ctDNA Reduction Is Potentially Predictive of OS Vs RECIST in Metastatic Uveal Melanoma

A linear correlation between magnitude of circulating tumor DNA (ctDNA) reduction and improved overall survival (OS) was observed in patients with metastatic uveal melanoma who were treated with tebentafusp, suggesting that ctDNA may be a better surrogate end point for OS compared with RECIST response criteria, according to findings from the phase 1/2 IMCgp100-102 trial (NCT02570308) that were presented during the 2021 ESMO Congress.

Uveal melanoma is a rare subtype of melanoma, characterized by frequent liver metastases, low tumor mutational burden (TMB), and common expression of gp100, a melanocytic protein. To date, checkpoint inhibitors have demonstrated poor efficacy in this subpopulation, and there is currently no established standard of care for patients.

Tebentafusp is a bispecific soluble therapy in which an affinity-enhanced T-cell receptor binds to CD3 and gp100. The agent is designed to redirect T cells to gp100-positive melanocytic cells.

Moreover, tebentafusp was the first agent to demonstrate an OS benefit in the randomized setting in uveal melanoma. Findings from the phase 2 IMCgp100-202 (NCT03070392) demonstrated an OS advantage with the therapy compared with investigator’s choice of first-line therapy in the intent-to-treat patient population (HR, 0.51; 95% CI, 0.37-0.71) and in patients with a best response of progressive disease (HR, 0.43; 95% CI, 0.27-0.68). Moreover, the results showed an overall response rate (ORR) of 9.1% by RECIST criteria and the HR for progression-free survival was 0.73 (95% CI, 0.58-0.94).

“[This] suggests that RECIST just isn’t a fantastic way to predict who will benefit from this drug,” said lead study author Alexander Noor Shoushtari, MD, a medical oncologist at Memorial Sloan Kettering Cancer Center, in a virtual presentation of the data.

To be eligible for the IMCgp100-102 trial, patients (n = 127) had to have HLA-A*02:01–positive metastatic uveal melanoma. Patients had to have measurable disease, and must have received at least 1 prior therapy in the metastatic setting.

The treatment-escalation schedule of once-weekly tebentafusp infusions comprised 20 mcg in week 1, 30 mcg in week 2, and 68 mcg from week 3 on. Patients received tebentafusp in the second-line or beyond setting.

Results from the study showed that the ORR was 5% and the duration of response was 8.7 months. The median OS was 16.8 months; the 1- and 2-year OS rates were 62% and 37%, respectively.

Compared with historical published data, where the median OS was 7.8 months and the 1- and 2-year OS rates were 37% and 15%, respectively, these data with tebentafusp were promising, Shoushtari explained.

“We hypothesized that ctDNA may be a better way to analyze those who might obtained benefit [from tebentafusp],” Shoushtari added.

Of the 127 patients who received treatment, 118 had detectable ctDNA at any time point, 116 had detectable ctDNA at baseline, and 99 had detectable ctDNA at baseline and by week 9.

Although uveal melanoma has the lowest median number of coding somatic mutations per mutational base compared with any other tumor type, 92% of patients had detectable ctDNA with mutations in known uveal melanoma oncogenes, such as GNA11, GNAQ, SF3B1, PLCB4, EIF1AX, CYSLTR2, and others.

“Uveal melanoma has a low TMB but has the benefit of having a really predictable set of mutations,” Shoushtari said.

At baseline, ctDNA levels were found to significantly correlate with tumor burden (R = .61; P = 1e-10) and serum lactate dehydrogenase (R = .75; P < 2.2e-16).

“The correlation was not perfect, suggesting some non-redundant information from each of these parameters,” Shoushtari added.

Additionally, 70% of all evaluable patients had any ctDNA reduction with the addition of tebentafusp. Regarding ctDNA clearance, 25% of patients had a less than 68% (<0.5 log) reduction in ctDNA, 30% had a 68% (0.5 log) to 99.9% (3.2 log) reduction in ctDNA, and 14% had cleared ctDNA.

ctDNA reduction was found to be a more sensitive measure of tebentafusp efficacy compared with tumor size in all evaluable patients and patients with a best response of progressive disease. In the cohorts, respectively, ctDNA reduction was associated with greater mean tumor shrinkage and less tumor growth.

Notably, a linear correlation was observed between ctDNA reduction and improved OS (R2= .88; P < .0001).

ctDNA reduction informed which patients were more likely to derive an OS benefit, irrespective of RECIST response, among all evaluable patients (HR, 0.56; 95% CI, 0.32-0.95; P = .03), patients with a best response of progressive disease (HR, 0.44; 95% CI, 0.2-0.94; P = .027), and patients with a best response of stable disease (HR, 0.48; 95% CI, 0.16-1.43; P = .16).

Additionally, 44% of all evaluable patients and 35% of patients with a best response of progressive disease had at least 0.5 log reduction of ctDNA. Twenty-eight percent of patients with a best response of stable disease had at least 1 log reduction ctDNA.

Of the 14% of patients who achieved ctDNA clearance, best RECIST responses included stable (n = 8) or progressive disease (n = 4). Other clearances included one patient with a partial response and one patient who was not evaluable for response.

Notably, all patients with ctDNA clearance were alive for at least 1 year (HR, 0.14; 95% CI, 0.03-0.57; P = 0).

Reference

1. Shoushtari A, Collins L, Arranz EE, et al. Early reduction in ctDNA, regardless of best RECIST response, is associated with overall survival (OS) on tebentafusp in previously treated metastatic uveal melanoma (uUM) patients. Presented at: 2021 ESMO Congress; September 16-21, 2021; virtual. Abstract 17570.