1321. Factors Associated With Nodal Positivity Following Neoadjuvant Therapy in Patients With Breast Cancer Who Are Initially Node-Negative on MRI
Authors * Denotes Presenting Author
  1. Berat Bersu Ozcan *; University of Texas Southwestern
  2. Firouzeh Arjmandi; University of Texas Southwestern
  3. Mike Bosh; University of Texas Southwestern
  4. Yin Xi; University of Texas Southwestern
  5. Sunati Sahoo; University of Texas Southwestern
  6. Heather McArthur; University of Texas Southwestern
  7. Basak Dogan; University of Texas Southwestern
We aimed to determine the rate of positive nodes on surgical pathology (ypN+) following neoadjuvant chemotherapy (NAC) in patients without evidence of nodal metastasis (cN0) on MRI and identify clinical and imaging factors associated with ypN+ status.

Materials and Methods:
We reviewed all newly diagnosed breast cancer patients who underwent preoperative breast MRI for locoregional evaluation at our institution between 2.1.2013-2.1.2018. Patients who did not receive neoadjuvant therapy or node positive (cN+) and recurrent breast cancers were excluded from our analysis. We collected patient age, race/ethnicity, immunohistochemistry features, primary tumor size and pretherapy axillary imaging results along with clinical TNM staging. Imaging response to neoadjuvant therapy in primary tumor, focality of the primary tumor, sentinel lymph node biopsy (SLNB) and axillary lymph node dissection (ALND) results along with final pathology (ypT, ypN) were recorded. Association between T stage, hormone status, multifocality and ypN0 versus ypN+ status was assessed with Chi-square test, Tumor size and Ki67% was compared with final ypN status by Mann-Whitney U test. Multivariable logistic regression with backward stepwise selection was used on all univariately significant variables and AUC was reported.

We identified 879 patients with a new breast cancer diagnosis who underwent preoperative breast MRI during the study period. Of those, 201 (22,9%) met the inclusion criteria and were analyzed. The mean age was 53.0 (SD±11.5) years. Surgical staging of the axilla was done in all patients [180(89.6%) SLNB alone, 4 (2.0%) ALND alone, and 17 (8.5%) both], which showed ypN+ disease in 29 [14.4%; ypN1 in 26(12.9%) and ypN2-N3 in 3 (1.5%)]. The median metastatic volume was 1 node (range,1-14). Higher rate of ER+ (20.4% vs 6.8%, p<.001), PR+ (21.8% vs 7.0%, p<.001), luminal A (36.8% vs 12.1%, p<.001), and multifocal (26.3% vs 11.7%, p=.03) cancers were ypN+ while high grade [93.9%(grade III) vs 81.8%-72.3%(grade I-II), p <.001] and triple negative cancers (95.5% vs 80.6%, p=.01) were ypN0. Ki67% was lower in ypN+ cancers (median, 24.5, IQR, 13.8-36.2 vs median, 51.5, IQR, 30.0-80.0, p<.001). Of 188 (89.5%) patients who had post-therapy MRI, 64 (34.2%) showed complete imaging response. Patients who showed complete response on MRI were more likely to have ypN0 axillae (93.8% vs 83.1%, p=.04). None of the cancers with complete pathologic response in the primary tumor were ypN+. Our model to predict axillary metastasis yielded AUC=0.75 (95%CI:0.65-0.85,p<.001) [sensitivity 91%, specificity 32%, positive predictive value 19%, negative predictive value 95%].

Multifocal, ER+, PR+, luminal A cancers, low invasive grade, and Ki67% are risk factors associated with nodal metastasis at post-NAC surgical staging in patients assessed as cN0 on pre-therapy MRI. Complete response on MRI is significantly associated with ypN0 status. Clinical trials are needed to determine whether patients with cN0 disease and complete response to NAC on MRI may be eligible to omit axillary surgery.