Abstract A50: Circulating tumor DNA (ctDNA) and magnetic resonance imaging (MRI) for monitoring and predicting response to neoadjuvant therapy (NAT) in high-risk early breast cancer patients in the I-SPY 2 TRIAL

Background: MRI measurements (Li et al., Magn Reson Imaging 2019; Hylton et al., Radiology 2016) and ctDNA (Magbanua et al., SABCS 2018) have both been independently shown to associate with response to NAT. We performed a retrospective study to examine correlation between ctDNA and MRI and to investigate whether information from these two measurements can be combined to improve early prediction of response.Methods: We analyzed serial ctDNA and MRI data from 84 high-risk (stage II/III) breast cancer patients collected at baseline (T0), 3 weeks after initiation of paclitaxel-based NAT (T1), between paclitaxel and anthracycline regimens (T2), and after NAT prior to surgery (T3). The response variable was pathologic complete response (pCR), defined as the absence of invasive tumor in the breast and lymph nodes after NAT. We examined correlations between MR functional tumor volume (FTV) and ctDNA using Spearman's rho (r). Mean FTV between ctDNA+/- groups were compared using t-test. Monte Carlo simulation was used to assess correlation between FTV and ctDNA trajectories in individual patients. We investigated the impact of adding ctDNA information to MR FTV-based predictors using receiver operating characteristic curves to calculate area under the curve (AUC), logistic regressions, and decision trees using recursive partitioning.Results: The mean levels of ctDNA (mutant molecules/mL plasma) were significantly correlated with FTV at all timepoints [T0 (r=0.49), T1 (r=0.42), T2 (r=0.42), T3 (r=0.43), all p<0.05]. The mean FTV in patients who had detectable ctDNA was significantly higher compared to those who were negative (all timepoints, all p<0.05). FTV and ctDNA trajectories in individual patients over the course of therapy were correlated (empirical 1-sided p=0.046). Adding continuous ctDNA information (mutant molecules/mL plasma) to FTV at T1 improved AUC in the pCR-prediction model, but the increase was not statistically significant (FTV: 0.59, FTV+ctDNA: 0.69, p=0.25). No improvements in AUCs were observed at other timepoints. Treated as a dichotomous variable, ctDNA positivity at T1 trended toward association with non-pCR in logistic regression models at T2 and T3 with MR-based prediction scores as a covariate (0.05).Citation Format: Mark Jesus M. Magbanua, Laura H. Hendrix, Terry Hyslop, William T. Barry, Eric P. Winer, Clifford Hudis, Deborah Toppmeyer, Lisa Anne Carey, Ann H. Partridge, Jean-Yves Pierga, Tanja Fehm, José Vidal-Martínez, Dimitrios Mavroudis, Jose A. Garcia-Saenz, Justin Stebbing, Paola Gazzaniga, Luis Manso, Rita Zamarchi, María Luisa Antelo, Leticia De Mattos-Arruda, Daniele Generali, Carlos Caldas, Elisabetta Munzone, Luc Dirix, Amy L. Delson, Harold Burstein, Misbah Qadir, Cynthia Ma, Janet H. Scott, François-Clément Bidard, John W. Park, Hope S. Rugo. Circulating tumor DNA (ctDNA) and magnetic resonance imaging (MRI) for monitoring and predicting response to neoadjuvant therapy (NAT) in high-risk early breast cancer patients in the I-SPY 2 TRIAL [abstract]. In: Proceedings of the AACR Special Conference on Advances in Liquid Biopsies; Jan 13-16, 2020; Miami, FL. Philadelphia (PA): AACR; Clin Cancer Res 2020;26(11_Suppl):Abstract nr A50.

Duke Scholars

Author Terry Hyslop Biostatistics & Bioinformatics, Division of Translational Bi ...