Matthias Dietzel MD, Presenter: Nothing to Disclose

Pascal Andreas Thomas Baltzer MD, Abstract Co-Author: Nothing to Disclose

Ramy Zoubi, Abstract Co-Author: Nothing to Disclose

Heike Habrecht MD, Abstract Co-Author: Nothing to Disclose

Hartmut Peter Burmeister MD, Abstract Co-Author: Nothing to Disclose

Oumar Camara, Abstract Co-Author: Nothing to Disclose

Werner Alois Kaiser MD, PhD, Abstract Co-Author: Researcher, Siemens AG Researcher, Bayer AG Researcher, General Electric Company Researcher, Suros Surgical Systems, Inc Researcher, C. R. Bard, Inc Researcher, Boston Scientific Corporation Researcher, Galil Medical Ltd Researcher, Koninklijke Philips Electronics NV Researcher, Confirma, Inc Researcher, CAD Sciences LLC Researcher, Carl Zeiss Stiftung

Presence of axillary lymph node metastases (LNM) is amongst the most powerful prognostic factors in breast cancer. Classification of LNM as „present“or „absent“ is a common approach in clinical practice. Yet, recent publications recommend using the quantity of metastatic lymph nodes as a more accurate predictor for survival outcome. However, this requires surgico pathological axillary staging. We performed this study to investigate the potential of breast MRI to non-invasively predict axillary tumor load using dedicated semiautomatic software.

During a consecutive period of 4 years, all patients with invasive breast cancers receiving preoperative breast MRI at our institution were enrolled into this IRB approved study. Standard of reference was surgico pathological staging (breast and axilla). For MRI imaging standardized protocols according to international recommendations were performed at our institution (T1w gradient echo before and after [n=7] application of Gd-DTPA at flow rate of 0.3ml/s and dose of 0.1mmol/kg). First, all breast MRI examinations were processed using a commercially available CAD-software (Computer assisted diagnosis). It allows semiautomatic analysis of the primary breast tumor based on standardized enhancement parameters of the early and late phase (Wash-in, Wash-out etc.). Subsequently axillary lmyph node ratio was determined (ALNR= Number of metastatic nodes / number of resected nodes). Finally, multiple logistic regression was performed using enhancements parameters calculated by the CAD-software as “independent” and ALNR as “dependent variables”.  

107 consecutive patients with nodal positive primary invasive breast cancers were enrolled (mean age 57 years; range 30-87 years). Mean ALNR was 0,28 (range 0,03-,0,97). Multiple logisitic regression identified significant potential of breast MRI to predicts axillary tumor load (P<0,05; r-square: 0,18).

Semiautomatic CAD-analysis of breast MRI correlates with axillary tumor load. As quantity of metastatic lymph nodes is amongst the most accurate important prognostic factors, our findings add further knowledge to the application of breast MRI as a non invasive prognostic tool.

Computer assisted diagnosis of breast MRI examinations provides prognostic information and might be used as future biomarker.

Dietzel, M, Baltzer, P, Zoubi, R, Habrecht, H, Burmeister, H, Camara, O, Kaiser, W, Computer-assisted Diagnosis (CAD) of Breast MRI Examination Predicts Axillary Tumor Load.  Radiological Society of North America 2011 Scientific Assembly and Annual Meeting, November 26 - December 2, 2011 ,Chicago IL. http://archive.rsna.org/2011/11016873.html