Comprehensive Analysis of Paraneoplastic Neurologic Syndrome and PNS-CARE Diagnostic Criteria in Clinical Practice

Abstract

Background and objectives: Paraneoplastic neurologic syndrome (PNS) diagnostic criteria were first proposed in 2004 and updated in 2021. The PNS-CARE score, derived from the updated criteria, is a composite model for assigning likelihood for patients with suspected PNS. In this study, we evaluated the utility and applicability of the 2021 PNS-CARE score and present our PNS cohort.

Methods: This is a retrospective study. We identified Mayo Clinic patients suspected to have PNS (1/2005-12/2020) and collected relevant information including demographics, PNS presentation, and clinical outcomes. Inclusion criteria were the following: (1) patients with a syndrome consistent with PNS and (2) patients with sufficient information available in charts. Exclusion criteria were the following: (1) evaluation only before 2005, (2) patients not evaluated by neurology, (3) presentation after immune checkpoint inhibitors, and (4) syndromes not included in 2021 criteria. All patients were evaluated for the 2021 and 2004 PNS criteria.

Results: We identified 484 patients suspected to have PNS at initial presentation, of whom 212 (44%) were considered to have PNS after completion of evaluation. Among these 212 patients, the most common autoantibodies were PCA1 (Yo)-IgG (17%), KLHL11-IgG (16%), and CRMP5-IgG (14%) and the most common phenotypes were rapidly progressive cerebellar syndrome (29%), brainstem encephalitis (14%), and limbic encephalitis (8%). The 2021 PNS criteria definite/probable categorization (PNS-CARE score ≥ 6) had a sensitivity and specificity of 93% and 100%, respectively, while the 2004 PNS criteria definite categorization had a sensitivity and specificity of 67% and 99%, respectively. We found 15 patients with a PNS-CARE score ≤5 who likely had PNS on our review. The most common presentation among these patients was KLHL11-IgG brainstem encephalitis (7/15, 47%) with likely burned-out testicular tumor.

Discussion: Our study validates the PNS-CARE score. A clearer understanding of typical PNS presentation and common underlying malignancies and autoantibodies can aid in earlier and more accurate diagnosis, which is crucial for downstream clinical decisions. Some patients with an intermediate-risk phenotype do not meet probable/definite criteria despite the presence of high-risk antibodies and/or underlying malignancy.

Figure 1. Flowchart Depicting Search Strategy for Identification of Patients Evaluated for Paraneoplastic Neurologic Syndromes

Search key words included the following: paraneoplastic neuropathy; paraneoplastic antibody level; paraneoplastic encephalomyelitis; paraneoplastic peripheral neuropathy; paraneoplastic cerebellar degeneration; paraneoplastic autonomic dysfunction; paraneoplastic antibody; paraneoplastic neuromyopathy and neuropathy; perineoplastic/paraneoplastic; paraneoplastic/perineoplastic condition; autoimmune encephalopathy; and myopathy autoimmune non-inflammatory. We excluded phenotypes that were not included by the updated 2021 article, including inflammatory myopathies (dermatomyositis, polymyositis, and necrotizing myopathies); myasthenia gravis; polyneuropathies associated with monoclonal gammopathies; and paraneoplastic retinopathy, optic neuritis, and cochlea-vestibulopathy.

Figure 2. Distribution of Antibodies, Malignancies, and Presenting Paraneoplastic Neurologic Syndromes in Our PNS+ Group (n = 212)

(A) Antibodies found in the PNS+ group. Other (n = 7) included 1 patient with each of the following: SOX1-IgG, neurofilament light-chain IgG, PQ calcium channel IgG in a patient with cerebellar syndrome, PCA2 (microtubule-associated protein 1B, MAP1B)-IgG, AMPA-IgG, and NMDA-IgG3 in 2 patients. (B) Malignancies found in the PNS+ group. Others (n = 8) included neuroendocrine tumor (n = 4), squamous cell carcinoma of the head and neck (n = 2), prostate adenocarcinoma (n = 1), and neuroblastoma (n = 1). (C) Paraneoplastic syndromes found in the PNS+ group. DNER = delta/notch-like epidermal growth factor-related receptor; GI = gastrointestinal; LEMS = Lambert-Eaton myasthenic syndrome; PN = peripheral neuropathy.

Figure 3. Representative MRI Changes in Patients With High-Risk Antibody Positivity

(A) Axial section of T2-FLAIR imaging showing T2 hyperintensity in bilateral, but left worse than right, mesial temporal lobes in a patient with limbic encephalitis, small-cell lung carcinoma, and ANNA1 (Hu)-IgG antibody. (B) Sagittal section of T2-FLAIR imaging showing a T2 hyperintensity of the dorsal aspect of the tegmentum extending to the pons, medulla, and upper cervical cord in a patient with brainstem encephalitis, seminoma, and KLHL11-IgG antibody. (C) Sagittal section of T2-FLAIR imaging showing generalized severe cerebellar atrophy in a patient with rapidly progressive cerebellar syndrome, ovarian carcinoma, and PCA1 (Yo)-IgG antibody. (D) Axial section of T2-FLAIR imaging showing (D) posterior thalamic region in a patient with brainstem encephalitis and limbic encephalitis, seminoma, and Ma2-IgG.

Figure 4. Heat Map of Associated Paraneoplastic Syndrome and Malignancy With the Most Frequently Found Autoantibodies

¹Cerebellar syndrome = rapidly progressive cerebellar syndrome; ²Unk src = unknown source, malignancy was identified by pathology on lymph nodes or metastatic lesions, but the primary malignancy was not identified; ³Concern = suspicious lesion identified on imaging, but the patient was either lost to follow-up or declined further evaluations. BrAd = breast adenocarcinoma; EnCar = endometrial carcinoma; GI = gastrointestinal; NSCLC = non–small-cell lung carcinoma; OvAd = ovarian adenocarcinoma; PN = peripheral neuropathy; ProAd = prostate adenocarcinoma; SCLC = small-cell lung carcinoma.

Figure 5. Schematic of Total Cohort Characterized According to the (A and C) 2004 PNS Criteria and (B and D) 2021 PNS-CARE Score