Virus-Specific T Cells and Response to Checkpoint Inhibitors in Progressive Multifocal Leukoencephalopathy.

Importance: Progressive multifocal leukoencephalopathy (PML) is a life-threatening demyelinating disease caused by reactivation of the JC virus (JCV) in immunocompromised patients. While immune checkpoint inhibitors (ICIs) show therapeutic potential, responses vary and predictive biomarkers are lacking. Objective: To determine whether pretreatment JCV- and/or BK virus-specific T cells in the blood are associated with treatment efficacy. Design, Setting, and Participants: This retrospective cohort study included 111 patients with PML who were treated with ICIs stratified by peripheral virus-specific T cell presence (ELISpot/flow cytometry) between August 2021 and May 2024, with a median (IQR) follow-up of 7 (1-13) months. Of 112 patients with definite PML across 39 centers, 1 patient refused participation; 111 patients were included. Exposure: Patients received pembrolizumab (n = 81), nivolumab (n = 28), or atezolizumab (n = 2) per availability and prescribing practices at participating centers. Main Outcome and Measures: Clinical outcomes, diagnostic parameters, and immune-related adverse events were compared; association of virus-specific T-cell responses with survival was analyzed using the Kaplan-Meier method. Results: The study cohort consisted of 111 patients (median [IQR] age, 61 [50-70] years; 74 male [66.6%]). Twenty-one patients had detectable virus-specific T cells prior to therapy, 22 were T cell-negative and 68 had an unknown T-cell status. T cell-positive patients showed significantly higher response rates and improved survival compared to both T cell-negative patients (18/21 [86%] vs 5/22 [23%]; P < .001; median survival time, none [95% CI, undefined] vs 136.5 days [95% CI, 19 to ∞]; P = .002) and those with unknown T-cell status (18/21 [86%] vs 29/68 [43%]; P = .001; median survival time, none vs 162 days [95% CI, 66 to ∞]; P = .004). They achieved better functional outcomes (median [IQR] modified Rankin Scale score, 3 [2-4] vs 4 [3-6]; P = .009) and lower JC viral load in cerebrospinal fluid (median [IQR], 0 copies/mL [0-502.5] vs 2500 copies/mL [0-6900]; P = .01) during follow-up compared to T cell-negative patients. Immune-related adverse events were most frequent in T cell-negative patients (10/20 [50%]), including the most severe events, and least frequent in T cell-positive patients (2/20 [10%]) (P = .02). Conclusions and Relevance: Preexisting functional virus-specific T cells were associated with better clinical response, longer survival, and lower toxicity in PML. These findings suggest the likely importance of preexisting antiviral immunity for successful ICI therapy.