Deep Learning-Based Spermatogenic Staging Assessment for Hematoxylin and Eosin-Stained Sections of Rat Testes

Abstract

In preclinical toxicology studies, a “stage-aware” histopathological evaluation of testes is recognized as the most sensitive method to detect effects on spermatogenesis. A stage-aware evaluation requires the pathologist to be able to identify the different stages of the spermatogenic cycle. Classically, this evaluation has been performed using periodic acid-Schiff (PAS)-stained sections to visualize the morphology of the developing spermatid acrosome, but due to the complexity of the rat spermatogenic cycle and the subtlety of the criteria used to distinguish between the 14 stages of the cycle, staging of tubules is not only time consuming but also requires specialized training and practice to become competent. Using different criteria, based largely on the shape and movement of the elongating spermatids within the tubule and pooling some of the stages, it is possible to stage tubules using routine hematoxylin and eosin (H&E)-stained sections, thereby negating the need for a special PAS stain. These criteria have been used to develop an automated method to identify the stages of the rat spermatogenic cycle in digital images of H&E-stained Wistar rat testes. The algorithm identifies the spermatogenic stage of each tubule, thereby allowing the pathologist to quickly evaluate the testis in a stage-aware manner and rapidly calculate the stage frequencies.

Deep learning-based image-analysis algorithm for classification and quantification of multiple histopathological lesions in rat liver

“…For each lesion, except for the hepatocellular hypertrophy, an accurate binary segmentation was achieved using a deep learning model based on a customized U-Net architecture 13 as shown in Fig. 2.…”

“…The diversity of histology due to the wide variety of animal species used in toxicity studies (such as rats, mice, dogs, monkeys, and mini pigs) and the large number of organs and tissues to be evaluated in a single study are major hurdles in training algorithms. Some AI-based image-analysis algorithms for laboratory animals using histopathological digital images have recently been reported, such as for detecting and quantifying testicular stage classification in rats 13 , of rodent cardiomyopathy 14 and hypertrophy and vacuolation of rat liver 15,16 . These reports showed that, for abnormal findings, the algorithms could detect and quantify only a single type of finding in each case, and none could immediately detect or classify multiple types of findings simultaneously on a WSI.…”

Deep Learning Approaches and Applications in Toxicologic Histopathology: Current Status and Future Perspectives

“…ML and DL examples in the literature have demonstrated that AI performs with good concordance to pathologist assessment, often with improved sensitivity and efficiency. [ 101 ] In addition, DL models have proven to be successful in quantifying morphological assessments of ROI or lesions from H & E slides in the heart,[ 90 ] testis[ 91 92 ] ovary,[ 81 89 ] eye/retina,[ 87 ] and liver. [ 85 86 ] In addition, DL-based morphological quantifications are a promising alternative for subjective assessments as well as for alleviating the burden of manual semi-quantitative analysis by pathologists.…”

Special Issue on Digital Pathology, Tissue Image Analysis, Artificial Intelligence, and Machine Learning: Approximation of the Effect of Novel Technologies on Toxicologic Pathology

“…12 Aside from these examples of the overarching use of AIbased morphometric assessments to entire studies, this special issue incorporates specific image analysis use-cases relevant for toxicologic pathology, many of which utilized AI-based tools. These include proprietary in-house built solutions, such as AI models built to count ovarian follicles, 13 or to quantify changes within retinal layer morphology, 14 and detection of endothelial tip cells in the oxygen-induced retinopathy model, 15 as well the utilization of commercially available application for spermatogenic staging, 16 analysis of rodent cardiomyocytes, 17 to support scoring of dextran sulfate sodium-induced colitis mouse model histology, 18 enumeration of cynomolgus bone marrow histology, 19 quantitative evaluation of hepatocellular cell hypertrophy in rats, 20 quantitate cell proliferation via common immunohistochemical biomarkers, 21 and for verification of changes observed in the Tg-rasH2 mouse used in carcinogenicity studies. 22 A fluorescence-based image analysis use-case (commercial software) is provided by Wilson et al 23 As novel applications at the periphery of the breadand-butter imaging work of a toxicologic pathologist are continuously emerging, Rousselle et al introduce a digital 3D topographic microscopy technique called scanning optical microscopy to evaluate re-endothelialization of vascular lumen after endovascular procedures.…”