The contribution of gray scale ultrasound histogram analysis findings of solid breast masses to malign-benign differentiation

Purpose: In our study, our purpose is revealing the efficiency of FLH (fat-to-lesion histogram) rates in differentiating between malign-benign in the patients who apply because of mass in breast, and determine the relation between FLH rates and BI-RADS scoring, and complementary effect in decision-making process for biopsy. Material-Method: 76 female patients were included in the study, with the ages varying between 18 and 75 having at least one solid mass in breast consulted with the pre-diagnosis of mass in breast in Afyon Kocatepe University, Faculty of Medicine, Radiology Department between the dates January 2015 and October 2015. The study was conducted by retrospective examination of the gray scale images recorded with Hitachi Preirus (Tokio, Japan) Ultrasound Device and 13-MHz lineer probe, and the histopathological data of the same mass in the Hospital Information Managing System- HIMS archive. The patients were classified according to their histopathological diagnoses and US BI-RADS scores. The patients who are diagnosed with a malign histopathology are classified as group I whereas those who are diagnosed with a benign histopathology are classified as group II without considering the BI-RADS scores. In addition, the malign-diagnosed patients who are scored as 4 are classified as group III whereas the benign-diagnosed patients who are scored as 4 are classified as group IV without considering the sub-groups (4a, 4b, 4c) of BI-RADS 4 scores. Histogram analysis of each patient's breast mass was made. Their mass sonographical features and histopathologies were recorded. The clinical findings obtained were examined in terms of the complementary effect of FLH rates to the BI-RADS criteria and in terms of the compliance between the BI-RADS classification and FLH rates. Findings: In the measurements of the patients who were grouped according to the histopathological diagnosis; group I average FLH value was as 2.88±0.65 (between 2.04 and 4.10), group II average FLH value was as 1.55±0.40 (between 0.86 and 2.36). Group III average FLH value was as 2.94±0.68 (between 2.06 and 4.10), group IV average FLH value was as 1.51±,0.41 (between 0.86 and 2.35). In terms of the complementary effect of the FLH rates to the BI-RADS criteria, the FLH rates' differences between group I and group II were found to be statistically significant. The FLH rates' differences between group III and group IV were found to be statistically significant. The statistically significant cut-off value was found as 2,03 , the sensitivity at this level was found as 1 , the specifity was found as 0,864 when ROC analysis was made for malign and benign groups in all lesions without caring for the BI-RADS scores. When ROC analysis was made for only BI-RADS 4 lesions, the area under the curve was found as 0,982. The significant cut-off value was determined as 2,02. The sensitivity at this level was determined as 1 and the specifity was determined as 0,857. In terms of being able to determine the malign lesions; in BI-RADS 4 lesions, PPV(positive predictive value) was found as 16,3% without using the cut-off value whereas the same PPV was found as 57,1% when we use the cut-off value as 2,02 and higher than this point. In terms of being able to determine the benign lesions; in BI-RADS 4 lesions, PPV was found as 83,1% without using the cut-off value whereas the same PPV was found as 100% when we use the cut-off value as lower than 2,02. Result: In our study, it was observed that the determination of FLH rates in breast mass lesions can be statistically significant in terms of malign-benign discrimination both in all lesions without caring for the BI-RADS groups and in particularly BI-RADS 4 lesions. It was observed that it is significant to decide for biopsy and malign-benign discrimination of mass lesions through only BI-RADS criteria, however this kind of classification may be insufficient in particularly BI-RADS 4b and 4c lesions from time to time. In this case, we consider that using the FLH rates additively will take us closer to the true diagnosis in decision-making process.