Diagnostic Discordance Among Histopathological Reviewers for Difficult-to-Diagnose Melanocytic Lesions

Abstract

Purpose: Histopathological examination is adequate for most melanocytic neoplasms. However, there is a subset of lesions with uncertain malignant potential that are difficult-to-diagnose where utilization of ancillary tests may be considered to aid in providing a more definitive diagnosis.

Study design: Under an IRB-approved protocol, 3,317 melanocytic lesions underwent independent diagnostic review by up to 9 unique board-certified dermatopathologists. Each lesion received ≥ 3 diagnoses via electronic whole slide images with zoom-in capability as benign, malignant, or unknown malignant potential (UMP). Lesions were subsequently binned according to the results of dermatopathology review in the following manner: concordant (all diagnoses of the same designation); conflicting (both benign and malignant designations); majority (a single designation with the highest number of diagnoses); and equivocal (equal number of different designations). Here we describe the identification of lesions suitable for ancillary testing.

Results: 30% of lesions were concordant benign and 46.3% were concordant malignant, indicating 76.3% of all lesions were not difficult-to-diagnose. 23.7% of lesions had some level of diagnostic difficulty and/or ambiguity defined by disagreement between reviewers and would be suitable for ancillary testing. Within these ~24%, 7.3% of lesions were majority benign, 4.8% of lesions were majority malignant, 2.7% were majority UMP, 6.9% of lesions were conflicting, and 1.6% of lesions were equivocal.

Conclusions: We demonstrate that in a large cohort of suspicious pigmented lesions, ~24% were difficult-to-diagnose, supporting the need for additional objective diagnostic tools. Patient care would benefit from ancillary testing to aid the diagnosing pathologist in identification of the most appropriate diagnosis to optimize patient management.

Potential conflicts of interest: This study was supported by Castle Biosciences, Inc. GAH and JAP have served as consultants for Castle Biosciences, Inc. MSG is an employee shareholder of Castle Biosciences, Inc. SIE is a consultant and shareholder for Castle Biosciences, Inc. Editorial assistance was provided by Brooke H. Russell, PhD, and Jason H Rogers, MSc, both employee shareholders of Castle Biosciences, Inc.