TOPLINE

Maintenance therapy was associated with longer overall survival (OS) than active surveillance in patients with metastatic colorectal cancer (mCRC) who achieved disease control after first-line bevacizumab-based combination chemotherapy. However, this association was not maintained after adjustment for baseline prognostic factors, suggesting that baseline differences may have contributed to the observed survival advantage.

METHODOLOGY

• Researchers conducted a single-centre retrospective observational study including 94 patients (median age, 62 years) with mCRC who achieved disease control (defined as complete response [CR], partial response [PR], or stable disease) after first-line bevacizumab-based combination chemotherapy between January 2015 and December 2025.
• Patients were divided into two groups: those who received maintenance therapy (capecitabine, capecitabine plus bevacizumab, or 5-fluorouracil/folinic acid plus bevacizumab; n = 44) and those who underwent active surveillance with clinical evaluation, laboratory investigations, and imaging every 3 months without systemic anticancer treatment until disease progression (n = 50). 
• A landmark analysis approach was employed to minimise immortal time bias, with the landmark time — defined as the date of the first radiologic response evaluation following the completion of first-line treatment — being approximately 2 weeks after the completion of bevacizumab-based combination chemotherapy. 
• The researchers assessed progression-free survival (PFS; defined as the time from the landmark date to disease progression or death from any cause) and OS (defined as the time from the landmark date to death or last follow-up). The median potential follow-up duration was 75.2 months. 

TAKEAWAY

• Median OS was significantly longer in the maintenance therapy group than in the active surveillance group (24.8 vs 18.0 months; log-rank P = .045; unadjusted hazard ratio [HR], 0.60; P = .047), but this association was not significant after multivariable adjustment (adjusted HR, 0.68; P = .220). 
• Median PFS was longer in the maintenance therapy group than in the active surveillance group (11.3 vs 6.7 months), but the difference was not significant (log-rank P = .077; adjusted HR, 0.62; P = .085). 
• In multivariable analysis, male sex (HR, 2.12; P = .019), the presence of comorbidities (HR, 2.27; P = .009), and two or more metastatic sites (HR, 2.49; P = .006) were independently associated with worse OS, whereas primary tumour resection (HR, 0.46; P = .022) and treatment response (CR plus PR; HR, 0.36; P = .001) were independently associated with improved OS. 
• In exploratory subgroup analyses, maintenance therapy was associated with a lower risk for death among patients with two or more metastatic sites (HR, 0.45; log-rank P = .016) and a lower risk for progression among those with RAS-mutant tumours (HR, 0.54; P = .042), those with comorbidities (HR, 0.50; P = .019), and those with two or more metastatic sites (HR, 0.52; P = .041); however, no treatment-by-subgroup interaction reached statistical significance. 

IN PRACTICE

"These real-world data suggest a potential survival advantage associated with maintenance therapy following first-line bevacizumab combination chemotherapy in patients with mCRC who achieve disease control. However, this association was observed only in the unadjusted analyses and was not maintained after multivariable adjustment, suggesting that baseline differences may have contributed to the observed findings," the authors concluded.

SOURCE

The study was led by Simay Çokgezer, Institute of Oncology, Istanbul University, Istanbul, Türkiye. It was published online on August 13, 2026, in the Journal of Clinical Medicine.

LIMITATIONS

The retrospective single-centre design limited causal inference and introduced potential selection bias and residual confounding because treatment was not randomised. The relatively small sample size reduced precision, especially in subgroup analyses, and limited multivariable adjustment. Missing data on BRAF mutation and microsatellite instability status, binary comorbidity assessment, lack of quality-of-life measures, and practice changes over time may have further limited interpretation.

DISCLOSURES

This study did not receive any external funding. The authors declared having no conflicts of interest.

This article was created using several editorial tools, including AI, as part of the process. Human editors reviewed this content before publication.