Ultrasound-guided diffuse optical tomography: an adjunct to ultrasound that can reduce unnecessary breast biopsies

Background: Breast biopsy is recommended for any lesion in which the radiologist deems the likelihood of malignancy to be greater than 2% and more than 75% of biopsies return with benign results. This study assesses whether the total hemoglobin concentration (HbT) and blood oxygen saturation (%sO2) of malignant and benign breast lesions, obtained using Ultrasound-guided diffuse optical tomography (US-guided DOT), can help reduce benign breast biopsy rates. Methods: 238 women consecutively recruited were examined with US-guided DOT before their US-guided biopsy. Three or four radiologists independently provided BI-RADS assessments using pre-procedure US and mammography (when available), first without and then with adjunctive US-guided DOT data. Radiologists and study personnel were blinded to biopsy outcomes. HbT and %sO2 were computed and correlated with histopathology. One-way ANOVA test was used on multiple groups followed by a post-hoc Tukey test. Student t-test was performed on any two groups. Results: 226 women (mean 52 years, 19–85 years) were analyzed, with histopathology showing: 70 invasive carcinomas (grade 3, n = 25, and grade 1&2, n = 45), 7 DCIS, 9 high-risk lesions, and 140 benign lesions. The integration of DOT-derived HbT and %sO2 reduced benign biopsies by 24.54% with a false-negative rate (FNR) of 1.78%. The HbT of the malignancy group (invasive carcinoma and DCIS) was significantly higher than that of all benign lesions (P < 0.001) and the %sO2 was significantly lower than in benign lesions (P = 0.033). The HbT of grade 3 carcinomas was significantly higher than that of grade 1&2 carcinomas (P = 0.042), while the %sO2 was significantly lower (P = 0.015). The HbT of carcinomas smaller than 2 cm (measured by ultrasound) was not significantly different from that of carcinomas larger than 2 cm (P = 0.781). However, %sO2 differed significantly between these two size groups (P = 0.023). Conclusions: Integration of HbT and %sO2 into the radiologists’ BI-RADS imaging assessments of ultrasound-visible breast lesions can reduce benign biopsy rates by 25%, while maintaining a FNR below 2%. The correlation of HbT and %sO2 with invasive carcinoma grade, and the size-independent HbT contrast and size-dependent %sO2 level of invasive carcinomas may guide preferential diagnosis of malignancies.