Can We Use Clinical Presentation and Genetics to Correctly Diagnose Parkinson Disease?

Michael S. Okun, MD reviewing Nalls MA et al. Lancet Neurol 2015 Oct.

A novel diagnostic model is tested.

Interest in developing inexpensive and noninvasive methods to diagnose Parkinson disease (PD) is growing. These authors created a model using Michael J. Fox Parkinson's Progression Marker Initiative data, which included 367 patients diagnosed with PD who had a dopamine transporter scan (DAT) suggestive of PD and 165 controls without neurological diagnoses. The model employed smell testing, genetic risk, family history, age, and gender. The researchers validated the model in five large independent cohorts. They also applied the model to cases where the clinical diagnosis was suggestive of PD, but the DAT scans revealed no evidence of dopaminergic deficit (SWEDD).

In the original cohort, the model's area under the curve for discriminating between PD and no PD was 0.923 (95% confidence interval, 0.900–0.946); sensitivity (0.834; 95% CI, 0.711–0.883) and specificity (0.903; 95% CI, 0.824–0.946) were high. External validation across the five independent cohorts showed adequate classification. The model initially classified 17 SWEDD patients with PD inaccurately; however, at 1-year follow-up, four of these cases converted to a PD diagnosis.

Citation(s):

Nalls MA et al. Diagnosis of Parkinson's disease on the basis of clinical and genetic classification: A population-based modelling study. Lancet Neurol 2015 Oct; 14:1002.