Abstract
Background: Phase 3 clinical data has shown higher proportions of patients with objective response, longer response duration, and longer overall survival with nivolumab versus docetaxel in patients with previously treated advanced non-small-cell lung cancer (NSCLC). We aimed to evaluate the long-term benefit of nivolumab and the effect of response and disease control on subsequent survival. Methods: We pooled data from four clinical studies of nivolumab in patients with previously treated NSCLC (CheckMate 017, 057, 063, and 003) to evaluate survival outcomes. Trials of nivolumab in the second-line or later setting with at least 4 years follow-up were included. Comparisons of nivolumab versus docetaxel included all randomised patients from the phase 3 CheckMate 017 and 057 studies. We did landmark analyses by response status at 6 months to determine post-landmark survival outcomes. We excluded patients who did not have a radiographic tumour assessment at 6 months. Safety analyses included all patients who received at least one dose of nivolumab. Findings: Across all four studies, 4-year overall survival with nivolumab was 14% (95% CI 11–17) for all patients (n=664), 19% (15–24) for those with at least 1% PD-L1 expression, and 11% (7–16) for those with less than 1% PD-L1 expression. In CheckMate 017 and 057, 4-year overall survival was 14% (95% CI 11–18) in patients treated with nivolumab, compared with 5% (3–7) in patients treated with docetaxel. Survival subsequent to response at 6 months on nivolumab or docetaxel was longer than after progressive disease at 6 months, with hazard ratios for overall survival of 0·18 (95% 0·12–0·27) for nivolumab and 0·43 (0·29–0·65) for docetaxel; for stable disease versus progressive disease, hazard ratios were 0·52 (0·37–0·71) for nivolumab and 0·80 (0·61–1·04) for docetaxel. Long-term data did not show any new safety signals. Interpretation: Patients with advanced NSCLC treated with nivolumab achieved a greater duration of response compared with patients treated with docetaxel, which was associated with a long-term survival advantage. Funding: Bristol-Myers Squibb.
| Original language | English (US) |
|---|---|
| Pages (from-to) | 1395-1408 |
| Number of pages | 14 |
| Journal | The Lancet Oncology |
| Volume | 20 |
| Issue number | 10 |
| DOIs | |
| State | Published - Oct 2019 |
ASJC Scopus subject areas
- Oncology
Access to Document
Other files and links
Fingerprint Dive into the research topics of 'Four-year survival with nivolumab in patients with previously treated advanced non-small-cell lung cancer: a pooled analysis'. Together they form a unique fingerprint.
Cite this
- APA
- Standard
- Harvard
- Vancouver
- Author
- BIBTEX
- RIS
Four-year survival with nivolumab in patients with previously treated advanced non-small-cell lung cancer : a pooled analysis. / Antonia, Scott J.; Borghaei, Hossein; Ramalingam, Suresh S.; Horn, Leora; De Castro Carpeño, Javier; Pluzanski, Adam; Burgio, Marco A.; Garassino, Marina; Chow, Laura Q.M.; Gettinger, Scott; Crinò, Lucio; Planchard, David; Butts, Charles; Drilon, Alexander; Wojcik-Tomaszewska, Joanna; Otterson, Gregory A.; Agrawal, Shruti; Li, Ang; Penrod, John R.; Brahmer, Julie.
In: The Lancet Oncology, Vol. 20, No. 10, 10.2019, p. 1395-1408.Research output: Contribution to journal › Article › peer-review
}
TY - JOUR
T1 - Four-year survival with nivolumab in patients with previously treated advanced non-small-cell lung cancer
T2 - a pooled analysis
AU - Antonia, Scott J.
AU - Borghaei, Hossein
AU - Ramalingam, Suresh S.
AU - Horn, Leora
AU - De Castro Carpeño, Javier
AU - Pluzanski, Adam
AU - Burgio, Marco A.
AU - Garassino, Marina
AU - Chow, Laura Q.M.
AU - Gettinger, Scott
AU - Crinò, Lucio
AU - Planchard, David
AU - Butts, Charles
AU - Drilon, Alexander
AU - Wojcik-Tomaszewska, Joanna
AU - Otterson, Gregory A.
AU - Agrawal, Shruti
AU - Li, Ang
AU - Penrod, John R.
AU - Brahmer, Julie
N1 - Funding Information: Our survival analyses are based on a minimum follow-up of 4 years, which is—to our knowledge—the longest follow-up to date for an anti-PD-1 or anti-PD-L1 therapy in a large population of patients with previously treated advanced NSCLC. Across all four studies, nivolumab showed long-term survival benefit, with estimated 4-year overall survival of 14% overall, 11% in patients with less than 1% PD-L1 expression, and 19% in patients with at least 1% PD-L1 expression. The proportions of patients alive with nivolumab mostly stabilised at 3 years, as was previously seen for ipilimumab in the large pooled analysis of long-term overall survival in melanoma. 11 The corresponding proportions of patients alive with nivolumab in the combined CheckMate 017 and 057 population were similar to those of the pooled analysis of all four studies. This finding indicates that the nivolumab group of CheckMate 017 and 057 was representative of patients treated with nivolumab across all four studies, while having the advantage of allowing for comparisons with docetaxel in the setting of a randomised study population. Although patients in CheckMate 003 were more heavily pretreated than were those in the other three studies and included patients treated with potentially non-optimal doses of nivolumab (ie, 1 mg/kg and 10 mg/kg), the median overall survival (9·9 months) 8 was similar to that across all four studies (10·3 months); thus, including patients who received 1 mg/kg and 10 mg/kg would be unlikely to affect the current findings. Additionally, the 4-year overall survival was high, remained unchanged at 5 years (16%), 8 and here we have reported the overall survival at 6 years, which was only slightly lower (15%). By contrast to our findings of nivolumab, 4-year overall survival was only 5% for patients treated with docetaxel in CheckMate 017 and 057, irrespective of PD-L1 expression. Notably, the most recent estimate of the proportion of patients alive at 5 years since diagnosis for metastatic lung cancer in the USA is 5% for 2008–15, 2 which largely preceded the availability of immunotherapy for lung cancer. Our results suggest that subsequent updates of overall survival statistics might show substantial improvements in the long-term survival of patients with this disease. Achievement of a complete or partial response was associated with greatly improved overall survival post-response in patients treated with nivolumab or docetaxel. However, consistent with the increased response duration with nivolumab versus docetaxel, the long-term overall survival benefit was much larger for responders to nivolumab than to docetaxel. This finding suggests that PD-1 inhibitors provide greater durability of response benefit than cytotoxic chemotherapy. Furthermore, even among 4-year survivors, nivolumab was associated with longer and deeper responses than was docetaxel, which probably reflects the different mechanisms of action of PD-1 inhibition and cytotoxicity. Depth of response to PD-1 inhibitor therapy has previously been associated with progression-free survival and overall survival in metastatic NSCLC. 17 By contrast, some studies 18,19 found—at best—weak associations of objective responses or depth of response with progression-free survival and overall survival in NSCLC treated with targeted therapy or chemotherapy. The results of our landmark analysis of CheckMate 017 and 057 suggest that response at 6 months was associated with increased long-term survival post-landmark in both treatment groups. However, patients with a complete response, partial response, or stable disease at 6 months had longer subsequent overall survival with nivolumab than with docetaxel. Furthermore, the HRs for subsequent death in patients with a complete or partial response or stable disease versus progressive disease were more favourable with nivolumab than with docetaxel. Post-landmark progression-free survival was also longer in patients treated with nivolumab who had a complete or partial response than in those who had stable disease, which is consistent with the long-term survival benefit of response to nivolumab. Nivolumab was well tolerated, with few treatment-related deaths and no new or unexpected safety signals. This finding is consistent with previous results from the individual studies included in the present analyses, including CheckMate 017 and 057, which have shown the superior safety and tolerability profile of nivolumab versus docetaxel. 3,4 However, although most potentially immune-related, treatment-related adverse events with nivolumab occurred in the first 2–3 years of treatment, our findings suggest that vigilance should continue even at 4 years, particularly for skin reactions and gastrointestinal adverse events. Randomised controlled trials have shown long-term overall survival benefit with PD-1 or PD-L1 inhibitor therapy versus docetaxel in patients with previously treated advanced NSCLC. 9,10,20,21 However, identifying who in this patient population is most likely to achieve long-term survival with immunotherapy is a key challenge. Associations of baseline characteristics (eg, smoking status, tumour histology, CNS metastases, and ECOG performance status) and biomarker status (PD-L1 expression and tumour mutational burden) with overall survival in previously treated patients who received PD-1 or PD-L1 inhibitor therapy for advanced NSCLC have been explored previously. 6,8–10,21–25 In the present post-hoc analyses, owing to restricted tissue availability (tissue samples were not mandatory in nivolumab studies in previously treated patients with NSCLC), only survival outcomes according to tumour histology and PD-L1 expression were assessed. Evaluation of potential baseline characteristics correlating with long-term responders to nivolumab are planned in a future analysis of CheckMate 017 and 057. In addition, we expect that patients who maintain their clinical response for 6 months or longer have a disease that is biologically different and more sensitive to treatment than patients who experience shorter durations of response. To better understand these biological differences, additional analyses, such as serial biopsies at progression, are needed; however, serial biopsies are difficult to do in patients with lung cancer. Herein, we focused on the effect of response on long-term survival. This relationship in patients with advanced NSCLC has been evaluated in large-scale, multi-trial analyses of patient populations receiving chemotherapy or targeted therapy (or both), and correlations between response and progression-free survival and between response and overall survival were seen at the patient level. 17,25,26 However, similar analyses have not been done with immunotherapy trials in patients with NSCLC. Given the long duration of responses to PD-1 or PD-L1 inhibitors in patients with previously advanced NSCLC, the association between response and overall survival could be expected to be particularly strong in this setting. However, few studies have explored this question, and the corresponding analyses used were restricted, given that they were based on shorter follow-up durations and smaller study populations than our present analyses. 2-year data from the phase 3 OAK study 21 in patients with previously treated NSCLC showed that response to atezolizumab or docetaxel was associated with increased overall survival, and the benefit of response to atezolizumab was more pronounced; however, no landmark analyses were done and overall survival post-response and post-progression were not evaluated. 21 Landmark analyses were previously used in two of the four studies included in our analyses (CheckMate 063 and 057) to examine the relationship between response and overall survival. A preliminary exploratory landmark analysis 27 from CheckMate 063 based on less than 18 months of follow-up showed that response to nivolumab at 3·5 months (the median time to response) was associated with increased overall survival versus stable disease or progressive disease. In CheckMate 057, multivariable Cox regression analyses of overall survival adjusted for age, sex, ECOG performance status, and time from diagnosis to randomisation showed a stronger association of response at 6 months to subsequent overall survival with nivolumab (HR for death between responders and non-responders 0·24 [95% CI 0·15–0·39]) versus docetaxel (0·55 [0·34–0·88]) confirming comparable estimates obtained from unadjusted univariable analysis. Additionally, sensitivity analyses of 4-month and 8-month landmarks yielded qualitatively similar results. 15 An alternative approach to landmark analysis, which also avoids immortal time bias, is an extended multivariate Cox regression analysis with response status as a time-varying covariate. 14,28 Using this approach, Santi and colleagues 15 showed that response with nivolumab was associated with a 57% reduction in the risk of death versus response to docetaxel (p<0·01). 15 These initial analyses, which were based on 2 years of follow-up, showed the appropriateness of landmark analyses for the evaluation of response-survival relationship. 15 To minimise heterogeneity, the pooled individual patient data analysis of the four studies only included patients treated for advanced NSCLC with nivolumab who had previously received at least one line of systemic chemotherapy. The most likely sources of heterogeneity potentially affecting outcomes in individual patients are tumour histology, which differed between studies, and PD-L1 expression. To document possible effects of these heterogeneities, overall survival and progression-free survival outcomes for PD-L1 and histology subgroups are provided, as well as outcomes in the individual trials. However, treatment comparisons between nivolumab and docetaxel were not at risk of heterogeneity, because both studies included in this analysis were identically designed randomised trials, except that CheckMate 017 was done in patients with squamous NSCLC and CheckMate 057 in patients with non-squamous NSCLC. Furthermore, the baseline characteristics show that the treatment groups were well matched for PD-L1 distribution in the pooled population. Furthermore, overall survival outcomes with nivolumab were virtually identical in the two randomised studies and across all four studies. Taken together, these data suggest both analyses are representative of a previously treated population with advanced NSCLC, with some expected variability in outcomes among subgroups. Our analyses have advantages and limitations. Landmark analyses have the advantage that comparisons across response status for a particular treatment are based on information collected before a fixed timepoint and thus avoid immortal time bias in the estimation of subsequent overall survival. 14 It is not a primary goal of a landmark analysis to determine a treatment difference from randomisation; for comparisons in randomised studies, landmark categories might not provide a consistent view of overall survival between treatment groups if overall survival at the landmark is different across groups. Nonetheless, overall survival estimates for all randomised patients in CheckMate 017 and 057 during the first 6 months were virtually identical in the two treatment groups. Although overall survival analyses can be potentially confounded by subsequent therapy, the use of subsequent immunotherapy in the nivolumab group of the CheckMate 017 and 057 population was low (4% of patients). A limitation of the overall survival post-progression analysis by best overall response is the exclusion of patients who remained in response or maintained stable disease at the time of database lock. Because the percentage of patients who fall into this category was higher for nivolumab than for docetaxel, these analyses probably underestimate the overall survival benefit post-progression for nivolumab versus docetaxel. Moreover, more patients died without documented progression in the docetaxel than in the nivolumab group. The exclusion of these patients from the post-progression survival analysis might further contribute to an underestimation of the survival advantage associated with nivolumab. In summary, our analyses provide evidence that response and disease control with nivolumab strongly benefit long-term survival, even after progression. A smaller overall survival benefit was associated with stable disease versus progressive disease at 6 months in the landmark analysis. Although response to docetaxel also favoured longer-term survival, the association between response and survival was less pronounced than with nivolumab. Additional analyses assessing the effect of various factors on long-term survival with immunotherapy versus chemotherapy are planned. Contributors All authors had full access to all of the data in these analyses, analysed and interpreted the data, and drafted or critically revised the manuscript for important intellectual content. AL did the statistical analyses. All authors approved the final version of the manuscript to submit for publication. Declaration of interests SJA is on the advisory board or scientific advisory board of Bristol-Myers Squibb, Novartis, Merck, CBMG, Boehringer Ingelheim, AstraZeneca/MedImmune, Memgen, FLX Bio, Nektar, and Veen; has done contracted research for Novartis; was previously employed by H Lee Moffitt Cancer Center & Research Institute (Tampa, FL, USA); and is currently employed by Duke Cancer Institute (Durham, NC, USA). HB reports grants from Bristol-Myers Squibb, Celgene, Lilly, Merck, and Millennium; personal fees for consultancy or advisory board services (or both) from Bristol-Myers Squibb, Celgene, Genentech, Lilly, EMD-Serono, Merck, Millennium, Pfizer, Boehringer Ingelheim, AstraZeneca, Genmab, Novartis, Regeneron, BioNTech, Cantargia AB, Amgen, and Axiom; has provided educational services for AbbVie and Axiom; and has served on a Data and Safety Monitoring Board for Takeda. SSR reports grants from AstraZeneca, Merck and Tesaro, and personal fees for consultant or advisory board services from Amgen, AbbVie, Bristol-Myers Squibb, Lilly, Genentech, Takeda, and Luxo. LH reports grants from Boehringer Ingelheim and Xcovery; non-financial support from Xcovery; and personal fees from AbbVie, AstraZeneca, Bristol-Myers Squibb, Merck, Genentech, Incyte, EMD Serono, and Tesaro. JDCC reports personal fees from Bristol-Myers Squibb, AstraZeneca, MSD, and Roche. AP reports personal fees from Bristol-Myers Squibb, AstraZeneca, Boehringer Ingelheim, MSD, Pfizer, and Roche, and non-financial support from Bristol-Myers Squibb, Boehringer Ingelheim, and Roche. MG reports personal fees from Bristol-Myers Squibb, AstraZeneca, Boehringer Ingelheim, Eli Lilly, MSD, Novartis, Otsuka Pharma, Celgene, Roche, Pfizer, Incyte, Inivata, and Takeda, and has served as principal investigator or conducted patient enrolment in clinical trials (or both) for Bristol-Myers Squibb, AstraZeneca, Eli Lilly, MSD, Novartis, Otsuka Pharma, Celgene, Roche, Pfizer, Tiziana Sciences, Clovis, Merck Serono, and Bayer. LQMC reports grants from Bristol-Myers Squibb, AstraZeneca/MedImmune, Genentech, Novartis, Merck, Lilly/ImClone, Incyte, Pfizer, VentiRx, Seattle Genetics and Dynavax; and personal fees from Bristol-Myers Squibb, Amgen, AstraZeneca/MedImmune, Genentech, Novartis, Merck, Pfizer, Sanofi-Genzyme, Seattle Genetics, Synthorx, Takeda, and Dynavax. SG served as a consultant or advisor (or both) to Bristol-Myers Squibb and NEKTAR; reports research funding from Bristol-Myers Squibb; and reports financial support to his institution from Iovance, Takeda/ARIAD and Genentech for clinical trials. DP reports honoraria from Bristol-Myers Squibb, AstraZeneca, Boehringer, Celgene, Novartis, Pfizer, MSD, Roche, prIME Oncology, and peer CME for services as a consultant, advisor, or lecturer; travel expenses from AstraZeneca, Roche, Novartis, prIME Oncology, and Pfizer, and financial support to his institution from AstraZeneca, Bristol-Myers Squibb, Boehringer Ingelheim, Eli Lilly, Merck, Novartis, Pfizer, Roche, Medimmune, Sanofi-Aventis, Taiho Pharma, Novocure, and Daiichi Sankyo for clinical trials. CB reports honoraria from Bristol-Myers Squibb, AstraZeneca, Boehringer Ingelheim, Merck, and Pfizer. AD reports honoraria from Medscape, Onclive, PeerVoice, Physicians Education Resources, Tyra Biosciences, Targeted Oncology, MORE Health, Research to Practice, Foundation Medicine, Peerview, AstraZeneca, Genentech and Bayer, and has served as a consultant or advisor to Ignyta, Loxo, TP Therapeutics, AstraZeneca, Pfizer, BluePrint Medicines, Genentech, Takeda, Helsinn Therapeutics, BeiGene, Hengrui Therapeutics, Exelixis, and Bayer. GAO reports grants from Bristol-Myers Squibb, AstraZeneca, Merck, Genentech, and Pfizer, and consultant fees for his institution from Merck, Novartis, Pfizer and Takeda. SA and JRP are employed by and own stocks in Bristol-Myers Squibb. AL is employed by Bristol-Myers Squibb. JB reports grants and travel expenses from Bristol-Myers Squibb and personal fees from AstraZeneca, Genentech, and Merck. MAB, LC, and JW-T declare no competing interests. Acknowledgments This analysis was supported by Bristol-Myers Squibb. Medical writing support was provided by Roland Tacke of Evidence Scientific Solutions, with funding from Bristol-Myers Squibb. We thank the patients and families who made these trials possible; the study investigators ( appendix pp 2–3 ) and clinical study teams who participated in the four clinical studies; the protocol manager, Jamie Yingst, for the studies; Dako, an Agilent Technologies company, for collaborative development of the PD-L1 IHC 28-8 pharmDx assay; and Bristol-Myers Squibb and ONO Pharmaceutical Company. Publisher Copyright: © 2019 Elsevier Ltd
PY - 2019/10
Y1 - 2019/10
N2 - Background: Phase 3 clinical data has shown higher proportions of patients with objective response, longer response duration, and longer overall survival with nivolumab versus docetaxel in patients with previously treated advanced non-small-cell lung cancer (NSCLC). We aimed to evaluate the long-term benefit of nivolumab and the effect of response and disease control on subsequent survival. Methods: We pooled data from four clinical studies of nivolumab in patients with previously treated NSCLC (CheckMate 017, 057, 063, and 003) to evaluate survival outcomes. Trials of nivolumab in the second-line or later setting with at least 4 years follow-up were included. Comparisons of nivolumab versus docetaxel included all randomised patients from the phase 3 CheckMate 017 and 057 studies. We did landmark analyses by response status at 6 months to determine post-landmark survival outcomes. We excluded patients who did not have a radiographic tumour assessment at 6 months. Safety analyses included all patients who received at least one dose of nivolumab. Findings: Across all four studies, 4-year overall survival with nivolumab was 14% (95% CI 11–17) for all patients (n=664), 19% (15–24) for those with at least 1% PD-L1 expression, and 11% (7–16) for those with less than 1% PD-L1 expression. In CheckMate 017 and 057, 4-year overall survival was 14% (95% CI 11–18) in patients treated with nivolumab, compared with 5% (3–7) in patients treated with docetaxel. Survival subsequent to response at 6 months on nivolumab or docetaxel was longer than after progressive disease at 6 months, with hazard ratios for overall survival of 0·18 (95% 0·12–0·27) for nivolumab and 0·43 (0·29–0·65) for docetaxel; for stable disease versus progressive disease, hazard ratios were 0·52 (0·37–0·71) for nivolumab and 0·80 (0·61–1·04) for docetaxel. Long-term data did not show any new safety signals. Interpretation: Patients with advanced NSCLC treated with nivolumab achieved a greater duration of response compared with patients treated with docetaxel, which was associated with a long-term survival advantage. Funding: Bristol-Myers Squibb.
AB - Background: Phase 3 clinical data has shown higher proportions of patients with objective response, longer response duration, and longer overall survival with nivolumab versus docetaxel in patients with previously treated advanced non-small-cell lung cancer (NSCLC). We aimed to evaluate the long-term benefit of nivolumab and the effect of response and disease control on subsequent survival. Methods: We pooled data from four clinical studies of nivolumab in patients with previously treated NSCLC (CheckMate 017, 057, 063, and 003) to evaluate survival outcomes. Trials of nivolumab in the second-line or later setting with at least 4 years follow-up were included. Comparisons of nivolumab versus docetaxel included all randomised patients from the phase 3 CheckMate 017 and 057 studies. We did landmark analyses by response status at 6 months to determine post-landmark survival outcomes. We excluded patients who did not have a radiographic tumour assessment at 6 months. Safety analyses included all patients who received at least one dose of nivolumab. Findings: Across all four studies, 4-year overall survival with nivolumab was 14% (95% CI 11–17) for all patients (n=664), 19% (15–24) for those with at least 1% PD-L1 expression, and 11% (7–16) for those with less than 1% PD-L1 expression. In CheckMate 017 and 057, 4-year overall survival was 14% (95% CI 11–18) in patients treated with nivolumab, compared with 5% (3–7) in patients treated with docetaxel. Survival subsequent to response at 6 months on nivolumab or docetaxel was longer than after progressive disease at 6 months, with hazard ratios for overall survival of 0·18 (95% 0·12–0·27) for nivolumab and 0·43 (0·29–0·65) for docetaxel; for stable disease versus progressive disease, hazard ratios were 0·52 (0·37–0·71) for nivolumab and 0·80 (0·61–1·04) for docetaxel. Long-term data did not show any new safety signals. Interpretation: Patients with advanced NSCLC treated with nivolumab achieved a greater duration of response compared with patients treated with docetaxel, which was associated with a long-term survival advantage. Funding: Bristol-Myers Squibb.
UR - http://www.scopus.com/inward/record.url?scp=85072654087&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85072654087&partnerID=8YFLogxK
U2 - 10.1016/S1470-2045(19)30407-3
DO - 10.1016/S1470-2045(19)30407-3
M3 - Article
C2 - 31422028
AN - SCOPUS:85072654087
VL - 20
SP - 1395
EP - 1408
JO - The Lancet Oncology
JF - The Lancet Oncology
SN - 1470-2045
IS - 10
ER -