A multicenter prospective study to predict pathologic complete response by vacuum-assisted breast biopsy based on MRI and US findings after neoadjuvant chemotherapy

Abstract Background: Surgical excision is still a standard treatment for patients with primary breast cancer because of inaccuracy of clinical complete response (cCR) by radiological findings regardless of high pathologic complete response rate (pCR) after neoadjuvant chemotherapy (NAC). How much vacuum-assisted breast biopsy (VAB) would contribute to improve the accuracy of prediction of pCR depending on the subtype of primary breast cancer has not been well assessed. Method: We conducted a multicenter prospective cohort study to assess the accuracy of prediction of pCR by VAB in combination with MRI and US findings for patients who achieved cCR after NAC in five Japanese hospitals. Patients with cT1-3N+M0 and primary breast cancer who achieved cCR (ycT0/is) in MRI and US findings after NAC were enrolled in this study. Primary outcome was an negative predictive value (NPV) and an false negative rate (FNR) for MRI and US findings, and for VAB in combination with these radiological findings in each subtype of ER-HER2+, ER-HER2-, and ER+HER2+ primary breast cancer. At least five US-guided 10G VAB samples were collected before surgery. Patients received anthracycline and taxane based regimen for NAC. Patients with HER2+ breast cancer received anti-HER2 agent in combination with taxane. Results: A total of 96 patients were enrolled in this study. Eighty-eight patients of them were included in the analysis: ER-HER2+ for 39 patients, ER-HER2- for 29 patients, and ER+HER2+ for 20 patients. A median age was 51 years (range, 26-72 years). Forty-four patients (50.0%) was premenopausal women. Numbers of patients for each tumor stage at diagnosis was cT1 for 25 patients (28.4%), cT2 for 59 patients (67.0%), and cT3 for 3 patients (3.4%). Fifty-six patients (63.6%) were node-negative at diagnosis (cN0) and 82 patients (93.2%) were node-negative after surgery (ycN0). Forty-eight patients (54.5%) had nuclear grade 3. Numbers of patients for each pathological tumor size after NAC was ypT0 for 47 patients (53.4%), ypTis for 19 (21.6%), ypT1 for 22 (25.0%), and ypT2 for one (1.1%). A median number of VAB sample was six (range 5-16). Residual tumor on VAB samples were invasive ductal carcinoma for 7 patients (8.0%), in situ lesion for 11 patients (12.5%). Fifty-nine patients (67.0%) underwent breast-conserving surgery. For radiological findings, NPV of ypT0/is was 84.6% for ER-HER2+, 65.5% for ER-HER2-, and 70.0% for ER+HER2+. VAB in addition to radiological findings yielded an NPV of 63.9% and an FNR of 81.2% for ER-HER2+, an NPV of 72.7% and an FNR of 46.2% for ER-HER2-, and an NPV of 66.7% and an FNR of 33.3% for ER+HER2+. A median size of residual tumor was 0.25cm (range 0.005-1.2cm) for ER-HER2+, 0.25cm (range 0.002-1.2cm) for ER-HER2-, and 0.55cm (range 0.1-0.6cm) for ER+HER2+. For patients with a clip maker, an NPV (90%) and an FNR (33.3%) were improved. Conclusions: Our study showed that VAB in addition to radiological findings was not accurate enough to predict pCR in any subtype of primary breast cancer. We could not support to omit surgery by VAB for patients who achieved cCR after NAC. Citation Format: Naoki Hayashi, Natsuki Teruya, Takashi Kuwayama, Yasuyuki Kojima, Sachiko Ohde, Sayuka Nakayama, Mizuho Tazo, Hiroyuki Takei, Takayuki Ueno, Takuji Iwase, Shinji Ohno, Hideko Yamauchi, Futoshi Akiyama, Hiroko Tsunoda, Koichiro Tsugawa, Seigo Nakamura. A multicenter prospective study to predict pathologic complete response by vacuum-assisted breast biopsy based on MRI and US findings after neoadjuvant chemotherapy [abstract]. In: Proceedings of the 2021 San Antonio Breast Cancer Symposium; 2021 Dec 7-10; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2022;82(4 Suppl):Abstract nr PD7-05.