Comparative efficacy of later-line therapies for metastatic colorectal cancer: a network meta-analysis of survival curves

ABSTRACT

Introduction

We evaluated the comparative efficacy of six later-line (≥3) therapies for metastatic colorectal cancer (mCRC) over placebo. We applied a novel statistical method of reconstructing pseudo-patient-level data (pseudo-IPD) to inform a network meta-analysis of survival curves that considers shape in addition to scale parameters.

Methods

A literature search yielded 10 phase II/III trials. We digitized all survival curves and applied a novel method incorporating curve coordinates, patients-at-risk, and events reported to generate pseudo-IPD. Using fitted random effects lognormal distributions, we estimated the survival proportions and HRs (95CrI) of progression-free (PFS) and overall survival (OS) over 12 months of follow-up.

Results

Compared to placebo, in ascending order, 12-month OS HRs were 0.50 (95% CrI = 0.35, 0.69; PFS = 0.11 (95% CrI = 0.06, 0.14)) for TAS+bevacizumab; 0.71 (95% CrI = 0.51, 0.97; PFS = 0.26 (95% CrI = 0.16, 0.41)) for regorafenib; 0.75 (95% CrI = 0.61, 0.91; (PFS = 0.24 (95% CrI = 0.17, 0.31)) for TAS-102; 0.80 (95% CrI = 0.79, 0.90; PFS = 0.18 (95% CrI = 0.13, 0.24)) for fruquintinib; 0.83 (95% CrI = 0.50, 0.99; PFS = 0.42 (95% CrI = 0.20, 0.75)) for atezolizumab+cobimetinib; and 1.03 (95% CrI = 0.55, 1.65; PFS = 0.67 (95% CrI = 0.29, 1.01)) for atezolizumab.

Conclusion

In this independent NMA of survival data, all later-line mCRC therapies but atezolizumab monotherapy exhibited superiority in 12-month PFS and OS over placebo. TAS+bevacizumab emerged as the most dominant option and may be the preferred choice, with fruquintinib, regorafenib, and TAS-102 monotherapy showing statistically significant but lower PFS and OS benefits.

Registration

PROSPERO: CRD42022371953

KEYWORDS:

• Colorectal cancer
• network meta-analysis
• comparative efficacy
• later-line therapy
• third-line treatment

Declaration of interest

The authors have no relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript. This includes employment, consultancies, honoraria, stock ownership or options, expert testimony, grants or patents received or pending, or royalties.

Author contributions

M Obeng-Kusi: Conceptualization, Methodology, Software, Writing – Original Draft J Martin: Method, Writing – Original Draft, Review & Editing D Roe: Conceptualization, Supervision, Writing – Review and Editing B Erstad: Conceptualization, Supervision, Writing – Review and Editing I Abraham: Conceptualization, Methodology, Supervision, Writing – Review and Editing

Data availability statement

The data that support the findings of this study are available on request from the corresponding author.

Supplementary material

Supplemental data for this article can be accessed online at https://doi.org/10.1080/14737167.2024.2365993

Reviewer disclosures

Peer reviewers on this manuscript have no relevant financial or other relationships to disclose.

Additional information

Funding

This paper was not funded.