Tumor Size Measurement of Breast Malignancies Presenting as Architectural Distortion on DBT Screening Mammography: How to Measure Distortion on Mammogram and Which Imaging Modality is Most Accurate?

Purpose: Accurate pre-operative measurement of breast tumor size is vital for appropriate treatment planning. Measuring breast cancers presenting as architectural distortion (AD) on mammography is a challenge for the radiologist due to the significant desmoplastic reaction surrounding the tumor and lack of an US correlate often. The purpose of this study was to evaluate the accuracy of two different approaches for tumor size measurement on mammography for cancers presenting as AD on screening mammography, as well as the accuracy of ultrasound (US) and MRI compared to the pathologic tumor size.

Materials and Methods: After Institutional Review Board approval, a retrospective review of breast cancers which presented as AD on digital breast tomosynthesis (DBT) screening mammography between 01/01/2016-06/30/18 was performed. Mammographic measurements of the AD were performed centrally excluding the surrounding spiculations (central) and then including these (total) by three radiologists. The mammographic, US and MRI size of these malignancies were analyzed for correlation with pathologic tumor size and tumor histology using Spearman correlation.

Results: There were total 84 breast malignancies detected in 78 patients in the study cohort. Both approaches of mammographic measurements (central vs total) demonstrated statistically significant comparable correlation with pathologic tumor size (central: r=0.46 vs total r=0.54, p<0.001). However, the central measurement had a better accuracy with mean tumor size being within 5 mm of pathologic tumor size for 69% cases compared to 29.8% for the total measurement. Overall US was the most accurate (r=0.6, p< 0.001) and MRI had no statistically significant correlation(p=0.014). The mammographic and US measurements of invasive ductal cancers (IDC) had a significant association with pathologic size (p < 0.001) but none of the imaging modalities could accurately predict invasive lobular cancer (ILC) size. The central tumor measurement for mammography and US mostly undermeasured the tumor size and the total AD mammographic measurement and MRI over measured it.

Conclusion: Ultrasound demonstrates the best correlation between preoperative and pathologic tumor size for malignancies presenting as AD. In the absence of an US correlate the central mammographic measurement of these tumors excluding the adjacent extending spiculations is more accurate than the total measurement of the AD. Accurate pre-operative prediction of tumor size is unreliable for ILC compared to IDC.

Clinical Relevance Statement: No clear guidelines exist on measurement of tumors presenting as architectural distortion on mammography. The aim of this study was to evaluate the most accurate technique for mammographic measurement of AD to best predict the tumor size.