Recessive Ataxia Differential Diagnosis Algorithm (RADIAL) Versus Specific Niemann-Pick Type C Suspicion Indices: A Retrospective Algorithm Comparison.

Early diagnosis of Niemann-Pick disease type C (NPC) is crucial to slow the progression of neurological manifestations. Different tools were developed to aid diagnosis of NPC, but to date, no study has compared their performance. We aimed to compare the RADIAL algorithm, intended for the differential diagnosis of autosomal recessive cerebellar ataxias (ARCAs) and NPC-specific suspicion indices (SIs). This study was a retrospective analysis of data from 834 patients with molecularly confirmed ARCAs, including 57 NPC cases (RADIAL cohort). We aimed to compare the algorithm performance of RADIAL (Top 1 and Top 3) with that of four SIs (Original, Refined, 2/3 and 2/7) in discriminating NPC cases and non-NPC cases. We also identified ARCAs closely related to NPC as those with low specificity to detect non-NPC cases and described differential and overlapping features with NPC. Overall, excellent sensitivity and specificity (> 0.90) were achieved with both RADIAL and SI tools for NPC cases. The highest sensitivity was attained with the 2/7 SI, Refined SI and Top 3 RADIAL algorithms. Top 1 and Top 3 RADIAL were the most specific tools, followed by the Original SI. The individual comparison of each ARCA revealed that Wilson disease, PLA2G6-associated neurodegeneration, and hypomyelinating leukodystrophy (POLR3A) are frequent NPC false positives (PLA2G6 and POL3A only with the SIs). Both RADIAL and SI diagnostic approaches showed strong discriminatory potential and may be useful screening tools in different clinical contexts.

Identifying Niemann-Pick type C in early-onset ataxia: two quick clinical screening tools.

Niemann-Pick disease type C (NP-C) is a rare multisystemic lysosomal disorder which, albeit treatable, is still starkly underdiagnosed. As NP-C features early onset ataxia (EOA) in 85-90 % of cases, EOA presents a promising target group for undiagnosed NP-C patients. Here, we assessed the ability of the previously established NP-C suspicion index (SI) and a novel abbreviated '2/3 SI' tool for rapid appraisal of suspected NP-C in unexplained EOA. This was a retrospective observational study comparing 'NP-C EOA' cases (EOA patients with confirmed NP-C) with non-NP-C EOA controls (EOA patients negative for NP-C gene mutations). NP-C risk prediction scores (RPS) from both the original and 2/3 SIs were calculated and their discriminatory performance evaluated. Among 133 patients (47 NP-C EOA cases; 86 non-NP-C EOA controls), moderate (40-69 points) and high (≥70 points) RPS were common based on original SI assessments in non-NP-C EOA controls [16 (19 %) and 8 (9 %), respectively], but scores ≥70 points were far more frequent [46 (98 %)] among NP-C EOA cases. RPS cut-off values provided 98 % sensitivity and 91 % specificity for NP-C at 70-point cut-off, and ROC analysis revealed an AUC of 0.982. Using the 2/3 SI, 90 % of NP-C EOA cases had scores of 2 or 3, and RPS analysis showed an AUC of 0.961. In conclusion, the NP-C SI and the new, quick-to-apply 2/3 SI distinguished well between NP-C and non-NP-C patients, even in EOA populations with high background levels of broadly NPC-compatible multisystemic disease features. While the original SI showed the greatest sensitivity, both tools reliably aided identification of patients with unexplained EOA who warranted further investigation for NP-C.