Why This Study Matters
MTAP is deleted in 13% of non-small cell lung cancer (NSCLC). MTAP loss leads to accumulation of MTA, a partial inhibitor of PRMT5, and MTAP-deleted cancer cells are selectively sensitized to further PRMT5 inhibition. Second-generation MTA-cooperative PRMT5 inhibitors are now in early-phase trials, which makes it important to know how patients with MTAP-deleted NSCLC fare on today's standard-of-care regimens. The authors note that data on this question were limited.
Study Design
Patients with advanced NSCLC treated at Dana-Farber Cancer Institute whose tumors underwent large-panel next-generation sequencing (OncoPanel, 447 genes) from October 26, 2016 to March 2, 2024 were included if MTAP status was known. MTAPdel meant a homozygous (two-copy) deletion; MTAPwt meant both copies intact. Patients with one-copy deletions were excluded. Progression-free survival (PFS) and overall survival (OS) were measured from treatment initiation and confirmed by chart review; the data analysis cut-off was December 12, 2024. Comparisons used Kaplan–Meier methods, Cox hazard ratios, and multivariable Cox models adjusted for histology, age, smoking pack-years, PD-L1 TPS (0–49% vs ≥50%), targetable co-mutation status, and treatment line.
400 patients in the cohort (93 MTAPdel, 307 MTAPwt), treated at 1 center (Dana-Farber), with 4 regimen groups analyzed. Sequencing used a 447-gene OncoPanel.
Randomization: not applicable (observational). Dosing: not applicable. Patients could appear in more than one regimen group, so group sizes should not be summed.
Patient Population
| Characteristic | MTAPdel (n=93) | MTAPwt (n=307) | p |
|---|---|---|---|
| Female | 65.6% | 51.5% | 0.018 |
| Stage IV at diagnosis | 78.8% | 60.9% | 0.002 |
| Brain metastases at diagnosis | 43.0% | 26.1% | 0.003 |
| ≥30 pack-years smoking | 33.3% | 49.0% | 0.008 |
| PD-L1 TPS <1% | 37.9% | 24.3% | 0.017 |
| Median TMB (mut./Mb) | 7.6 | 9.9 | 0.012 |
| CDKN2A two-copy deletion | 91.4% | 4.2% | <0.001 |
| CDKN2B two-copy deletion | 89.2% | 3.9% | <0.001 |
| EGFR mutation (top-30 gene analysis) | 39.8% | 20.8% | <0.001 (q<0.030) |
| Any targetable mutation | 41.9% | 30.9% | 0.060 |
Age, self-reported race and ethnicity, histology, bone/liver/adrenal metastases, and ECOG performance status at diagnosis did not differ significantly between cohorts.
Primary Endpoint Results
The authors compared PFS and OS between MTAPdel and MTAPwt cohorts within each regimen. The strongest signals were for PFS on anti-PD-(L)1-based therapy, both in univariable and multivariable analysis. Hazard ratios below are MTAPdel vs MTAPwt (univariable unless noted).
| Regimen (n MTAPdel vs MTAPwt) | Median PFS (months) | PFS HR (95% CI); p | Median OS (months) | OS HR (95% CI); p |
|---|---|---|---|---|
| Platinum doublet (32 vs 106) | 4.8 vs 5.4 | 1.4 (0.9–2.0); 0.108 | 11.9 vs 17.4 | 1.6 (1.0–2.5); 0.054 |
| Anti-PD-(L)1 monotherapy (28 vs 128) | 2.0 vs 4.9 | 2.0 (1.2–3.3); 0.003 | 12.7 vs 16.7 | 1.2 (0.8–2.0); 0.367 |
| Platinum doublet + anti-PD-(L)1 (40 vs 104) | 4.2 vs 7.5 | 2.0 (1.4–3.3); <0.001 | 14.9 vs 19.1 | 1.3 (0.8–2.0); 0.194 |
| Docetaxel-based (9 vs 43) | 1.6 vs 2.0 | 1.2 (0.6–2.5); 0.622 | 4.5 vs 6.4 | 1.3 (0.6–3.3); 0.474 |
After multivariable adjustment, MTAPdel was independently associated with shorter PFS on anti-PD-(L)1 monotherapy (HR 1.70 [95% CI 1.02–2.82]; p=0.040) and on platinum doublet + anti-PD-(L)1 (HR 1.89 [95% CI 1.20–2.97]; p=0.006). The trend toward shorter OS on platinum doublet alone in univariable analysis was, per the authors, more likely driven by later treatment lines and older age in the MTAPdel group.
Subgroup and Sensitivity Analyses
In a sensitivity analysis excluding all cases with tumor purity <40%, median PFS remained shorter in the MTAPdel cohort on anti-PD-(L)1 monotherapy (1.4 vs 4.9 months; p=0.005) and platinum doublet + anti-PD-(L)1 (3.7 vs 7.5 months; p<0.001). Multivariable HRs were 1.74 (95% CI 1.00–2.90; p=0.050) and 2.18 (95% CI 1.33–3.51; p=0.002), respectively. For platinum doublet + anti-PD-(L)1, median OS was 12.3 vs 19.1 months (p=0.036) on univariable analysis, but MTAPdel was not independently associated with shorter OS on multivariable analysis.
Among MTAPdel patients with more extensive 9p loss, only 4 received anti-PD-(L)1 monotherapy and only 4 received platinum doublet + anti-PD-(L)1, too few for firm conclusions. Subgroup analyses by PD-L1 level: not reported as separate statistics in the source.
Safety Profile
Not specified in source. This retrospective outcomes study did not report adverse-event rates for the regimens studied. [Background from the paper's introduction, not a result of this study: first-generation PRMT5 inhibitors in phase I trials had treatment-related adverse events such as fatigue (28–45%), thrombocytopenia (27–41%), anemia (24–41%), and nausea (29–39%), with dose discontinuation rates of 4–13%.]
Interpretation and Broader Context
The authors conclude that MTAP-deleted advanced NSCLC has features of more aggressive disease (lower PD-L1 expression, more stage IV disease and brain metastases at diagnosis) and worse PFS on anti-PD-(L)1-based therapy, even after adjustment. They note directional concordance with an earlier cohort study (Ebot et al., 2022) of first-line pembrolizumab-based therapy, which found shorter survival with pembrolizumab monotherapy but not pembrolizumab plus chemotherapy. They propose several possible biological explanations, including co-deletion of 9p21.3 interferon genes and CDKN2A, and effects of MTA on T-cell function, without testing them in this dataset.
In summary, our findings suggest that MTAPdel NSCLC is an underserved and more aggressive subset of NSCLC that will require dedicated drug development to develop effective monotherapies and rational combination strategies.
— Huang et al., Clin Cancer Res 2026
Limitations
Limitations the authors state include a retrospective single-center design, MTAP status inferred from DNA sequencing (protein expression data were not available in most cases), different tumor-purity cut-offs between cohorts (MTAPwt <40% excluded; MTAPdel lowered to 20%), a small docetaxel cohort (9 MTAPdel patients), and possible non-independence of OS across sequential regimens. OS was not significantly different in any group.