Combined utility of blood glucose and white blood cell in predicting outcome after acute ischemic stroke: The ENCHANTED trial

Interpretable machine learning for prediction of clinical outcomes in acute ischemic stroke

Background and aims

Predicting the prognosis of acute ischemic stroke (AIS) is crucial in a clinical setting for establishing suitable treatment plans. This study aimed to develop and validate a machine learning (ML) model that predicts the functional outcome of AIS patients and provides interpretable insights.

Methods

We included AIS patients from a multicenter stroke registry in this prognostic study. ML-based methods were utilized to predict 3-month functional outcomes, which were categorized as either favorable [modified Rankin Scale (mRS) ≤ 2] or unfavorable (mRS ≥ 3). The SHapley Additive exPlanations (SHAP) method was employed to identify significant features and interpret their contributions to the predictions of the model.

Results

The dataset comprised a derivation set of 3,687 patients and two external validation sets totaling 250 and 110 patients each. Among them, the number of unfavorable outcomes was 1,123 (30.4%) in the derivation set, and 93 (37.2%) and 32 (29.1%) in external sets A and B, respectively. Among the ML models used, the eXtreme Gradient Boosting model demonstrated the best performance. It achieved an area under the receiver operating characteristic curve (AUC-ROC) of 0.790 (95% CI: 0.775-0.806) on the internal test set and 0.791 (95% CI: 0.733-0.848) and 0.873 (95% CI: 0.798-0.948) on the two external test sets, respectively. The key features for predicting functional outcomes were the initial NIHSS, early neurologic deterioration (END), age, and white blood cell count. The END displayed noticeable interactions with several other features.

Conclusion

ML algorithms demonstrated proficient prediction for the 3-month functional outcome in AIS patients. With the aid of the SHAP method, we can attain an in-depth understanding of how critical features contribute to model predictions and how changes in these features influence such predictions.

Association of White Blood Cell Count With Clinical Outcome Independent of Treatment With Alteplase in Acute Ischemic Stroke

Introduction

Higher white blood cell (WBC) count is associated with poor functional outcome in acute ischemic stroke (AIS). However, little is known about whether the association is modified by treatment with intravenous alteplase.

Methods

WAKE-UP was a randomized controlled trial of the efficacy and safety of magnetic resonance imaging [MRI]-based thrombolysis in unknown onset stroke. WBC count was measured on admission and again at 22–36 h after randomization to treatment (follow-up). Favorable outcome was defined by a score of 0 or 1 on the modified Rankin scale (mRS) 90 days after stroke. Further outcome were stroke volume and any hemorrhagic transformation (HT) that were assessed on follow-up CT or MRI. Multiple logistic regression analysis was used to assess the association between outcome and WBC count and treatment group.

Results

Of 503 randomized patients, WBC count and baseline parameters were available in 437 patients (μ = 64.7 years, 35.2% women) on admission and 355 patients (μ = 65.1 years, 34.1% women) on follow-up. Median WBC count on admission was 7.6 × 10⁹/L (interquartile range, IQR, 6.1–9.4 × 10⁹/L) and 8.2 × 10⁹/L (IQR, 6.7–9.7 × 10⁹/L) on follow-up. Higher WBC count both on admission and follow-up was associated with lower odds of favorable outcome, adjusted for age, National Institutes of Health (NIH) Stroke Scale Score, temperature, and treatment (alteplase vs. placebo, adjusted odds ratio, aOR 0.85, 95% confidence interval [CI] 0.78–0.94 and aOR 0.88, 95% CI 0.79–0.97). No interaction between WBC count and treatment group was observed (p = 0.11). Furthermore, WBC count on admission and follow-up was significantly associated with HT (aOR 1.14, 95% CI 1.05–1.24 and aOR 1.13, 95% CI 1.00–1.26). Finally, WBC count on follow-up was associated with larger stroke volume (aOR 2.57, 95% CI 1.08–6.07).

Conclusion

Higher WBC count is associated with unfavorable outcome, an increased risk of HT, and larger stroke volume, independent of treatment with alteplase. Whether immunomodulatory manipulation of WBC count improves stroke outcome needs to be tested.

Trial Registration

ClinicalTrials.gov Identifier: NCT01525290.