Background: We investigated the performance of optical imaging evaluating deoxyhemoglobin concentration alteration (DeHCA) in breast tissues.
Methods: We enrolled all consecutive patients from January 2015 to October 2016 with clinically suspicious and/or BI-RADS grade 3-5 lesions at mammography or ultrasound (US). Patients underwent optical imaging (ComfortScan) to evaluate for DeHCA. The reference standard was pathology from a surgical specimen for malignant lesions, pathology from a surgical specimen or core needle biopsy for benign lesions, and negative follow-up for contralateral negative breasts. Non-parametric statistics, receiver operating characteristic, and linear discrimination analyses were performed.
Results: Of 334 enrolled patients, 168 (50%) were excluded for technical problems and 166 (50%) (median age 52 years) were analyzed totaling 331 breasts and 176 lesions. Of these, 75 were benign (median size 19 mm) and 101 malignant (median size 20 mm). The median DeHCA score in malignant lesions (0.95, interquartile range [IQR] 1.00-0.87) was higher (p < 0.001) than in benign lesions (0.80, IQR 0.95-0.70). Using the optimal cutoff (0.85), DeHCA score was less accurate than mammography, US, and their combination, with 78% sensitivity, 52% specificity, 40% positive predictive value (PPV), and 85% negative predictive value (NPV); using a 0.8 cutoff, sensitivity reached 93% and NPV 91%, but specificity fell to 32% and PPV to 37%. The accuracy of DeHCA score linearly combined with mammography or US was higher than that of DeHCA score alone (p < 0.001) and not significantly higher than that of mammography or US alone.
Conclusions: DeHCA score was significantly higher in malignant than in benign lesions, but its accuracy was significantly lower than that of mammography or US. Future refinements are needed.
Keywords: Breast cancer; Breast ultrasound; Deoxyhemoglobin concentration alteration; Mammography; Optical imaging.