Objective:

To establish the diagnostic accuracy of tuberous sclerosis complex (TSC) clinical features to better guide the screening process for the families of people with TSC (PwTSC).

Background:

Up to 10-15% of PwTSC do not carry an identifiable pathological genetic variant and are diagnosed clinically. In such situations, family members cannot be screened using genetic testing. 

Design/Methods:

We used the TSC Natural History Database, a longitudinal database of PwTSC from 22 North American centers. We used a definite genetic TSC diagnosis as our gold standard. We estimated the sensitivity (95% confidence interval [CI]) of TSC-related skin, structural brain, renal, and cardiac manifestations, as well as combinations of these manifestations. Using a series of sensitivity analyses to test alternate assumptions, we estimated positive and negative predictive values (PPV and NPV).

Results:

Among 1,300 PwTSC, 50.3% were female, and the mean age at diagnosis was 3.7 years. The sensitivity of any skin or structural brain manifestation was 98.7% (95% CI: 98.0%, 99.2%). The PPV and NPV were 83.2 (95% CI: 81.6%, 84.6%) and 98.4% (95% CI: 97.8%, 98.8%), respectively, while assuming 50% prevalence and 80% specificity. Including cardiac manifestations marginally increased the sensitivity, PPV, and NPV to 99.5% (95% CI: 98.9%, 99.7%), 83.3% (95% CI: 81.8%, 84.7%), and 99.4% (95% CI: 99.0%, 99.6%), respectively. Other combinations of TSC manifestations had lower or similar diagnostic accuracy.

Conclusions:

Screening for brain and skin manifestations in family members of genetic-negative PwTSC is sufficient, with excellent sensitivity and NPV. Further cardiac screening may be indicated in select cases.