This study evaluates the effectiveness of transfer learning with pre-trained convolutional neural networks (CNNs) for the automated binary classification of surgical suture quality (high-quality/low-quality) using photographs of three suture types: interrupted open vascular sutures (IOVS), continuous over-and-over open sutures (COOS), and interrupted laparoscopic sutures (ILS). To address the challenge of limited medical data, eight state-of-the-art CNN architectures—EfficientNetB0, ResNet50V2, MobileNetV3Large, VGG16, VGG19, InceptionV3, Xception, and DenseNet121—were trained and validated on small datasets (100–190 images per type) using 5-fold cross-validation. Performance was assessed using the F1-score, AUC-ROC, and a custom weighted stability-aware score (Scoreadj). The results demonstrate that transfer learning achieves robust classification (F1 > 0.90 for IOVS/ILS, 0.79 for COOS) despite data scarcity. ResNet50V2, DenseNet121, and Xception were more stable by Scoreadj, with ResNet50V2 achieving the highest AUC-ROC (0.959 ± 0.008) for IOVS internal view classification. GradCAM visualizations confirmed model focus on clinically relevant features (e.g., stitch uniformity, tissue apposition). These findings validate transfer learning as a powerful approach for developing objective, automated surgical skill assessment tools, reducing reliance on subjective expert evaluations while maintaining accuracy in resource-constrained settings.
Wound infections, exacerbated by antibiotic resistance and biofilm formation, necessitate alternative therapeutic options. One promising approach is topical bacteriotherapy with live probiotics. The work aimed to summarize existing research on the multifaceted effects of live probiotics in infected wound healing. The review examines how probiotics directly influence pathogenic wound microflora by competing for adhesion sites and nutrients, secreting antagonistic compounds, altering the wound environment, and promoting co-aggregation. Probiotics’ antibiofilm activity is highlighted, including inhibition of quorum sensing systems and disruption of the extracellular biofilm matrix. Moreover, the review discusses the findings of in vitro and in vivo studies on the effects of live probiotics on macroorganisms. Probiotics have been shown to activate signaling pathways that regulate macrophage shift from a pro-inflammatory (M1) phenotype to a pro-repair (M2) phenotype, cytokine synthesis, keratinocyte and fibroblast migration and proliferation, angiogenesis, and collagen matrix formation. Thus, the multifaceted effects of probiotics cannot be replicated by metabolites or inactivated cells alone, highlighting the need for effective delivery systems for live probiotic cultures in infected wound management.
5 November 2026 in Siberian Scientific Medical Journal
Infected wounds are a severe medical problem that requires a comprehensive approach to diagnosis, treatment, and prevention. Adequate and modern model systems are needed for a study of the pathophysiology of wound infections, as well as for the development and testing of new therapeutic strategies. This review provides a comprehensive analysis of current models aimed at recreating the healing processes of infected skin wounds. The search for sources was conducted in the PubMed database using the keywords «infected wound», «model», « in vivo », « in vitro », and « ex vivo », excluding the words «burn» and «diabetic». The search depth was 22 years: from 2003 to 2025. A comparative study of in vitro , ex vivo , and in vivo models has been conducted, including their methodological features, advantages, disadvantages, and translational effectiveness. Particular attention is paid to the analysis of each model in the context of reproducing the pathophysiological processes occurring in infected wounds. The result of the work is a comparative characteristic of the models, which makes it possible to assess their economic feasibility, deadlines, labor costs, complexity of implementation and reproducibility. To achieve more relevant and accurate results, the integration of data obtained in various model systems is promising.