Predicting the conversion from clinically isolated syndrome to multiple sclerosis: An explainable machine learning approach.

INTRODUCTION: Predicting the conversion of clinically isolated syndrome (CIS) to clinically definite multiple sclerosis (CDMS) is critical to personalizing treatment planning and benefits for patients. The aim of this study is to develop an explainable machine learning (ML) model for predicting this conversion based on demographic, clinical, and imaging data. METHOD: The ML model, Extreme Gradient Boosting (XGBoost), was employed on the public dataset of 273 Mexican mestizo CIS patients with 10-year follow-up. The data was divided into a training set for cross-validation and feature selection, and a holdout test set for final testing. Feature importance was determined using the SHapley Additive Explanations library (SHAP). Then, two experiments were conducted to optimize the model's performance by selectively adding variables and selecting the most contributive variables for the final model. RESULTS: Nine variables including age, gender, schooling, motor symptoms, infratentorial and periventricular lesion at imaging, oligoclonal band in cerebrospinal fluid, lesion and symptoms types were significant. The model achieved an accuracy of 83.6 %, AUC of 91.8 %, sensitivity of 83.9 %, and specificity of 83.4 % in cross-validation. In the final testing, the model achieved an accuracy of 78.3 %, AUC of 85.8 %, sensitivity of 75 %, and specificity of 81.1 %. Finally, a web-based demo of the model was created for testing purposes. CONCLUSION: The model, focusing on feature selection and interpretability, effectively stratifies risk for treatment decisions and disability prevention in MS patients. It provides a numerical risk estimate for CDMS conversion, enhancing transparency in clinical decision-making and aiding in patient care.

Congenital Heart Defects in Pregnancies Conceived by Assisted Reproductive Technology: Comparing Functional and Structural Defects.

Introduction Congenital heart defects (CHD) are one of the most common congenital anomalies, and their association with assisted reproductive technology (ART) is controversial in different populations. The purpose of this study was to evaluate this association and to provide information about the necessity of specialized echocardiography during pregnancy with ART. Methods This retrospective study was performed on all pregnancies conceived by ART and referred for fetal echocardiography to the Rasoul Akram and Akbar Abadi hospitals in Tehran, Iran. A total of 109 patients were enrolled in the study (56 in the ART group and 53 in the non-ART). Two-dimensional and color Doppler echocardiography were performed on all patients to identify heart problems and anomalies and medical records of the patients were reviewed. The outcome was considered the presence of functional and structural heart defects on echocardiography. Results The study groups were similar in terms of maternal age and GA. The ART group consisted of 31 singletons (55%) and 25 multiples (45%). All pregnancies in the non-ART group were singletons. Following in vitro fertilization (33%), ovulation induction (25%) was the next most used method. The findings of echocardiography were one atrial septal defect (ASD) in ART and one in non-ART, six ventricular septal defects (VSD) in ART and three in non-ART, and one ASD and VSD in the ART group. These structural abnormalities showed no difference in the two groups (P value = 0.58). There was no significant difference in rhythm between the two groups (P = 0.51). Echocardiographic indices of both groups did not differ statistically except in the TR-PG index (P value = 0.02). Conclusions The structural defects of the two groups were not different, and no heart dysfunction was observed in ART fetuses. There was no association between ART and CHD in our study. We concluded that echocardiography by pediatric cardiologists is not necessary for these fetuses.