Human interpretation of 3D histopathology image datasets of whole prostate biopsies reveals more cribriform pattern (Gleason pattern 4) carcinoma than is seen in standard 2D histology images

Abstract Background: Prostate biopsies of patients suspected of having cancer are taken to identify and grade the cancer, if present. These biopsies are currently diagnosed based on thin (up to 4 micron thick) sections of 1 mm thick biopsies. Consequently, only a minor portion of the tissue is examined (a set of 4 micron thick sections from 1 mm thick biopsies is less than 1% of the tissue). We developed a microscope and process that can nondestructively generate 3D image datasets of whole biopsies, allowing complete examination of the tissue specimen as a series of virtual sections. We hypothesize that clinically important grades of the cancer can be found by examining all levels of the biopsies. Methods: This study is based on 161 biopsies of low to intermediate grade cancer taken from radical prostatectomies that had been collected for an active surveillance project of the Canary Foundation. The formalin-fixed biopsy tissue was deparaffinized, stained with fluorescence analogues of hematoxylin and eosin (H&E), optically cleared, imaged in 3D, and false colored to resemble H&E stains. Two sets of images were created. One set was based on current practice – sections at 3 levels with 20 micron spacing between levels. The second set fully imaged the biopsy in 3D with at least 500 virtual levels per biopsy (at a spacing of ~1 micron per level). The images were acquired using a custom-developed 3D open-top light-sheet (OTLS) microscope and were diagnosed and graded by two pathologists using a web-based viewing platform. Results: Biopsies that had been fully imaged in 3D had more foci of intermediate risk grade cancer (Grade Group/GG 2 to 3) than biopsies that had been imaged based on current standard of practice. With 3D pathology, 13% of the biopsies had the cribriform variant of Gleason pattern 4 cancer when 2D pathology did not, while only 4.3% of the biopsies had cribriform carcinoma in 2D but not in 3D. However, when looking at overall GG, a similar number of biopsies were upgraded from low-grade (GG = 1) to higher-grade cancer (GG > 1) when comparing 3D to 2D images, and vice versa. Conclusion: Due to thorough sampling of tissue, viewing 3D datasets of whole prostate biopsies reveals more biopsies with cribriform pattern carcinoma than is seen in standard 2D pathology images. However, the grade of the PCa in the overall set of 161 biopsies is not changed. This is most likely due to poorly formed glands (Gleason pattern 4) being “downgraded” to well formed glands (Gleason pattern 3) when viewed as 3D vs 2D images. Based on these findings, more research is warranted to examine whether human interpretation of 3D pathology datasets can better predict patient prognosis and, thus, lead to treatment more appropriate for the grade of the cancer in a patient than is provided by standard 2D histopathology. Citation Format: Lawrence D. True, Peter A. Humphrey, Vanessa Roybal, Suet-Ling Sarah Chow, Weisi Xie, Jonathan T. C. Liu. Human interpretation of 3D histopathology image datasets of whole prostate biopsies reveals more cribriform pattern (Gleason pattern 4) carcinoma than is seen in standard 2D histology images [abstract]. In: Proceedings of the AACR Special Conference: Advances in Prostate Cancer Research; 2023 Mar 15-18; Denver, Colorado. Philadelphia (PA): AACR; Cancer Res 2023;83(11 Suppl):Abstract nr B050.