Anti-SARS-CoV-2 mRNA vaccines can deeply affect cell-mediated immune responses in immunocompromised recipients, including cancer patients receiving active treatments. The clinical implications of changes in peripheral blood lymphocyte subsets following the third dose of mRNA-BNT162b2 vaccination (tozinameran) in patients on immune checkpoint blockade are not fully understood. We conducted a prospective analysis of the Vax-On-Third-Profile study to evaluate the impact of circulating lymphocyte dynamics on disease outcomes in this subgroup of patients. Recipients of booster dosing who had received before vaccination at least one course of an anti-PD-1/PD-L1 treatment for an advanced solid tumor were eligible. Immunophenotyping of peripheral blood was performed before the third dose of tozinameran (timepoint-1) and four weeks later (timepoint-2) to quantify the absolute counts of lymphocyte subpopulations, including CD3+CD4+ T cells, CD3+CD8+ T cells, B cells, and NK cells. Logistic regression was used to analyze the relationship between lymphocyte subsets and durable clinical benefit (DCB). The log-rank test and Cox regression model were applied to evaluate the relationship between lymphocyte subpopulations and both vaccine-related time-to-treatment failure (V-TTF) and overall survival (OS). We included a total of 56 patients with metastatic disease who were given a third dose of tozinameran between 23 September and 7 October 2021 (median age: 66 years; male: 71%). Most recipients had a diagnosis of lung cancer and were being treated with pembrolizumab or nivolumab. Compared to baseline, the third immunization resulted in an incremental change in the median counts of all lymphocyte subpopulations, which was statistically significant only for NK cells (p < 0.001). A significant correlation was found between NK cell counts and DCB at timepoint-2 (p < 0.001). Multivariate logistic regression analysis of DCB confirmed the predictive significance of high-level NK cell counts (p = 0.020). In multivariate Cox regression analysis, high-level NK cell counts independently predicted longer V-TTF [HR 0.34 (95% CI 0.14-0.80), p = 0.014] and OS [HR 0.36 (95% CI 0.15-0.89), p = 0.027]. Our data suggest expansion of NK cell counts as the most noteworthy change in circulating lymphocytes after the third dose of tozinameran in cancer patients receiving PD-1/PD-L1-targeted agents. This change correlated with enhanced therapeutic efficacy, improving the rate of disease control, and prolonging survival outcomes. Similar findings have not been previously reported, implying that they have proof-of-concept value and warrant further confirmation.