Characteristic | Overall | Development set | External testing set |
|---|
Number of patients | 364 | 336 | 28 |
Age (y ± SD)* | 64.83 ± 6.72 | 64.88 ± 6.71 | 64.18 ± 6.83 |
Female sex (%) | 126 (34.62%) | 117 (34.82%) | 9 (32.14%) |
CSVD type |
 aSVD (%) | 217 (59.62%) | 203 (61.42%) | 14 (50.00%) |
 CAA (%) | 106 (29.12%) | 99 (29.46%) | 7 (25.00%) |
 CADASIL (%) | 41 (11.26%) | 34 (10.12%) | 7 (25.00%) |
- *mean ± standard. The numbers in parentheses are the percentages. Unless otherwise specified the data indicate the number of participants. The susceptibility-weighted MR sequences (SWS) of patients from the Third Affiliated Hospital, Sun Yat-sen University (SYSUTH) and Maoming People’s Hospital (MMPH) served as the development set, and the SWS of patients from the First Affiliated Hospital of SHANTOU University Medical College (STUMFH) served as the external test set for testing the geographic performance of the model. The tenfold cross-validation was used to evaluate the stability of the deep learning (DL) model
- y year, SD standard deviation, CSVD cerebral small vessel disease, aSVD arteriolosclerosis, CAA cerebral amyloid angiopathy, CADASIL cerebral autosomal dominant arteriopathy with subcortical infarcts and leukoencephalopathy
