Dynamic Changes of Circulating Tumor DNA Predict Clinical Outcome in Patients With Advanced Non–Small-Cell Lung Cancer Treated With Immune Checkpoint Inhibitors

Sabrina Weber,Paul van der Leest,Hylke C. Donker,Thomas Schlange,Wim Timens,Menno Tamminga,Samantha O. Hasenleithner,Ricarda Graf,Tina Moser,Benjamin Spiegl,Marie-Laure Yaspo,Leon W. M. M. Terstappen,Grigory Sidorenkov,T. Jeroen. N. Hiltermann,Michael R. Speicher,Ed Schuuring,Ellen Heitzer,Harry J. M. Groen
DOI: https://doi.org/10.1200/po.21.00182
2021-11-01
JCO Precision Oncology
Abstract:PURPOSE Immune checkpoint inhibitors (ICIs) are increasingly being used in non–small-cell lung cancer (NSCLC), yet biomarkers predicting their benefit are lacking. We evaluated if on-treatment changes of circulating tumor DNA (ctDNA) from ICI start (t 0 ) to after two cycles (t 1 ) assessed with a commercial panel could identify patients with NSCLC who would benefit from ICI. PATIENTS AND METHODS The molecular ctDNA response was evaluated as a predictor of radiographic tumor response and long-term survival benefit of ICI. To maximize the yield of ctDNA detection, de novo mutation calling was performed. Furthermore, the impact of clonal hematopoiesis (CH)–related variants as a source of biologic noise was investigated. RESULTS After correction for CH-related variants, which were detected in 75 patients (44.9%), ctDNA was detected in 152 of 167 (91.0%) patients. We observed only a fair agreement of the molecular and radiographic response, which was even more impaired by the inclusion of CH-related variants. After exclusion of those, a ≥ 50% molecular response improved progression-free survival (10 v 2 months; hazard ratio [HR], 0.55; 95% CI, 0.39 to 0.77; P = .0011) and overall survival (18.4 v 5.9 months; HR, 0.44; 95% CI, 0.31 to 0.62; P < .0001) compared with patients not achieving this end point. After adjusting for clinical variables, ctDNA response and STK11/ KEAP1 mutations (HR, 2.08; 95% CI, 1.4 to 3.0; P < .001) remained independent predictors for overall survival, irrespective of programmed death ligand-1 expression. A landmark survival analysis at 2 months (n = 129) provided similar results. CONCLUSION On-treatment changes of ctDNA in plasma reveal predictive information for long-term clinical benefit in ICI-treated patients with NSCLC. A broader NSCLC patient coverage through de novo mutation calling and the use of a variant call set excluding CH-related variants improved the classification of molecular responders, but had no significant impact on survival.
oncology
What problem does this paper attempt to address?