Clinical Course of Neurologic Adverse Events Associated With Immune Checkpoint Inhibitors: Focus on Chronic Toxicities

Abstract

Background and objectives: The clinical course and the risk of chronicity of neurologic immune-related adverse events (n-irAEs) associated with immune checkpoint inhibitors (ICIs) are not well documented. This study aimed to characterize the clinical course of n-irAEs and assess the prevalence of chronic events.

Methods: This nationwide, multicenter, retrospective study included patients with n-irAEs identified at 7 Italian hospitals. The clinical course of n-irAEs was categorized into fulminant (if resulted in death within 12 weeks), monophasic (if resolved within 12 weeks), and chronic (if persisted beyond 12 weeks). Chronic n-irAEs were further subdivided into active (if there was indirect evidence of ongoing inflammation [i.e., required ongoing immunosuppression, relapsed on steroid tapering, or exhibited neurologic progression]) and inactive (if patients had neurologic sequelae without ongoing inflammation). Comparisons between groups and time-to-death analyses were performed.

Results: Sixty-six patients were included (median age: 69 years [IQR 62-75]; 53 [80%] men). n-irAEs involved the peripheral nervous system in 48 patients (73%), the central nervous system in 14 (21%), and both in 4 (6%). Twelve patients (18%) had a fulminant course, with the risk being significantly higher in those with concurrent myocarditis (OR 5.4; 95% CI [1.02-28.31]). Among 54 patients with a nonfulminant course, 23 (43%) had a monophasic n-irAE and 31 (57%) had a chronic n-irAE, of which 16 of 31 (52%) were chronic active (due to ongoing immunosuppression [69%], relapses at corticosteroid tapering [19%], or neurologic disease progression [12%]) and 15 of 31 (48%) were chronic inactive. In patients with chronic inactive n-irAEs, neurologic sequelae included cerebellar ataxia (33%), neuromuscular weakness (27%), visual loss (13%), sensory disturbances (13%), focal neurologic signs (7%), and cognitive impairment (7%). Compared with patients with monophasic events, those with chronic n-irAEs had a higher rate of severe neurologic disability at the last evaluation (p < 0.01), shorter survival (p < 0.01), and higher overall mortality (p < 0.01), primarily due to cancer progression.

Discussion: More than half of the patients with n-irAEs who survived the acute phase developed a chronic condition. Patients with chronic n-irAEs were at higher risk of death, mainly due to cancer progression. Future studies are needed to further characterize chronic n-irAEs and identify optimal long-term management strategies.

Figure 1. Flowchart of Patients With Neurologic Immune-Related Adverse Events Included in the Study and Their Clinical Phenotypes

^a1 patient with melanoma and immune-related polyradiculoneuropathy who died 2 months after n-irAE onset because of intracerebral hemorrhage (bleeding of brain metastases); 1 patient with melanoma and immune-related cranial neuropathy (bilateral VII and VIII cranial nerve involvement) who died 2.5 months after n-irAE onset because of bowel obstruction. ^bIn patients with myositis, the presence of myasthenic features was documented by decremental response to repetitive nerve stimulation (7 patients) or by clinical response to acetylcholinesterase inhibitors (7 patients). CIDP = chronic inflammatory demyelinating polyneuropathy; GBS = Guillain-Barrè syndrome; LETM = long-extensive transverse myelitis; MG = myasthenia gravis; n-irAE = neurologic immune-related adverse event; PNS = peripheral nervous system.

Figure 2. MRI of the Brain and Muscle Biopsy of 2 Patients With n-irAEs

(A and B) Brain MRI fluid-attenuated inversion recovery sequences in a patient with anti-Ma2 encephalitis after receiving pembrolizumab for lung cancer, showing hyperintensity and swollen appearance of the right uncus of the hippocampus (A: axial view, arrow; B: coronal view, arrowheads). (C and D) Skeletal muscle biopsy of a patient with myositis and myasthenia after combination treatment with anti–PD-1 and CTLA-4 inhibitors for lung cancer (C: hematoxylin and eosin–stained section showing diffuse and abundant endomysial inflammatory infiltrate associated with muscle fiber necrosis and regeneration; D: CD8 immunohistochemistry showing CD8⁺ T-cell inflammatory infiltrate).

Figure 3. Clinical Course in Different Types of n-irAEs

*Other types of CNS involvement included 3 demyelinating syndromes (1/3 chronic active and 2/3 chronic active) and 1 CNS vasculitis. **Multifocal n-irAEs included myositis and unilateral optic neuritis (chronic inactive), cerebellitis and cranial neuritis (chronic inactive), cerebellitis and cranial neuritis (monophasic), brain demyelinating lesions and inflammatory polyradiculopathy (monophasic).

Figure 4. Kaplan-Meier Curves Showing Cumulative Probability of Survival After n-irAEs, Stratified by Clinical Course

Log-rank test: p < 0.01. Compared with patients with monophasic n-irAEs, patients with both chronic active and inactive n-irAEs had shorter survival.