Melanoma Shrinks With Injection After Anti-PD-1 Progression

Intratumoral injection of a toll-like receptor (TLR) agonist showed potential for reversing PD-(L)1 resistance in melanoma, a preliminary clinical trial showed.

A total of 23.5% of patients with metastatic melanoma responded to intratumoral CMP-001 and systemic pembrolizumab (Keytruda), including seven complete responses. All of the patients had disease progression or stable disease during prior anti-PD(L)-1 therapy.

The results also showed that noninjected lesions regressed following the treatment, suggesting abscopal antitumor activity, Mohammed Milhem, MD, of the University of Iowa in Iowa City, reported during a press briefing that preceded the Society for Immunotherapy of Cancer virtual meeting.

“Intratumoral CMP-001 plus pembrolizumab reversed resistance to PD-1 blockade and resulted in durable, systemic responses in patients with advanced melanoma and prior disease progression on anti-PD-1 therapy,” he said. “Systemic responses to CMP-001 plus pembrolizumab were observed with a similar mean regression of injected and noninjected target lesions. CMP-001 alone and in combination with pembrolizumab was well tolerated, with the most common treatment-related adverse events being low-grade flu-like symptoms and injection-site pain.”

A member of the virtual audience questioned whether at least some patients might have responded to pembrolizumab monotherapy.

“All patients had received at least 12 weeks of treatment with the PD-1 inhibitor and had demonstrated that their tumors were growing,” said Milhem. However, confirmation of progression was not required in the study but will be required in an ongoing trial of the combination.

“If you gave another 4 to 6 weeks of pembrolizumab, we would wonder if there would be a response or not,” he acknowledged.

More than 90% of the study participants had progressive disease, as opposed to stable disease, Milhem added.

Although checkpoint inhibitors have revolutionized treatment of advanced melanoma, patients who progress during or after treatment have few effective options. CMP-001 consists of a TLR9 agonist within an immunogenic virus-like particle (VLP) that induces anti-VLP antibodies and stimulates plasmacytoid dendritic cells to stimulate interferon-alpha induction. In preclinical models, the construct increased tumor regression as compared with a construct that omitted the TLR9 agonist.

Milhem reported findings from a trial involving patients with metastatic nonuveal melanoma and lesions amenable to intratumoral injection. During the dose-escalation/expansion component of the trial, 159 patients received CMP-001 and pembrolizumab. Investigators evaluated two dosing schedules for CMP-001: once weekly for 7 weeks or 2 weeks, followed by additional dosing every 3 weeks until treatment discontinuation. During a preplanned second part of the trial, 40 patients received CMP-001 only.

During the dose-escalation/expansion phase, 98 patients received CMP-001 with a polysorbate 20 (PS20) concentration of 0.01%, and 61 received CMP-001 with a PS20 concentration of 0.00167%.

The primary objectives of the trial were to determine the recommended phase II dose and schedule of CMP-001 in combination with pembrolizumab (part 1) and to determine the safety of single-agent CMP-001 monotherapy (part 2). Secondary objectives included safety and antitumor activity of the combination and single-agent CMP-001.

Patients who received the combination with the higher concentration of PS20 had a twofold higher response rate, 23.5% vs 11.5%. Overall, 23 of 98 patients had objective responses with the higher concentration of PS20 versus eight of 61 in the group that received the lower concentration. Complete responses occurred in seven patients and one patient, respectively, in the two groups. An additional 16 of the 98 patients had stable disease versus 18 of 61.

The CMP-001 monotherapy group had an objective response rate of 17.5% (seven of 40 patients), all partial responses. An additional 13 patients had stable disease. Responses were less durable with single-agent CMP-001 (median of 5.6 months) versus the combination (19.9 months).

An analysis of 22 patients with both injected and noninjected target lesions showed that some degree of tumor shrinkage occurred in all noninjected target lesions, and tumor shrinkage exceeded 30% in 20 of the 22 cases. In several instances, the degree of tumor shrinkage in noninjected lesions exceeded that of the injected lesions, including some patients whose injected lesions did not regress.