VET3-TGI yields 71% response rate in refractory cutaneous malignancies

By Tom Whitfield

In the phase 1/1b STEALTH-001 trial, intratumoral VET3-TGI produced a 71% objective response rate in seven patients with advanced cutaneous malignancies refractory to immune checkpoint inhibitors. Blinded independent central review found five responses in that subgroup: one complete response and four partial responses.

Across 12 efficacy-evaluable patients with advanced solid tumors, the ORR was 42%, including one CR and four PRs, and the disease control rate was 75%. Responses appeared in both injected and non-injected lesions, and no dose-limiting toxicities occurred during the study.

Merkel cell carcinoma patients (n=3) achieved one CR and two PRs. Melanoma patients (n=3) achieved two PRs, with one patient’s best response classified as progressive disease; the single patient with cutaneous squamous cell carcinoma also had progressive disease.

At the September 2026 data cutoff, median response duration had not been reached. All responders remained on study treatment, with treatment durations ranging from 10 to 19 months.

The findings were presented at the 2026 Society for Immunotherapy of Cancer Annual Meeting, November 4-8 in Phoenix, as abstract 572.

What makes the response pattern notable?

“…we very consistently observe responses in non-injected lesions,” said Yana G. Najjar, MD, lead study author, in an exclusive interview with OncLive. Najjar, an associate professor of medicine and director of the Clinical and Translational Research Center at UPMC Hillman Cancer Center in Pittsburgh, also said the approach adds efficacy evidence for therapeutic payloads.

Responses outside the injection site are a central efficacy observation, although the 71% estimate remains preliminary evidence from a seven-patient, phase 1/1b subgroup. Najjar said the need is substantial because patients with advanced cutaneous malignancies have limited options after progression on immune checkpoint blockade.

How was STEALTH-001 designed?

STEALTH-001, registered as NCT06444815, is a first-in-human, open-label trial that opened with a 3+3 dose-escalation phase in an all-comer solid-tumor population. Monotherapy VET3-TGI was tested from 1 × 10⁸ to 3 × 10⁹ plaque-forming units, with a 28-day DLT observation window.

Investigators are now enrolling a monotherapy expansion cohort restricted to cutaneous malignancies. A combination cohort pairs VET3-TGI, at or below the highest monotherapy dose, with atezolizumab (Tecentriq).

Patients received intratumoral injections on days 1, 8, 15, and 22. Optional maintenance boosters could occur as often as every two weeks. Atezolizumab was administered at 840 mg every two weeks for four doses, followed by 1,680 mg every four weeks.

Safety, DLTs and the recommended phase 2 dose were primary endpoints. Tumor response, survival, and viral and immune activity were secondary endpoints.

VET3-TGI is a Western Reserve vaccinia virus modified to express CXCR3, allowing attachment to lymphocytes, homing to tumors and avoidance of neutralization, Najjar explained. The construct encodes two therapeutic payloads: a TGF-β inhibitor that originated at UPMC Hillman Cancer Center and interleukin-12.

Who received treatment?

As of the September 2026 cutoff, 20 patients had received intratumoral VET3-TGI: 16 for monotherapy and four in combination with atezolizumab. All 20 comprised the safety-evaluable group.

The seven-patient cutaneous subgroup had a median of three prior lines of therapy, with a range of one to nine. Their disease and prior treatment were characterized as follows:

• All seven patients had received anti–PD-(L)1 therapy and remained refractory to immunotherapy.
• 42.9% had received both a CTLA-4 inhibitor and an anti–PD-(L)1 agent.
• 57.1% had stage IV disease, 85.7% had lactate dehydrogenase above the upper limit of normal, and 57.1% had visceral metastases.

What did the translational analyses show?

Paired biopsies collected at baseline and on treatment showed increased immune activity in all five evaluable injected lesions and all three evaluable non-injected lesions. In injected lesions, median fold increases on days 15 to 22 were 7.2× for cytotoxic CD8+GRB+ T cells and 10.4× for proliferating CD8+Ki67+ T cells.

In non-injected lesions, the corresponding median increases at day 15 were 2.8× and 3.1×. Investigators characterized these descriptive findings as a biologic rationale for response durability, suggesting conversion of immunologically “cold,” checkpoint inhibitor–refractory tumors into “hot” tumors rather than a transient effect limited to injection sites.

What did the safety data show?

No protocol-defined DLTs occurred during the 28-day observation window, and dose escalation proceeded through 3 × 10⁹ PFU. Treatment-related adverse effects were common among the 20 safety-evaluable patients, although most were low grade.

The most frequent treatment-related adverse effects of any grade were cytokine release syndrome, or CRS, at 50%; chills at 40%; nausea at 40%; vomiting at 35%; and decreased lymphocyte count at 25%. Every CRS event was grade 1, reported in 20%, or grade 2, reported in 30%; no grade 3 or 4 CRS occurred.

Grade 3 treatment-related effects included decreased lymphocyte count in 10% of patients. Nausea, vomiting, anemia, hyponatremia and increased alanine aminotransferase levels each occurred in 5%. Grade 4 effects were limited to decreased lymphocyte count in 10% and decreased white blood cell count in 5%. No treatment-related deaths occurred.

Najjar said lymphopenia improved over time and that nothing in the safety profile fell outside expectations for oncolytic viruses. She added that CRS management has become more familiar to melanoma specialists as bispecific T-cell engagers and oncolytic viruses have entered the field.

What happens next?

Enrollment in the cutaneous-malignancy cohorts continues to inform phase 2 study design. “From my vantage point as an investigator, it seems clear that the signal is in cutaneous malignancies,” Najjar said.

Intravenous STEALTH-001 arms are under evaluation and will be reported separately. The immediate clinical question is whether the response signal persists as the cutaneous-malignancy cohort expands and whether immune activation away from injection sites translates into durable control.