Abstract P1-02-06: Serial monitoring of circulating tumor DNA in patients with metastatic breast cancer

Background: For patients with MBC, there is currently no evidence that changing therapy on the basis of biomarker results improves outcome. Clinical benefit of treatment is defined as complete response, objective response, or stable disease as determined by RECIST criteria on radiological evaluation. Serial measurements of serum biomarkers such as CA2729 and CTCs have proven unsuccessful in predicting clinical benefit. Circulating tumor DNA(ctDNA) has emerged as a potential biomarker that may predict response to therapy or progression of disease. The present retrospective study was conducted to evaluate the relationship between change in ctDNA with clinical benefit determined by clinical and radiological evaluations of patients with MBC patients.

Methods: We conducted a retrospective, single-institutional study to determine if serial monitoring of ctDNA allele frequency levels predict clinical benefit of a treatment. 55 patients with measurable MBC who had serial monitoring of ctDNA between August 2014 and May 2016 were included. The median age was 55.9 (27–94) years). Clinical outcomes were determined as per standard guidelines. The analysis was performed on all cases that had serial measurements of ctDNA with no change in therapy in between and the repeat blood draw was done within 30 days of repeat radiographic evaluation. The dataset contained 125 observations from 48 unique patients. The relationship between the change in ctDNA and clinical benefit was analyzed using a generalized linear model with a random subject effect to account for the intrapatient dependence occurring from obtaining multiple evaluations from the same patient. A logit link function was used akin to logistic regression and a compound symmetric correlation structure was assumed.

Results: 68.8% of the cases were hormone receptor-positive, 18.8% HER2-positive, and 27.1% TNBC. The treatments received were 58.4% hormonal therapy, 31.2% chemotherapy, 26.4% included anti-HER2 therapy, 2 cases were on targeted therapy, and 1 case was not on any treatment. Three patients had stage 4 disease in complete remission. ctDNA analysis was repeated on average 4 days prior to radiological evaluation. The average time between repeat assessments was 108.5 days. 93% of the patients had a genomic alteration detected at some point during their course of disease. The most common mutations detected were TP53 41.7%, PIK3CA 35.4%, ESR1 18.8%, and ERBB2 amplifications 6.3%. A dichotomized change in ctDNA is a significant predictor of clinical benefit (p < 0.0001). The intrapatient correlation is estimated to be 0.273 for the transformed variable. The model yields a predicted probability of clinical benefit of 26.9% when the increase in ctDNA is greater than or equal to 0.5 and when the increase in ctDNA is less than 0.5, the a predicted probability of clinical benefit is 78.4%. The concordance of change in ctDNA and change in CA 27-29 was 76.2%.

Conclusions: Serial evaluation of serum ctDNA may be useful to evaluate molecular response to treatment which may correlate with clinical benefit and potentially guide treatment decisions. Early indication that a chosen therapy is not effective may lead to avoidance of overtreatment and initiation of an alternative regimen. Further, prospective studies are needed.

Citation Format: Patel A, Mukherjee A, Hwang D, Ensor J, Patel TA, Chang JC, Rodriguez AA. Serial monitoring of circulating tumor DNA in patients with metastatic breast cancer [abstract]. In: Proceedings of the 2016 San Antonio Breast Cancer Symposium; 2016 Dec 6-10; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2017;77(4 Suppl):Abstract nr P1-02-06.