Abstract 5573: Use of circulating tumor DNA (ctDNA) for early assessment of treatment response in patients with non-small cell lung cancer (NSCLC): A real-world (RW) analysis incorporating baseline ctDNA level and molecular response

Abstract Background: Data suggests that changes in ctDNA quantity correlate with response to therapy in patients with advanced solid malignancies. Furthermore, absolute baseline (pre-treatment) ctDNA level has been shown to be associated with patient prognosis. However, there is little information on how these variables can be combined to better interpret ctDNA results and enhance predictive power of treatment response. Here, we develop a model that incorporates the effects of both baseline ctDNA level as well as linear and nonlinear relative ctDNA change. Methods: We queried the Guardant INFORM database, which comprises aggregated commercial payer health claims and de-identified records from over 225,000 individuals with comprehensive ctDNA testing via Guardant360 (G360). From September 2018-March 2022, patients with NSCLC treated with immune checkpoint inhibitors (ICI) (monotherapy or in combination) who received a ctDNA test within 15 weeks prior to treatment initiation and a second test 3-15 weeks after treatment initiation were retrospectively evaluated. Cox proportional hazards (CPH) were used for RW time to next treatment (TTNT) and time to treatment discontinuation (TTD) analyses. Gender, age, line of therapy (LOT) and comorbidities were included as covariates. ctDNA change from baseline to on-treatment (using the Guardant360 Response algorithm) was modeled as a continuous variable using a regularized cubic spline and baseline ctDNA level (using maximum variant allele fraction) was modeled as an interaction effect. Model validation and calibration were carried out via bootstrap resampling, ANOVA was used for multiple design hypothesis testing. Results: 82 patients met the study criteria. 31 patients with large ctDNA decrease (e.g. >=50% decrease), also known as molecular responders, had significantly longer TTNT and TTD compared to other patients, regardless of baseline ctDNA level. 51 patients with small ctDNA decrease or ctDNA increase showed ctDNA level dependent trends in TTNT and TTD, with decreased risk for shorter TTNT and TTD for lower baseline ctDNA level. We found that the interaction between baseline ctDNA level and ctDNA change was significant (TTNT: p = .0069, TTD: p = .0236), and that ctDNA change had a nonlinear effect on both TTNT and TTD (TTNT: p = .0014, TTD: p = .0349). TTNT (TTD) bootstrap-adjusted concordance indices were 0.69 (0.65) and hazard ratios [95% CI] were 5.36 [1.79, 16.10] (2.37 [1.13, 4.98]). Conclusions: These data demonstrate a statistically significant interaction between baseline ctDNA level and treatment-induced ctDNA level change in association with RW clinical outcomes. This facilitates the development of patient-based models incorporating baseline ctDNA level with ctDNA molecular response for association with RW clinical outcomes. Citation Format: Sean Gordon, Bojan Losic, Katie Quinn, Kyle Chang, Jiemin Liao, Han-Yu Chuang. Use of circulating tumor DNA (ctDNA) for early assessment of treatment response in patients with non-small cell lung cancer (NSCLC): A real-world (RW) analysis incorporating baseline ctDNA level and molecular response. [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2023; Part 1 (Regular and Invited Abstracts); 2023 Apr 14-19; Orlando, FL. Philadelphia (PA): AACR; Cancer Res 2023;83(7_Suppl):Abstract nr 5573.