Journal Publication

Vactosertib-pembrolizumab combination shows activity in liver-metastasis-free colorectal cancer

A phase Ib/IIa study found a 12.6% response rate in non-MSI-high metastatic colorectal cancer, rising to 22.5% in patients without liver metastases.

A multicenter phase Ib/IIa study of vactosertib plus pembrolizumab reported outcomes in non-MSI-high metastatic colorectal and gastric cancer.

By AppliedXLJuly 21, 2026Trial NCT03724851

Executive Summary

A phase Ib/IIa study combining an oral TGF-beta receptor kinase inhibitor with an anti-PD-1 antibody found a response signal in metastatic colorectal cancer that depended heavily on whether patients had liver metastases.
The data narrow where this combination shows activity: colorectal cancer patients without liver involvement had responses several times more often than those with liver metastases, while gastric cancer patients did not respond at all.
Adding an immune checkpoint inhibitor to a TGF-beta blocker in a checkpoint-inhibitor-naive, heavily pretreated population came with adverse events in most patients, chiefly skin-related, alongside the efficacy signal.
The result lands in a target-indication pairing with a documented history of terminations, and a positive, reproducible subgroup signal stands out against that record even though the overall study was open-label and uncontrolled.

The result

The study, registered as NCT03724851, evaluated vactosertib, an oral small-molecule inhibitor of the TGF-beta receptor kinase TGFBR1, combined with pembrolizumab, an anti-PD-1 antibody, in patients with non-MSI-high metastatic colorectal cancer or metastatic gastric cancer who had exhausted standard therapies. Among 103 colorectal cancer patients evaluable for efficacy, the objective response rate by RECIST v1.1 was 12.6%. Patients without liver metastases responded at 22.5%, versus 6.3% for those with liver metastases. The median overall survival across the colorectal cohort was 13.2 months. No objective responses were observed among the 12 gastric cancer patients evaluated. NCT03724851+1

How it was done

The trial was an open-label, single-arm, multicenter study conducted in South Korea, run in dose-escalation and dose-expansion phases across three arms testing vactosertib at 200 mg once daily, 200 mg twice daily, and 300 mg twice daily, each given five days per week alongside pembrolizumab 200 mg intravenously every three weeks. Enrollment reached 120 patients, with 108 colorectal and 12 gastric cancer patients included in the safety analysis and 115 patients (103 colorectal, 12 gastric) in the efficacy analysis. All patients were checkpoint-inhibitor-naive prior to enrollment. The primary endpoints were maximum tolerated dose and safety and tolerability; objective response rate, progression-free survival, and overall survival were prespecified secondary endpoints. NCT03724851

The safety picture

Treatment-related adverse events occurred in 70.8% of the 120 patients treated, with pruritus and rash the most frequently reported events. The published abstract characterized the combination's safety profile as manageable, consistent with the tolerability focus built into the primary endpoint. The trial did not report a control arm against which to weigh the frequency of these events, since the design used pembrolizumab plus vactosertib in every treated patient without a comparator group. Vactosertib+1

Where this sits

TGFBR1-targeted combinations remain sparsely tested specifically in colorectal cancer: across TGFBR1 x colorectal cancer trials, three companies have recorded failures in this pairing, and the phase 1 subset produced one completion against three terminations. The closest direct comparator drug by shared target linkage is Merck's pembrolizumab, which is approved and marketed across multiple tumor types including in combination regimens, and which anchors most of the direct comparator set tied to this program's PD-1 component. Within the broader TGF-beta mechanism class, ontunisertib (Agomab Therapeutics) and LY3200882 (Eli Lilly) are the closest same-target peers in early-phase testing, both outside colorectal cancer specifically. NCT03724851

This analysis was produced using AI-assisted reporting systems, AppliedXL data, and official public records. These systems undergo editorial review, quality checks, and regular audits by human experts. Errors may still occur, as with any automated system. Always consult the linked primary sources. Read our AI Editorial Policy.