Bioengineering

Background/PurposeDiagnosing wrist ligament injuries is challenging; early detection and treatment are important to prevent osteoarthritis progression. Interosseous proximity maps, a proxy measure for joint space, can be generated from volumetric imaging data and may provide important information about wrist health. Artificial intelligence (AI) could enhance accuracy of noninvasive diagnosis based on imaging-derived metrics. This work demonstrates feasibility of AI training using synthetic proximity map data generated from finite element models (FEMs). MethodsPersonalized wrist FEMs for two asymptomatic participants were created from four-dimensional computed tomography-derived anatomic and kinematic data. Monte Carlo sampling varied 22 ligament material properties and simulated 7,500 unique injury scenarios generating 9,000,000 labeled red, green, and blue (RGB) images of interosseous proximity vector fields from FEM-derived motions. Images were associated with 17 descriptive metrics, including gross wrist angles and bone surface pairs, and used to develop mixed-input convolutional neural networks (CNNs). Model performance was evaluated for identifying specific ligament injuries. ResultsAverage area under receiver operating characteristic curve (AUROC) for CNNs was 0.757 across all injury types and kinematics. In a subset with clinically-relevant functional angles, the average AUROC was 0.824. Best-performing individual ligament AUROCs ranged from 0.807 to 0.999. Sensitivities and specificities exceeded 0.99 for some ligament injury simulations under specific wrist angles and bone surface pairs. ConclusionThis study demonstrates the feasibility of using synthetic data from FEMs to train AI models for classifying wrist ligament injuries. Proximity-based RGB images may be a relevant biomarker of ligamentous injury.

PurposeThe ability to detect, monitor, and precisely time the morphokinetic stages of human pre -implantation embryo development plays a critical role in assessing their viability and potential for successful implantation. Therefore, there is a need for accurate and accessible tools to analyse embryos. This work describes a highly accurate, machine learning model designed to predict 17 morphokinetic stages of pre-implantation human development, an improvement on existing models. This model provides a robust tool for researchers and clinicians, enabling the automation of morphokinetic stage prediction, standardising the process, and reducing subjectivity between clinics. MethodA computer vision model was built on a publicly available dataset for embryo Morphokinetic stage detection. The dataset contained 273,438 labelled images based on Embryoscope/+(C) embryo images. The dataset was split 70/10/20 into training/validation/test sets. Two different deep learning architectures were trained and tested, one using EfficientNet-V2-Large and the other using EfficientNet-V2-Large with the addition of fertilisation time as input. A new postprocessing algorithm was developed to reduce noise in the predictions of the deep learning model and detect the exact time of each morphokinetic stage change. ResultsThe proposed model reached an overall test accuracy of 87% across 17 morphokinetic stages on an independent test set. ConclusionThe proposed model shows a 17% accuracy improvement, compared to the best models on the same dataset. Therefore, our model can accurately detect morphokinetic stages in static embryo images as well as detecting the exact timings of stage changes in a complete time-lapse video.

Background: Midurethral sling placement is often performed during prolapse repair to treat or prevent stress urinary incontinence. However, some women experience persistent or new-onset stress or urgency urinary incontinence after surgery. It is unclear how prolapse repair, with or without a concomitant midurethral sling, alters urethral morphology and support, and how these changes relate to urinary continence outcomes. Objectives: To compare postoperative urethral morphology (dimensions, angles, shape) and support (position, mobility) after transvaginal prolapse repair with vs without a concurrent midurethral sling, and to explore associations between postoperative urethral characteristics and urinary outcomes (stress, urgency symptoms). Study Design: This ancillary analysis used magnetic resonance imaging and urinary outcome data from the Defining Mechanisms of Anterior Vaginal Wall Descent Study conducted across 8 clinical sites within the United States Pelvic Floor Disorders Network. Eighty-two women (median age, 65 years) underwent transvaginal prolapse repair (vaginal mesh hysteropexy or vaginal hysterectomy with uterosacral ligament suspension) with or without a concurrent midurethral sling between April 2013 and February 2015. Postoperative imaging at rest and during strain was performed 30-42 months after surgery (or earlier if they chose reoperation) between June 2014 and May 2018. Prolapse recurrence, defined as descent beyond the vaginal introitus during strain, was recorded. The urethra was segmented from postoperative scans to create 3-dimensional models for measuring urethral diameters, length, surface area, volume, angles, shape (principal component scores from a statistical shape model), position, and mobility (rest-to-strain displacement). Preoperative and 24-48-month postoperative urinary continence outcomes were assessed using validated questionnaires: the Urogenital Distress Inventory, Urinary Impact Questionnaire, and the Incontinence Severity Index. Comparisons of urethral and urinary outcomes by (1) midurethral sling and (2) stress urinary incontinence were made using Wilcoxon rank-sum tests, principal component analysis, and multivariate models as appropriate. Associations between urethral and urinary outcomes were evaluated with Spearmans rank correlation. Results: Forty-six women (22 hysteropexy, 24 hysterectomy) were in the sling group, and 36 (19 hysteropexy, 17 hysterectomy) were in the no-sling group. Among the 48 women without prolapse recurrence (28 sling, 20 no-sling), those with a sling (vs without) had larger urethral dimensions (all P<.03), a more anterior-superior position of the proximal urethra (indicating better bladder neck support) (P=.04), a straighter urethral shape (P=.006), and reported less bothersome postoperative stress incontinence (P=.02). Overall, 14 women (17%) experienced postoperative stress incontinence. Stress urinary incontinence was linked to a more acute proximal urethral sagittal angle (more aligned with axial plane) (P=.01), and a lower proximal urethra position (P=.04) and mid-urethra position (P=.03). Poorer stress and urgency urinary outcomes were associated with a shorter urethral length (P=.01), a more posterior-inferior urethral position (all P<.05), increased C or S-shaped urethral concavity (P=.008; P=.006), and smaller rest-to-strain displacement of the proximal (P=.03) and distal (P=.009) urethra. Conclusions: Urethral morphology and support differed with concomitant midurethral sling (vs no sling) and stress urinary incontinence after vaginal surgery. Urethral characteristics were also associated with postoperative urinary symptoms. Urethral configuration may influence urinary outcomes and could be considered during prolapse and stress urinary incontinence repairs.

Objective: To develop and validate a robust deep-learning model capable of fine-grained phase recognition in total hysterectomy, particularly the complex periuterine dissection phase. Design: Multicentre retrospective observational study. Setting: Japan. Sample: Surgical videos (n = 764) from 43 institutions. Methods: We developed a robust and generalisable deep-learning model for surgical phase recognition in total hysterectomy, applicable to laparoscopic and robot-assisted procedures. Overall, 1,591,334 still images were annotated across nine surgical phases. A convolutional neural network (Xception architecture) was trained on 200 cases using four-fold cross-validation, with institutional separation between training and testing sets. Main outcome measures: Model performance was assessed using accuracy, precision, recall, and F1 score. Subgroup analysis and logistic regression evaluated the association between background clinical factors and recognition accuracy. Results: The model achieved an overall phase recognition accuracy of 0.78 (95% CI: 0.74--0.80), with a precision of 0.75 (95% CI: 0.72--0.78) and a recall of 0.76 (95% CI: 0.74--0.78). Performance was consistent across laparoscopic and robot-assisted procedures and across most surgical phases. Accuracy plateaued after training on 120 cases. No clinical factors significantly impacted performance. Trends toward lower accuracy were observed for cases with cervical myoma and pouch of Douglas adhesions. Conclusions: This model demonstrated high accuracy across diverse institutions and patient backgrounds. Its potential applications include surgical education, real-time intraoperative support, and training efficiency enhancement.

PurposePrediction of Ectasia Screening Index (ESI), an estimator provided by the Casia2 for identifying keratoconus, from raw Optical Coherence Tomography (OCT) data with Convolutional Neural Networks (CNN). MethodsThree CNN architectures (ResNet18, DenseNet121 and EfficientNetB0) were employed to predict the ESI. Mean Absolute Error (MAE) was used as the performance metric for predicting the ESI by the adapted CNN models on the test set. Scans with an ESI value higher than a certain threshold were classified as Keratoconus, while the remaining scans were classified as Not Keratoconus. The models performance was evaluated using metrics such as accuracy, sensitivity, specificity, Positive Predictive Value (PPV) and F1 score on data collected from patients examined at the eye clinic of the Homburg University Hospital. The raw data from the Casia2 device, in 3dv format, was converted into 16 images per examination of one eye. For the training, validation and testing phases, 3689, 1050 and 1078 scans (3dv files) were selected, respectively. ResultsIn the prediction of the ESI, the MAE values for the adapted ResNet18, DenseNet121 and EfficientNetB0, rounded to two decimal places, were 7.15, 6.64 and 5.86, respectively. In the classification task, the three networks yielded an accuracy of 94.80%, 95.27% and 95.83%, respectively; a sensitivity of 92.07%, 94.64% and 94.17%, respectively; a specificity of 96.61%, 95.69% and 96.92%, respectively; a PPV of 94.72%, 93.55% and 95.28%, respectively; and a F1 score of 93.38%, 94.09% and 94.72%, respectively. ConclusionsOur results show that the prediction of keratokonus based on the ESI values estimated from raw data outperforms previous approaches using processed data. Adapted EfficientNetB0 outperformed both the other adapted models and those in state-of-the-art studies, with the highest accuracy and F1 score.

ObjectiveTo evaluate the feasibility of selective removal of endometriotic lesions using the Cavitron Ultrasonic Surgical Aspirator (CUSA(R) Clarity) in ovarian endometriomas, with a focus on histological preservation of normal ovarian tissue. MethodsWe analyzed tissue from a woman in her early 30s who underwent laparoscopic surgery for an ovarian endometrioma measuring approximately 7 cm after preoperative dienogest therapy. Resected cyst wall specimens were divided into five parts, each assigned to a Tissue Select(R) setting (0-4). Samples were scraped with CUSA, followed by histological and immunohistochemical evaluation (H&E, Sirius Red, CK7, CD10). ResultsEndometriotic lesions (epithelial and stromal cells) were effectively removed across all settings. At higher Tissue Select settings (3-4), preservation of surrounding tissue was superior, with minimal vacuolization compared to lower settings (0-2). Primordial follicles were observed approximately 600 m beneath the surface, highlighting the importance of limiting cavitation depth. ConclusionCUSA Clarity enabled selective removal of endometriotic lesions with relative preservation of normal ovarian tissue, particularly at higher Tissue Select settings. This novel approach may represent a fertility-preserving alternative to cystectomy or laser ablation in the management of ovarian endometriomas. Further studies are warranted.

PurposeTo develop a deep learning model - Cornea nnU-Net Extractor (CUNEX) - for full-thickness corneal segmentation of anterior segment optical coherence tomography (AS-OCT) images and evaluate its utility in artificial intelligence (AI) research. MethodsWe trained and evaluated CUNEX using nnU-Net on 600 AS-OCT images (CSO MS-39) from 300 patients: 100 normal, 100 keratoconus (KC), and 100 Fuchs endothelial corneal dystrophy (FECD) eyes. To assess generalizability, we externally validated CUNEX on 1,168 AS-OCT images from an infectious keratitis dataset acquired from a different device (Casia SS-1000). We benchmarked CUNEX against two recent models, CorneaNet and ScLNet. We then applied CUNEX to our dataset of 194,599 scans from 37,499 patients as preprocessing for a classification model evaluating whether segmentation improves AI prediction, including age, sex, and disease staging (KC and FECD). ResultsCUNEX achieved Dice similarity coefficient (DSC) and intersection over union (IoU) scores ranging from 94-95% and 90-99%, respectively, across healthy, KC, and FECD eyes. This was similar to ScLNet (within 3%) but better than CorneaNet (8-35% lower). On external validation, CUNEX maintained high performance (DSC 83%; IoU 71%) while ScLNet (DSC 14%; IoU 8%) and CorneaNet (DSC 16%; IoU 9%) failed to generalize. Unexpectedly, segmentation minimally impacted classification accuracy except for sex prediction, where accuracy dropped from 81 to 68%, suggesting sex-related features may lie outside the cornea. ConclusionCUNEX delivers the first open-source generalizable corneal segmentation model using the latest framework, supporting its use in clinical analysis and AI workflows across diseases and imaging platforms. It is available at https://github.com/lkandakji/CUNEX.

Intervertebral disc degeneration is the most recognized cause of low back pain, characterized by the decline of tissue structure and mechanics. Image-based mechanical parameters (e.g., strain, stiffness) may provide an ideal assessment of disc function that is lost with degeneration but unfortunately remains underdeveloped. Moreover, it is unknown whether strain or stiffness of the disc may be predicted by MRI relaxometry (e.g. T1 or T2), an increasingly accepted quantitative measure of disc structure. In this study, we quantified T1 and T2 relaxation times and in-plane strains using displacement-encoded MRI within the disc under physiological levels of compression and bending. We then estimated shear modulus in orthogonal image planes and compared these values to relaxation times and strains within regions of the disc. Intratissue strain depended on the loading mode, and shear modulus in the nucleus pulposus was typically an order of magnitude lower than the annulus fibrosis, except in bending, where the apparent stiffness depended on the loading. Relative shear moduli estimated from strain data derived under compression generally did not correspond with those from bending experiments, with no correlations in the sagittal plane and only 4 of 15 regions correlated in the coronal plane, suggesting that future inverse models should incorporate multiple loading conditions. Strain imaging and strain-based estimation of material properties may serve as imaging biomarkers to distinguish healthy and diseased discs. Additionally, image-based elastography and relaxometry may be viewed as complementary measures of disc structure and function to assess degeneration in longitudinal studies.

Stargardt disease and age-related macular degeneration are the leading causes of blindness in the juvenile and geriatric populations, respectively. The formation of atrophic regions of the macula is a hallmark of the end-stages of both diseases. The progression of these diseases is tracked using various imaging modalities, two of the most common being fundus autofluorescence (FAF) imaging and spectral-domain optical coherence tomography (SD-OCT). This study seeks to investigate the use of longitudinal FAF and SD-OCT imaging (month 0, month 6, month 12, and month 18) data for the predictive modelling of future atrophy in Stargardt and geographic atrophy. To achieve such an objective, we develop a set of novel deep convolutional neural networks enhanced with recurrent network units for longitudinal prediction and concurrent learning of ensemble network units (termed ReConNet) which take advantage of improved retinal layer features beyond the mean intensity features. Using FAF images, the neural network presented in this paper achieved mean ({+/-} standard deviation, SD) and median Dice coefficients of 0.895 ({+/-} 0.086) and 0.922 for Stargardt atrophy, and 0.864 ({+/-} 0.113) and 0.893 for geographic atrophy. Using SD-OCT images for Stargardt atrophy, the neural network achieved mean and median Dice coefficients of 0.882 ({+/-} 0.101) and 0.906, respectively. When predicting only the interval growth of the atrophic lesions with FAF images, mean ({+/-} SD) and median Dice coefficients of 0.557 ({+/-} 0.094) and 0.559 were achieved for Stargardt atrophy, and 0.612 ({+/-} 0.089) and 0.601 for geographic atrophy. The prediction performance in OCT images is comparably good to that using FAF which opens a new, more efficient, and practical door in the assessment of atrophy progression for clinical trials and retina clinics, beyond widely used FAF. These results are highly encouraging for a high-performance interval growth prediction when more frequent or longer-term longitudinal data are available in our clinics. This is a pressing task for our next step in ongoing research.

Recent trends have seen an increase in the use of orthopaedic acetabular implants produced by additive manufacturing. These implants utilise novel trabecular metals with seemingly large variations in their structural designs. There is no available biomolecular comparisons between these different implants. We have therefore created novel techniques to image and analyse 5 different implants. Our results have shown differing responses of Human Bone Marrow Stem Cells cultured on these 5 different implants.

STRUCTURED ABSTRACTO_ST_ABSImportanceC_ST_ABSSexual dysfunction can occur after midurethral sling (MUS) and transvaginal prolapse surgery. It remains unclear whether these procedures impact the clitoris, despite its role in sexual function and proximity to the MUS and vagina. ObjectivesTo compare postoperative sexual function and clitoral features by MUS and vaginal surgery approach after transvaginal prolapse repair with/without concomitant MUS. DesignCross-sectional ancillary study of magnetic resonance imaging (MRI) and sexual function data from the Defining Mechanisms of Anterior Vaginal Wall Descent study. SettingEight clinical sites in the US Pelvic Floor Disorders Network. Participants: 88 women with uterovaginal prolapse who underwent vaginal mesh hysteropexy or vaginal hysterectomy with uterosacral ligament suspension with/without MUS between 2013-2015. Data were analyzed between September 2021-June 2023. ExposuresBetween June 2014-May 2018, participants underwent pelvic MRI 30-42 months after surgery, or earlier if reoperation was desired. Sexual activity and function at baseline and 24-48-month follow-up were evaluated using the Pelvic Organ Prolapse/Incontinence Sexual Questionnaire, IUGA-Revised (PISQ-IR). Clitoral features were obtained from postoperative MRI-based 3-dimensional models. Main Outcomes and MeasuresPISQ-IR scores and clitoral features (size, position). ResultsEighty-two women (median [range] age, 65 [47-79] years) were analyzed: 45 MUS (22 hysteropexy, 23 hysterectomy) and 37 No-MUS (19 hysteropexy, 18 hysterectomy). Postoperatively, 25 MUS, 12 No-MUS, 20 hysteropexy, and 17 hysterectomy patients were sexually active (SA). Overall, within the MUS and vaginal surgery groups, sexual function remained unchanged or improved (most PISQ-IR change from baseline scores were [&ge;]0) among SA and NSA women. Among SA women after surgery, the MUS group (vs No-MUS) had a poorer PISQ-IR arousal/orgasm (SA-AO) score (median, 3.5 vs 4.3; P=.02). The hysteropexy group (vs hysterectomy) had less improvement in PISQ-IR SA-AO score (median, 0.0 vs 0.3; P=.01). Women with MUS (vs without) had a smaller clitoral glans thickness (median, 9.0 mm vs 10.0 mm; P=.008) and clitoral body volume (median, 2783.5 mm3 vs 3587.4 mm3; P=.01). Conclusions and RelevanceSA women with MUS (vs without) or hysteropexy (vs hysterectomy) experienced poorer postoperative sexual function. MUS was linked to a smaller clitoris. Future studies should explore surgery-induced changes in clitoral anatomy and sexual function. KEY POINTSO_ST_ABSQuestionC_ST_ABSHow do sexual function and clitoral anatomy differ by midurethral sling placement and vaginal surgery approach? FindingsThis cross-sectional study compared patient-reported sexual function outcomes and 30-42-month postoperative magnetic resonance imaging-based 3-dimensional clitoral models of 82 women after vaginal prolapse surgery with or without concomitant midurethral sling. Midurethral sling (vs no sling) and vaginal mesh hysteropexy (vs vaginal hysterectomy) were associated with poorer postoperative sexual function outcomes. Additionally, midurethral sling was associated with a smaller clitoral glans and body. MeaningMidurethral sling and vaginal mesh hysteropexy were associated with, and may adversely alter, postoperative sexual function and/or clitoral anatomy. VISUAL ABSTRACT/PROMOTIONAL IMAGE O_FIG O_LINKSMALLFIG WIDTH=200 HEIGHT=113 SRC="FIGDIR/small/26351291v1_ufig1.gif" ALT="Figure 1"> View larger version (33K): org.highwire.dtl.DTLVardef@e99cbborg.highwire.dtl.DTLVardef@130e79eorg.highwire.dtl.DTLVardef@1b64611org.highwire.dtl.DTLVardef@1b228d1_HPS_FORMAT_FIGEXP M_FIG C_FIG

PurposeTo develop a deep learning framework, RPEGENE-Net, capable of predicting gene expression profiles of retinal pigment epithelium (RPE) cells using live-cell microscopy images. MethodsA dataset of live-cell images of RPE cells, treated with various drug regimens and captured at magnifications of 40x, 100x, 200x, and 400x, was used. Gene expression of six key genes involved in epithelial-mesenchymal transition or EMT (including -SMA, ZEB1, TGF-{beta}, CD90, {beta}-catenin, Snail) and treatment classes (Aflibercept, Bevacizumab, Dexamethasone, Aflibercept + Dexamethasone, and untreated control) were analyzed. After preprocessing the image data and gene expression values, we trained and evaluated twelve state-of-the-art deep learning architectures, including three variants of DenseNet, five variants of ResNet, EfficientNet_b5, Inception_v3, RegNet_y_400mf, and a vision transformer model (Swin_b). A two-stage pipeline was implemented, combining autoencoder-based pretraining to extract meaningful features with fine-tuning specifically optimized for gene expression regression and treatment type classification tasks. Features extracted from the second stage across four magnifications were concatenated to generate the final prediction, leveraging multi-scale morphological information for improved accuracy. ResultsDenseNet121 demonstrated superior performance, achieving the highest Pearson correlation coefficients for four genes: -SMA (0.79), ZEB1(0.84), TGF-{beta} (0.83), and Snail (0.86). ResNet34 outperformed other models for CD90 (0.87) and {beta}-catenin (0.85) predictions. The average mean absolute error (MAE) and average root mean square error (RMSE) on test dataset were 0.0244 and 0.1228, respectively. The R2 scores ranged from 0.50 (-SMA) to 0.74 (TGF-{beta}), indicating strong alignment between predicted and actual gene expression values. A multi-level approach, combining data from 40x, 100x, 200x, and 400x magnifications yielded higher R2 scores for almost all genes compared to single-magnification models. For the classification task, DenseNet121 achieved F1 score, precision, recall, and accuracy of 0.98, with a specificity of 0.99. ConclusionsRPEGENE-Net provides a simple, cost-effective method to predict gene expression from live-cell images, with potential applications in experimental studies, RPE transplantation quality control, and broader cell-based research. Multi-magnification imaging enhances model performance, supporting its utility as a scalable tool for diverse gene expression studies.

AimThis study aims to compare three deep learning-based retinal image registration methods RetinaRegNet, EyeLiner, and GeoFormer on the FIRE dataset to determine which approach provides optimal registration accuracy and computational efficiency across varying image overlap conditions (Classes S, A, and P) using mean landmark error as the primary outcome measure. MethodsThe three pipelines were evaluated under consistent conditions. RetinaRegNet incorporates diffusion features, dual keypoint sampling (SIFT and random), two stage outlier removal, and a multilevel registration hierarchy progressing from homography to polynomial transforms. EyeLiner integrates anatomical segmentation with SuperPoint feature extraction, LightGlue matching, and thin-plate spline warping. GeoFormer builds on LoFTR through cross-attention mechanisms and RANSAC-based refinement. Registration performance was quantified using mean landmark error (MLE). ResultsAcross all 134 FIRE image pairs, RetinaRegNet achieved the lowest overall MLE (3.12 pixels), outperforming EyeLiner (3.66 pixels) and GeoFormer (6.06 pixels). Class-specific analysis showed that RetinaRegNet delivered the highest accuracy in Class S images (1.70 pixels), competitive performance in Class A (5.24 pixels), and the strongest results in the most challenging Class P cases (4.57 pixels). GeoFormer demonstrated the shortest processing time at 0.32 seconds per image pair, compared with 4.92 seconds for EyeLiner and 31.23 seconds for RetinaRegNet. In Class P, RetinaRegNet achieved a 59.2% improvement in accuracy relative to GeoFormer (4.57 vs 11.20 pixels). DiscussionOverall, the evaluation reveals a clear trade-off between registration precision and computational speed. RetinaRegNet achieves the lowest MLE for complex clinical cases despite higher computational cost, EyeLiner balances precision and speed for routine use, while GeoFormer prioritizes rapid throughput where processing speed is critical.

Epiretinal membrane (ERM) is a vitreoretinal interface disease that, if not properly addressed, can lead to vision impairment and negatively affect quality of life. For ERM detection and treatment planning, Optical Coherence Tomography (OCT) has become the primary imaging modality, offering non-invasive, high-resolution cross-sectional imaging of the retina. Deep learning models have also led to good ERM detection performance on OCT images. Nevertheless, most deep learning models cannot be easily understood by clinicians, which limits their acceptance in clinical practice. Post-hoc explanation methods have been utilised to support the uptake of models, albeit, with partial success. In this study, we trained a sparse BagNet model, an inherently interpretable deep learning model, to detect ERM in OCT images. It performed on par with a comparable black-box model and generalised well to external data. In a multitask setting, it also accurately predicted other changes related to the ERM pathophysiology. Through a user study with ophthalmologists, we showed that the visual explanations readily provided by the sparse BagNet model for its decisions are well-aligned with clinical expertise. We propose potential directions for clinical implementation of the sparse BagNet model to guide clinical decisions in practice.

With the advancement in technology and computation capabilities, identifying retinal damage through state-of-the-art CNNs architectures has led to the speedy and precise diagnosis, thus inhibiting further disease development. In this study, we focus on the classification of retinal damage caused by diabetic macular edema (DME), Choroidal Neovascularization (CNV),DRUSEN, and NORMAL in optical coherence tomography (OCT) images. The emphasis of our experiment is to investigate the component of depth in the neural network architecture. We introduce a shallow convolution neural network - LightOCT, outperforming the other deep model configurations, with the lowest value of LVCEL and highest accuracy (+98% in each class). Next, we experimented to find the best fit optimizer for LightOCT. The results proved that the combination of LightOCT and Adam gave the most optimal results. Finally, we compare our approach with transfer learning models, and LightOCT outperforms the state-of-the-art models in terms of computational cost, least training time and gives comparable results in the criteria of accuracy. We would like to focus our future efforts on improving the accuracy metrics using shallow models, such that the trade-off between training time and accuracy is reduced..

Deep learning has shown promise in diabetic retinopathy screening using fundus images. However, many existing models operate as "black boxes," providing limited interpretability at the lesion level. This study aimed to develop an explainable deep learning model capable of detecting four diabetic retinopathy-related lesions--hemorrhages, hard exudates, cotton wool spots, and microaneurysms--and evaluate its performance using both conventional per-lesion metrics and a novel syntactic agreement framework. A total of 1,087 fundus images were obtained from publicly available datasets (EyePACS and APTOS), which contained 585 images graded as mild-to-moderate nonproliferative diabetic retinopathy (DR1 or DR2). All images were manually annotated for the presence of the four lesions. A U-Net-based segmentation model was trained to generate binary predictions for each lesion type. The performance of the model was evaluated using sensitivity, specificity, precision, and F1 score, along with five syntactic agreement criteria that evaluated the lesion-set consistency between the predicted and ground truth outputs at the image level. The model achieved high sensitivity and F1 scores for hemorrhages and hard exudates, showed moderate performance for cotton wool spots, and failed to detect any microaneurysms (0% sensitivity), with 92.9% of the microaneurysms cases misclassified as hemorrhages. Despite this limitation, the image-level agreement remained high, with any-lesion match and hemorrhage match rates exceeding 95%. These findings suggest that although individual lesion classification was imperfect, the model effectively recognized abnormal images, highlighting its potential as a screening tool. The proposed syntactic agreement framework offers a complementary evaluation strategy that aligns more closely with clinical interpretation and may help bridge the gap between artificial intelligence-based predictions and real-world ophthalmic decision-making.

Multiple Sclerosis (MS) is a chronic autoimmune disorder of the central nervous system that can lead to significant neurological disability. Retinal imaging--particularly Scanning Laser Ophthalmoscopy (SLO) and Optical Coherence Tomography (OCT)--provides valuable biomarkers for early MS diagnosis through non-invasive visualization of neurodegenerative changes. This study proposes a foundation-based bi-modal classification framework that integrates SLO images and OCT-derived retinal thickness maps for MS diagnosis. To facilitate this, we introduce two modality-specific foundation models--SLOFound and TMFound--fine-tuned from the RETFound-Fundus backbone using an independent dataset of 203 healthy eyes, acquired at Noor Ophthalmology Hospital with the Heidelberg Spectralis HRA+OCT system. This dataset, which contains only normal cases, was used exclusively for encoder adaptation and is entirely disjoint from the classification dataset. For the classification stage, we use a separate dataset comprising IR-SLO images from 32 MS patients and 70 healthy controls, collected at the Kashani Comprehensive MS Center in Isfahan, Iran. We first assess OCT-derived maps layer-wise and identify the Ganglion Cell-Inner Plexiform Layer (GCIPL) as the most informative for MS detection. All subsequent analyses utilize GCIPL thickness maps in conjunction with SLO images. Experimental evaluations on the MS classification dataset demonstrate that our foundation-based bi-modal model outperforms unimodal variants and a prior ResNet-based state-of-the-art model, achieving a classification accuracy of 97.37%, with perfect sensitivity (100%). These results highlight the effectiveness of leveraging pre-trained foundation models, even when fine-tuned on limited data, to build robust, efficient, and generalizable diagnostic tools for MS in medical imaging contexts where labeled datasets are often scarce.

PurposeAugmented reality (AR) may allow vitreoretinal surgeons to leverage microscope-integrated digital imaging systems to analyze and highlight key retinal anatomic features in real-time, possibly improving safety and precision during surgery. By employing convolutional neural networks (CNNs) for retina vessel segmentation, a retinal coordinate system can be created that allows pre-operative images of capillary non-perfusion or retinal breaks to be digitally aligned and overlayed upon the surgical field in real-time. Such technology may be useful in assuring thorough laser treatment of capillary non-perfusion or in using pre-operative optical coherence tomography (OCT) to guide macular surgery when microscope-integrated OCT (MIOCT) is not available. MethodsThis study is a retrospective analysis involving the development and testing of a novel image registration algorithm for vitreoretinal surgery. Fifteen anonymized cases of pars plana vitrectomy with epiretinal membrane peeling, along with corresponding preoperative fundus photographs and optical coherence tomography (OCT) images, were retrospectively collected from the Mayo Clinic database. We developed a TPU (Tensor-Processing Unit)-accelerated CNN for semantic segmentation of retinal vessels from fundus photographs and subsequent real-time image registration in surgical video streams. An iterative patch-wise cross-correlation (IPCC) algorithm was developed for image registration, with a focus on optimizing processing speeds and maintaining high spatial accuracy. The primary outcomes measured were processing speed in frames per second (FPS) and the spatial accuracy of image registration, quantified by the Dice coefficient between registered and manually aligned images. ResultsWhen deployed on an Edge TPU, the CNN model combined with our image registration algorithm processed video streams at a rate of 14 FPS, which is superior to processing rates achieved on other standard hardware configurations. The IPCC algorithm efficiently aligned pre-operative and intraoperative images, showing high accuracy in comparison to manual registration. ConclusionThis study demonstrates the feasibility of using TPU-accelerated CNNs for enhanced AR in vitreoretinal surgery.

While thrombosis is the most common indication for graft failure and immediate hysterectomy of a transplanted uterus, there is no optimal method to monitor graft perfusion. In this feasibility study, a near-infrared spectroscopy probe that monitors local tissue oxygenation (StO2) was attached to four uterine cervices and three donor cervices at the time of hysterectomy and transplantation respectively. Real-time StO2 measurements were gathered at key surgical steps of the hysterectomies and uterus transplantations. The average StO2 for the sequential steps of a hysterectomy of baseline, ovarian vessel ligation, contralateral ovarian vessel ligation, uterine vessel ligation, contralateral uterine vessel ligation, and colpotomy was 70.2%, 56.7%, 62.1%, 50.5%, 35.8%, and 8.5% respectively. The average StO2 for the sequential steps of uterus transplantation of iliac vein anastomosis, iliac artery anastomosis, contralateral iliac vein anastomosis, contralateral iliac artery anastomosis, and vaginal anastomosis was 8.9%, 27.9%, 56.9%, 65.9%, and 65.2% respectively. As uterine blood supply decreases in a hysterectomy, the measured StO2 also decreases, and vice versa for uterus transplantation. Tissue oximetry may be a reliable, non-invasive means of monitoring perfusion of a uterine graft. Additional studies are needed to determine if these devices complement current assessments of uterine graft viability and salvage thrombosed grafts.

The effects of specific footwear features on biomechanical parameters are often confounded by simultaneous changes in other shoe conditions, making it difficult to identify the isolated effect of material and design properties on relevant biomechanical outcomes. This study aimed to propose a tool, namely the Modular Footwear Setup (MFS), to assess the effects of midsole modifications on lower limb joint kinematics and in-shoe plantar pressure. The MFS uses a micro-hook-and-loop fastening system and a custom alignment device to enable fast, strong, and reliable midsole attachment/detachment to/from the upper. Accuracy and repeatability of the MFS in replicating the biomechanical outcomes of a control shoe featuring the same upper and midsole were tested in 10 healthy participants (5M,5F; age=33.2{+/-}9.2 yrs; BMI=21.5{+/-}2.8 kg/m2). Participants were asked to walk wearing both the MFS and the standard control shoe in three sessions. Kinematics of lower limb joints were measured via inertial measurement units, while capacitive pressure insoles were used to measure in-shoe plantar pressure. Intraclass correlation coefficient (ICC) was used to assess the repeatability of kinematic and pressure measurements between sessions. Statistical Parametric Mapping analysis did not identify significant differences in joint kinematics between conditions. While the MFS exhibited slightly lower peak pressure at the rearfoot, pressure parameters were not statistically different in the other foot regions. The MFS demonstrated good-to-excellent inter-session repeatability (ICC 0.84-0.97) for peak and mean pressure. Participants reported similar levels of comfort and stability in both shoes. The findings of the present study suggest the MFS has the potential to be a reliable and accurate tool for evaluating the effect of midsole features on relevant biomechanical parameters. This modular approach may improve data-driven footwear design by providing a consistent platform for testing the effects of midsole designs and materials across various applications, including therapeutic, safety, and athletic shoes.