Explainable machine learning for long-term outcome prediction in two-center stroke patients after intravenous thrombolysis

 Will you stop predicting failure to recover and just do the GODDAMNED RESEARCH THAT GET SURVIVORS RECOVERED? This is useless for getting survivors recovered! I'd fire you all.

Explainable machine learning for long-term outcome prediction in two-center stroke patients after intravenous thrombolysis

Zheng Ping^1*†, She Huiyu^2†, Li Min^1†, Bai Qingke¹, Lu Qiuyun^3‡ and Chen Xu^3‡

• ¹Department of Neurosurgery, Shanghai Pudong New Area People’s Hospital, Shanghai, China
• ²The Center for Pediatric Liver Diseases, Children’s Hospital of Fudan University, Shanghai, China
• ³Department of Neurology, Shanghai Eighth People’s Hospital, Shanghai, China

Objective: Neurological outcome prediction in patients with ischemic stroke is very critical in treatment strategy and post-stroke management. Machine learning techniques with high accuracy are increasingly being developed in the medical field. We studied the application of machine learning models to predict long-term neurological outcomes in patients with after intravenous thrombolysis.

Methods: A retrospective cohort study was performed to review all stroke patients with intravenous thrombolysis. Patients with modified Rankin Score (mRs) less than two at three months post-thrombolysis were considered as good outcome. The clinical features between stroke patients with good and with poor outcomes were compared using three different machine learning models (Random Forest, Support Vector Machine and Logistic Regression) to identify which performed best. Two datasets from the other stroke center were included accordingly for external verification and performed with explainable AI models.

Results: Of the 488 patients enrolled in this study, and 374 (76.6%) patients had favorable outcomes. Patients with higher mRs at 3 months had increased systolic pressure, blood glucose, cholesterol (TC), and 7-day National Institute of Health Stroke Scale (NIHSS) score compared to those with lower mRs. The predictability and the areas under the curves (AUC) for the random forest model was relatively higher than support vector machine and LR models. These findings were further validated in the external dataset and similar results were obtained. The explainable AI model identified the risk factors as well.

Conclusion: Explainable AI model is able to identify NIHSS_Day7 is independently efficient in predicting neurological outcomes in patients with ischemic stroke after intravenous thrombolysis.

Introduction

Stroke is one of the common neurological diseases, among which ischemic stroke stays about 70–80% of adult stroke, and its incidence rate is rising every year (Benjamin et al., 2019). The average stroke incidence is 120–180/100,000/year, being greater for men (Thayabaranathan et al., 2022). The traditional treatments of ischemic stroke are mechanical thrombectomy and intravenous thrombolysis (Berkowitz et al., 2014; Alet et al., 2020), while the novel treatment strategy include cellular therapy and non-invasive brain stimulation (Richards and Cramer, 2023). However, due to the limited time window for the thrombolysis interventions, generally within 4.5 h and extended 6 h (Qingke et al., 2021), the early prediction of clinical outcomes in stroke patients is essential in post-stroke management.

The prognostic prediction models have been established thereafter (Jiang et al., 2021; Kerleroux et al., 2021). Although object prediction systems, such as ASTRAL (Michel et al., 2010), DRAGON (Wang et al., 2017), and THRIVE (Flint et al., 2014), have been reported to assess the efficiency of intravenous thrombolysis in ischemic stroke patients, most of these scales are based on traditional algorithms with limited clinical features. With recent developments in artificial intelligence, medical machine learning has produced several exciting findings (Jamin et al., 2021; Jayatilake and Ganegoda, 2021). Considering its extended impact on ischemic stroke management, machine learning (ML) models for outcome prediction in patients with intravenous thrombolysis were developed based on the comparison of clinical data according to the modified Rankin score (mRs) at 90 days after thrombolysis. Although some ML studies have identified the risk and protective factors in ischemic stroke (Livne et al., 2018; Heo et al., 2019; Lee et al., 2020; Yang et al., 2020; Ramos et al., 2022), however, the procedure of modeling is hard to be translated to the clinical session. Therefore, in our study, the predictive p-value in each model was further compared and validated in two external datasets. The explainable AI model was added to identify the risk factors as well.

More at link.