Development and internal validation of a clinical risk score to predict incident renal and pulmonary tumors in people with tuberous sclerosis complex
Frédéric Loubert1, Andrew House2, Catherine Larochelle3, Philippe Major4, Mark Keezer5
1Université de Montréal, 2London Health Science Centre, 3CRCHUM, 4CHU Ste-Justine PO SA178275-D, 5Centre Hospitalier Universite de Montreal
Objective:
Develop and internally validate a clinical risk score to predict incident renal angiomyolipoma (AML) and pulmonary lymphangioleiomyomatosis (LAM) in people with tuberous sclerosis complex (TSC).
Background:
TSC affects approximately 1/6000 live births. The primary sources of morbidity and mortality in children are manifestations affecting the brain and heart. Adolescents and adults, on the other hand, are at the greatest risk of developing AML and LAM.
Design/Methods:
We will leverage data from 2,420 participants in the TSC Alliance Natural History Database for these analyses. This is the largest international research database for TSC. We will use Cox regression models to predict AML and LAM development using 13 brain, heart, and cutaneous manifestations of TSC as potential predictors. We will use the results of these regression models to develop clinical risk scores. These will then be internally validated using bootstrap resampling, measuring discrimination and calibration.
Results:
We have acquired access to all required data and our analyses are underway, to be completed by the end of 2022. So far, descriptive analyses have shown that 426 out of 2,420 individuals in our data developed AML, while 77 developed LAM. Preliminary analyses have demonstrated excellent calibration for AML, and discrimination that is either good or strong for AML and LAM (c-statistics from 0.72 to 0.81).
Conclusions:
It is currently unknown whether the clinical TSC features during the childhood of an individual can accurately estimate the risk of developing AML or LAM during adolescence and adulthood. Our work will be an important step towards identifying individuals who would benefit from prophylactic treatments (e.g. with mTOR inhibitors) and lead to more tailored schedules for screening imaging. We expect that our work will allow for more personalized medicine in people with TSC.