Abstract P1-15-11: Predicting pathological complete response (pCR) to neoadjuvant chemotherapy (NACT) based on pre- and post-NACT digital mammography and digital breast tomosynthesis findings

Abstract

Abstract Background: In invasive breast cancer patients being treated with neoadjuvant chemotherapy (NACT), achieving pathological complete response (pCR) is a useful goal of treatment. Monitoring response to NACT and predicting pCR is helpful in planning further therapy and providing robust prognostic information. Digital mammography (DM) and additional digital breast tomosynthesis (DBT) features are important tell-tales of tumor characteristics and behaviour. Following NACT, the mammographic features- both DM and DBT- of responding tumors can vary considerably. In this prospective study, we correlated the DM and DBT features of pre-NACT and post-NACT mammograms to investigate if these can reliably predict pCR to NACT. Methods: Following approval by institutional ethics committee, starting January 2016, 200 consecutive invasive breast carcinoma patients (mean age 51.2 years, all palpable breast masses) undergoing diagnostic breast imaging had their DM and DBT reviewed by two radiologists independently, who were blinded of the cyto/histology and the original DM and DBT reporting. Of these, 47 patients who were treated with NACT and had pre- and post-NACT DM and DBT were recruited. After a core-biopsy, radio-opaque marker(s) were placed in tumor core/margin. The pre- and post-NACT DM and DBT findings were compared and correlated with the extent of response of the primary breast tumor to NACT. DM and DBT characteristics predictive of (in-breast) pCR of index breast lesion were identified. Results: Of the 47 patients who underwent NACT, 44 received both anthracycline and taxane, and 3 received only an anthracycline based combination chemotherapy. Twelve patients underwent breast conservative surgery and the remaining underwent mastectomy. pCR was seen in 17 (36.2%) patients based on the surgical specimen histology. On clinical examination, 19 (40.4%) patients had clinical complete response (cCR) of the breast tumor, 11 (64.7%) of whom had pCR as well. Five patients had radiological complete response (rCR, no breast lesion visualised on post-NACT imaging)- 2 patients on DM alone, 2 patients on DBT alone, and one patient on both DM and DBT. Radio-opaque clips had some obscuring effects in 3 of these 5 patients, especially on DBT, in form of reduced visibility of breast lesion on DBT, c.w. corresponding DM images. All 5 patients with rCR had pCR (sensitivity=29.4%, specificity=100%), in contrast to only 11 (57.9%) patients with cCR having pCR. Patients with pCR had benign appearing (forced Bi-RADS 2 and 3) lesions on mammography more commonly on DM (p<0.001) than on DBT (p=0.042) (41.2% vs 23.5%). Post NACT lesion morphology varied significantly between patients with and without pCR on DM (p=0.038) but not on DBT (p=0.182). Pre-NACT forced Bi-RADS score, lesion morphology or margin characteristics on DM and DBT did not vary significantly amongst patients with and without pCR. Conclusions: Post-NACT DM and DBT features can predict pCR with high specificity but with low sensitivity. Pre-NACT DM and DBT features did not reliably predict response to NACT, and pCR in this study. DM may be better than DBT for assessing response to NACT in the presence of radio-opaque markers/clips. Citation Format: Agarwal G, Sonthineni C, Mohindra N, Jain N, Neyaz Z, Agrawal V, Krishnani N, Maylivahnan S, Mishra A, Lal P. Predicting pathological complete response (pCR) to neoadjuvant chemotherapy (NACT) based on pre- and post-NACT digital mammography and digital breast tomosynthesis findings [abstract]. In: Proceedings of the 2018 San Antonio Breast Cancer Symposium; 2018 Dec 4-8; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2019;79(4 Suppl):Abstract nr P1-15-11.