FOCUS4-N and Maintenance Capecitabine

Why FOCUS4-N Asked a Maintenance-Treatment Question

First-line chemotherapy can control metastatic colorectal cancer, but continuing intensive treatment indefinitely may be difficult for patients. One alternative is to pause treatment and monitor for progression; another is to give a less intensive maintenance treatment. FOCUS4-N directly compared these strategies in a defined group of patients who had already completed an initial period of chemotherapy.

The trial was part of the FOCUS4 platform, which screened patients with metastatic colorectal cancer and allocated eligible patients to randomized treatment comparisons. FOCUS4-N tested maintenance capecitabine against active monitoring after 16 weeks of first-line chemotherapy, among patients whose disease had responded or remained stable. Its findings therefore address a practical decision at that point in treatment—not the effect of capecitabine for every person starting treatment for metastatic disease.

What Was Compared and What Was Measured?

Patients were randomized after the induction period, once their disease was assessed as responding or stable. In the capecitabine group, patients received maintenance capecitabine; in the comparison group, patients underwent active monitoring. Randomization supports a fair comparison of these strategies among the eligible patients who reached that decision point. The treatment contrast is not capecitabine versus no treatment throughout the entire course of illness: both groups had received first-line chemotherapy before randomization.

The reported FOCUS4-N results favored capecitabine for PFS. Median PFS was 3.9 months in the capecitabine group and 1.9 months in the active-monitoring group. The hazard ratio was 0.40 (95% CI 0.21 to 0.75; p=0.002). This is evidence that capecitabine delayed progression or death in the trial’s randomized population. It does not mean that every patient gained two months, or that the PFS difference was the same for every patient.

The OS result told a different part of the story. Median OS was 14.8 months with capecitabine and 15.2 months with active monitoring; the hazard ratio was 0.93 (95% CI 0.69 to 1.27; p=0.66). The confidence interval includes 1, and the result does not provide clear evidence of an OS difference. It also does not establish that the treatments have identical effects on OS: the interval is compatible with a range of effects.

Reading the Hazard Ratio Carefully

A hazard is the instantaneous event rate among patients who have not yet experienced the event, conditional on having remained event-free up to that time. A hazard ratio compares these rates between groups over follow-up. Under a proportional-hazards interpretation, a hazard ratio of 0.40 means the capecitabine group had an estimated hazard of progression or death 0.40 times the active-monitoring hazard, or an estimated 60% lower hazard. It is not a 60% reduction in the probability of progression, nor does it mean that PFS duration increased by 60%.

The usual Cox-model interpretation assumes that the hazard ratio is reasonably constant over time. The trial’s randomization also supports comparable groups at baseline, while time-to-event analysis requires appropriate handling of censoring. In particular, patients lost to follow-up or still event-free at the analysis cutoff should not be treated as if their event time were known; valid inference relies on censoring not being systematically informative in a way the analysis fails to address. A Kaplan–Meier curve can help reveal whether group separation is delayed, temporary, or sustained. A single hazard ratio can be less informative if curves cross or the relative effect changes substantially over time.

Why PFS and OS Can Differ

PFS records progression or death after randomization, so it can reflect the effect of treatment during the maintenance period. OS includes everything that happens after randomization, including treatment received after progression. A strategy that delays progression need not produce a detectable OS difference, particularly when later treatment, subsequent clinical events, or limited statistical precision affect the OS comparison. FOCUS4-N’s PFS finding should therefore be reported as a PFS finding, not converted into a claim that patients lived longer.

The median is another useful but limited summary. A median PFS of 3.9 months means that the estimated PFS curve reached 0.50 at about 3.9 months; it does not describe the full curve. The two medians differ by 2.0 months, but that subtraction is not a hazard ratio and does not show that each patient’s time to progression was extended by two months. The hazard ratio uses event-time information across follow-up and depends on its modeling assumptions.

Benefit, Burden, and the Treatment Decision

A maintenance strategy should be assessed using both disease control and treatment burden. Capecitabine can cause adverse effects, and FOCUS4-N reported more grade 3 or worse adverse events in the capecitabine group than in the active-monitoring group. Hand–foot syndrome is a recognized capecitabine toxicity. Thus, the PFS improvement is not a complete summary of the patient experience: toxicity, monitoring, treatment preferences, and the value placed on time without progression all matter.

Assumptions, Diagnostics, and Alternatives

Several checks help keep the interpretation aligned with what the trial can support. Examine whether the PFS curves show the pattern suggested by a single hazard ratio; consider whether censoring and follow-up are adequately described; and check whether outcome definitions and analysis populations are clear. Also distinguish a randomized treatment effect among patients eligible at the post-induction decision point from an effect among all patients at diagnosis. Patients who progressed or could not complete induction were not part of this specific randomized comparison.

The principal alternative in FOCUS4-N was active monitoring rather than another active maintenance regimen. This makes the comparison clinically direct for the question of whether to give capecitabine at that stage, but it does not rank capecitabine against every possible maintenance approach. Nor does the result alone determine the best choice for an individual patient; expected benefit, adverse effects, patient priorities, and the available treatment plan are relevant.

A common mistake is to say, “The PFS median was two months longer, so capecitabine gave each patient two extra months and improved survival.” The correction is that the 2.0-month figure is a difference between group medians, not an individual treatment effect. The randomized PFS analysis supports delayed progression or death on average in the study population; the OS result did not show clear evidence of longer survival. Another mistake is to interpret the nonsignificant OS result as proof of no effect. Its confidence interval includes both potentially lower and higher hazards, so the conclusion is uncertainty, not equivalence.