Real-time Analysis of Skin Biopsy Specimens With 2-Photon Fluorescence Microscopy

Abstract

ImportanceNonmelanoma skin cancers (NMSCs) are primarily diagnosed through paraffin section histologic analysis of skin biopsy specimens that requires days to weeks before a formal diagnosis is reported. Two-photon fluorescence microscopy (TPFM) has the potential for point-of-care diagnosis of NMSC and other dermatologic conditions, which could enable same-visit diagnosis and treatment.ObjectiveTo demonstrate that TPFM imaging of NMSC can occur within minutes of obtaining biopsies and provide similar histological features to those of conventional histology and evaluate TPFM diagnostic performance with respect to conventional histology.Design, Setting, and ParticipantsThis comparative effectiveness pilot study examined 29 freshly excised biopsies from confirmed NMSC lesions in patients presenting for treatment. Biopsies underwent imaging immediately with TPFM on site at Rochester Dermatologic Surgery (Victor, New York) between October 2019 and August 2021. The imaged biopsies were subsequently submitted for paraffin histology to produce coregistered images. Twelve of these coregistered image pairs (41.4%) were used as a training set. Fifteen (51.7%) were used in a masked evaluation by a board-certified dermatopathologist. Two (6.9%) were excluded from the study before evaluation because they could not be coregistered.Main Outcomes and MeasuresSensitivity, specificity, and accuracy of TPFM for NMSC biopsies were evaluated compared with conventional histology.ResultsFourteen of the 15 biopsy specimens (93.3%) in the evaluation set were identically diagnosed with TPFM and paraffin histology. The TPFM had 100% sensitivity (95% CI, 48%-100%), 100% specificity (95% CI, 69%-100%), and 100% accuracy (95% CI, 78%-100%) for basal cell carcinoma diagnosis. For squamous cell carcinoma diagnosis, TPFM had 89% sensitivity (95% CI, 52%-100), 100% specificity (95% CI, 54%-100%), and 93% accuracy (95% CI, 68%-100%). For overall NMSC diagnosis, TPFM had a 93% sensitivity (95% CI, 66%-100%), 100% specificity (95% CI, 3%-100%), and 93% accuracy (95% CI, 68%-100%). Examination of the 1 discordant pair revealed mismatched imaging planes as the source of error.Conclusions and RelevanceThe results of this comparative effectiveness pilot study suggest that TPFM captures histological characteristics of NMSC that are present in conventional histology, which reveals its potential as a rapid, point-of-care diagnostic alternative that does not need extensive sample preparation or retraining for image evaluation. Further validation of TPFM imaging performed for a larger cohort is needed to fully evaluate its diagnostic accuracy and potential effect within the field.