J Formos Med Assoc. 2025 Feb 15:S0929-6646(25)00059-2. doi: 10.1016/j.jfma.2025.02.017. Online ahead of print.

NO ABSTRACT

PMID:39956680 | DOI:10.1016/j.jfma.2025.02.017

Biomed Pharmacother. 2025 Feb 15;184:117891. doi: 10.1016/j.biopha.2025.117891. Online ahead of print.

ABSTRACT

Cyclin-dependent kinase 8 (CDK8) plays a crucial role in the transforming growth factor beta (TGF-β) signaling pathway, which is critical to the pathology of idiopathic pulmonary fibrosis (IPF). CDK8 promotes the epithelial-mesenchymal transition (EMT) and excessive extracellular matrix (ECM) deposition, making it a promising target for IPF treatment. This study focused on optimizing F059-1017, a previously identified CDK8 inhibitor, to enhance its potency. Through integrated structure-based modifications, a series of compounds was synthesized, and their inhibitory effects on CDK8 were tested. Results indicated that substituting with cyclopentanone significantly improved the inhibitory activity, and compound 4j demonstrated the best potency (IC50 = 16 nM). Notably, compared to F059-1017, its potency increased 35-fold, and kinase profiling revealed that the compound was selective for CDK8. Compound 4j inhibited the TGF-β1-induced EMT, cell migration, and morphological changes in A549 cells at a concentration of 0.1 μM and inhibited ECM and EMT protein expressions. In addition, the compound blocked TGF-β1-induced transcriptional changes and inhibited Smad3 and RNA polymerase II phosphorylation. These results highlight the potential of the optimized CDK8 inhibitor as a prospective drug for IPF treatment.

PMID:39955852 | DOI:10.1016/j.biopha.2025.117891

J Math Biol. 2025 Feb 16;90(3):31. doi: 10.1007/s00285-025-02191-3.

ABSTRACT

In this paper we introduce a formalism that allows to describe the response of a part of a biochemical system in terms of renewal equations. In particular, we examine under which conditions the interactions between the different parts of a chemical system, described by means of linear ODEs, can be represented in terms of renewal equations. We show also how to apply the formalism developed in this paper to some particular types of linear and non-linear ODEs, modelling some biochemical systems of interest in biology (for instance, some time-dependent versions of the classical Hopfield model of kinetic proofreading). We also analyse some of the properties of the renewal equations that we are interested in, as the long-time behaviour of their solution. Furthermore, we prove that the kernels characterising the renewal equations derived by biochemical system with reactions that satisfy the detail balance condition belong to the class of completely monotone functions.

PMID:39956846 | DOI:10.1007/s00285-025-02191-3

J Allergy Clin Immunol. 2025 Feb 14:S0091-6749(25)00170-8. doi: 10.1016/j.jaci.2025.01.044. Online ahead of print.

NO ABSTRACT

PMID:39956282 | DOI:10.1016/j.jaci.2025.01.044

Reg Anesth Pain Med. 2025 Feb 16:rapm-2024-105955. doi: 10.1136/rapm-2024-105955. Online ahead of print.

ABSTRACT

BACKGROUND/IMPORTANCE: Driving under the influence of drugs (DUID) refers to operating a vehicle after consuming drugs or medications other than alcohol that impair the ability to drive safely. There is no consensus on legal limits for drug intoxication while driving in the USA. Balancing the benefits of prescription medications, such as opioids, with traffic safety remains an ongoing public health challenge.

OBJECTIVE: This article examines DUID policy and provides recommendations for policy improvement and unification grounded in scientific evidence on opioid-related impairment and driving risks.

EVIDENCE REVIEW: A literature review of epidemiologic data, psychomotor effects, and public policy related to opioid use and driving was conducted. A total of 38 epidemiological studies, 21 studies on psychomotor effects, and pertinent laws and policies were reviewed.

FINDINGS: Epidemiological data reveal an increasing prevalence of opioid-positive drivers and an association between opioid use and elevated risk of motor vehicle collisions. Psychomotor studies show mixed results, with some indicating impairment in opioid users and others suggesting minimal effects on driving ability. State laws regarding DUID remain heterogeneous, with trends toward expanded testing powers, lower impairment thresholds, and limitations on prescription-based defenses. The lack of standardized opioid testing limits and inconsistent policy approaches across states hinder effective management of opioid-related impaired driving.

CONCLUSIONS: A balanced public health approach can reduce opioid-involved crashes through education, prevention, enhanced enforcement tools, and rehabilitation. In drafting future DUID laws, policymakers must analyze evolving opioid research when balancing the pain relief of opioids with public roadway safety.

PMID:39956556 | DOI:10.1136/rapm-2024-105955

Trop Med Int Health. 2025 Feb 16. doi: 10.1111/tmi.14091. Online ahead of print.

ABSTRACT

BACKGROUND: Drug-related adverse events cause poorer treatment outcomes amongst people with multi-drug-resistant tuberculosis, exacerbating a major global public health problem. The Harnessing new mHealth technologies to Strengthen the Management of Multi-Drug-Resistant Tuberculosis in Vietnam (V-SMART) trial tests whether a mobile health (mHealth) application (app) can optimise management of drug-related adverse events, within routine health services in Vietnam. Implementation of digital health within routine services is complex and driven by behaviour change as well as a range of health system factors. Understanding implementation is key to informing the evidence base for digital health prior to scale up, despite its potential appeal.

METHODS: Through a process evaluation of the V-SMART trial, we aim to (i) understand the multi-drug-resistant tuberculosis service delivery context and how trial procedures are implemented within services; (ii) describe 'dose' and 'reach' of the app; and (iii) understand health worker and patient perspectives of app implementation and identify areas for improvement. To achieve this, we will (i) conduct process maps (patient flow maps) to describe implementation of the mHealth intervention within routine multi-drug-resistant tuberculosis health services including adverse event management pathways at different levels of the health system; (ii) measure app usage by all participating health workers and people with multi-drug-resistant tuberculosis over time; and (iii) conduct a total of up to 45 semi-structured interviews in seven provinces, with people with multi-drug-resistant tuberculosis, health workers, and policymakers, to identify determinants of app uptake and suggestions for future person-centred app design. Interview topic guides are informed by the Theoretical Framework for Acceptability, Normalisation Process Theory, and the Tailored Implementation of Chronic Diseases framework respectively.

DISCUSSION: The process evaluation will strongly complement the parent trial impact evaluation, and the economic evaluation. Moreover, it will inform future tailored approaches to scaling up digital health as part of broader health system strengthening initiatives.

PMID:39956136 | DOI:10.1111/tmi.14091

Eur J Immunol. 2025 Feb;55(2):e202451225. doi: 10.1002/eji.202451225.

ABSTRACT

Regulatory T cells (Tregs) are a subset of T cells defined by the expression of Forkhead box protein P3 (FoxP3) playing a crucial role in regulating effector T cell activity. Tregs accumulate in the tumor microenvironment facilitating tumor growth. Thus, targeting FoxP3+ Tregs could improve cancer immunotherapies. Here, we conducted a high-throughput, phenotypic screening of a drug repurposing library to identify compounds downregulating FoxP3 expression in human primary T cells. We identified the tyrosine kinase inhibitor bosutinib and the STAT3 inhibitor nifuroxazide effectively downregulating FoxP3 expression. To identify more potent compounds, structural analogs of these two compounds were searched and validated. These analogs were found to reduce FoxP3 expression in a similar- or more potent manner than the original hits. All compounds inhibited Treg suppressive functions and reduced the expression of Treg activation markers. Importantly, bosutinib disrupted FAK and CaMKII signaling more potently in Tregs, whilst nifuroxazide and its analog NA16 targeted STAT3 protein levels more effectively in Tregs. Additionally, bosutinib and NA16 targeted effector Tregs more effectively than other Treg subsets. In summary, bosutinib, nifuroxazide, and their analogs inhibited FoxP3 expression, Treg suppressive abilities, and Treg activation effectively, which could serve as tools for the improvement of current cancer immunotherapies.

PMID:39955647 | DOI:10.1002/eji.202451225

Asian Pac J Allergy Immunol. 2025 Feb 16. doi: 10.12932/AP-131223-1749. Online ahead of print.

ABSTRACT

BACKGROUND: Atopic dermatitis (AD) and food allergy (FA) often originate early in life. Gut microbiota interactions with the host immune system influence allergy development, yet the distinct gut microbiome and functional profiles in individuals with AD, FA, or both AD+FA remain underexplored.

OBJECTIVE: We investigated microbial colonization and proteomic profiles in infants with AD, FA, and AD+FA compared to age- and sex-matched controls from the Allergy Development in Early Life and Associated Factors in the Thai Birth Cohort (ALICE).

METHODS: Gut microbiomes from stool samples were analyzed using 16S sequencing, and proteomic analysis was conducted by liquid chromatography-tandem mass spectrometry.

RESULTS: The study included 16 AD, 5 FA, 5 AD+FA subjects, and 26 controls. AD+FA group exhibited the most severe dysbiosis. Enrichment of proteins involved in methionine biosynthesis in Bifidobacterium scardovii and high Erysipelotrichaceae colonization suggest a link to high-fat diets, known to reduce intestinal short-chain fatty acid and serotonin levels, contributing to allergies. Erysipelotrichaceae in AD+FA groups also expressed proteins related to histidine degradation. Low Bifidobacteriaceae levels were noted in FA and AD+FA, with more pathogenic strains colonized. Increased Bacteroidaceae in FA and AD+FA and Enterobacteriaceae in FA were detected. Pathways involving vitamin B1, a ligand for proliferator-activated receptor-γ (PPAR-γ) from Enterobacteriaceae could promote TH2 cells, type 2 innate lymphoid cells, and M2 macrophages, likely contribute to allergic inflammation.

CONCLUSIONS: AD+FA phenotype exhibited the most distinctive gut microbiome alterations, highlighting unique dysbiosis patterns. Microbiome biosynthesis pathways involving metabolism of methionine, histidine, serotonin, and vitamin B1 point to new targets for modifying or treating AD and FA.

PMID:39955638 | DOI:10.12932/AP-131223-1749

Cancer. 2025 Feb 15;131(4):e35768. doi: 10.1002/cncr.35768.

ABSTRACT

BACKGROUND: Borrmann type-4 (B-4) advanced gastric cancer is challenging to diagnose through routine endoscopy, leading to a poor prognosis. The objective of this study was to develop an artificial intelligence (AI)-based system capable of detecting B-4 gastric cancers using upper endoscopy.

METHODS: Endoscopic images from 259 patients who were diagnosed with B-4 gastric cancer and 595 controls who had benign conditions were retrospectively collected from Seoul National University Hospital for training and testing. Internal validation involved prospectively collected endoscopic videos from eight patients with B-4 gastric cancer and 148 controls. For external validation, endoscopic images and videos from patients with B-4 gastric cancer and controls at the Seoul National University Bundang Hospital were used. To calculate patient-based accuracy, sensitivity, and specificity, a diagnosis of B-4 was made for patients in whom greater than 50% of the images were identified as B-4 gastric cancer.

RESULTS: The accuracy of the patient-based diagnosis was highest in the internal image test set, with accuracy, sensitivity, and specificity of 93.22%, 92.86%, and 93.39%, respectively. The accuracy of the model in the internal validation videos, the external validation images, and the external validation videos was 91.03%, 91.86%, and 86.71%, respectively. Notably, in both the internal and external video sets, the AI model demonstrated 100% sensitivity for diagnosing patients who had B-4 gastric cancer.

CONCLUSIONS: An innovative AI-based model was developed to identify B-4 gastric cancer using endoscopic images. This AI model is specialized for the highly sensitive detection of rare B-4 gastric cancer and is expected to assist clinicians in real-time endoscopy.

PMID:39955610 | DOI:10.1002/cncr.35768

Cell Rep Methods. 2025 Feb 24;5(2):100990. doi: 10.1016/j.crmeth.2025.100990. Epub 2025 Feb 14.

ABSTRACT

The need for a deeper understanding of adverse drug reaction (ADR) mechanisms is vital for improving drug safety and repurposing. This study introduces Drug Adverse Reaction Mechanism Explainer (DREAMER), a network-based framework that uses a comprehensive knowledge graph to uncover molecular mechanisms underlying ADRs and disease phenotypes. By examining shared phenotypes of drugs and diseases and their effects on protein-protein interaction networks, DREAMER identifies proteins linked to ADR mechanisms. Applied to 649 ADRs, DREAMER identified molecular mechanisms for 67 ADRs, including ventricular arrhythmia and metabolic acidosis, and emphasized pathways like GABAergic signaling and coagulation proteins in personality disorders and intracranial hemorrhage. We further demonstrate the application of DREAMER in drug repurposing and propose sotalol, ranolazine, and diltiazem as candidate drugs to be repurposed for cardiac arrest. In summary, DREAMER effectively detects molecular mechanisms underlying phenotypes, emphasizing the importance of network-based analyses with integrative data for enhancing drug safety and accelerating the discovery of novel therapeutic strategies.

PMID:39954672 | DOI:10.1016/j.crmeth.2025.100990

Cell Rep Methods. 2025 Feb 10:100990. doi: 10.1016/j.crmeth.2025.100990. Online ahead of print.

ABSTRACT

The need for a deeper understanding of adverse drug reaction (ADR) mechanisms is vital for improving drug safety and repurposing. This study introduces Drug Adverse Reaction Mechanism Explainer (DREAMER), a network-based framework that uses a comprehensive knowledge graph to uncover molecular mechanisms underlying ADRs and disease phenotypes. By examining shared phenotypes of drugs and diseases and their effects on protein-protein interaction networks, DREAMER identifies proteins linked to ADR mechanisms. Applied to 649 ADRs, DREAMER identified molecular mechanisms for 67 ADRs, including ventricular arrhythmia and metabolic acidosis, and emphasized pathways like GABAergic signaling and coagulation proteins in personality disorders and intracranial hemorrhage. We further demonstrate the application of DREAMER in drug repurposing and propose sotalol, ranolazine, and diltiazem as candidate drugs to be repurposed for cardiac arrest. In summary, DREAMER effectively detects molecular mechanisms underlying phenotypes, emphasizing the importance of network-based analyses with integrative data for enhancing drug safety and accelerating the discovery of novel therapeutic strategies.

PMID:39954672 | DOI:10.1016/j.crmeth.2025.100990

Expert Opin Drug Discov. 2025 Feb 15. doi: 10.1080/17460441.2025.2465373. Online ahead of print.

ABSTRACT

INTRODUCTION: Bladder cancer presents a significant health problem worldwide, with environmental and genetic factors contributing to its incidence. Histologically, it can be classified as carcinoma in situ, non-muscle invasive and muscle-invasive carcinoma, each one with distinct genetic alterations impacting prognosis and response to therapy. While traditional transurethral resection is commonly performed in carcinoma in situ and non-muscle invasive carcinoma, it often fails to prevent recurrence or progression to more aggressive phenotypes, leading to the frequent need for additional treatment such as intravesical chemotherapy or immunotherapy. Despite the advances made in recent years, treatment options for bladder cancer are still lacking due to the complex nature of this disease. So, animal models may hold potential for addressing these limitations, because they not only allow the study of disease progression but also the evaluation of therapies and the investigation of drug repositioning.

AREAS COVERED: This review discusses the use of animal models over the past decade, highlighting key discoveries and discussing advantages and disadvantages for new drug discovery.

EXPERT OPINION: Over the past decade animal models have been employed to evaluate new mechanisms underlying the responses to standard therapies, aiming to optimize bladder cancer treatment. The authors propose that molecular engineering techniques and AI may hold promise for the future development of more precise and effective targeted therapies in bladder cancer.

PMID:39954010 | DOI:10.1080/17460441.2025.2465373

Cancer Chemother Pharmacol. 2025 Feb 15;95(1):34. doi: 10.1007/s00280-025-04759-8.

ABSTRACT

PURPOSE: In 20-30% of the patients, fluoropyrimidines (5-FU) based chemotherapy leads to severe toxicity, which is associated with dihydropyridine dehydrogenase (DPD) deficiency. Therefore, DPYD genotyping became standard practice before treatment with fluoropyrimidines. Nevertheless, only 17% of the patients with severe toxicity have a DPYD variant. Therefore, an urgent need persists to investigate other strategies contributing to prediction and prevention of toxicity. Endogenous DPD substrates are considered as potential biomarkers to predict toxicity, yet contradictional data exist on demonstrating uracil as a reliable biomarker. Thymine as biomarker for toxicity has been investigated less. The aim of this study was to determine the association between the concentrations of uracil, thymine dihydrouracil (DHU) and dihydrothymine (DHT), with the systemic drug exposure of 5-FU and DPD enzyme activity in patients treated with 5-FU.

METHODS: We included 36 patients with gastrointestinal malignancy who received 5-FU infusion. DPYD genotyping was conducted before start of treatment. Blood samples for determining 5-FU, uracil and thymine concentrations during infusion and DPD enzyme activity were taken.

RESULTS: We found a significant correlation between the 5-FU systematic exposure and baseline thymine concentrations (R2 = 0.1468; p = 0.0402). DPD enzyme activity was significantly correlated with baseline thymine concentrations but no correlation was found between DPD enzyme activity and 5-FU systemic drug exposure.

CONCLUSION: 5-FU dose individualization based on thymine concentrations could be a promising addition to DPYD genotyping to predict 5-FU-induced toxicity. Larger prospective trials are needed to examine thymine as predictor for toxicity in daily practice.

TRIAL REGISTRATION: Trial NL7539 at 'Overview of Medical Research in the Netherlands' (ID NL-OMON21471). Date of registration 19-02-2019.

PMID:39955449 | DOI:10.1007/s00280-025-04759-8

J Genet Genomics. 2025 Feb 13:S1673-8527(25)00038-4. doi: 10.1016/j.jgg.2025.02.001. Online ahead of print.

ABSTRACT

Founder events influence recessive diseases in highly endogamous populations. Several Indian populations have experienced significant founder events due to strict endogamy. However, the clinical implications of it remain underexplored. Therefore, we perform whole-exome sequencing of 281 individuals from four South Indian populations, characterized by high IBD scores. Our study reveals a high inbreeding rate of 59% across the populations. We identify ∼29.2% of the variants that are exclusively present in a single population and uncovered 1284 unreported exonic variants, underscoring the underrepresentation of Indian populations in global databases. Among these, 23 are predicted to be deleterious, all present in heterozygous state may be pathogenic when homozygous, an expected phenomenon in endogamous populations. Approximately 16%-33% of the identified pathogenic variants showed significantly higher occurrence rates compared to the South Asian populations from 1000 Genomes dataset. Pharmacogenomic analysis revealed distinct allele frequencies of variants in CYP450 and non-CYP450 genes, highlighting heterogeneous drug responses and associated risks. We report a high prevalence of ankylosing spondylitis in Reddy population, linked to HLA-B*27:04 allele and strong founder effect. Our findings highlight the need for extensive genomic research in understudied Indian populations for better understanding of disease risk and evolving strategies for precision and preventive medicine.

PMID:39955025 | DOI:10.1016/j.jgg.2025.02.001

J Psychiatr Res. 2025 Feb 6;183:106-111. doi: 10.1016/j.jpsychires.2025.02.005. Online ahead of print.

ABSTRACT

INTRODUCTION: Although post-COVID major depressive disorder (MDD) is frequent, the physiological mechanisms associated with it remain unclear. This study aimed to assess the association between 10 residual blood markers of inflammation and the presence of MDD 4 months after the acute phase of COVID-19.

METHODS: This is a cross-sectional study of the COMEBAC cohort that followed patients 4 months after hospitalization for COVID-19 at Bicêtre Hospital. Patients with lingering symptoms or who had been in critical care (n = 177) were invited to a day hospital for assessment of MDD and peripheral inflammation. Ten peripheral inflammatory markers were examined: plasmatic C-reactive protein; leukocyte, monocyte, neutrophil, and lymphocyte counts; the neutrophil to lymphocyte ratio; the systemic inflammatory index (i.e., the (platelet x neutrophil) to lymphocyte ratio); cortisol, ferritin, and hemoglobin levels. Current MDD was assessed through structured interviews with a psychiatrist, depressive symptoms through self-questionnaires. Peripheral inflammatory markers were compared between patients with post-COVID MDD and patients without a lifetime history of psychiatric disorders (controls).

RESULTS: Out of 177 patients, 24 (13.6%) had MDD. No significant differences in peripheral inflammatory markers were observed between patients with post-COVID MDD and controls. Furthermore, peripheral inflammatory markers were not correlated with symptoms of depression.

CONCLUSION: We found no association between post-COVID MDD and 10 peripheral inflammatory markers 4 months after COVID-19 infection. Other potential mechanisms warrant investigation.

PMID:39954540 | DOI:10.1016/j.jpsychires.2025.02.005

Lancet Respir Med. 2025 Feb 12:S2213-2600(25)00017-7. doi: 10.1016/S2213-2600(25)00017-7. Online ahead of print.

NO ABSTRACT

PMID:39954705 | DOI:10.1016/S2213-2600(25)00017-7

BMC Oral Health. 2025 Feb 15;25(1):246. doi: 10.1186/s12903-025-05588-0.

ABSTRACT

OBJECTIVES: This study aimed to develop a cephalometric classification method based on multitask learning for eight diagnostic classifications.

METHODS: This study was retrospective. A total of 3,310 lateral cephalograms were collected to construct a dataset. Eight clinical classifications were employed, including sagittal and vertical skeletal facial patterns, maxillary and mandibular anteroposterior positions, inclinations of upper and lower incisors, as well as their anteroposterior positions. The images were manually annotated for initially classification, which was verified by senior orthodontists. The data were randomly divided into training, validation, and test sets at a ratio of approximately 8:1:1. The multitask learning classification model was constructed based on the ResNeXt50_32 × 4d network and consisted of shared layers and task-specific layers. The performance of the model was evaluated using classification accuracy, precision, sensitivity, specificity and area under the curve (AUC).

RESULTS: This model could perform eight clinical diagnostic classifications on cephalograms within an average of 0.0096 s. The accuracy of the six classifications was 0.8-0.9, and the accuracy of the two classifications was 0.75-0.8. The overall AUC values for each classification exceeded 0.9.

CONCLUSIONS: An automatic diagnostic classification model for lateral cephalograms was established based on multitask learning to achieve simultaneous classification of eight common clinical diagnostic items. The multitask learning model achieved better classification performance and reduced the computational costs, providing a novel perspective and reference for addressing such problems.

PMID:39955570 | DOI:10.1186/s12903-025-05588-0

BMC Med Inform Decis Mak. 2025 Feb 15;25(1):81. doi: 10.1186/s12911-025-02923-x.

ABSTRACT

The diagnosis of pancreatic cancer presents a significant challenge due to the asymptomatic nature of the disease and the fact that it is frequently detected at an advanced stage. This study presents a novel approach combining graph-based data representation with DARTS-optimised MobileViT models, with the objective of enhancing diagnostic accuracy and reliability. The images of the pancreatic CT were transformed into graph structures using the Harris Corner Detection algorithm, which enables the capture of complex spatial relationships. Subsequently, the graph representations were processed using MobileViT models that had been optimised with Differentiable Architecture Search (DARTS), thereby enabling dynamic architectural adaptation. To further enhance classification accuracy, advanced machine learning algorithms, including K-Nearest Neighbours (KNN), Support Vector Machines (SVM), Random Forest (RF), and XGBoost, were applied. The MobileViTv2_150 and MobileViTv2_200 models demonstrated remarkable performance, with an accuracy of 97.33% and an F1 score of 96.25%, surpassing the capabilities of traditional CNN and Vision Transformer models. This innovative integration of graph-based deep learning and machine learning techniques demonstrates the potential of the proposed method to establish a new standard for early pancreatic cancer diagnosis. Furthermore, the study highlights the scalability of this approach for broader applications in medical imaging, which could lead to improved patient outcomes.

PMID:39955532 | DOI:10.1186/s12911-025-02923-x

Int J Comput Assist Radiol Surg. 2025 Feb 15. doi: 10.1007/s11548-024-03290-0. Online ahead of print.

ABSTRACT

OBJECTIVE: BRAFV600E is the most common mutation found in thyroid cancer and is particularly associated with papillary thyroid carcinoma (PTC). Currently, genetic mutation detection relies on invasive procedures. This study aimed to extract radiomic features and utilize deep transfer learning (DTL) from ultrasound images to develop a noninvasive artificial intelligence model for identifying BRAFV600E mutations.

MATERIALS AND METHODS: Regions of interest (ROI) were manually annotated in the ultrasound images, and radiomic and DTL features were extracted. These were used in a joint DTL-radiomics (DTLR) model. Fourteen DTL models were employed, and feature selection was performed using the LASSO regression. Eight machine learning methods were used to construct predictive models. Model performance was primarily evaluated using area under the curve (AUC), accuracy, sensitivity and specificity. The interpretability of the model was visualized using gradient-weighted class activation maps (Grad-CAM).

RESULTS: Sole reliance on radiomics for identification of BRAFV600E mutations had limited capability, but the optimal DTLR model, combined with ResNet152, effectively identified BRAFV600E mutations. In the validation set, the AUC, accuracy, sensitivity and specificity were 0.833, 80.6%, 76.2% and 81.7%, respectively. The AUC of the DTLR model was higher than that of the DTL and radiomics models. Visualization using the ResNet152-based DTLR model revealed its ability to capture and learn ultrasound image features related to BRAFV600E mutations.

CONCLUSION: The ResNet152-based DTLR model demonstrated significant value in identifying BRAFV600E mutations in patients with PTC using ultrasound images. Grad-CAM has the potential to objectively stratify BRAF mutations visually. The findings of this study require further collaboration among more centers and the inclusion of additional data for validation.

PMID:39955452 | DOI:10.1007/s11548-024-03290-0

NPJ Digit Med. 2025 Feb 15;8(1):105. doi: 10.1038/s41746-025-01496-3.

ABSTRACT

Primary cutaneous squamous cell carcinoma (cSCC) is responsible for ~10,000 deaths annually in the United States. Stratification of risk of poor outcome at initial biopsy would significantly impact clinical decision-making during the initial post operative period where intervention has been shown to be most effective. Using whole-slide images (WSI) from 163 patients from 3 institutions, we developed a self supervised deep-learning model to predict poor outcomes in cSCC patients from histopathological features at initial diagnosis, and validated it using WSI from 563 patients, collected from two other academic institutions. For disease-free survival prediction, the model attained a concordance index of 0.73 in the development cohort and 0.84 in the Mayo cohort. The model's interpretability revealed that features like poor differentiation and deep invasion were strongly associated with poor prognosis. Furthermore, the model is effective in stratifying risk among BWH T2a and AJCC T2, known for outcome heterogeneity.

PMID:39955424 | DOI:10.1038/s41746-025-01496-3