Factors affecting the precision of lesion sizing with contrast-enhanced spectral mammography

AIM: To evaluate the precision of the pre-surgical measurement of the size of breast cancer by contrast-enhanced spectral mammography (CESM). MATERIAL AND METHODS: This was a retrospective study of 204 breast cancers. Variables related to tumour biology and anthropometric variables were recorded and considered when evaluating the efficacy of CESM in predicting tumour size. Microscopic measurement of the largest diameter of the tumour at pathology was chosen as the reference standard. RESULTS: The mean size of tumours at pathology was 20.7 +/- 15.8 mm, while at CESM it was 23.6 +/- 16.7 mm (Bland-Altman 2.9 mm overestimation, 2.9 mm; 95% confidence interval [CI]:-10.3-16.2 mm). Spearman's correlation coefficient was 0.83 (p<0.0001). The concordance analysis indicated that 37.8% of the measurements were concordant, 47% were overestimated, and 15.2% were underestimated. Tumour size, nodal involvement, breast density, and breast size significantly modified the sizing accuracy. CONCLUSION: Quality of tumour size prediction with CESM is good, and this appears to be a promising imaging technique in the surgical planning of breast cancer. Biological tumour features, and anthropological characteristics of the patients do, however, affect the diagnostic precision and should be taken into account. (C) 2017 The Royal College of Radiologists. Published by Elsevier Ltd. All rights reserved.