(333) Uncovering follicular morphometric alterations in pediatric and adolescent patients exposed to low levels of alkylating chemotherapy using OvoPath.

Monica M. Laronda^1,2,3, Isaac Vieco-Marti^4,5, Hannah McDowell^1,2,3, Shravya Pant^1,2,3, Sofia Granados-Aparici^4,5.

¹ Stanley Manne Children’s Research Institute, Ann & Robert H. Lurie Children’s Hospital of Chicago, Chicago, Illinois

² Department of Pediatrics, Feinberg School of Medicine, Northwestern University, Chicago, Illinois

³ Division of Pediatric Surgery, Ann & Robert H. Lurie Children’s Hospital of Chicago, Chicago, Illinois

⁴ Centro de Investigación Biomédica en Red de Cáncer, Instituto de Salud Carlos III, 28029 Madrid, Spain.

⁵ Pathology Department, Medical School, University of Valencia- INCLIVA, Valencia

Abstract Text:

Ovarian follicle stages are classified by the number and morphology of granulosa cells. Due to the subjective nature of defining follicle morphology, manual assessment of normality leads to inter-observer differences, making standardized evaluations difficult for the reproductive community. Our team has standardized methods to classify follicles in human ovarian tissue using our open-source analysis software, OvoPath (https://github.com/Sograap/OvoPath). To aid in follicle segmentation, we have integrated a fine-tuned U-net. Ovarian biopsy images were divided into 512x512 pixel tiles, and a dataset of 2641 tiles was used (2112 training, 264 validation, 265 test sets). U-Net training for H&E sections was performed for 23 epochs, with micro metrics for Intersection Over Union of 0.9292 and 0.931 in the validation and test sets, respectively. These values indicate good follicle area segmentation and will improve user ease. Next, we aimed to assess the utility of these new OvoPath features. A high dose of alkylating chemotherapy increases one’s risk of premature ovarian insufficiency, as it diminishes the ovarian reserve. However, many fertility preservation procedures, including ovarian tissue cryopreservation, occur after an initial low chemotherapy dose and it is unclear how these low, acute levels may affect follicles and their surrounding microenvironment. Here, we employed our OvoPath’s morphometric analysis pipeline to assess granular details of follicle stages within prepubertal versus post-pubertal ovarian tissue before or after exposure to acute therapy. Ovarian biopsies from 32 individuals undergoing ovarian tissue cryopreservation (0.53 - 22.81yo) were used. Interestingly, significant differences in follicle morphology were identified between chemotherapy-exposed and naïve pre- and post-pubertal patients with OvoPath analyses. Ovaries from pre-pubertal patients that were exposed to acute chemotherapy treatments had significantly smaller primordial and transitional follicles compared to ovaries from alkylating chemotherapy naïve patients (1430µm², 1539µm² and 1323µm², 1460µm²). Additionally, both primordial and transitional follicles had significantly fewer granulosa cells per follicle in ovaries from exposed patients (8.95, 11.61 cells and 10.28,13.08 cells). Primordial and transitional follicles contained fewer granulosa cells (9.46, 10.36 cells and 11.19, 12.13 cells) in ovaries from post pubertal patients had previous chemotherapy treatment. Interestingly, an opposite trend was observed in follicle area where primordial follicles were larger (1512µm², 1416µm²) in ovaries from post-pubertal individuals with previous chemotherapy treatment compared to ovaries from post pubertal individuals with no previous treatment. These data indicate that OvoPath uncovered granular details of each follicle and revealed previously overlooked morphometric features that may indicate reduced follicle health and quality. To further define the biological significance of these observations, we assessed follicular DNA damage. Oocytes from exposed post-pubertal patients had increased DNA damage compared to oocytes from chemotherapy treatment naïve patients (0.026, 0.259 γH2AX⁺ oocytes/total oocytes). Future studies will identify if exposure to low levels of acute alkylating agents results in premature activation or death, thereby influencing the observed morphologic differences identified using OvoPath. Moreover, the improvements implemented for OvoPath regarding follicle segmentation may help streamline follicle analysis. Altogether, these data support the use of unbiased and standardized in-depth digital analyses using OvoPath for the assessment of follicle normality and the impact of diseases or exogenous conditions on folliculogenesis.

MML is a Gesualdo Family Foundation Research Scholar, and this work was supported by NIH U01HD110336 (HM, MML), NIH T32HD094699 (HM) and NIH F31HD116552 (HM), Center for Reproductive Science Marcia L. Storch Scholarship (SP), and the Office of Undergraduate Research (SP).