Comparative neurologic toxicity profiles of chemotherapy versus immune checkpoint inhibitors in melanoma: a propensity score-matched analysis

Abstract Purpose Immune checkpoint inhibitors (ICIs) have transformed melanoma treatment, yet comparative real-world data characterizing ICI-associated neurologic adverse events against the well-established neurotoxicity profile of conventional chemotherapy remain limited. We sought to contextualize the neurologic risk landscape of ICIs relative to chemotherapy as a historical comparator, to inform treatment-specific supportive care and surveillance strategies. Methods We performed a propensity score–matched analysis using the TriNetX Global Collaborative Network, encompassing 162 healthcare organizations. Adult patients with melanoma receiving chemotherapy (dacarbazine, paclitaxel, carboplatin, or temozolomide; n = 9787) or ICI therapy (pembrolizumab, nivolumab, or ipilimumab; n = 14,627) from January 2014 through November 2025 were identified. After 1:1 nearest-neighbor matching on 52 baseline covariates, 6887 pairs were analyzed over 1–1095 days post-index therapy. Primary outcomes were peripheral neuropathy, encephalopathy, and laboratory-anchored myositis (ICD-10 myositis codes with creatine kinase ≥ 500 U/L). Fractures served as a negative control. A pre-specified conservative sensitivity analysis excluded bidirectional treatment switching, restricted the ICI cohort to anti-PD-1 predominant exposure, tightened outcome definitions, and extended follow-up to 5 years. Results Among 13,774 matched patients (mean age 67.1 years; 48.5% male), peripheral neuropathy occurred in 7.2% of chemotherapy-treated versus 3.2% of ICI-treated patients (HR 2.47; 95% CI 2.10–2.92). Encephalopathy occurred in 1.0% versus 1.4% (HR 0.70; 95% CI 0.51–0.96). Laboratory-confirmed myositis occurred in 1.0% versus 2.1% (HR 0.47; 95% CI 0.35–0.64). No cases of Guillain–Barré syndrome or CNS vasculitis were detected. The negative control showed no significant difference ( P = .20). In the conservative sensitivity analysis (6640 matched pairs; 5-year follow-up), the peripheral neuropathy association was substantially strengthened (HR 5.45; 95% CI 4.16–7.15) and the laboratory-anchored myositis association was essentially unchanged (HR 0.46; 95% CI 0.33–0.64), while the encephalopathy association attenuated to null (HR 1.03; 95% CI 0.79–1.34). Conclusions Chemotherapy carries a substantially higher peripheral neuropathy risk, while ICI therapy confers elevated laboratory-confirmed myositis risk. These robust findings support treatment-specific neurologic surveillance and supportive care strategies in melanoma management, particularly neuropathy monitoring during chemotherapy and neuromuscular/CK surveillance during ICI therapy.