Oops, you're using an old version of your browser so some of the features on this page may not be displaying properly.

MINIMAL Requirements: Google Chrome 24+Mozilla Firefox 20+Internet Explorer 11Opera 15–18Apple Safari 7SeaMonkey 2.15-2.23

Poster display session: Biomarkers, Gynaecological cancers, Haematological malignancies, Immunotherapy of cancer, New diagnostic tools, NSCLC - early stage, locally advanced & metastatic, SCLC, Thoracic malignancies, Translational research

2680 - Fractional polynomial network meta-analysis: a different approach to indirectly assess the comparative efficacy of 2L+ cancer immunotherapy (CIT) treatments for metastatic NSCLC


20 Oct 2018


Poster display session: Biomarkers, Gynaecological cancers, Haematological malignancies, Immunotherapy of cancer, New diagnostic tools, NSCLC - early stage, locally advanced & metastatic, SCLC, Thoracic malignancies, Translational research



Tumour Site


Christian Schulz


Annals of Oncology (2018) 29 (suppl_8): viii493-viii547. 10.1093/annonc/mdy292


C. Schulz1, P. Chu2, C.G. Berardo3, B. Karthuria4, J. Foo5, C. Morel5, C. Watkins6, M. Ballinger7, D.R. Gandara8

Author affiliations

  • 1 Department Of Internal Medicine Ii, University of Regensburg, 93042 - Regensburg/DE
  • 2 Health Technology Assessment Group, F. Hoffman-La Roche Ltd., 4070 - Basel/CH
  • 3 Global Product Strategy, F. Hoffman-La Roche Ltd., 4070 - Basel/CH
  • 4 Product Development Medical Affairs, F. Hoffman-La Roche Ltd., 4070 - Basel/CH
  • 5 Medical Oncology, Mapi Group, 3995 AX - Houten/NL
  • 6 Statistics, Clarostat Consulting, Cheshire/GB
  • 7 Product Development - Oncology, Genentech Inc. - Roche - USA, 94080 - South San Francisco/US
  • 8 Internal Med: Hematology-oncology, University of California Davis Cancer Center, 95817 - Sacramento/US


Login to access the resources on OncologyPRO.

If you do not have an ESMO account, please create one for free.

Abstract 2680


CIT is associated with delayed onset of clinical effect and long-term (LT) survival, making traditional proportional hazards models less applicable. This study used a novel method, Bayesian fractional polynomial (FP) network meta-analysis (NMA) of time-to-event data, to address challenges in evaluating overall survival (OS) and switching with CIT.


A systematic literature review identified randomized, controlled, Phase II-IV trials of US- or EU-approved 2L+ treatments for advanced/metastatic NSCLC that enrolled ≥ 10 adult patients (pts) per treatment arm. An FP-NMA was used to estimate time-to-event outcomes: OS (5-y horizon) and progression-free survival (PFS; 2.5-y horizon). Expected survival time reflected the area under the curve over the time horizon. Adjusted analyses accounted for treatment switching. Expected OS times were ranked by median ranking with 95% credible intervals and by surface under the cumulative ranking curve (SUCRA).


Of 25,115 screened publications, 35 studies of chemotherapy (chemo), PD-(L)1 inhibitors, targeted and other non-chemo agents, and placebo were included in the FP-NMA. In this model, all CIT PD-(L)1 inhibitors (nivolumab [nivo], atezolizumab [atezo] and pembrolizumab [pembro] in rank order) had similar estimated 5-year OS and ranked above all other treatments (Table). When adjusted for switching, all 3 CITs remained the highest-ranking treatments, with atezo ranked highest, followed by nivo and pembro. PFS and PD-(L)1 inhibitors’ population subgroup results will be shown.


This FP-NMA showed that PD-(L)1 inhibitors had the highest expected 5-year OS amongst 2L treatments and that factors such as switching affected OS benefit. These results, which vary from prior models, suggest that FP-NMA may be a more relevant and viable method for assessing LT clinical benefit in pts treated with CIT.

Clinical trial identification

Legal entity responsible for the study

F. Hoffmann-La Roche AG.


F. Hoffmann-La Roche AG.

Editorial Acknowledgement

Jeff Frimpter, Health Interactions.


C. Schulz: Advisory Role/Expert Testimony: AstraZeneca, Boehringer Ingelheim, Celgene, Lilly, MSD, Novartis, Roche; Honoraria: AstraZeneca, Boehringer Ingelheim, Celgene, Lilly, Novartis, Roche; Financing of Scientific Research: AstraZeneca, BMS, CellAct, Lilly, Novartis, Roche, Pfizer. P. Chu, C.G. Berardo, B. Karthuria: Employee: F. Hoffman-La Roche Ltd. J. Foo, C. Morel: Employee: Mapi Group (an ICON plc company), which received financial support from Roche. C. Watkins: Employee: Clarostat Consulting, which received financial support by Roche. M. Ballinger: Employee: Genentech Inc., a member of the Roche Group. D.R. Gandara: Grants and personal fees: Genentech Inc., a member of the Roche Group.Table: 1470P

Ranking outcomes of expected 5-year overall survival (intention-to-treat primary population of 850 pts from the OAK study; extended network including pembrolizumab)

Unadjusted AnalysisSwitching-Adjusted Analysis
Median Rank (95% CrI)SUCRAMedian Rank (95% CrI)SUCRA
Nivolumab 3 mg/kg2 (1, 9)0.881Atezolizumab 1200 mg2 (1, 8)0.903
Atezolizumab 1200 mg3 (1, 10)0.865Nivolumab 3 mg/kg3 (1, 9)0.872
Pembrolizumab 2 mg/kg3 (1, 13)0.829Pembrolizumab 2 mg/kg3 (1, 14)0.818
Docetaxel 40 mg/m2 qw7 (1, 18)0.592Ramucirumab + docetaxel 60 mg/m27 (1, 18)0.594
Ramucirumab + docetaxel 60 mg/m27 (1, 18)0.589Docetaxel 40 mg/m2 qw7 (1, 18)0.575
Erlotinib 150 mg8 (3, 14)0.570Erlotinib 150 mg8 (3, 14)0.564
Ramucirumab + docetaxel 75 mg/m28 (2, 17)0.547Ramucirumab + docetaxel 75 mg/m28 (2, 17)0.55
Erlotinib 300 mg8 (1, 18)0.533Erlotinib 300 mg8 (1, 18)0.536
Pemetrexed or docetaxel9 (2, 17)0.519Pemetrexed or docetaxel9 (2, 17)0.518
Docetaxel q3w pooled9 (4, 15)0.514Nintedanib + docetaxel 75 mg/m29 (2, 17)0.513
Nintedanib + docetaxel 75 mg/m29 (2, 17)0.513Docetaxel q3w pooled9 (4, 16)0.508
Paclitaxel poliglumex9 (2, 18)0.489Paclitaxel poliglumex10 (2, 18)0.477
Docetaxel 75 mg/m211 (7, 15)0.412Docetaxel 75 mg/m211 (7, 15)0.408
Docetaxel qw pooled12 (5, 17)0.379Docetaxel qw pooled12 (6, 17)0.369
Pemetrexed 900 mg/m215 (4, 18)0.259Pemetrexed 900 mg/m214 (4, 18)0.288
Pemetrexed 500 mg/m215 (8, 17)0.239Pemetrexed 500 mg/m214 (8, 17)0.256
Placebo16 (7, 18)0.195Placebo16 (7, 18)0.178
Pemetrexed 1000 mg/m218 (9, 18)0.075Pemetrexed 1000 mg/m218 (9, 18)0.075

CrI, Credible interval; SUCRA, surface under the cumulative ranking curve; q3w, every 3 weeks; qw, once a week.

This site uses cookies. Some of these cookies are essential, while others help us improve your experience by providing insights into how the site is being used.

For more detailed information on the cookies we use, please check our Privacy Policy.

Customise settings
  • Necessary cookies enable core functionality. The website cannot function properly without these cookies, and you can only disable them by changing your browser preferences.