Ovarian lesions are common and often incidentally detected. A critical shortage of expert ultrasound examiners has raised concerns of unnecessary interventions and delayed cancer diagnoses. Deep learning has shown promising results in the detection of ovarian cancer in ultrasound images; however, external validation is lacking. In this international multicenter retrospective study, we developed and validated transformer-based neural network models using a comprehensive dataset of 17,119 ultrasound images from 3,652 patients across 20 centers in eight countries. Using a leave-one-center-out cross-validation scheme, for each center in turn, we trained a model using data from the remaining centers. The models demonstrated robust performance across centers, ultrasound systems, histological diagnoses and patient age groups, significantly outperforming both expert and non-expert examiners on all evaluated metrics, namely F1 score, sensitivity, specificity, accuracy, Cohen’s kappa, Matthew’s correlation coefficient, diagnostic odds ratio and Youden’s J statistic. Furthermore, in a retrospective triage simulation, artificial intelligence (AI)-driven diagnostic support reduced referrals to experts by 63% while significantly surpassing the diagnostic performance of the current practice. These results show that transformer-based models exhibit strong generalization and above human expert-level diagnostic accuracy, with the potential to alleviate the shortage of expert ultrasound examiners and improve patient outcomes.

In renal histopathology, the routine clinical use of several histological stains presents challenges for the direct application of stain-specific deep learning-based analysis tools to whole-slide images. We present an approach to the in silico histological staining of kidney tissue where samples stained with hematoxylin and eosin (H&E) are virtually restained with periodic acid-Schiff (PAS). Our approach is underpinned by cycle-consistent generative adversarial neural networks trained on the National Unified Renal Translational Research Enterprise data set & horbar;the first UK-wide Biobank for chronic kidney disease & horbar;which features diverse data from 16 nephrology centers. Our work is divided into the following 4 main components: (1) we developed a virtual staining model, which infers PAS staining from H & E; (2) 2 board-certified pathologists assessed the virtual staining by attempting to distinguish it from real examples; (3) we trained a glomerular segmentation model using 3 independent renal segmentation data sets (Kidney Precision Medicine Project, Human BioMolecular Atlas Program [Kidney], and data by Jayapandian et al); and (4) we demonstrated the utility of virtual staining by inferring PAS staining from previously unseen H&E test images and applying our PAS-specific glomerular segmentation model. Each pathologist was able to identify 52.5% and 75.8% of the virtually stained images, respectively, showing an overlap in the variability of the authentic and synthetic staining. We discussed the utility of virtual staining in digital pathology, the need for pathology-specific testing with respect to chronic damage, and minimal changes and steps for incorporating more stains. Furthermore, alongside this article, we included complete glomerular annotations for 20 Kidney Precision Medicine Project H&E-stained slides.

The deep learning field is converging towards the use of general foundation models that can be easily adapted for diverse tasks. While this paradigm shift has become common practice within the field of natural language processing, progress has been slower in computer vision. In this paper we attempt to address this issue by investigating the transferability of various state-of-the-art foundation models to medical image classification tasks. Specifically, we evaluate the performance of five foundation models, namely Sam, Seem, Dinov2, BLIP, and OpenCLIP across four well-established medical imaging datasets. We explore different training settings to fully harness the potential of these models. Our study shows mixed results. Dinov2 consistently outperforms the standard practice of ImageNet pretraining. However, other foundation models failed to consistently beat this established baseline indicating limitations in their transferability to medical image classification tasks.

Ovarian lesions are common and often incidentally detected. A critical shortage of expert ultrasound examiners has raised concerns of unnecessary interventions and delayed cancer diagnoses. Deep learning has shown promising results in the detection of ovarian cancer in ultrasound images; however, external validation is lacking. In this international multicenter retrospective study, we developed and validated transformer-based neural network models using a comprehensive dataset of 17,119 ultrasound images from 3,652 patients across 20 centers in eight countries. Using a leave-one-center-out cross-validation scheme, for each center in turn, we trained a model using data from the remaining centers. The models demonstrated robust performance across centers, ultrasound systems, histological diagnoses and patient age groups, significantly outperforming both expert and non-expert examiners on all evaluated metrics, namely F1 score, sensitivity, specificity, accuracy, Cohen’s kappa, Matthew’s correlation coefficient, diagnostic odds ratio and Youden’s J statistic. Furthermore, in a retrospective triage simulation, artificial intelligence (AI)-driven diagnostic support reduced referrals to experts by 63% while significantly surpassing the diagnostic performance of the current practice. These results show that transformer-based models exhibit strong generalization and above human expert-level diagnostic accuracy, with the potential to alleviate the shortage of expert ultrasound examiners and improve patient outcomes.