Diagnostic Accuracy of Stereoscopic Optical Palpation for Tumour Margin Assessment in Breast-Conserving Surgery

Abstract

Abstract Purpose Optical elastography has been developed for intraoperative tumour margin assessment during breastconserving surgery (BCS) based on the elevated stiffness of tumour. It aims to assist the surgeon in fully removing a cancer in one operation, thus reducing re-excision surgeries and potentially decreasing the recurrence rate. In this study, we investigate the use of a recently proposed costeffective, camerabased optical elastography technique, stereoscopic optical palpation (SOP), for breast cancer detection and present the diagnostic accuracy of the technique for tumour margin assessment. Methods SOP was performed on the margins of intact, freshly excised specimens from 48 patients undergoing BCS. Pairs of photographs were acquired on 169 margins and, within two minutes of capturing each image, detailed maps were generated showing areas of mechanical pressure at the tissue surface (stress maps) over a 29 × 29 mm2 region. To determine the diagnostic accuracy of SOP in tumour margin assessment, 247 regions of 10 × 10 mm2 were selected, analysed and coregistered with standard histopathology to provide a ground truth. All regions were randomly divided into 10 groups and an automatic classifier was developed and evaluated using 10fold crossvalidation to calculate the accuracy of cancer detection using SOP. Results In the regions of interest included in the analysis, the prevalence of cancer within 1 mm of the margin, based on histopathology, was 11.3% (n = 28). Based on the stress maps acquired using SOP and the automatic classifier, the sensitivity of cancer detection within 1 mm of the tissue boundary (positive margin) is 82.1% (n = 23) and the specificity of identifying benign tissue within 1 mm of the tissue boundary (negative margin) is 83.6% (n = 183). The mean stress threshold determined to identify positive margins is 10.1 kPa. Conclusion This study demonstrates that SOP can achieve an accuracy of 83.4% in identifying cancer within 1 mm of the tissue boundary, which surpasses conventional intraoperative imaging techniques for margin assessment, including radiography and ultrasound. Due to its straightforward implementation and low cost, SOP holds promise as an effective point-of-care tool to assist in tumour margin assessment for BCS.