Artificial Intelligence Techniques-based Prognostic Research In Patients With Intracerebral Hemorrhage

Section Ⅰ:Constructing A Deep Learning-Based Prognostic Prediction Model for Hemorrhagic StrokeBackground:The prognosis following a hemorrhagic stroke is usually extremely poor.Rating scales have been developed to predict the outcomes of patients with intracerebral hemorrhage(ICH).To date,however,the prognostic prediction models have not included the full range of relevant imaging features.We constructed a clinic-imaging fusion model based on convolutional neural networks(CNN)to predict the short-term prognosis of ICH patients.Methods:This was a multi-center retrospective study,which included 1,990 patients with ICH.Clinical and imaging data were collected from the Chinese Intracranial Hemorrhage Image Database(CICHID)for analysis.Two CNN-based deep learning models were constructed to predict the neurofunctional outcomes at discharge;these were validated and tested using a nested 5-fold cross-validation approach.The models’ predictive efficiency was compared with the original ICH scale and the ICH grading scale.Poor neurological outcome was defined as a Glasgow Outcome Scale(GOS)score of 1-3.Additionally,we used multi-modal attention module to derive visual explanation by localizing the image area that most influences the decision made by our CNN model.Results:The training and test sets included 1,599 and 391 patients,respectively.For the test set,the clinic-imaging fusion model had the highest area under the curve(AUC=0.903),followed by the imaging-based model(AUC=0.886),the ICH scale(AUC=0.777),and finally the ICH grading scale(AUC=0.747).The multi-modal attention module can emphasis on the perihematomal edema area in CT image.Conclusions:The CNN prognostic prediction model based on neuroimaging features was more effective than the ICH scales in predicting the neurological outcomes of ICH patients at discharge.The CNN model’s predictive efficiency improved when clinical data were included.The image features of perihematomal edema area are relevant to the prognostic value.Therefore,it is necessary to further verify the relationship between the perihematomal edema formation and the neurological prognosis of ICH.Section Ⅱ:Construction and Validation of Perihematomal Edema Expansion as A New Imaging Biomarker to Predict Clinical Outcome in Patients with Intracerebral HemorrhageBackground:Perihematomal edema(PHE)has been considered a potential marker of secondary brain injury after intracerebral hemorrhage(ICH).Enrolling early PHE expansion to redefine prognosis-associated ICH expansion merits exploration.Additionally,we attempt to generate a definition of delayed perihematomal edema expansion(DPE)and analyze its time course,risk factors,and clinical outcomes,to provide references for risk stratification and prognostic evaluation.Methods:A multi-cohort data was derived from the Chinese Intracranial Hemorrhage Image Database(CICHID).A non-contrast computed tomography(NCCT)-based deep learning model was constructed for fully automated segmentation hematoma and PHE.Time course of hematoma and PHE evolution correlated to initial hematoma volume was volumetrically assessed.Logistic regression analysis was utilized to identify risk factors for poor outcomes.Receiver operating characteristic curve(ROC)analysis was performed to calculate the predictive values of PHE expansion and hematoma expansion(HE).Predictive values for DPE were calculated through ROC analysis and were tested in an independent cohort.Results:For early PHE expansion analysis,223 target patients were enrolled.Multivariable analysis showed the early PHE expansion is the independent risk factors for poor prognosis.The predictive value of absolute PHE expansion(AUC=0.776,sensitivity=67.9%,specificity=77.0%)was higher than that of absolute HE(AUC=0.573,sensitivity=41.7%,specificity=87.1%)and HE(>6 ml)(AUC=0.594,sensitivity=23.8%,specificity=95.0%).The best cutoff for early absolute/relative PHE expansion resulting in a poor outcome was 5.96 ml and 31%.For DPE analysis,312 patients were enrolled.The best cutoff for DPE to predict poor outcome was 3.34 mL of absolute PHE expansion from 4-7 days to 8-14 days(AUC=0.784,sensitivity=72.2%,specificity=81.2%),and 3.78 mL of absolute PHE expansion from 8-14 days to 15-21 days(AUC=0.682,sensitivity=59.3%,specificity=92.1%)in the derivation sample.Patients with DPE was associated with worse outcome(OR:12.340,95%CI:6.378-23.873,P<0.01),and the larger initial hematoma volume(OR:1.021,95%CI:1.000-1.043,P=0.049)was the significant risk factor for DPE formation.Conclusions:Early PHE expansion or early combined ICH expansion(hematoma combined with PHE)was associated with a poor outcome,characterized by a better predictive value than HE.A new definition of DPE was generated and is confirmed to be related to poor outcomes in ICH.Section Ⅲ:A Machine Learning Approach for Predicting Perihematomal Edema Expansion in Patients with Intracerebral HemorrhageBackground:Preventing the expansion of perihematomal edema(PHE)represents a novel strategy for the improvement of neurological outcomes in intracerebral hemorrhage(ICH)patients.Our goal was to predict early and delayed PHE expansion using a machine learning approach.Methods:We enrolled 550 patients with spontaneous ICH to study early PHE expansion,and 389 patients to study delayed expansion.Two imaging researchers rated the shape and density of hematoma in non-contrast computed tomography(NCCT).We trained a radiological machine learning(ML)model,a radiomics ML model,and a combined ML model,using data from radiomics,traditional imaging and clinical indicators.We then validated these models on an independent dataset by using a nested 4-fold cross-validation approach.We compared models with respect to their predictive performance,which was assessed using the receiver operating characteristic curve.Results:For both early and delayed PHE expansion,the combined ML model was most predictive(early/delayed AUC values were 0.840/0.705),followed by the radiomics ML model(0.799/0.663),the radiological ML model(0.779/0.631)and the imaging readers(reader 1:0.668/0.565,reader 2:0.700/0.617).Conclusions:We validated a machine learning approach with high interpretability for the prediction of early and delayed PHE expansion.This new technique may assist clinical practice for the management of neurocritical patients with ICH.