False-Negative Histopathologic Diagnosis of Prostatic Adenocarcinoma

Abstract

Context.— Histopathologic diagnosis of adenocarcinoma of the prostate is based on light-microscopic examination of hematoxylin-eosin–stained tissue sections. Multiple factors, including preanalytic and analytic elements, affect the ability of the pathologist to accurately diagnose prostatic adenocarcinoma. False-negative diagnosis, that is, failure to diagnose prostatic adenocarcinoma, may have serious clinical consequences. It is important to delineate and understand those factors that may affect and cause histopathologic false-negative diagnoses of prostatic adenocarcinoma. Objectives.— To review common factors involved in histopathologic underdiagnosis of prostatic adenocarcinoma, including the following: (1) tissue processing and sectioning artifacts, (2) minimal adenocarcinoma, (3) deceptively benign appearing variants of acinar adenocarcinoma, (4) single cell adenocarcinoma, and (5) treatment effects. Data Sources.— Data sources included published, peer-reviewed literature and personal experiences of the senior author. Conclusions.— Knowledge of the reasons for histopathologic false-negative diagnosis of adenocarcinoma of the prostate is an important component in the diagnostic assessment of prostate tissue sections. Diagnostic awareness of the histomorphologic presentations of small (minimal) adenocarcinoma; deceptively benign appearing variants including atrophic, foamy gland, microcystic, and pseudohyperplastic variants; single cell carcinoma; and treatment effects is critical for establishment of a definitive diagnosis of adenocarcinoma and the prevention of false-negative diagnoses of prostate cancer.

A method for utilizing automated machine learning for histopathological classification of testis based on Johnsen scores

“…We consider that an AI-based algorithm would also be suitable for quantification such as that in Gleason grading for prostate cancer. Gleason grading is an established scoring system for prostate cancer that is widely used by pathologists worldwide 20 . In fact, the development of a medical grade AI-based algorithm for evaluating prostate core needle biopsies using Gleason scores has already been reported 21 .…”

Independent real‐world application of a clinical‐grade automated prostate cancer detection system

“…If the consensus of the central pathologists and Paige Prostate agreed, then this classification was assigned as the ground truth for the part‐specimen; otherwise, additional histologic sections of the corresponding part‐specimen were cut and subjected to IHC analysis and reinterpretation by the pathologists to assign the final ground truth (Figure 1). IHC analysis is regularly performed for prostate biopsies when a definitive diagnosis cannot be made from H&E alone [3] and was also performed in 113 randomly selected slides related to 20 patients where the diagnoses were concordant between the local pathologist, central pathologists, and Paige Prostate to ensure the accuracy of the ground truth diagnoses (supplementary material, Supplementary materials and methods). IHC was performed utilizing ready‐to‐use antibodies against high‐molecular‐weight cytokeratin (ready‐to‐use, clone 34βE12; DAKO, Glostrup, Denmark), p63 (ready‐to‐use, clone 4A4; DAKO), and P504S (ready‐to‐use, clone 13H4; DAKO) on an Agilent Autostainer Link 48 system (DAKO) following DAKO PT link (DAKO) antigen retrieval using Tris–EDTA buffer (pH 9.0).…”

“…Diagnostic surgical pathology remains the 'gold standard' for cancer diagnosis, despite the challenges posed by the shortage of diagnostic pathologists [1], the difficulty in detecting small quantities of cancer in biopsy material, and non-trivial levels of inter-observer variability [2,3]. These problems may be mitigated by applying artificial intelligence (AI)-based systems on digital whole-slide images (WSIs) of histopathology sections to provide greater diagnostic accuracy [4][5][6][7][8][9], reduce inter-observer variability, and alleviate the pressure on pathologists who face increasing caseloads and a decreasing qualified workforce [4][5][6][7][8][9].…”

“…One of the potential applications of AI is to increase the accuracy and the efficiency of histopathologic assessment of transrectal ultrasound-guided (TRUS) prostate biopsies, where the false-negative diagnosis rate for prostate needle biopsies is 1-4% [3,10,11]. To address this unmet need, AI-based algorithms to detect prostate cancer in diagnostic biopsies have been developed [7,[12][13][14].…”

New Criteria for Histopathological Classification of Testis Based on Johnsen Score for Male Infertility Using Automated Deep Learning Software

“…We consider that an AI-based algorithm could also be suitable for quanti cation such as that in Gleason grading for prostate cancer. Gleason scoring is an established scoring system for prostate cancer that is widely used by pathologists worldwide 16 . In this regard, the development of a medical grade AI-based algorithm for evaluating prostate core needle biopsies using Gleason scores has been reported 17 .…”