Pre-surgical ctDNA clearance reached 66% with neoadjuvant nivolumab against 38% with placebo in CheckMate 77T

Pre-surgical ctDNA clearance in 66% of evaluable nivolumab patients against 38% on placebo

Among the 98 nivolumab-treated patients with evaluable biomarkers, 76 (78%) had detectable and evaluable circulating tumour DNA both before and after neoadjuvant treatment, and 50 of those 76 (66%) cleared it before surgery. In the placebo group, 64 of 92 patients (70%) met the same evaluability condition, and 24 of them (38%) cleared it.

The denominators matter for reading those percentages. Clearance is only measurable in a patient whose ctDNA was detectable to begin with and whose post-treatment sample could be evaluated, so the 66% and the 38% describe 76 and 64 patients respectively, not the 229 and 232 who were randomised. Detection at baseline itself was not universal: 83 of 98 nivolumab-treated patients (85%) and 75 of 92 placebo-treated patients (82%) had detectable ctDNA before neoadjuvant treatment began, and detection rose to 90 of 98 (92%) and 78 of 92 (85%) at the completion of neoadjuvant treatment.

Half of the nivolumab patients who cleared ctDNA had a pathologic complete response

Of the 50 nivolumab-treated patients with pre-surgical clearance, 25 (50%) had a pathologic complete response at surgery. Of the 24 placebo-treated patients with clearance, 3 (12%) did.

The two measures are taken at different moments and from different material. Clearance is a blood measurement made before the operation, and pathologic complete response is read from the resection specimen afterwards, so the second is a check on what the first anticipated. What the figures establish is that clearance and pathologic complete response occurred together far more often in the nivolumab group, and that clearance on placebo was much less likely to be accompanied by a complete pathologic result. The abstract carries no survival figure attached to either comparison.

Every patient whose molecular residual disease returned during adjuvant treatment relapsed

Four of 48 patients (8%) in the nivolumab group and 9 of 44 (20%) in the placebo group were negative for molecular residual disease after surgery and before adjuvant treatment began, then turned positive during the adjuvant period. All of them had disease recurrence.

Thirteen patients is a small number on which to rest a statement about a test, and the analysis reports the observation rather than a predictive value derived from it. The finding is nevertheless unambiguous within its own cohort: in this trial, no patient who converted from molecular residual disease negative to positive on adjuvant treatment avoided recurrence.

“These findings provide insights into predictive markers for outcomes with perioperative nivolumab in resectable NSCLC.”Cascone et al., Nature (2026)

A hazard ratio of 0.48 in tumours carrying KEAP1, STK11, CDKN2A or SMARCA4 alterations

Event-free survival appeared prolonged with nivolumab (n=60) against placebo (n=45) in patients whose tumours carried single alterations or co-alterations in any of KEAP1, STK11, CDKN2A or SMARCA4, with a hazard ratio of 0.48 and a 95% confidence interval of 0.28 to 0.83. The authors report the result with the verb “seemed”, and the abstract gives no median event-free survival for either arm of that subgroup.

A machine-learning model trained on the biomarker-evaluable patients placed pre-surgical ctDNA clearance, non-N2 disease, pathologic complete response, squamous histology and nivolumab treatment among its top predictors of prolonged event-free survival. The abstract names those predictors without ordering them and gives no performance measure for the model itself.

The analysis covers 41% of the randomised patients

Biomarkers were evaluable in 98 of 229 patients assigned to nivolumab and 92 of 232 assigned to placebo, which the authors state is 41% of those randomised. Every figure above is therefore drawn from patients who had usable samples, and the analysis does not establish that they resemble the 59% who did not.

CheckMate 77T reported previously that perioperative nivolumab significantly improved event-free survival against placebo in resectable non-small cell lung cancer. This report does not revisit that primary result. It describes which patients within the trial the benefit tracked with, on samples collected during the trial, and the strength of the evidence is that of a biomarker analysis of a subset rather than of a randomised comparison.