% CI 0.39–1.12) were not independently associated with DFS. Results were consistent in advanced-stage disease (stage III–IV). Conclusions: In this cohort of patients with CRC, neither T2DM nor lipid-derived indices reflecting atherogenic dyslipidemia and insulin resistance were independently associated with OS or DFS. These findings help refine the clinical interpretation of lipid-derived biomarkers in CRC, suggesting limited prognostic utility beyond established oncologic factors. Full article

(This article belongs to the Special Issue Diagnosis and Management of Colorectal Diseases, 2nd Edition)

► Show Figures

Figure 1

14 pages, 1751 KB

Open AccessArticle

Morphological Variability of Sphenoid Sinus Pneumatization and Its Impact on Adjacent Neurovascular Structures

by Panagiotis Papadopoulos-Manolarakis, George Triantafyllou, Christos Georgalas, Ioannis Paschopoulos, George Stranjalis and Maria Piagkou

Diagnostics 2026, 16(5), 809; https://doi.org/10.3390/diagnostics16050809 - 9 Mar 2026

Viewed by 595

Abstract

Background/Objectives: The sphenoid sinus (SS) exhibits marked morphological variability, influencing the relationship of critical neurovascular skull base structures. This study aimed to characterize sphenoid sinus pneumatization (SSP) patterns and assess their impact on the course of the internal carotid artery (ICA), optic [...] Read more.

Background/Objectives: The sphenoid sinus (SS) exhibits marked morphological variability, influencing the relationship of critical neurovascular skull base structures. This study aimed to characterize sphenoid sinus pneumatization (SSP) patterns and assess their impact on the course of the internal carotid artery (ICA), optic nerve (ON), Vidian nerve (VN), and maxillary nerve (MN) within a Greek adult population. Methods: A retrospective analysis of 253 adult skull base computed tomography (CT) scans was performed. The degree and direction of SSP were classified according to established radiological criteria. Anterior, lateral, and posterior extensions were evaluated. The course of adjacent neurovascular structures was categorized as typical, protruding, or dehiscent. Associations between pneumatization types and neurovascular variants were analyzed. Results: The sellar complete type was the predominant SS pattern (63.2%), followed by sellar incomplete (27.7%) and presellar (8.7%) types; agenesis was rare (0.4%). Posterior (63.6%) and lateral (46.6%) extensions were most common. Lateral and posterior pneumatization significantly correlated with protrusion and/or dehiscence of adjacent neurovascular structures, particularly the ICA, ON, and VN. LW extension was strongly associated with ON protrusion (96%), while PP and full-lateral extensions correlated with VN protrusion (56.1% and 79.9%, respectively). No significant sex- or side-related differences were identified. Conclusions: SSP demonstrates extensive morphological variability that significantly affects the anatomical course and osseous coverage of neighboring neurovascular structures. Comprehensive preoperative CT evaluation of SS anatomy is essential for planning endoscopic transsphenoidal and extended skull base procedures to minimize the risk of neurovascular injury. Full article

(This article belongs to the Special Issue Brain/Neuroimaging 2025–2026)

► Show Figures

Figure 1

19 pages, 591 KB

Open AccessArticle

Neurocognitive Correlates of Diagnostic Heterogeneity in Children with ADHD: The Differential Contributions of Cognitive Disengagement Syndrome, Symptom Severity, and Anxiety

by İbrahim Adak, Esin Özdeniz Varan, Nergis Eyüpoğlu, Ayşim Alpman, Zeynep Durmuş, Oğuz Bilal Karakuş, İpek Süzer Gamlı and Özalp Ekinci

Diagnostics 2026, 16(5), 808; https://doi.org/10.3390/diagnostics16050808 - 9 Mar 2026

Viewed by 601

Abstract

Background/Objectives: Attention-Deficit/Hyperactivity Disorder (ADHD) shows substantial cognitive heterogeneity, complicating individualized clinical formulation. This study examined whether Cognitive Disengagement Syndrome (CDS), anxiety, and ADHD symptom severity are associated with memory functions and visuospatial skills in children with ADHD. Methods: The sample included 120 children [...] Read more.

Background/Objectives: Attention-Deficit/Hyperactivity Disorder (ADHD) shows substantial cognitive heterogeneity, complicating individualized clinical formulation. This study examined whether Cognitive Disengagement Syndrome (CDS), anxiety, and ADHD symptom severity are associated with memory functions and visuospatial skills in children with ADHD. Methods: The sample included 120 children aged 6–12 years with ADHD (ADHD + CDS: n = 40; ADHD-only: n = 80). Memory was assessed with the Oktem Verbal Memory Processes Test (OVMPT) and Wechsler Memory Scale–Visual Reproduction (WMS–VR), and visuospatial skills with WISC-IV Block Design and Judgment of Line Orientation (JLO). ADHD symptoms were rated using combined parent–teacher Turgay-Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition-Based Disruptive Behavior Disorders Scale (T-DSM-IV-S) scores; CDS symptoms with the Barkley Child Attention Scale; and anxiety with the SCARED-Child Form. Group comparisons, correlation analyses, and multivariable linear regression models were conducted. Results: The ADHD + CDS group performed worse on WISC-IV Block Design than the ADHD-only group (p = 0.005). In the ADHD + CDS group, inattention severity showed a strong negative association with WMS–VR short-term memory (r = −0.560, p < 0.001). In the ADHD-only group, inattention severity was negatively associated with OVMPT Spontaneous Recall (ρ = −0.319, p = 0.004) and JLO total score (ρ = −0.348, p = 0.002). Anxiety severity in the ADHD-only group was positively associated with OVMPT Total Learning (ρ = 0.350, p = 0.001), Highest Learning (ρ = 0.370, p = 0.001), and WMS–VR short-term memory (ρ = 0.304, p = 0.006). In regression analyses, the presence of CDS independently and negatively predicted WMS–VR short-term memory (β = −0.187, p = 0.018) and Block Design performances (β = −0.226, p = 0.016). Inattention symptom severity was also independently and negatively associated with Block Design performance (β = −0.243, p = 0.013). Conclusions: CDS status and symptom dimensions contribute to cognitive variability in pediatric ADHD, with CDS showing independent associations with timed visuospatial construction and short-term visual memory. Inattention severity emerged as a robust dimensional predictor of cognitive inefficiency across domains, supporting the clinical utility of symptom-based cognitive profiling in ADHD diagnostic evaluations. In addition, mild anxiety symptoms demonstrated meaningful associations with some learning and memory performances within the ADHD-only group, indicating that affective factors may modulate cognitive outcomes in ADHD. Taken together, these findings support considering CDS status and symptom dimensions jointly when characterizing cognitive variability in ADHD. Full article

(This article belongs to the Special Issue Advances in the Diagnosis and Management of Neuropsychiatric Disorders—2nd Edition)

► Show Figures

Figure 1

19 pages, 1345 KB

Open AccessArticle

A Novel Dual-Modality Dual-View Hybrid Deep Learning–Machine Learning Framework for the Prediction of Carotid Plaque Vulnerability via Late Fusion

by Wenxuan Zhang, Chao Hou, Xinyi Wang, Hongyu Kang, Shuai Li, Yu Sun, Yongping Zheng, Wei Zhang and Sai-Kit Lam

Diagnostics 2026, 16(5), 807; https://doi.org/10.3390/diagnostics16050807 - 9 Mar 2026

Viewed by 734

Abstract

Background: Ultrasound imaging is an ideal tool for regular carotid plaque screening to identify individuals at high risk of stroke for clinical intervention. However, no existing study leverages multi-modal multi-view ultrasound imaging for AI-enabled auto-classification of carotid plaque vulnerability. This study aims [...] Read more.

Background: Ultrasound imaging is an ideal tool for regular carotid plaque screening to identify individuals at high risk of stroke for clinical intervention. However, no existing study leverages multi-modal multi-view ultrasound imaging for AI-enabled auto-classification of carotid plaque vulnerability. This study aims to develop and validate an effective AI model for carotid plaque vulnerability classification through the applications of dual-modal (B-Mode and contrast-enhanced mode) dual-view (longitudinal and cross-sectional) settings to maximize the utility and potential of ultrasound imaging. Methods: Hybrid deep-learning (DL) and machine-learning (ML) methods were employed to balance between model discriminability and interpretability. B-Mode ultrasound (BMUS) and contrast-enhanced ultrasound (CEUS) images from 241 patients were retrospectively analyzed using the proposed hybrid-DL-ML variants. Results: Our findings suggest the hybrid VGG-RF model developed from a dual-modal dual-view setting outperforms those developed from other settings for identifying vulnerable carotid plaques. The VGG-RF model emerged as the best-performing model, achieving an optimal performance with an AUC of 0.908, precision of 0.765, recall of 0.929, specificity of 0.886, and F1 score of 0.839. The inherent interpretability of the VGG-RF model divulged that long-axis views of BMUS and CEUS images were the major contributing features for discriminating vulnerable carotid plaques against their counterparts. Conclusions: The present study underscored the effectiveness of AI models developed from dual-modal dual-view settings of ultrasound images. Notably, the hybrid VGG-RF model was benchmarked as the best-performing model among other studied hybrid DL-ML variants. Further studies on a larger cohort in a prospective setting are warranted to validate the findings of the current study. Full article

(This article belongs to the Special Issue Artificial Intelligence in Medical Imaging: Innovations and Diagnostic Applications)

► Show Figures

Figure 1

24 pages, 1806 KB

Open AccessReview

Fetal Growth Restriction: Contemporary Evidence to Guide Delivery Timing and Intrapartum Management

by Ana Carolina Rabachini Caetano, Ana Cristina Perez Zamarian, Luciano Marcondes Machado Nardozza, Seizo Miyadahira, Giselle Darahem Tedesco, Lara Dariolli Rossi, Gustavo Yano Callado, Edward Araujo Júnior and Alessandra Cristina Marcolin

Diagnostics 2026, 16(5), 806; https://doi.org/10.3390/diagnostics16050806 - 9 Mar 2026

Viewed by 1328

Abstract

Fetal growth restriction (FGR), a condition in which the fetus fails to achieve its growth and developmental potential, affects 5% to 10% of pregnancies and is associated with high rates of perinatal morbidity and mortality. There is currently insufficient high-quality evidence to define [...] Read more.

Fetal growth restriction (FGR), a condition in which the fetus fails to achieve its growth and developmental potential, affects 5% to 10% of pregnancies and is associated with high rates of perinatal morbidity and mortality. There is currently insufficient high-quality evidence to define the optimal approach for diagnosing fetal growth restriction. In 2016, with the aim of standardizing clinical practice and enabling comparability across scientific studies, an expert opinion-based consensus was published. This document proposed unified terminology and clear diagnostic criteria for early- and late-onset fetal growth restriction (FGR). Because no effective treatment is available, careful assessment of fetal well-being and appropriate timing of delivery are the main tools for managing these fetuses. This decision should be based on gestational age and the severity of abnormalities identified on fetal surveillance tests, balancing the risks of prematurity against the risks of severe permanent sequelae or fetal death. The objective of this update is to analyze the most recent evidence on when and how to deliver pregnancies complicated by fetal growth restriction, emphasizing that specific abnormalities on fetal surveillance examinations warrant delivery at different gestational ages. To this end, a literature search of the PubMed/Medline and Latin America and the Caribbean Literature on Health Sciences (LILACS) databases was conducted using the terms fetal growth restriction, management, and delivery over the past ten years. Results were grouped into gestational age at delivery, mode of delivery, and methods of labor induction. The main fetal surveillance abnormalities prompting delivery in each gestational-age range were discussed, leading to the development of management flowcharts. Despite the lack of consensus in the literature and the limited number of randomized clinical trials guiding clinical decisions in FGR, the available evidence was summarized to assist clinicians in managing pregnancies complicated by FGR. It should be emphasized that there are few randomized clinical trials to guide management decisions in FGR. Full article

(This article belongs to the Special Issue Obstetrics in the 21st Century: New Challenges and Translational Perspectives in Diagnostics)

► Show Figures

Figure 1

15 pages, 753 KB

Open AccessArticle

Dynamic Liver Function Tests in Paediatric Liver Disease

by Thora Wesenberg Helt, Jon Nielsen, Gabriella Ficerai-Garland, Robin de Nijs, Christina Louise Winther, Søren Møller, Viktoria Setterberg, Vibeke Brix Christensen and Lise Borgwardt

Diagnostics 2026, 16(5), 805; https://doi.org/10.3390/diagnostics16050805 - 9 Mar 2026

Viewed by 800

Abstract

Background/Objectives: Liver function is difficult to estimate accurately. Conventional liver function tests can be normal, even in severe diseases. Dynamic liver function tests, including indocyanine green (ICG) clearance and hepatobiliary scintigraphy (HBS), are useful in adults. We aimed to evaluate the association [...] Read more.

Background/Objectives: Liver function is difficult to estimate accurately. Conventional liver function tests can be normal, even in severe diseases. Dynamic liver function tests, including indocyanine green (ICG) clearance and hepatobiliary scintigraphy (HBS), are useful in adults. We aimed to evaluate the association between ICG clearance and HBS in children with liver disease and to identify liver disease markers associated with liver function measured with ICG clearance and HBS. Methods: Children aged 0–18 years followed at Copenhagen University Hospital, Rigshospitalet between November 2015 and August 2024 were eligible for inclusion if they had acute or chronic liver disease, suspected liver disease, or previous liver transplantation (LTx). All underwent ICG clearance and HBS. Results: We included 131 children with a total of 200 visits. The median visit age was 11.4 [6.6; 15.6] years. The ICG-plasma disappearance rate had the strongest correlation with the hepatic extraction fraction (ρ = 0.64, p < 0.001). ICG clearance and HBS were associated with liver injury, reduced synthetic function, cholestasis, cirrhosis, and portal hypertension, while only ICG clearance was associated with the portal blood flow. LTx was associated with increased HBS parameters, but not with ICG clearance. Conclusions: ICG clearance and HBS are correlated, and both are associated with most conventional liver function markers. This suggests their usefulness in evaluating children with liver disease. However, further evaluation of the predictive and clinical value of ICG clearance and HBS in disease progression is needed. Full article

(This article belongs to the Special Issue Diagnosis and Management of Liver Diseases, Third Edition)

► Show Figures

Figure 1

26 pages, 3042 KB

Open AccessArticle

Thermoacoustic Ultrasound Assessment of Liver Steatosis—A Novel Approach for MASLD Diagnosis

by Jang Hwan Cho, Christopher M. Bull, Michael Thornton, Jing Gao, Jonathan M. Rubin and Idan Steinberg

Diagnostics 2026, 16(5), 804; https://doi.org/10.3390/diagnostics16050804 - 9 Mar 2026

Viewed by 828

Abstract

Background/Objectives: Metabolic Dysfunction-Associated Steatotic Liver Disease (MASLD) is a global health crisis, but current diagnostics are limited. Liver biopsy is invasive, magnetic resonance imaging-proton density fat fraction (MRI-PDFF) is expensive, and quantitative ultrasound methods are low-accuracy, especially in patients with a high [...] Read more.

Background/Objectives: Metabolic Dysfunction-Associated Steatotic Liver Disease (MASLD) is a global health crisis, but current diagnostics are limited. Liver biopsy is invasive, magnetic resonance imaging-proton density fat fraction (MRI-PDFF) is expensive, and quantitative ultrasound methods are low-accuracy, especially in patients with a high body mass index (BMI). This study introduces a novel thermo-acoustic (TA) method that generates ultrasound signals based on tissue electrical conductivity, where lean tissue (high in water and electrolytes) absorbs more radio-frequency (RF) energy than fatty tissue, providing a direct molecular contrast for fat. Methods: A prospective, cross-sectional feasibility study compared a new thermo-acoustic fat fraction (TAFF) score with the reference standard MRI-PDFF in 40 subjects with suspected fatty liver disease. Bland–Altman analysis, Deming regression, and Binary classification performance were tested. To establish system stability, a dedicated Repeatability and Reproducibility (R&R) study (N = 14) evaluated inter-operator and intra-operator consistency using an Intraclass Correlation Coefficient (ICC) derived from a two-way random-effects ANOVA model. Results: TAFF estimates demonstrated a substantial correlation (r = 0.89) with MRI-PDFF and an average absolute error of 3.04% fat fraction. Classification performance was high, with an Area Under the Receiver Operating Characteristic Curve (AUROC) of 0.92 at the 12% fat fraction threshold and 0.99 at the 20% fat fraction threshold. The R&R study confirmed robust stability (intraclass correlation = 0.89) and a negligible mean inter-operator difference of 0.36%. Estimation errors showed no statistically significant correlation with BMI or other body habitus measurements. Conclusions: These findings support thermoacoustics’ potential as an accurate, non-invasive, point-of-care solution that can serve as a new imaging biomarker. By providing predictive values closely aligned with MRI-PDFF across the full MASLD spectrum, TAFF can complement currently available ultrasound methods to address the cost and access constraints of MRI for the assessment, diagnosis, and monitoring of MASLD. Full article

(This article belongs to the Section Medical Imaging and Theranostics)

► Show Figures

Figure 1

11 pages, 1110 KB

Open AccessArticle

Investigation of the Optimal Duration and Modality for Postoperative Surveillance of Intraductal Papillary Mucinous Neoplasm (IPMN): A Single-Center Retrospective Study

by Akane Ozawa, Atsushi Nara, Kota Yokoyama, Junichi Tsuchiya, Daisuke Ban and Ukihide Tateishi

Diagnostics 2026, 16(5), 803; https://doi.org/10.3390/diagnostics16050803 - 8 Mar 2026

Viewed by 459

Abstract

Background/Objectives: Although multiple guidelines exist for the management of intraductal papillary mucinous neoplasms (IPMN), the duration and modality of postoperative surveillance remain inconsistent. We aimed to retrospectively review medical images of patients with IPMN after surgery and to investigate the optimal surveillance [...] Read more.

Background/Objectives: Although multiple guidelines exist for the management of intraductal papillary mucinous neoplasms (IPMN), the duration and modality of postoperative surveillance remain inconsistent. We aimed to retrospectively review medical images of patients with IPMN after surgery and to investigate the optimal surveillance duration and modality. Methods: In this study, we included 191 patients with IPMN who underwent surgery at a single institution between January 2006 and May 2024. Patients were followed from the postoperative period until July 2025. Image interpretation reports written by diagnostic radiologists were examined to determine the time to recurrence detection and the imaging modality used. Results: Sixteen patients (8.3%) were eligible during the observation period. Seven patients experienced intrapancreatic recurrence, and ten patients experienced extrapancreatic recurrence (one patient was included in both categories). The mean time to identification of intrapancreatic lesions was 63.9 months; five of seven cases were detected using contrast-enhanced computed tomography (CT). The mean time to identification of extrapancreatic lesions was 12.0 months, which was significantly shorter than that for intrapancreatic lesions (p = 0.005). Eight of ten extrapancreatic recurrences were detected using contrast-enhanced CT. Conclusions: Extrapancreatic lesions appeared earlier after IPMN surgery than intrapancreatic lesions. Contrast-enhanced CT was the most commonly used modality for detecting recurrent lesions, suggesting its usefulness in postoperative surveillance. Full article

(This article belongs to the Section Clinical Diagnosis and Prognosis)

► Show Figures

Figure 1

17 pages, 1641 KB

Open AccessArticle

Large-Scale Validation of a Dual Cross-Attention Network for Automated Sleep Staging Using Wearable Photoplethysmography Signals

by Ruochen Li, Yutao He, Yanan Bie, Jiawei Guo, Lichao Wang, Yao Zhao, Jun Zhong and Wei Zhu

Diagnostics 2026, 16(5), 802; https://doi.org/10.3390/diagnostics16050802 - 8 Mar 2026

Viewed by 634

Abstract

Background: Sleep staging is vital for diagnosing sleep disorders, but the clinical gold standard, polysomnography, is too intrusive for routine home monitoring. While photoplethysmography (PPG) offers a wearable alternative, achieving high diagnostic accuracy remains challenging due to signal noise and individual variability. Methods: [...] Read more.

Background: Sleep staging is vital for diagnosing sleep disorders, but the clinical gold standard, polysomnography, is too intrusive for routine home monitoring. While photoplethysmography (PPG) offers a wearable alternative, achieving high diagnostic accuracy remains challenging due to signal noise and individual variability. Methods: We developed DCA-Sleep, a deep learning framework using a Dual Cross-Attention (DCA) mechanism to capture long-range temporal dependencies from raw single-channel PPG. To overcome data scarcity, a cross-modality transfer learning strategy was implemented, pre-training the model on six electrocardiogram (ECG) datasets before extensive validation on a combined cohort of 9738 subjects across nine public datasets (including MESA and CFS). Results: DCA-Sleep demonstrated superior robustness, achieving an average F1-score of 0.731 and a Cohen’s Kappa of 0.652 on the MESA dataset, significantly outperforming state-of-the-art baselines. The model showed high sensitivity in detecting Wake and Deep Sleep stages, which are critical for clinical assessment. Conclusions: This study provides a large-scale validation of a PPG-based staging tool, confirming its reliability as a non-invasive, scalable solution for long-term sleep monitoring and clinical screening. Full article

(This article belongs to the Special Issue AI-Powered Solutions for Personalized Healthcare Monitoring with Wearables)

► Show Figures

Figure 1

16 pages, 4901 KB

Open AccessArticle

Quantitative Comparison of Two Novel Swept-Source Optical Coherence Tomography Angiography Devices

by Michael Hafner, Daniel J. P. Deschler, Alexander Kufner, Lisa M. Katscher, Siegfried G. Priglinger and Maximilian J. Gerhardt

Diagnostics 2026, 16(5), 801; https://doi.org/10.3390/diagnostics16050801 - 8 Mar 2026

Viewed by 805

Abstract

Background: Swept-source optical coherence tomography angiography (SS-OCTA) enables rapid assessment of retinal microvasculature. However, cross-platform comparability remains limited by device-specific acquisition and image quality characteristics. This study prospectively compared two novel SS-OCTA systems, DREAM (200 kHz) and BMizar (400 kHz). Methods: [...] Read more.

Background: Swept-source optical coherence tomography angiography (SS-OCTA) enables rapid assessment of retinal microvasculature. However, cross-platform comparability remains limited by device-specific acquisition and image quality characteristics. This study prospectively compared two novel SS-OCTA systems, DREAM (200 kHz) and BMizar (400 kHz). Methods: Fifty eyes from 25 healthy participants underwent 3 mm × 3 mm macular OCTA imaging with both devices in a single session. Images were analysed using OCTAVA to extract foveal avascular zone (FAZ) area, vessel area density (VAD), total vessel length (TVL), node counts, fractal dimension (FD), median vessel length (MVL) in SCP, and mean vessel diameter (MVD) in DCP. Image quality was assessed using FAZ-noise rate, contrast-to-noise ratio (CNR), and FAZ noise-floor standard deviation. Paired comparisons were performed using Wilcoxon signed-rank tests and Cliff’s delta. Results: BMizar acquisition time was shorter than DREAM for the evaluated 3 × 3 mm protocol (median 5.36 s vs. 9.93 s), reflecting differences in A-scan rate and protocol implementation; acquisition time is therefore reported descriptively. In the SCP, DREAM yielded lower VAD (41.9% vs. 48.8%) and fewer nodes (1547 vs. 1879) but exhibited markedly less background noise (noise-floor SD 4.1 vs. 57.9) and substantially higher CNR (16.7 vs. 0.82). DREAM also showed longer MVL (45 vs. 39 µm) and higher FD (1.98 vs. 1.97; δ = 0.90). In the DCP, DREAM demonstrated smaller FAZ areas (0.27 vs. 0.42 mm²), thinner MVD (14 vs. 25 µm), higher node counts (3144 vs. 2301), longer TVL (223.6 vs. 206.2 mm), and higher FD (1.98 vs. 1.97), whereas VAD was higher on BMizar (32.96% for DREAM vs. 49.93% for BMizar). FAZ-noise rates were consistently higher for BMizar in both plexuses. Conclusions: Both devices provide reliable SS-OCTA imaging, but with distinct strengths. DREAM delivers higher vascular continuity and more reliable FAZ and DCP quantification, whereas BMizar achieves faster acquisition at the cost of noise, inflating SCP density and distorting FAZ-based metrics. Awareness of these characteristics is essential to ensure the valid use of OCTA biomarkers in clinical and research applications. Full article

(This article belongs to the Special Issue Applications of Optical Coherence Tomography and Optical Coherence Tomography Angiography in Ophthalmology)

► Show Figures

Figure 1

14 pages, 1532 KB

Open AccessReview

Rare Causes of Portal Vein Thrombosis and Occlusion: A Narrative Review

by Lavinia Alice Bălăceanu, Claudia Georgeta Iacobescu, Teodora Burloiu, Marian-Vlad Lăpădat, Ion Dina and Ion Daniel Baboi

Diagnostics 2026, 16(5), 800; https://doi.org/10.3390/diagnostics16050800 - 8 Mar 2026

Viewed by 742

Abstract

Portal vein thrombosis (PVT) refers specifically to the presence of a thrombus within the main portal vein trunk or its intrahepatic branches. In contrast, portal vein occlusion encompasses a broader spectrum of conditions, including tumor invasion, external compression and disorders that predispose to [...] Read more.

Portal vein thrombosis (PVT) refers specifically to the presence of a thrombus within the main portal vein trunk or its intrahepatic branches. In contrast, portal vein occlusion encompasses a broader spectrum of conditions, including tumor invasion, external compression and disorders that predispose to thrombosis, such as thrombophilia or inflammatory states. Advanced liver disease, particularly cirrhosis, is the most common cause of PVT, primarily due to portal hypertension, altered hemostasis and hemodynamic changes, followed by malignancies and inherited or acquired thrombophilic conditions. In contrast to these common etiologies, our clinical experience has highlighted rare causes of portal vein obstruction associated with typical presentations, which pose diagnostic challenges. Examples include acute PVT during transjugular intrahepatic portosystemic shunt (TIPS) placement and non-thrombotic porto-mesenteric obstruction related to portal venous gas. While these events may appear unexpected, they represent a recognizable group of uncommon causes rather than isolated incidents. PVT can present as an acute or chronic condition: acute thrombosis is characterized by recent thrombus formation and potential intestinal ischemia, whereas chronic thrombosis is associated with long-standing obstruction, cavernous transformation and portal hypertension. This narrative review integrates a comprehensive literature search with clinical experience, with particular emphasis on uncommon etiologies of portal vein obstruction. Full article

(This article belongs to the Section Medical Imaging and Theranostics)

► Show Figures

Figure 1

19 pages, 537 KB

Open AccessArticle

Bioelectrical Activity of Masticatory Muscles and Postural Stability Across TMD Subtypes

by Aleksandra Dolina, Justyna Pałka, Magdalena Zawadka, Marcin Wójcicki, Monika Litko-Rola, Jacek Szkutnik and Piotr Gawda

Diagnostics 2026, 16(5), 799; https://doi.org/10.3390/diagnostics16050799 - 8 Mar 2026

Viewed by 429

Abstract

Background: Existing evidence suggests an association between temporomandibular disorders (TMDs) and alterations in body posture and balance; however, the mechanism underlying this relationship remains unknown. The present study aimed to investigate the associations between specific TMD subtypes, indices of bioelectrical activity of [...] Read more.

Background: Existing evidence suggests an association between temporomandibular disorders (TMDs) and alterations in body posture and balance; however, the mechanism underlying this relationship remains unknown. The present study aimed to investigate the associations between specific TMD subtypes, indices of bioelectrical activity of the masticatory muscles, and parameters of body posture and balance. Methods: The study followed a case–control study design. A total of 81 participants were enrolled, including 33 controls and 48 individuals with TMD, classified into myofascial (n = 14), articular (n = 17), and mixed (n = 17) subtypes. Diagnosis of temporomandibular disorders was carried out by prosthodontic specialists using the Polish adaptation of the Diagnostic Criteria for Temporomandibular Disorders. Masticatory muscle bioelectrical activity was assessed by surface electromyography. For statistical analysis, the Asymmetry Index and Functional Clenching Activity Indices were used. Static balance was evaluated with a pedobarographic platform. The sway area, velocity, and length of the Center of Pressure, as well as the foot contact area, were recorded and automatically calculated by the system. Measurements were performed under different mandibular conditions, with both eyes open and eyes closed. Correlation analyses were performed using Spearman Rank Order Correlation. Pearson’s Chi-squared test was used for the analysis of categorical variables. Results: Weak to moderate negative correlations were primarily observed, indicating that higher indices of masticatory muscle bioelectrical activity were associated with better postural balance, with distinct correlation patterns identified across different TMD subtypes. Conclusions: This exploratory study identified multiple correlations between masticatory muscle activity and postural or balance parameters, suggesting possible subtype-specific patterns in TMDs. However, the evidence remains preliminary and should be interpreted with caution, warranting further confirmatory and longitudinal research. Full article

(This article belongs to the Special Issue Diagnostic Approaches to Temporomandibular Disorders)

► Show Figures

Figure 1

13 pages, 1890 KB

Open AccessArticle

Photon-Counting CT Angiography Enables Superior Preoperative Perforator Depiction for Fibular Transplant Surgery Requiring Less Contrast Agent Compared to Energy-Integrating CT

by Ramin Saam Dazeh, Jan-Lucca Hennes, Tobias Prester, Viktor Hartung, Henner Huflage, Andreas Vollmer, Thorsten Alexander Bley, Philipp Gruschwitz and Kristina Krompaß

Diagnostics 2026, 16(5), 798; https://doi.org/10.3390/diagnostics16050798 - 8 Mar 2026

Viewed by 743

Abstract

Background/Objectives: The objective of this study was to ascertain whether photon-counting CT angiography (PCD-CTA) can optimize image quality for the visualization of perforating arteries for planning fibular transplant procedures in comparison to energy-integrating CT angiography (EID-CTA). Methods: In this retrospective single-center [...] Read more.

Background/Objectives: The objective of this study was to ascertain whether photon-counting CT angiography (PCD-CTA) can optimize image quality for the visualization of perforating arteries for planning fibular transplant procedures in comparison to energy-integrating CT angiography (EID-CTA). Methods: In this retrospective single-center study, all patients who underwent preoperative CT of the peripheral runoff for planning between October 2021 and July 2023 were consecutively included. PCD-CTA was performed in standard resolution mode as 55 keV images with 90 mL of iodine-containing contrast agent or alternatively, an EID-CTA as a low-kV scan with 110 mL of contrast agent. The raw data were reformatted using comparable soft vascular and sharp regular convolution kernels, slice thickness/increment, and field of view. Contrast-to-noise ratio was calculated for objective image quality. Subjective evaluation was based on a rating by three radiologists using a five-point Likert scale (criteria: overall image quality, luminal attenuation, vessel sharpness, and perforator depiction). Results: Of the 26 patients who were screened, 9 could be included in each group, while 8 were excluded due to incomplete reconstructions. The reduction in contrast agent dose resulted in a non-significant decrease in luminal attenuation on PCD-CTA (452.5 ± 53.6 HU vs. 465.5 ± 99.6 HU; p = 0.375). The image noise was considerably lower for PCD-CTA (21.1 ± 1.0 HU vs. 32.9 ± 1.6 HU; p < 0.001). This resulted in a significantly higher contrast-to-noise ratio (CNR) for sharp kernel reconstructions (22.4 ± 3.5 vs. 14.5 ± 3.8; p < 0.001). No significant differences were observed for the soft vascular kernel. Subjective evaluation revealed a significant enhancement in overall image quality, vascular sharpness, and perforator depiction for PCD-CTA with sharp reconstructions. In contrast, soft kernel reconstructions and luminal attenuation demonstrated no substantial difference. Interrater agreement was good to excellent. Conclusions: PCD-CTA with sharp kernel reformatting has been demonstrated to yield superior image quality and perforator delineation of the fibular artery in comparison to standard EID-CTA. Full article

(This article belongs to the Special Issue Photon-Counting CT in Clinical Application)

► Show Figures

Figure 1

11 pages, 2678 KB

Open AccessCase Report

Mediastinal Ectopic Pancreas Mimicking Lymphoma with Discordant Histology and Flow Cytometry: A Diagnostic Challenge

by Guilin Ren, Hongfeng Wang, Haiqin Deng, Jianbin Chen, Li Wang, Qian Zhan, Jinxing Wu and Liwan Dai

Diagnostics 2026, 16(5), 797; https://doi.org/10.3390/diagnostics16050797 - 8 Mar 2026

Viewed by 456

Abstract

Background: Mediastinal ectopic pancreas (EP) is an exceptionally rare entity that can mimic malignancy. Diagnosis is typically established post-operatively; pre-operative confirmation is challenging. Case Presentation: We describe a 28-year-old man presenting with life-threatening airway obstruction due to a progressive mediastinal mass, requiring emergency [...] Read more.

Background: Mediastinal ectopic pancreas (EP) is an exceptionally rare entity that can mimic malignancy. Diagnosis is typically established post-operatively; pre-operative confirmation is challenging. Case Presentation: We describe a 28-year-old man presenting with life-threatening airway obstruction due to a progressive mediastinal mass, requiring emergency tracheal stenting. Diagnostic workup revealed a critical discordance: while CT-guided core biopsy confirmed benign ectopic pancreatic tissue, concurrent flow cytometry identified a monoclonal B-cell population with a high Ki-67 index (~86%), raising concern for a high-grade lymphoid process. However, no morphological evidence of lymphoma was found, and PET-CT showed only moderate metabolic activity (SUVmax 4.6), making an untreated aggressive lymphoma less consistent. The patient declined surgical resection. Management proceeded with a conservative strategy of structured clinical surveillance based on the benign histology. At 6-month follow-up, the patient remained clinically stable without chemotherapy, supporting the diagnosis of benign ectopic pancreas and suggesting the flow cytometric findings represented reactive “pseudo-monoclonality” secondary to inflammation. Conclusions: This case highlights mediastinal EP as a rare airway emergency and illustrates a major diagnostic pitfall: flow cytometric clonality and high proliferative fractions can occur in inflammatory settings and must not override benign architectural histology. When discordance persists and definitive tissue cannot be obtained, management should emphasize multidisciplinary review, deliberate specimen triage, and structured surveillance with predefined triggers for repeat higher-yield biopsy or surgical sampling and airway-stent reassessment. Full article

(This article belongs to the Section Pathology and Molecular Diagnostics)

► Show Figures

Figure 1

17 pages, 5672 KB

Open AccessArticle

Prevalence of Unfilled MB2 Canals and Their Association with Apical Periodontitis: A CBCT-Based Cross-Sectional Study in a German Population

by Maythem Al Fartousi and Christian Ralf Gernhardt

Diagnostics 2026, 16(5), 796; https://doi.org/10.3390/diagnostics16050796 - 7 Mar 2026

Cited by 1 | Viewed by 1019

Abstract

Background/Objectives: The presence of untreated second mesio-buccal canals (MB2) in maxillary first molars is usually associated with endodontic treatment failure. Previous CBCT-based investigations have evaluated the quality of root canal fillings and the prevalence of apical lesions in endodontically treated teeth. However, [...] Read more.

Background/Objectives: The presence of untreated second mesio-buccal canals (MB2) in maxillary first molars is usually associated with endodontic treatment failure. Previous CBCT-based investigations have evaluated the quality of root canal fillings and the prevalence of apical lesions in endodontically treated teeth. However, evidence specifically addressing untreated MB2 canals and their association with apical periodontitis remains limited. Therefore, the aim of this cross-sectional study was to evaluate the prevalence of unfilled MB2 canals in endodontically treated maxillary first molars and their association with apical periodontitis. Methods: CBCT scans of 75 patients from an endodontic practice were retrospectively analyzed. Maxillary first molars (teeth 16 and 26) were evaluated for the presence and filling status of root canals (MB1, MB2, palatal, distal) and the presence of periapical radiolucency using the CBCT periapical index. Two calibrated examiners independently assessed all images. The association between unfilled MB2 canals and apical periodontitis was analyzed using chi-square tests, and odds ratios with 95% confidence intervals were calculated. Results: The mean patient age was 53.4 ± 15.5 years (range: 14–80). An MB2 canal was present in 84% (63/75) of eligible teeth. Among teeth with an MB2 canal, only 20.6% (13/63) were endodontically filled, while 79.4% remained untreated. Apical periodontitis was observed in 65.3% (49/75) of all teeth. A significant association was found between unfilled MB2 canals and apical periodontitis (p < 0.001), with an odds ratio of 0.095 (95% CI: 0.022–0.402), indicating that filled MB2 canals significantly reduced the possible risk of periapical pathology. Conclusions: A high prevalence of unfilled MB2 canals was observed in this German population (79.4%). Furthermore, unfilled MB2 canals were strongly associated with apical periodontitis. Therefore, clinicians should utilize all available diagnostic tools, including CBCT and dental microscopes, to maximize MB2 canal identification and improve endodontic treatment outcomes. Full article

(This article belongs to the Special Issue Advances in Dental Diagnostics)

► Show Figures

Figure 1

18 pages, 1187 KB

Open AccessArticle

Application of Multivariate Adaptive Regression Splines to Estimate Fatty Liver Index in Healthy Young Taiwanese Men

by Po-Chung Chen, Chung-Chi Yang, Dee Pei, Ta-Wei Chu and Jyh-Gang Leu

Diagnostics 2026, 16(5), 795; https://doi.org/10.3390/diagnostics16050795 - 7 Mar 2026

Viewed by 618

Abstract

Background: Non-alcoholic fatty liver disease (NAFLD) represents the most widespread chronic liver disorder globally, impacting roughly 30% of the general population. Numerous factors have been linked to NAFLD, including obesity, type 2 diabetes, diet, physical inactivity, age, sex, genetic factors, and metabolic [...] Read more.

Background: Non-alcoholic fatty liver disease (NAFLD) represents the most widespread chronic liver disorder globally, impacting roughly 30% of the general population. Numerous factors have been linked to NAFLD, including obesity, type 2 diabetes, diet, physical inactivity, age, sex, genetic factors, and metabolic syndrome. Previous research predominantly treated NAFLD as a categorical outcome, providing less granular data compared to the continuous fatty liver index (FLI). This investigation enrolled healthy young Taiwanese men and applied multivariate adaptive regression spline (MARS) modeling to develop a predictive equation. Our aims were twofold: 1. To assess the predictive accuracy of traditional multiple linear regression (MLR) versus MARS. 2. To construct a MARS-derived equation for estimating FLI in this demographic. Methods: Data originated from the Taiwan MJ Cohort, comprising 5496 men aged 20–50 years not using medications for metabolic syndrome. MARS was used to formulate the FLI estimation equation. Model performance was compared using symmetric mean absolute percentage error (SMAPE), relative absolute error (RAE), root relative squared error (RRSE), and root mean squared error (RMSE). Results: Evaluation indicated that MARS yielded lower estimation errors than MLR, demonstrating its superior performance. The derived equation is: FLI = 65.224 − 0.436 × B1 − 0.490 × B2 + 0.252 × B3 − 2.962 × B4 + 2.231 × B5 − 0.292 × B6 + 0.189 × B7 − 0.361 × B8 − 0.699 × B9 + 0.160 × B10 − 2.715 × B11 + 0.799 × B12 − 0.153 × B13 + 0.084 × B14 − 35.274 × B15 − 4.424 × B16. Conclusions: Using MLR as a benchmark, our analysis revealed that MARS delivered better predictive performance. The presented equation explains 62.7% of the variance in FLI (r² = 0.627). Based on standardized variable importance scores (nsubsets metric), CRP emerged as the most influential predictor, followed by WBC, UA, HDL-C, AST, age, ALT, FPG, SBP, and LDL in this cohort of healthy young Taiwanese men. Full article

(This article belongs to the Special Issue Metabolic Diseases: Diagnosis, Management, and Pathogenesis)

► Show Figures

Figure 1

15 pages, 6568 KB

Open AccessArticle

From Plastics to Prognosis: ANO4 Susceptibility Links Phthalate Exposure to Prostate Cancer Progression

by Chi-Fen Chang, Shu-Pin Huang, Yei-Tsung Chen, Lih-Chyang Chen, Chao-Yuan Huang, Chia-Cheng Yu, Victor C. Lin, Te-Ling Lu and Bo-Ying Bao

Diagnostics 2026, 16(5), 794; https://doi.org/10.3390/diagnostics16050794 - 7 Mar 2026

Viewed by 570

Abstract

Background/Objective: Di-2-ethylhexyl phthalate and its bioactive metabolite mono-2-ethylhexyl phthalate (MEHP) are ubiquitous endocrine-disrupting chemicals implicated in carcinogenesis. However, the molecular mechanisms linking MEHP exposure, host genetic susceptibility, and prostate cancer progression remain incompletely defined. Methods: We integrated transcriptomic profiling of MEHP-exposed human prostate [...] Read more.

Background/Objective: Di-2-ethylhexyl phthalate and its bioactive metabolite mono-2-ethylhexyl phthalate (MEHP) are ubiquitous endocrine-disrupting chemicals implicated in carcinogenesis. However, the molecular mechanisms linking MEHP exposure, host genetic susceptibility, and prostate cancer progression remain incompletely defined. Methods: We integrated transcriptomic profiling of MEHP-exposed human prostate epithelial cells with a genetic association study of 630 patients with prostate cancer receiving androgen deprivation therapy. MEHP-responsive genes were identified from public microarray datasets and subjected to pathway enrichment analyses. Germline single-nucleotide polymorphisms (SNPs) in MEHP-regulated genes were evaluated for their association with progression-free survival, overall survival, and cancer-specific survival. The clinical and functional relevance of the key genes was further assessed using large-scale public prostate cancer expression datasets. Results: MEHP exposure induced widespread transcriptional reprogramming, prominently suppressing focal adhesion and cell–matrix interaction pathways. Genetic analyses identified multiple prognostically relevant SNPs within MEHP-responsive genes, with anoctamin 4 (ANO4) variants showing consistent associations across all clinical endpoints. The minor allele of rs17485225 in ANO4 was significantly associated with reduced all-cause and prostate cancer-specific mortality. Pooled analyses revealed reduced ANO4 expression levels in prostate cancer tissues and improved survival in patients with high ANO4 expression levels. Pathway analyses linked low ANO4 expression levels with enhanced cell cycle activity and compromised cell adhesion. Conclusions: Our findings suggest that ANO4 may act as a mediator of MEHP-associated prostate cancer progression and support a gene–environment interaction model in which environmental toxicant exposure and germline variation converge on focal adhesion dysregulation to potentially contribute to aggressive disease. Full article

(This article belongs to the Special Issue Predictive Biomarkers in Oncology)

► Show Figures

Figure 1

27 pages, 898 KB

Open AccessReview

Diagnostic and Therapeutic Challenges in Rare and Non-Tubal Ectopic Pregnancies: A Narrative Review

by Stefan Ivanovic, Milica Ivanovic, Dragana Maglic, Milica Mandic, Lidija Tulic, Katarina Ivanovic, Milos Milincic, Nikola Jovic and Rastko Maglic

Diagnostics 2026, 16(5), 793; https://doi.org/10.3390/diagnostics16050793 - 7 Mar 2026

Viewed by 753

Abstract

In relation to the most commonly described ampullary ectopic pregnancies in contemporary gynecological practice, rare localizations of ectopic pregnancies represent a diagnostic and therapeutic challenge whose clinical significance far exceeds their frequency. In contrast to tubal ectopic pregnancy, these implantation localizations are characterized [...] Read more.

In relation to the most commonly described ampullary ectopic pregnancies in contemporary gynecological practice, rare localizations of ectopic pregnancies represent a diagnostic and therapeutic challenge whose clinical significance far exceeds their frequency. In contrast to tubal ectopic pregnancy, these implantation localizations are characterized by specific anatomical relationships and early trophoblastic invasion into highly vascularized tissues, which is why classical diagnostic algorithms and therapeutic patterns are often not applicable in clinical practice. Clinical uncertainty is further increased by the fact that a large proportion of these pregnancies in early gestation cannot be precisely mapped and initially present as pregnancies of unknown location. This narrative review integrates contemporary evidence and guidelines of relevant professional societies with the aim of highlighting patterns of diagnostic errors, systemic weaknesses of existing approaches, and key points for safe clinical decision-making. Special emphasis is placed on the role of disciplined transvaginal ultrasound evaluation, terminological precision, and timely recognition of high-risk and nonspecific implantations. Analysis of the available literature indicates that therapeutic decisions must be individualized and guided by the implantation site and assessment of hemorrhagic risk, rather than gestational age or absolute β-hCG values. Understanding these principles represents the basis for reducing serious complications and for the development of future diagnostic and therapeutic algorithms, thereby improving treatment outcomes. Full article

(This article belongs to the Section Clinical Diagnosis and Prognosis)

► Show Figures

Figure 1

18 pages, 1559 KB

Open AccessReview

Stress Echocardiography in the Diagnosis and Evaluation of Pulmonary Hypertension: Practical Recommendations, Haemodynamic Phenotyping, and Application in Adults and Children

by Dafni Charisopoulou, George Koulaouzidis, Panagiota Kleitsioti, Nikolaos Antoniou, Christos Mantzios, Orestis Grammenos and Sotiria Iliopoulou

Diagnostics 2026, 16(5), 792; https://doi.org/10.3390/diagnostics16050792 - 6 Mar 2026

Viewed by 901

Abstract

Pulmonary hypertension (PH) is a complex condition in which early diagnosis remains challenging, particularly in patients with exertional symptoms and normal or borderline resting haemodynamics. Although right heart catheterisation is the diagnostic gold standard, transthoracic echocardiography is the recommended first-line non-invasive test. However, [...] Read more.

Pulmonary hypertension (PH) is a complex condition in which early diagnosis remains challenging, particularly in patients with exertional symptoms and normal or borderline resting haemodynamics. Although right heart catheterisation is the diagnostic gold standard, transthoracic echocardiography is the recommended first-line non-invasive test. However, resting echocardiography provides only a static assessment and may underestimate disease severity in early or latent pulmonary vascular disease due to preserved pulmonary vascular compliance and adaptive right ventricular responses. Because pulmonary haemodynamics are intrinsically flow-dependent, pathological abnormalities may only emerge during increased cardiac output. Stress echocardiography, performed using exercise or pharmacological stress, enables dynamic evaluation of pulmonary pressure responses, cardiac output augmentation, right ventricular contractile reserve, and ventricular interaction. Increasing evidence indicates that stress echocardiography can unmask abnormal pulmonary pressure–flow relationships, impaired pulmonary vascular reserve, and reduced right ventricular–pulmonary arterial coupling that are not apparent at rest, thereby improving functional and haemodynamic characterisation in selected patients. This Diagnostic Review outlines the physiological basis for stress echocardiographic assessment of pulmonary circulation, proposes practical recommendations for patient selection and testing protocols, and provides a framework for interpretation centered on pressure–flow relationships rather than absolute pulmonary pressure thresholds. Particular attention is given to clinical scenarios with high diagnostic yield, including unexplained exertional dyspnoea, systemic sclerosis, suspected heart failure with preserved ejection fraction, at-risk relatives of patients with pulmonary arterial hypertension, selected athletes, and paediatric populations. Stress echocardiography should not be considered a standalone diagnostic test for PH but, when performed in experienced centers and integrated within structured diagnostic pathways, it represents a valuable non-invasive adjunct to guide referral for invasive haemodynamic confirmation. Full article

(This article belongs to the Special Issue Beyond the Image: Cardiac Imaging at the Service of the Patient)

► Show Figures

Figure 1

22 pages, 1119 KB

Open AccessArticle

Development of microRNA-Based Glioblastoma Biomarkers Using Blood Plasma Specimens

by Sophia Giliberto, Kenny K. Ablordeppey, Jacob Goldman, Melinda Yin, Rahul Chowdhury, Jacob Till, Kira Sheinerman, Sydney D. Finkelstein, Samuil Umansky, Alidad Mireskandari, Gyanendra Kumar, Erica L. Carpenter and Stephen J. Bagley

Diagnostics 2026, 16(5), 791; https://doi.org/10.3390/diagnostics16050791 - 6 Mar 2026

Viewed by 795

Abstract

Background: Noninvasive biomarkers for the detection and monitoring of glioblastoma (GBM) are needed to improve clinical outcomes for patients. The objective of this pilot study was to evaluate the expression of a panel of 48 pre-selected microRNAs (miRNAs) in plasma specimens from GBM [...] Read more.

Background: Noninvasive biomarkers for the detection and monitoring of glioblastoma (GBM) are needed to improve clinical outcomes for patients. The objective of this pilot study was to evaluate the expression of a panel of 48 pre-selected microRNAs (miRNAs) in plasma specimens from GBM patients versus healthy controls to identify candidate miRNA biomarkers for noninvasive diagnosis of GBM. Methods: Selection of candidate miRNA biomarkers was based on a comprehensive literature review and data mining. RNA was extracted from plasma samples obtained prior to resection from patients with GBM (n = 30) and age- and sex-matched healthy controls (n = 30), as well as from matched FFPE GBM tissue samples when available (n = 3). Expression levels of 48 miRNAs were assessed in all samples, and expression data was processed using proprietary software to generate potential biomarkers and train linear classifiers. Results: Overall miRNA expression patterns were similar between matched plasma and FFPE tumor tissues in patients with GBM. miRNA levels were examined in pairs to determine the ratio between two miRNAs, which served to normalize the data. The top five miRNA pairs for distinguishing between GBM and healthy control plasma included miR-17-5p/miR-19b-3p (AUC 0.93, 95% CI = 0.870, 0.970), miR-20a-5p/miR-19b-3p (AUC 0.93, 95% CI = 0.870, 0.970), miR-93-5p/miR-92a-3p (AUC 0.92, 95% CI = 0.875, 0.965), miR-17-5p/miR-92a-3p (AUC 0.91, 95% CI = 0.865, 0.955), and miR-93-5p/miR-19b-3p (AUC 0.90, 95% CI = 0.850, 0.950). For the development of a multi-biomarker combination classifier consisting of up to three miRNA pair biomarkers, miRNA pairs with an AUC ≥ 0.8 were selected to build equal-weight linear classifiers. All possible combinations of three high-performing miRNA pairs were tested across the 60 samples. The top classifier (miR-20a-5p/miR-451a, miR-582-5p/miR-222-3p, and miR-17-5p/miR-222-3p) achieved an AUC value of 0.992, sensitivity of 0.93, specificity of 1, and accuracy of 0.97. Conclusions: These findings support the continued development of a plasma-based miRNA molecular diagnostic approach for the detection of GBM. The strong discriminatory performance observed in this study, including high AUC values, highlights the potential of circulating miRNA signatures as a minimally invasive diagnostic tool. As a pilot analysis, this work establishes a foundation for future prospective studies in larger, independent cohorts—including relevant disease control populations—to further define clinical performance, specificity, and utility in diagnostic and monitoring settings. Collectively, these results represent an important step toward the translation of plasma-based miRNA profiling into clinical application for GBM. Full article

(This article belongs to the Section Pathology and Molecular Diagnostics)

► Show Figures

Figure 1

23 pages, 3651 KB

Open AccessArticle

A Unified Framework for Survival Prediction: Combining Machine Learning Feature Selection with Traditional Survival Analysis in Heart Failure and METABRIC Breast Cancer

by Fangya Tan, Jian-Guo Zhou, Shuqiao Li, Bowen Long, Srikar Bellur, Yang Zhou and Mark Newman

Diagnostics 2026, 16(5), 790; https://doi.org/10.3390/diagnostics16050790 - 6 Mar 2026

Viewed by 974

Abstract

Background: The clinical use of machine learning (ML) in survival analysis is often limited by the “black box” nature of complex algorithms, which makes their results difficult to interpret in practice. In this study, we propose a unified and clinically grounded framework [...] Read more.

Background: The clinical use of machine learning (ML) in survival analysis is often limited by the “black box” nature of complex algorithms, which makes their results difficult to interpret in practice. In this study, we propose a unified and clinically grounded framework that integrates ML-based feature selection with traditional survival analysis. This approach aims to bridge the gap between strong predictive performance and clear, clinically meaningful interpretation. Methods: High-impact prognostic clinical features were identified using ML models GBM-Cox, RSF, and LASSO-Cox with 5-fold stratified cross-validation and subsequently validated using Cox Proportional Hazards and Kaplan–Meier analysis. The framework was evaluated across two distinct disease domains, Heart Failure and the METABRIC breast cancer cohort, to assess robustness and generalizability. Results: In the Heart Failure dataset, age group, serum creatinine, and blood pressure stratified patients into distinct risk groups. The high-risk group had significantly higher mortality (HR: 2.61; 95% CI: 1.42–4.78; p = 0.0013). In the METABRIC cohort, age at diagnosis, HER2 status, and Nottingham Prognostic Index (NPI) showed strong survival separation (p < 0.001). The high-risk group had an HR of 2.73 (95% CI: 2.34–3.19) and the faced a significantly shorter median survival (104.7 vs. 252.3 months), representing a 12.3-year reduction in life expectancy compared to low-risk group. This prognostic separation emphasizes the predictive power of selected baseline variables. Performance remained stable across cohorts, with C-index values (0.665–0.794) consistent with standard clinical benchmarks. Conclusions: Integrating cross-validated machine learning feature selection with Cox-based survival analysis enables stable and clinically interpretable risk stratification across diseases. By translating ML selected predictors into hazard ratios and absolute survival differences, this framework provides a reproducible and clinically grounded approach for survival risk assessment. Full article

(This article belongs to the Section Machine Learning and Artificial Intelligence in Diagnostics)

► Show Figures

Figure 1

Journal Menu

Journal Browser

Diagnostics, Volume 16, Issue 5 (March-1 2026) – 175 articles

Further Information

Guidelines

MDPI Initiatives

Follow MDPI