Although immune checkpoint inhibitors (ICIs) have improved systemic cancer care, in rare cases, patients develop neuromuscular immune related adverse events (irAEs), which carry high morbidity and mortality. This case study describes an 85-year-old male who developed the triad of myositis, myasthenic crisis, and myocarditis in association with ICI treatment for metastatic urothelial carcinoma (T3N0M1). The patient developed progressive weakness and gait impairment five weeks after initiating treatment with pembrolizumab. Later, he developed reduced ambulation, fatigue, shortness of breath, and orthopnea with concomitant ECG changes. Clinical exam demonstrated dysphagia, ptosis, proximal muscle weakness, fatigable weakness, and areflexia, ultimately resulting in respiratory failure, intubation, and inability to stand. Lab workup revealed elevated creatinine kinase, myoglobin, lactate dehydrogenase, and aldolase, but was notably negative for acetylcholine receptor and muscle specific kinase antibodies. Labs, clinical timeline, and clinical features supported his diagnoses. Interestingly, the cytokine panel demonstrated a high-normal soluble IL-2 receptor alpha which supports autoimmune pathology and elevated IL-10. The patient was treated with pyridostigmine, prednisone, and plasma exchange (PLEX) for 10 exchanges. EMG 10 days into treatment supported evidence of inflammatory myositis. After methylprednisone and PLEX treatment, the patient’s symptoms improved to ambulation with a walker and occasional BiPap support. His metastatic disease remained stable for several weeks while holding pembrolizumab due to toxicity. This case illustrates how early intervention and aggressive treatment for neuromuscular ICI-induced triad can stabilize patients and improve neurologic recovery. Due to the high morbidity and mortality for these irAEs, the authors support increased clinical scrutiny when evaluating these rare conditions and consideration of investigating the acute inflammatory processes. With improved pathophysiologic understanding of these rare conditions, the community can improve outcomes and refine treatments.

Background: Discordance between pre-hospital (EMS) and experienced stroke team (ED) determination of Vision-Aphasia-Neglect (VAN) assessment threatens VAN’s utility in routing patients with large vessel occlusion (LVO) to thrombectomy centers. Discordance may be due to a fluctuating exam in the setting of stroke or stroke mimics. We sought to evaluate the rate and etiology of stroke mimics in patients with discordant versus concordant VAN assessments. Methods: The EMS run reports were linked with stroke registries from three CSCs in New Orleans, Louisiana from 2/2019-7/2019. EMS-VAN was compared to that of CSCs in all patients who presented within 6 hours of last seen normal. Stroke mimics were compared in patients with concordance vs discordance between VAN determination by EMS and experienced stroke teams. Results: EMS documented VAN in 63.2% of stroke codes brought to CSCs. Concordance was 66.9%; the k, CI was 0.35, 0.20 - 0.50. The proportion of stroke mimics was largest in concordant VAN negative patients (64.4%) with seizures and altered mental status (23.7% each) the most frequent causes. The proportion of stroke mimics was smallest in concordant VAN positive patients (17.3%) with seizures as the most frequent cause (44.4%). Similar high frequency of seizure mimic was determined in VAN discordant patients; seizure accounted for 47.8% of discordant mimics and 20% of all discordant patients. Discussion: VAN assessment as performed by trained EMS providers had only fair correlation with experienced stroke team members. The rapidly evolving exam of patients with seizures may explain a significant portion of stroke mimics among patients whose exams change between EMS and ED stroke team assessments.

Background: The Louisiana pre-hospital routing protocol for acute stroke patients is modeled after Mission Lifeline Stroke. The Vision-Aphasia-Neglect (VAN) assessment was adopted for large vessel occlusion (LVO) screening after education was provided to pre-hospital providers (EMS). We sought to establish the inter-rater reliability (IRR) of VAN between EMS and experienced stroke team members at comprehensive stroke centers (CSC) and assessed validity of EMS and CSC VAN. Methods: The EMS run reports were linked with stroke registries from three CSCs in New Orleans, Louisiana from 2/2019-7/2019. EMS-VAN was compared to that of CSCs in all patients who presented within 6 hours of last seen normal. Validity analyses were performed comparing EMS-VAN and CSC-VAN with radiologically confirmed (CTA) LVO as the gold standard. Kappa (k) and 95% confidence interval (CI) are reported as a measure of IRR. Likelihood of VAN non documentation by EMS and VAN discordance between EMS and CSC was determined as Odds Ratios (OR) and CI. Results: EMS documented VAN in 63.2% of stroke codes brought to CSCs. Advanced age (OR, CI: 1.02, 1.01 - 1.04) and pre-existing disability (OR, CI: 1.53, 1.32 - 1.76) increased the likelihood that VAN was not documented by EMS. Both EMS and CSC VAN assessments were documented for 166 patients Concordance was 66.9%; the k, CI was 0.35, 0.20 - 0.50. Discordance was more likely when EMS assessed a patient as positive than when assessed as negative, but no clinical predictor of discordance was identified. The accuracy of EMS-VAN was 60.2% for distinguishing LVO positive and negative patients; CSC-VAN was 65.5% accurate. Discussion: VAN assessment as performed by trained EMS providers had only fair correlation with experienced stroke team members. Further education is needed to increase the inter-rater reliability and accuracy of the VAN assessment in order to improve its utility in meeting the goal of routing patients with LVO to the thrombectomy centers.