Can AI predict epithelial lesion categories via automated analysis of cervical biopsies: The TissueNet challenge?

The French Society of Pathology (SFP) organized its first data challenge in 2020 with the help of the Health Data Hub (HDH). The organization of this event first consisted of recruiting nearly 5000 cervical biopsy slides obtained from 20 pathology centers. After ensuring that patients did not refuse to include their slides in the project, the slides were anonymized, digitized, and annotated by expert pathologists, and finally uploaded to a data challenge platform for competitors from around the world. Competing teams had to develop algorithms that could distinguish 4 diagnostic classes in cervical epithelial lesions. Among the many submissions from competitors, the best algorithms achieved an overall score close to 95%. The final part of the competition lasted only 6 weeks, and the goal of SFP and HDH is now to allow for the collection to be published in open access for the scientific community. In this report, we have performed a "post-competition analysis" of the results. We first described the algorithmic pipelines of 3 top competitors. We then analyzed several difficult cases that even the top competitors could not predict correctly. A medical committee of several expert pathologists looked for possible explanations for these erroneous results by reviewing the images, and we present their findings here targeted for a large audience of pathologists and data scientists in the field of digital pathology.

A Partial Label-Based Machine Learning Approach For Cervical Whole-Slide Image Classification: The Winning TissueNet Solution.

Cervical cancer is the fourth most common cancer in women worldwide. To determine early treatment for patients, it is critical to accurately classify the cervical intraepithelial lesion status based on a microscopic biopsy. Lesion classification is a 4-class problem, with biopsies being designated as benign or increasingly malignant as class 1-3, with 3 being invasive cancer. Unfortunately, traditional biopsy analysis by a pathologist is time-consuming and subject to intra- and inter-observer variability. For this reason, it is of interest to develop automatic analysis pipelines to classify lesion status directly from a digitalized whole slide image (WSI). The recent TissueNet Challenge was organized to find the best automatic detection pipeline for this task, using a dataset of 1015 annotated WSI slides. In this work, we present our winning end-to-end solution for cervical slide classification composed of a two-step classification model: First, we classify individual slide patches using an ensemble CNN, followed by an SVM-based slide classification using statistical features of the aggregated patch-level predictions. Importantly, we present the key innovation of our approach, which is a novel partial label-based loss function that allows us to supplement the supervised WSI patch annotations with weakly supervised patches based on the WSI class. This led to us not requiring additional expert tissue annotation, while still reaching the winning score of 94.7%. Our approach is a step towards the clinical inclusion of automatic pipelines for cervical cancer treatment planning.Clinical relevance- The explanation of the winning Tis-sueNet AI algorithm for automated cervical cancer classification, which may provide insights for the next generation of computer assisted tools in digital pathology.

Plasma concentrations of lipoproteins and risk of lower-limb peripheral artery disease in people with type 2 diabetes: the SURDIAGENE study.

AIMS/HYPOTHESIS: The lipid profile has not been fully investigated in individuals with peripheral artery disease (PAD). We aimed to evaluate the relationship between plasma concentrations of lipoproteins and the prevalence of lower-limb PAD at baseline and its incidence during follow-up in people with type 2 diabetes. METHODS: Plasma concentrations of total cholesterol, HDL-cholesterol, triacylglycerol and apolipoprotein (Apo) A-I, ApoA-II, ApoB-100 and Apo(a) were measured at baseline using colorimetric or MS methods in the SURDIAGENE cohort. Total cholesterol/HDL-cholesterol ratio, non-HDL-cholesterol and LDL-cholesterol were estimated using computation formulas. Logistic and Cox proportional hazard regression models were fitted to estimate OR or HR, with related 95% CI, for baseline prevalence or incidence of major PAD (lower-limb amputation or requirement of revascularisation) during follow-up by increasing lipoprotein tertiles, after adjustment for key confounders. RESULTS: Among 1468 participants (women 42%, mean ± SD age 65 ± 11 years, duration of diabetes 14 ± 10 years at baseline), 129 (8.8%) had a baseline history of major PAD. Major PAD was less prevalent at baseline in the highest (vs lowest) tertile of HDL-cholesterol (OR 0.42 [95% CI 0.26, 0.71], p = 0.001) and ApoA-I (OR 0.39 [95% CI 0.23, 0.67], p = 0.0007), and more frequent in the highest tertile of total cholesterol/HDL-cholesterol ratio (OR 1.95 [95% CI 1.18, 3.24], p = 0.01). Among 1339 participants without a history of PAD at baseline, incident PAD occurred in 97 (7.2%) during a median (25th-75th percentile) duration of follow-up of 7.1 (4.4-10.7) years, corresponding to 9685 person-years and an incidence rate of 9.8 (95% CI 8.0, 12.0) per 1000 person-years. The risk of incident PAD was lower in the top (vs bottom) tertile of HDL-cholesterol (HR 0.54 [95% CI 0.30, 0.95], p = 0.03) or ApoA-I (HR 0.50 [95% CI 0.28, 0.86], p = 0.01) and higher in the top tertile of total cholesterol/HDL-cholesterol ratio (HR 2.81 [95% CI 1.61, 5.04], p = 0.0002) and non-HDL-cholesterol (HR 1.80 [95% CI 1.06, 3.12], p = 0.03). CONCLUSIONS/INTERPRETATION: We reported independent associations between HDL-cholesterol, ApoA-I, total cholesterol/HDL-cholesterol ratio or non-HDL-cholesterol and the prevalence or the incidence of major PAD in people with type 2 diabetes. Our findings provide a picture of lipoprotein profile in people with type 2 diabetes. Graphical abstract.