A Multi-Center Study: Developing a Nomogram for Predicting Genetic Results of Trio-Based Whole-Exome Sequencing (Trio-WES) in Diagnosing Children with Syndromic Neurodevelopmental Disorders (s-NDDs)

Objective: The diagnostic efficacy of empirical trio-based whole-exome sequencing (trio-WES) for unexplained syndromic neurodevelopmental disorders (NDDs) remains unsatisfactory. This study aimed to explore the diagnostic value of phenotypic indicators and establish a nomogram for predicting the genetic results of trio-WES for diagnosing patients with unexplained s-NDDs. Methods: We retrospectively collected phenotypic and genotypic data from 265 children with s-NDDs who received trio-WES at Sun Yat-sen Memorial Hospital (training cohort), 38 and 97s-NDDs patients with trio-WES test at Shenshan Medical Center (validation cohort-1) and Weierkang Children’s Rehabilitation Center (validation cohort-2), respectively. Logistic analysis was employed to identify the independent predictors of a positive trio-WES diagnosis in the training cohort for model establishment. The predictive performance and robustness of the model were assessed using receiver operating characteristic (ROC) and confusion matrix analyses, respectively. Results: The analysis revealed that the severity of neurodevelopmental delays, head circumference abnormality, and complexity of neurodevelopmental comorbidities were independent predictive indicators for distinguishing s-NDDs patients with positive trio-WES results. The nomogram combining the three predictors showed good predictive performance with an area under the ROC (AUC) in the training cohort of 0.827 (95% CI: 0.775– 0.879), yielding a confusion matrix with sensitivity, specificity, accuracy, precision, and F1 score of 78.23%, 78.01%, 78.11%, 75.78%, and 0.77, respectively. The model also had an excellent prediction in the external validation cohorts (AUC: 0.953; 95% CI: 0.881– 0.998, sensitivity: 88.89%; specificity: 80.00%; accuracy: 84.21%; precision: 80.00%; and F1 score: 0.84 in validation cohort-1 and AUC: 0.910; 95% CI: 0.843– 0.978, sensitivity: 83.33%; specificity: 85.45%; accuracy: 84.54%; precision: 81.40%; and F1 score: 0.82 in validation cohort-2). Conclusion: The model can serve as a useful tool for assisting decision-making in applying trio-WES in the diagnostic strategy for s-NDDs, helping to implement personalized pre-diagnosis assessments for affected families.