Giovanna Mariscotti, Abstract Co-Author: Nothing to Disclose

Manuela Durando, Abstract Co-Author: Nothing to Disclose

Pier Paolo Campanino, Presenter: Nothing to Disclose

Maddalena Rigo, Abstract Co-Author: Nothing to Disclose

Elisa Regini, Abstract Co-Author: Nothing to Disclose

Mattia Robella, Abstract Co-Author: Nothing to Disclose

Laura Bergamasco, Abstract Co-Author: Nothing to Disclose

Paolo Fonio, Abstract Co-Author: Nothing to Disclose

Giovanni Gandini MD, Abstract Co-Author: Nothing to Disclose

To evaluate the accuracy of DM, DM combined with DBT, US and MRI in predicting residual tumour size and pathological response after NC for locally advanced breast cancer.

44 patients (mean age 49.2 years; range 31-71) with locally advanced breast cancer who underwent NC were enrolled in the study. We retrospectively evaluated size and response of tumours to NC by DM, DM combined with DBT, US and MRI before, during and at the end of treatment. We assumed as gold standard the tumour size measured at pathology. Patients were divided into responders (with pathologic complete (pCR) or partial response (pPR)) and non-responders (NR). Measurements were considered concordant if they were ±10 mm. Tumour size assessments were statistically analyzed with paired t-test, regression line and Pearson’s linear correlation coefficient and Bland-Altman Plots; categorical variables were arranged in contingency tables and analyzed with chi square test or Fisher’s test; 95% Confidence Intervals were estimated for all percentages.

For pCR patients (16/44), size estimates by all modalities showed an exponential decrease during treatment time (r=0.9; p≤0.005). The size agreement with pathology was 29 (95%CI 10-55)% for US, 36(14-62)% for DM, 33(12-62)% for DM+DBT, 54(27-79)% for MRI. For pPR patients (18/44), size estimates by imaging showed a linear decrease during treatment (r=0.9; p≤0,008). The size agreement was 69(41-89)% for US, 54(27-79)% for DM, 70(38-92)% for DM+DBT, 87(62-98)% for MRI. For NRs (10/44), US, DM and DM+DBT overestimated tumour size, while MRI measurements agreed with pathology. For the responders, the agreement between pCR predictions at mid-treatment and pathological responses was 7.1(0.4-90.5)% for both US and DM, 11.1(15.7-65.9)% for DM+DBT, 38.5(15.7-65.9)% for MRI; pPR prediction was 54.5(25.9-81)% for US, DM and DM+DBT, 84.6(57.8-97.3)% for MRI. NR prediction at mid-treatment was 80(47-99)% for US, 80(33-98)% for DM, 82(47-99)% for DM+DBT, 86(47-99)% for MRI.

Predictions of response and residual tumour size made on MRI showed a better agreement with pathology than DM, DM+DBT, US. DBT in addition to DM improved conventional imaging in pPR and NR predictions.

Breast MRI can be considered the most reliable imaging modality for pathological response evaluation after neoadjuvant chemotherapy, but the addition of DBT improves conventional imaging performances.

Mariscotti, G, Durando, M, Campanino, P, Rigo, M, Regini, E, Robella, M, Bergamasco, L, Fonio, P, Gandini, G, Evaluation with Digital Mammography (DM), DM Combined with Digital Breast Tomosynthesis (DBT), Ultrasound (US) and Dynamic Breast MRI of Pathological Response after Neoadjuvant Chemotherapy (NC) Treatment of Breast Carcinoma.  Radiological Society of North America 2013 Scientific Assembly and Annual Meeting, December 1 - December 6, 2013 ,Chicago IL. http://archive.rsna.org/2013/13024628.html