Abstract P2-02-11: Accuracy of MRI and ultrasonography in predicting pathologic complete response in breast cancer patients treated with neoadjuvant chemotherapy

Abstract

Background: Surgical treatment could be omitted if we could predict pathologic complete response (pCR) after neoadjuvant chemotherapy (NAC) in primary breast cancer patients by radiological findings accurately. The aim of this study is to assess the accuracy of MRI and ultrasonography (US) in predicting pathologic complete response in breast cancer patients treated with NAC. Methods: Five hundred sixty nine primary breast cancer patients who underwent breast conserving surgery after NAC between 2004 and 2008 were reviewed. Sensitivity and specificity of clinical complete response (cCR) by MRI alone and in combination with US were retrospectively assessed. pCR in primary tumor was defined as no residual invasive carcinoma. Spotted, linear or trabecular lesions without vascularity detected by US and/or enhanced on late phase of MRI findings were suspected to be residual in situ carcinoma. Results: The median age of the 569 patients analyzed was 50 years (range, 26-76 years). Of the patients, 86 (15.1%) had pCR. Of 79 patients who were diagnosed cCR by MRI alone, 72 patients (91.1%) were accurately predicted. Of 490 patients who were diagnosed to have residual invasive carcinoma, 14 patients (2.9%) had pCR. Of 75 patients who were diagnosed cCR by MRI in combination with US, 68 patients (90.7%) was accurately predicted. Of the 494 patients who were diagnosed to have residual invasive carcinoma, 18 patients (3.6%) had pCR. The combination of MRI and US had a sensitivity of 79.1%, specificity of 98.6%. prediction of pCR by MRI alone and in combination with US MRI aloneMRI + US cCRnon-cCRcCRnon-cCR (n = 79)(n = 490)(n = 75)(n = 494)pCR72146818(n = 86)(91.1%)(2.9%)(90.6%)(3.6%)non-pCR74767476(n = 483)(8.9%)(97.1%)(9.4%)(96.4%) For 85 patients who were diagnosed cCR by MRI and/or US, positive predictive value was 89.4% in ER-positive breast cancer, 100% in ER-negative/HER2-positive breast cancer, and 81.5% in triple negative breast cancer. In terms of prediction of residual in situ carcinoma, of the 86 patients with pCR, 43 (50%) had residual in situ carcinoma. The median size of residual tumor was 2.0 cm (range, 0.01-8.0 cm). Twenty five of 37 patients (67.6%) who are diagnosed cCR without any residual in situ carcinoma by both MRI and US were accurately predicted. Twelve of the 37 patients (31.6%) had residual carcinoma (in situ carcinoma, n = 10, and invasive carcinoma, n = 2) on surgical specimen. Twenty one of 38 patients (55.3%) who are diagnosed cCR with residual in situ carcinoma by both MRI and US had pCR with residual in situ carcinoma accurately. Twelve of the 38 patients (31.6%) had pCR without any in situ carcinoma and 5 (13.2%) had residual invasive carcinoma on surgical specimen. Conclusions: Our results indicated that MRI in combination with US was highly useful to predict pCR in breast cancer patients treated with NAC, although it was hard to predict a presence of residual in situ carcinoma. Pathological intervention using a vacuum-assisted breast biopsy in addition to these radiological modalities is expected for higher accuracy of predicting pCR. We are going to conduct a prospective multi-institutional study to assess predictive value of the method. Citation Information: Cancer Res 2013;73(24 Suppl): Abstract nr P2-02-11.

Role of Magnetic Resonance Imaging in Detection of Pathologic Complete Remission in Breast Cancer Patients Treated With Neoadjuvant Chemotherapy: A Meta-analysis

“…12,21,[28][29][30][31][33][34][35]40,41,43,[46][47][48]59,63,65,66,68,70,73,78 In 3 studies, 32,42,76 the definition of pCR was not reported. In 29 studies, 10,11,19,20,[36][37][38][39]44,45,[50][51][52][53][54][56][57][58][60][61][62]64,67,69,71,72,74,…”