Assessment of breast cancer tumor size depends on method, histopathology and tumor size itself*.

Purpose: Mammography (MG), breast (BU) and axillary ultrasound (AU), and clinical examination (CE) are commonly used for clinical staging. These different methods were compared in order to assess the accuracy of clinical tumor staging (cT).
Method: About 503 breast cancer (BC) patients were prospectively measured by MG, ultrasound and clinical examination. Pearson's correlation to pathological tumor size (pT) was tested and the deviation of MG, BU and CE to pT was analyzed in subgroups defined by pT, grading (G), estrogen receptor (ER), progesteron receptor (PR), proliferation (MIB-1) and HER2/neu. Association of AU to pN was examined by chi(2)-test. Receiver operating characteristics (ROC) were used to test the prediction of a pT > 2 cm.
Results: Mammography correlated best with pT (r = 0.752). Mammography (mean (MG) = 2.17 cm) overestimated tumors in size (mean (pT) = 2.04 cm) rather than ultrasound (mean (BU) = 1.86 cm) and clinical examination (mean (cT) = 1.70 cm). pT of invasive ductal BC could be estimated significantly better than pT of invasive lobular BC. Smaller tumors were better to assess than larger ones. Tumors with a grading G1 were easier to estimate than tumors with G2/3. Best predictor of a pT > 2 cm was the mammography with an area under the curve of 0.876. The combination of all three modalities by linear regression performed even better with an AUC of 0.906.
Conclusions: The dimension of invasive ductal carcinomas, small and low grading tumors is significantly better to estimate. Concerning treatment decisions, we propose a combination of all three modalities, as the best predictive value was seen for the complementary use of mammography, ultrasound and clinical examination.