Abstract
BACKGROUND: Cancer diagnostic probe (CDP) had been developed to detect involved breast cavity side margins in real-time (Miripour et al. Bioeng Transl Med. e10236.). Here, we presented the results of the in vivo human model CDP studies on non-neoadjuvant cases. METHODS: This study is a prospective, blind comparison to a gold standard, and the medical group recruited patients. CDP and frozen data were achieved before the permanent pathology experiment. The main outcome of the study is surgical margin status. From November 2018 to April 2020, 202 patients were registered, and 188 were assigned for the study. Breast-conserving surgery at any age or gender, re-surgery due to re-currency, or involved margins are acceptable. Patients must be non-neoadjuvant. The reliability of CDP scoring had been evaluated by the pathology of the scored IMs. Then, three models of the study were designed to compare CDP with the frozen sections. Receiver operating characteristic (ROC) curves and AUC were measured based on the permanent postoperative pathology gold standard. RESULTS: A matched clinical diagnostic categorization between the pathological results of the tested IMs and response peaks of CDP on 113 cases, was reported (sensitivity = 97%, specificity = 89.3%, accuracy = 92%, positive predictive value (PPV) = 84.2%, and negative predictive value (NPV) = 98%). Study A showed the independent ability of CDP for IM scoring (sensitivity = 80%, specificity = 90%, accuracy = 90%, PPV = 22.2%, and NPV = 99.2%). Study B showed the complementary role of CDP to cover the missed lesions of frozen sections (sensitivity = 93.8%, specificity = 91%, accuracy = 91%, PPV = 55.6%, and NPV = 99.2%). Study C showed the ability of CDP in helping the pathologist to reduce his/her frozen miss judgment (specificity = 92%, accuracy = 93%, PPV = 42.1%, and NPV = 100%). Results were reported based on the post-surgical permanent pathology gold standard. CONCLUSION: CDP scoring ability in intra-operative margin detection was verified on non-neoadjuvant breast cancer patients. Non-invasive real-time diagnosis of IMs with pathological values may make CDP a distinct tool with handheld equipment to increase the prognosis of breast cancer patients.