Letter to the editor:
The search of predictive biomarkers for immune checkpoint inhibitor (ICIs) therapy remains a topic of major interest in oncology research. Previous studies have suggested that the development of immune-related adverse events (irAEs) can be associated with efficacy of ICIs in different cancer types[
- Kawai T.
- Sato Y.
- Makino K.
- Yamada Y.
- Nomiya A.
- Nakamura M.
- et al.
Immune-related adverse events predict the therapeutic efficacy of pembrolizumab in urothelial cancer patients.
- Rogado J.
- Sánchez-Torres J.M.
- Romero-Laorden N.
- Ballesteros A.I.
- Pacheco-Barcia V.
- Ramos-Leví A.
- et al.
Immune-related adverse events predict the therapeutic efficacy of anti–PD-1 antibodies in cancer patients.
- Fujii T.
- Colen R.R.
- Bilen M.A.
- Hess K.R.
- Hajjar J.
- Suarez-Almazor M.E.
- et al.
Incidence of immune-related adverse events and its association with treatment outcomes: the MD Anderson Cancer Center experience.
]. However, statistical caveats such as the immortal-time bias or the publication bias may have overestimated this association[
Immortal time bias in pharmacoepidemiology.
]. Recently, Kfoury et al
have analysed 577 patients treated with anti-PD-(L)1 antibodies at Gustave Roussy Institute to evaluate the association of irAEs and efficacy outcomes after correctly accounting for immortal-time bias[
- Kfoury M.
- Najean M.
- Lappara A.
- Voisin A.-L.
- Champiat S.
- Michot J.-M.
- et al.
Analysis of the association between prospectively collected immune-related adverse events and survival in patients with solid tumor treated with immune-checkpoint blockers, taking into account immortal-time bias.
]. Authors concluded that development of irAEs was associated with improved overall survival (OS) using a time-dependent Cox model (hazard ratio: 0.56, 95% confidence interval (CI) 0.41-0.75; p=0.0001) but not using a 12-week landmark analysis (p=0.26). Similar results were obtained in terms of progression-free survival (PFS).
Aiming to validate those findings, we have reviewed information from 403 patients treated with anti-PD-(L)1 antibodies in clinical trials or as per standard practice at Vall D’Hebron Institute of Oncology (VHIO) from 2015 to 2021. Different tumour types were included in this analysis (29.8% head and neck, 20.8% melanoma, 19.1% colorectal, 10% lung, and 20.3% others). To estimate the association between irAEs grade
2 and survival endpoints (PFS and OS), three different methods were used: i) unadjusted analysis using the standard Cox model, ii) time-dependent Cox model and iii) landmark analysis at 4 and 12-weeks[
Landmark analysis at the 25-year landmark point.
]. Survival endpoints were calculated since the first dose of ICIs treatment, and all analyses were stratified by tumor type.
Among the 403 patients analyzed, 95 (23.6%) developed irAEs grade
2 with a median time to onset of 2.3 months (95%CI 1.2 -3.8) in a median follow-up period of 27.1 months. Efficacy results are summarized in Table 1
. In terms of OS, the time-dependent Cox model analysis showed an association between irAEs grade
2 and OS (HR: 0.63; 95%CI 0.44-0.92). The landmark analysis was strongly influenced by the selected time-point, at 4-weeks the estimation showed a trend towards a protective effect in patients developing irAEs grade
2 (HR: 0.68; 95%CI 0.36-1.31), but not at 12 weeks (HR=1.06). On the other hand, the unadjusted analysis overestimated the risk reduction in patients with irAEs grade
2 (HR: 0.42; 95%CI 0.29-0.60). For the PFS endpoint, the only analysis that showed a significant association was the unadjusted analysis. An inconclusive association was found in the time-dependent Cox model analysis (HR: 0.85; 95%CI 0.61-1.16) and, again, heterogenous results were obtained using different time-points in the landmark analysis.
Table 1Hazard ratio with 95%CI of the association between immune-related adverse events grade 2 and survival endpoints.
In conclusion, our data validates the results presented by Kfoury et al providing more evidence of the association between irAEs and OS outcomes. Results in the PFS endpoint could not be validated. In addition, our results confirm the impact of different statistical methods on final conclusions. Unadjusted analyses do not properly address immortal-time bias and previous studies that used this standard methodology have overestimated the magnitude of the association between irAEs and efficacy outcomes. The landmark analysis emerged as an alternative to deal with this bias, although, two important drawbacks have to be considered: i) there is significant reduction in the statistical power and ii) the selection of the time-point can strongly influence the results. The time-dependent Cox model analysis is the recommended statistical approach to evaluate the association between irAEs and efficacy outcomes.
Confict of interest statement
G.V. has received research honoraria for speakers bureau from MSD, GSK and Pierre Fabrer and advisory role from Astrazeneca. A.H.C has received financial interest (Travel, Accommodations, Expenses) from Merk Serono and Kyowa Kirin International. R.B, O.S and D.M none. O.M nothing has received research honoraria from Roche, Kyowa Kirin and Ferrer and travel grant from Kyowa Kirin. I.B has received funding as principal investigator from Astrazeneca, Bicycle therapeutics, Boehringer Ingellheim, Bristol Myers Squibb, Celgene, Dragonfly, GlaxoSmithkline, Gliknik, Immutep, ISA Pharmaceuticals, Janssen Oncology, Kura, Merck Serono, Nanobiotics, Novartis, Northern Biologics, Orion Pharma, Odonate Therapeutics, Regeneron, Pfizer, Sanofi, Pharmamar, Seattle Genetics, Shattuck Labs and VCN Biosciences. Consulting fees fro, Achilles Therapeutics, Bristol Myers Squibb, Cancer Expert Now, eTheRNA immunotherapies, Merck Serono, Merck Sharp & Dohme (MSD), Rakuten pharma, Boehringer Ingellheim and PCI biotech. Payment or honoraria from Bristol Myers Squibb, Merck Serono, Merck Sharp & Dohme (MSD). Support for attending meetings and/or travel from Merck Sharp &Dohme (MSD) and Merck Serono. Leadership or fiduciary role in other board, society, committee or advocacy group, paid or unpaid: ESMO Head and Neck track, EORTC Head and Neck group and Cancer Core Europe Clinical Taskforce. Other financial or non-financial interests: Bristol Myers Squibb educational grant. E.M.C has received advisory board from Amgen, Bristol-Myers Squibb, Merck Sharp & Dohme, Novartis, Pierre Fabre, Roche and Sanofi. Honoraria from Amgen, Bristol-Myers Squibb, Merck Sharp & Dohme, Novartis, Pierre Fabre and Roche. Clinical trial participation (principal investigator) from Amgen, Bristol-Myers Squibb, GlaxoSmithKline, Merck Sharp & Dohme, Novartis, Pierre Fabre, Roche and Sanofi. E.G. has received consulting or advisory role from Roche, Ellipses pharma, Neomed Therapeutics, Janssen, Boehringer Ingelheim, Seattle Genetics, TFS, Alkermes, Thermo Fisher Scientific, Bristol-Myers Squibb, MabDiscovery, Anaveon and F-Star Therapeutics. Speakers Bureau from MSD, Roche, Thermo Fisher Scientific and Lilly. Research funding (recipient institution) from Novartis, Roche, Thermo Fisher Scientific, AstraZeneca/MEdImmune, Taiho Oncology, BeiGene. Other relationship (recipient institution) from Affimed Gmbh, Amgen SA, Anaveon AG, AstraZeneca AB, Biontech Gmbh, Catalym Gmbh, CytomX Therapeutics, F.Hoffmann La Roche Ltd, F-Satar Beta Limited, Genentech Inc, Genmab B.v, Hutchison MEdipharma Limited, Imcheck Therapeutics, Immunoscore Ltd, Janssen-Cilag SA, Medimmune Llc, Merck kgga, Novartis Farmaceutica, S.a, peptomyc, Ribon Therapeutics, Roche Farma SA, Seattle Genetics Inc, Symphogen A/S and Taiho Pharma Usa Inc. R.D. has received advisory role for Roche, Boehringer Ingelheim, received a speaker’s fee from Roche, Boehringer Ingelheim, Ipsen, Amgen, Servier, Sanofi, Libbs, Merck Sharp & Dohme, Lilly and AstraZeneca and research grants from Merck and Pierre Fabre.