L858R with a TP53 co-mutation halved real-world PFS on front-line osimertinib

Median real-world progression-free survival of 10.1 months with both markers against 21.4 with neither

The two-marker comparison is the central result. Patients whose tumours carried EGFR L858R together with a TP53 co-mutation had a median real-world progression-free survival of 10.1 months on front-line osimertinib. Patients whose tumours carried an exon 19 deletion and wild-type TP53 had a median of 21.4 months. The hazard ratio was 2.2, with a P value below 0.001.

The size of the group makes the comparison worth reading closely. Of the 606 patients, 186 (30.7%) carried both unfavourable alterations, so the group with the shorter survival is close to a third of the cohort rather than a rare subgroup. The endpoint is real-world progression-free survival, and the abstract does not state how progression was ascertained.

Median overall survival of 21.3 months against 53.4 months in the same patients

Overall survival separated by a wider absolute margin than progression-free survival. Median overall survival was 21.3 months in patients with both markers and 53.4 months in patients with neither, a hazard ratio of 2.3 with a P value below 0.001. The difference of just over 32 months is larger than the whole median survival of the two-marker group.

An overall survival difference observed in a database cohort carries whatever treatment patients received after osimertinib as well as the markers being studied, and the abstract reports no subsequent-therapy data. What the figures establish is the survival of these patients as observed, not the share of it attributable to the markers themselves.

“Having EGFR L858R or a TP53 co-mutation were independent predictors of inferior rwPFS and OS with front-line osimertinib.”Kim et al., Lung Cancer (2026)

Hazard ratios of 1.4 and 1.5 for each marker taken on its own

Taken separately, L858R against exon 19 deletion gave a hazard ratio of 1.4 for real-world progression-free survival (P equal to 0.001) and 1.3 for overall survival (P equal to 0.01). A TP53 co-mutation against wild-type TP53 gave 1.5 for progression-free survival and 1.6 for overall survival, both with P values below 0.001.

Both associations held in multivariable Cox regression after accounting for clinical covariates, which is the basis on which the authors call them independent predictors. The abstract names neither the covariates included nor the confidence intervals around any of the four hazard ratios, so the precision of each estimate cannot be read from it.

A database cohort of 606 patients, with a real-world progression endpoint

Patients were studied from a United States clinical-genomic database rather than enrolled in a trial, and the endpoints were real-world progression-free survival and overall survival determined by Kaplan-Meier methods. TP53 co-mutations were present in 384 of the 606 patients, or 63%, which makes the TP53-wildtype group the smaller of the two throughout the comparisons above.

The authors propose the combination of the two markers as a way of identifying patients for novel trials or for approved intensified therapies. That is a proposal about trial design and treatment selection rather than a demonstration that intensified treatment helps these patients, and this study tested no intensified regimen.