Changes in Blood Tumor DNA Before and After Lung Cancer Surgery Reveal Warning Signs of Recurrence

For patients with resectable lung cancer, surgically removing tumors visible to the naked eye does not necessarily mean that cancer cells have disappeared. An exploratory analysis of CheckMate 77T published in *Nature* showed that blood-based monitoring of circulating tumor DNA (ctDNA) may provide earlier indications of treatment response and recurrence risk, offering a new approach to risk stratification for perioperative immunotherapy.

CheckMate 77T is a phase 3, double-blind trial that enrolled patients with resectable stage IIA to IIIB non-small cell lung cancer. Before surgery, participants received platinum-based doublet chemotherapy and were randomly assigned to receive either nivolumab or placebo; after completing surgery, they received the corresponding adjuvant treatment. This study focused on 190 patients with evaluable tumor and blood samples, including 98 in the nivolumab group and 92 in the placebo group.

Among patients whose ctDNA could be assessed both before and after treatment, 50 patients, or 66%, in the nivolumab group had no detectable tumor DNA that had originally been detectable before surgery, compared with 24 patients, or 38%, in the placebo group. Looking further, among patients who achieved ctDNA clearance, 50% in the nivolumab group had no viable cancer cells in their surgical specimens, meaning they achieved a pathological complete response, compared with 12% in the placebo group. Almost no patients without ctDNA clearance achieved a pathological complete response.

Serial postoperative monitoring revealed another signal. All 13 patients who were negative for molecular residual disease (MRD) before starting adjuvant treatment and subsequently became positive ultimately experienced disease recurrence; this change occurred in 4 patients in the nivolumab group and 9 in the placebo group. The results indicate that ctDNA may detect renewed growth of small numbers of cancer cells earlier than imaging, but the sample of 13 patients is extremely small, and it cannot be inferred from this that all MRD-positive patients will inevitably experience recurrence.

ctDNA clearance also cannot be directly equated with a cure, nor has it been shown to independently identify which patients must receive or can omit immunotherapy. This biomarker analysis covered only 41% of all randomized patients; some samples could not be included because of insufficient tissue, missing paired plasma, or assay failure. Differences in patients’ performance status and tumor PD-L1 expression may also have affected the observed associations. Before ctDNA can be used to adjust postoperative treatment, prospective trials are still needed to directly verify whether such decisions can improve survival.

The overall trial had previously shown that perioperative nivolumab can prolong event-free survival; this analysis adds insight into how blood-based biomarkers may characterize treatment depth and recurrence risk. Overall survival data remain immature. Safety was consistent with previous findings: treatment-related grade 3–4 adverse events occurred in 32% of the nivolumab group and 25% of the placebo group, with no new surgery-related safety signals.