Abstract PD7-01: Towards personalized medicine - patient-derived breast tumor grafts as predictors of relapse and response to therapy. Preliminary results

Background: Predicting the risk of relapse in patients with non-metastatic breast cancer is important for medical decisions and patient counseling. For patients who receive neoadjuvant chemotherapy, pathologic response and especially pathologic complete response (pCR) has been associated with risk of relapse; however this association is imperfect. Our prior work in patient-derived xenograft (PDX) models indicated that tumor engraftment in mice correlated with risk of recurrence. To understand further the prognostic utility of PDX, we designed a prospective clinical trial to determine the correlation between PDX generation with residual disease following neoadjuvant chemotherapy and relapse. Preliminary data are presented. Methods: Women with newly diagnosed non-metastatic hormone receptor low-positive (HR-low, ER and/or PR ≤ 10%) or negative breast cancer planned to receive systemic chemotherapy prior to definitive surgery were eligible. Tumor tissue at diagnosis was orthotopically implanted in NOD/SCID mice. The primary objective of the study is to correlate the ability of a tumor to engraft in mice with pathologic responses and clinical outcomes. Results: Between 12/2016 and 2/2020, 58 patients enrolled (triple negative breast cancer (TNBC), n=37; HR-low or negative/Her2+, n=11; or HR-low/Her2-, n=8; mixed, n=1). PDXs were successfully established from 16 patients. Patients uniformly received intensive preoperative chemotherapy per standard of care. Among the 12 patients whose tumors engrafted in mice (PDX(+)) and underwent surgery, the pCR rate was 41.7%. In the subgroup of patients with postoperative follow up >6 months (n=9), 3 patients had achieved a pCR, 1 of whom recurred; and 6 patients had residual disease, 3 of whom recurred (overall relapse rate 44.4%). All patients who relapsed (TNBC n=3; HR-low/Her2- n=1), experienced a very early relapse ( Among the 38 patients whose tumors did not engraft (PDX(-)) and underwent surgery, the overall pCR rate was 60.6%. In the subgroup of patients with postoperative follow up >6 months (n=30), 18 patients had achieved a pCR of whom none recurred, and 12 had residual disease of whom 1 patient (with TNBC) had locoregional recurrence 35.1 months after her definitive surgery (relapse rate 3.3%). She underwent repeat surgery followed by radiation therapy and 5.5 months after her recurrence, she remains disease-free. Achievement of pCR was not statistically different between the PDX(+) and PDX(-) group. In the entire cohort, the presence of residual disease was associated with high risk of relapse (22.2%) as compared to the achievement of pCR (4.8%, p = 0.16). However, successful tumor engraftment was associated not only with a higher risk of relapse (44.4% vs 3.3%, p = 0.0068, odds ratio 20.45), but also an early and exceptionally aggressive relapse. Conclusion. Our functional studies not only identify a patient population with a high risk of relapse with greater precision than the achievement of pCR, but also identify patients whose relapse has a particularly aggressive natural history. We established models that recapitulate this aggressive disease and in-depth genomic studies are underway. These studies will lead us to better patient stratification on the basis of risk of relapse and novel therapeutic strategies focused on patients with exceptionally aggressive breast cancers that our current diagnostic methods cannot identify. Citation Format: Christos Vaklavas, Zhengtao Chu, Rachel E Factor, Alana L Welm, Bryan E Welm, Cindy B Matsen. Towards personalized medicine - patient-derived breast tumor grafts as predictors of relapse and response to therapy. Preliminary results [abstract]. In: Proceedings of the 2020 San Antonio Breast Cancer Virtual Symposium; 2020 Dec 8-11; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2021;81(4 Suppl):Abstract nr PD7-01.