Introduction: Accurate preoperative assessment of breast cancer size is important for choosing appropriate surgical treatment. Mammography and ultrasonography are the most widely used breast imaging techniques. The aim of this study was to compare the tumour size measured by these two modalities with the pathological size of native specimen.
Methods and subjects: From 2001 to 2007, a retrospective review was conducted of 299 patients operated on at Masaryk Memorial Cancer Institute for the diagnosis of invasive breast carcinoma detected on ultrasonography or mammography as a nucleus shadow lesion where the preoperative size was estimated. Pearson's correlation to pathological size was tested and the mean deviation was analysed in the whole group of patients as well as in subgroups defined by pathological size (pT), histogical type and grading.
Results: Ultrasonography was accurate in determination of the tumour size (i.e. within the deviation of 5mm) in 195 patients (74%), it underestimated in 45 cases (17%) and overestimated in 24 cases (9%). Pearson's correlation coefficient (r) was 0.610 and mean deviation minus 0.115 cm. Mammography estimated accurate results in 162 patients (81%), the size was underestimated in 14 cases (7%) and overestimated in 25 (12%). Pearson's correlation coefficient (r) was 0,645 and mean deviation 0.08 cm. Generally lower accuracy in assessing the size was noted in lobular carcinomas.
Conclusions: Despite some limitations of our study, we can resume that in most cases (approximately 75%) the size assessment of invasive tumour lesion by both ultrasound and mammography is relatively reliable within the deviation of 5mm from the pathological size. But there's still been some portion of cases remaining where the estimation is not accurate, therefore we are not completely able to avoid redundant removal of tissue or, on contrary, demand of consecutive operations due to positive resection margins.