Development of a Nomogram for Predicting Tuberous Sclerosis Complex Genotypes in Children Using Advanced Diffusion MRI and Clinical Data.

Abstract

Rationale and objectives: Tuberous sclerosis complex (TSC) is a multisystem genetic disorder. Focusing on central nervous system manifestations, this study developed an imaging-clinical model combining advanced diffusion MRI parameters with neurological clinical features to distinguish TSC1 vs. TSC2 genotypes.

Materials and methods: Eighty-eight patients newly diagnosed with TSC were enrolled. All underwent a stratified genetic testing strategy comprising whole-exome sequencing, whole-genome sequencing, and tissue-specific deep sequencing. Diffusion spectrum imaging provided parameters from diffusion tensor imaging (DTI), diffusion kurtosis imaging (DKI), neurite orientation dispersion and density imaging (NODDI), and mean apparent propagator MRI (MAP-MRI). A combined prediction model was constructed using logistic regression and validated via bootstrap resampling.

Results: A younger age of onset, autism, neuropsychiatric disorders, intracellular volume fraction, and q-space inverse variance were independently associated with TSC2 mutations. The combined model achieved an AUC of 0.879 (95% CI: 0.841-0.917) in the training set and 0.864 (95% CI: 0.803-0.926) in the validation set. By DeLong's test, it significantly outperformed the clinical model (AUC: 0.637, 95% CI: 0.552-0.723; p < 0.001), while the difference from the imaging model (AUC: 0.833, 95% CI: 0.763-0.903) was not statistically significant (p = 0.068). However, net reclassification (NRI = 0.702, p < 0.001) and integrated discrimination improvement (IDI = 0.097, p < 0.001) both supported the combined model's superior classification ability.

Conclusion: Integrating advanced diffusion MRI parameters with clinical data significantly improves prediction of TSC1 vs. TSC2 genotypes. This combined approach offers valuable support for early diagnosis and personalized treatment in TSC.