The CADASIL scale: a screening tool to select patients for NOTCH3 gene analysis

Background and Purpose—Cerebral autosomal-dominant arteriopathy with subcortical infarcts and leukoencephalopathy (CADASIL) phenotype is highly variable, and, although the full clinical–neuroimaging picture may be suggestive of the disease, no characteristic is pathognomonic. Thus, a genetic test remains the diagnostic gold standard, but because it is costly and time-consuming, a pregenetic screening appears desirable. We aimed at developing the CADASIL scale, a screening tool to be applied in the clinical setting. Methods—A preliminary scale was created assigning weighted scores to common disease features based on their frequencies obtained in a pooled analysis of selected international CADASIL series. The accuracy of the scale versus the genetic diagnosis was tested with receiver operating characteristic analysis after the application of this scale to 61 CADASIL and 54 NOTCH3-negative patients (no pathogenic mutation on exons 2–23 of the NOTCH3 gene). To improve the scale accuracy, we then developed an ad hoc optimization algorithm to detect the definitive scale. A third group of 39 patients affected by sporadic small-vessel disease was finally included in the algorithm to evaluate the stability of the scale. Results—The cutoff score of the definitive CADASIL scale had a sensitivity of 96.7% and a specificity of 74.2%. This scale was robust to contamination of patients with sporadic small-vessel disease. Conclusions—The CADASIL scale is a simple and sufficiently accurate screening tool to select patients with a high probability to be affected by the disease and therefore to be subjected to the genetic testing.