Spontaneous Intracerebral Hemorrhage
Conditions
Keywords
Normative analysis; Lesion-symptom mapping; Radiomics; Etiological classification; Machine learning
Brief summary
The investigators retrospectively collected participants with spontaneous cerebral hemorrhage(sICH) from January 2015 to December 2019 for training and internal validation. Clinical and imaging data were collected. Modified Rankin Scale (mRS) scores were determined good outcome as mRS = 0-2, poor outcome as mRS = 3-6. The location features of sICH were extracted by symptom mapping. Noncontrast computed tomography images of patients and hematoma masks were registered with standard human brain templates to identify specific affected brain regions. Then a probability map of hemorrhage for different causes and prognosis is generated. PyRadiomics was used to extract the radiomic features, integrate radiomic and clinical features into multiple logistic regression models, and develop and validate optimal etiological and prognostic models. Further tests were performed in an independent cohort. The area under the working characteristic curve (AUC), sensibility, specificity were used to evaluate the reliability of the model.
Interventions
None listed
Sponsors
Study design
Eligibility
Inclusion criteria
* The confirmation of intracranial hemorrhage was achieved through NCCT.
Exclusion criteria
* They presented with secondary hemorrhage resulting from head trauma, hemorrhagic transformation of ischemic infarction, brain tumors, or exhibited abnormalities in blood coagulation, liver, kidney function, or were subject to drug-induced cerebral hemorrhage * Only postoperative CT images were available * Lost follow-up records * CT image artifacts
Design outcomes
Primary
| Measure | Time frame | Description |
|---|---|---|
| Modified Rankin Scale (mRS) scores | follow-ups at twelve months post sICH event | Good outcome as mRS = 0-2, poor outcome as mRS = 3-6 |
Countries
China