
Beyond the Scan: Predicting Immunotherapy Response With ctDNA in NSCLC
Hear how a single ctDNA blood draw 3 to 9 weeks into immunotherapy may reveal response earlier than imaging in NSCLC, with Valsamo Anagnostou, MD, PhD.
Valsamo “Elsa” Anagnostou, MD, PhD, of Johns Hopkins University School of Medicine, joins Targeted Talks to discuss a study on
Radiographic imaging, the current gold standard, can be slow and imprecise at capturing immunotherapy response. Anagnostou explains why a rapid molecular readout is needed to tailor treatment. In the prospective study of 109 patients, undetectable ctDNA 3 to 9 weeks after starting anti–PD-(L)1 therapy, alone or in combination, signaled clinical benefit. Persistent ctDNA signaled molecular disease progression.
A single on-treatment measurement predicted outcomes as well as the traditional 2-time point approach, which compares ctDNA at baseline with a later time point. That matters for patients who miss a baseline draw or have no detectable ctDNA at baseline. Clearing ctDNA even once in the window predicted outcome, regardless of variability in readouts within the window.
Anagnostou also covers:
- What’s still needed: Evidence that adapting treatment to ctDNA improves survival. The BR.36 trial, which she chairs, tests this in metastatic NSCLC with high PD-L1 expression. All patients start pembrolizumab. At 6 weeks, patients with detectable ctDNA are randomized to continue pembrolizumab or intensify with chemotherapy.
- Challenges: Choosing assays that fit the clinical setting, and handling technical and biological noise. Some mutations detected in blood come from blood cells rather than the tumor.
- What clinicians can do now: Enroll patients in ctDNA-adaptive trials. Serial liquid biopsy is already common practice in oncogene-driven NSCLC, such as EGFR-mutant and ALK-rearranged disease. It is also used informally when imaging is unclear.
Related to this article







