From the Journals

NSCLC survival on immunotherapy much lower in ‘real world’


Real-world use of the immune checkpoint inhibitors for first-line treatment of advanced non–small cell lung cancer (NSCLC) provides nowhere near the same survival advantage as seen in clinical trials, according to a retrospective cohort study of nearly 20,000 Medicare patients.

For example, the median overall survival (OS) in the “real world” was 11.4 months for patients treated with pembrolizumab (Keytruda, Merck) monotherapy – approximately 15 months shorter than the median OS among pembrolizumab-treated participants in the KEYNOTE-024 trial.

Indeed, OS was shorter for Medicare patients treated with an immune checkpoint inhibitor alone than it was for patients treated with a chemoimmunotherapy regimen of platinum plus pemetrexed plus pembrolizumab, at a median of 12.9 months – which in itself was approximately 10 months shorter than survival outcomes with this triplet therapy in the KEYNOTE-189 trial.

“These results, based on the nationwide experience for patients on Medicare, may inform discussions between physicians and patients with respect to expectations for outcomes among older patients with NSCLC,” lead author Kenneth Kehl, MD, assistant professor of medicine, Harvard Medical School, Boston, said in a statement.

Deborah Schrag, MD, chief, division of population sciences, Dana-Farber Cancer Institute, Boston, and Harvard Medical School, agreed, adding in the same statement that “this information empowers patients and clinicians with realistic expectations and equips them to make informed decisions.”

The study was published online May 21 in JAMA Network Open and was done in conjunction with the Health Data Analytics Institute, an analytics firm that applies artificial intelligence for measuring health risks.

Systemic therapy

For the study, the team analyzed Medicare data for 19,529 patients (median age, 73.8 years) who had all initiated first palliative-intent systemic therapy for lung cancer between January 2016 and December 2018. Some 3,079 patients received pembrolizumab monotherapy, 5,159 patients received a platinum-based regimen plus pemetrexed, 9,866 received a platinum plus a taxane, and 1,425 received platinum, pemetrexed, and pembrolizumab.

The authors noted that uptake of pembrolizumab-containing regimens in the Medicare population was rapid.

In the second quarter of 2016, pembrolizumab was used in only 0.7% of first-line treatments for advanced NSCLC, but increased to 42.4% of first-line treatments 2 years later, in the third quarter of 2018.

“The primary outcome was OS, which was measured using the restricted mean survival time (RMST),” Dr. Kehl and colleagues noted.

After propensity-score stratification, patients who received pembrolizumab had an adjusted RMST of 11 months compared with an adjusted RMST of 11.1 months for those who received the combination of platinum plus pemetrexed.

Survival was statistically worse for patients who received pembrolizumab than it was for those treated with a platinum/taxane combination, although the magnitude of difference between the two groups was small, at 0.7 months (P < .001). Patients who received the platinum/pemetrexed/pembrolizumab triplet had an adjusted RMST of 11.7 months, which was significantly better than the adjusted RMST of 11.2 months for patients who received the platinum/pemetrexed doublet, but the magnitude of the difference between these two groups was small, at 0.5 months (P = .02), the investigators added.

Different patient groups

Patients who received immunotherapy alone may have been more ill than those who received chemotherapy, the authors suggested. Patients who were 70 years of age or older, who were female, and who had a higher baseline mortality risk were more likely to receive single-agent pembrolizumab than chemotherapy, they noted. “Indeed, immunotherapy may be construed as a potential first-line treatment for patients who would otherwise have been deemed too frail for treatment at all, including patients older than 80 years,” they observed.

It is also possible that the Medicare patients included in the current analysis may differ substantively from advanced NSCLC participants enrolled in clinical trials, they wrote. For example, the median age of the Medicare cohort was approximately 10 years older than the median age of participants in both KEYNOTE-024 and KEYNOTE-189, the authors pointed out.

“If clinicians recommend immunotherapy disproportionately to Medicare patients with poor performance status or greater comorbidity – perhaps even if PD-L1 (programmed cell death-ligand-1) expression levels are below thresholds associated with the most substantial immunotherapy benefit – it may not be surprising that large survival improvements associated with immunotherapy were not observed in this analysis,” Dr. Kehl and colleagues suggested.

It is possible that durable benefit from immunotherapy, at least among some subgroups of patients included in the Medicare analysis, might have become more evident with additional follow-up beyond 18 months, they noted. However, they added, in “both KEYNOTE-024 and KEYNOTE-189, pembrolizumab was associated with substantial improvements in overall survival by that point.

“These results may inform prognostic considerations in practice and reinforce the importance of understanding patient selection dynamics in assessing the value and clinical utility of transformative treatment strategies,” they cautioned.

Dr. Kehl has reported receiving personal fees from Aetion, Roche, and IBM. Dr. Schrag has reported receiving personal fees from JAMA for editorial services and travel reimbursement/speaker fees from Pfizer.

A version of this article first appeared on

Next Article: