Related Experiment Video
Updated: Oct 17, 2025
Computerized Adaptive Testing System of Functional Assessment of Stroke
Published on: January 7, 2019
A Web-based System to Assist With Etiology Differential Diagnosis in Children With Arterial Ischemic Stroke
Anjini Karthik1, Bin Jiang1, Ying Li1
1Department of Radiology, Neuroradiology Section, Stanford University, Stanford, CA.
Insights
A new web tool aids in diagnosing pediatric arterial ischemic stroke by analyzing clinical and imaging data. It provides a differential diagnosis for arteriopathy, improving diagnostic accuracy in children.
Area of Science:
- Pediatric Neurology
- Vascular Neurology
- Medical Informatics
Background:
- Childhood arteriopathy diagnosis is challenging.
- Pediatric arterial ischemic stroke (AIS) requires accurate etiological classification.
- Existing diagnostic methods can be complex and time-consuming.
Purpose of the Study:
- To develop and evaluate a web-based classification system for diagnosing pediatric AIS.
- To create an evidence-backed tool that provides a differential diagnosis for arteriopathy subtypes.
- To improve the accuracy and timeliness of diagnosing arteriopathy in children.
Main Methods:
- A prospective cohort of 355 children with AIS was analyzed.
- A logistic regression model used clinical and imaging data from a derivation cohort (174 patients) to identify arteriopathy subtypes.
- A web interface was developed to provide a probabilistic differential diagnosis, tested on a separate cohort (34 patients).
Main Results:
- The tool's differential diagnosis completely agreed with expert opinions in 20.6% of cases.
- Partial agreement was observed in 41.2% of patients, with overlap in 29.4%.
- The tool disagreed with expert diagnoses in only 8.8% of patients.
Conclusions:
- The developed web tool assists in classifying pediatric AIS, yielding overlapping differential diagnoses in many cases.
- The tool facilitates high-quality, timely diagnoses of arteriopathy in pediatric patients.
- Further validation in an independent cohort is recommended.
Background And Purpose:
The diagnosis of childhood arteriopathy is complex. We present a Web-based, evidence-backed classification system to return the most likely cause(s) of a pediatric arterial ischemic stroke. This tool incorporates a decision-making algorithm that considers a patient's clinical and imaging features before returning a differential diagnosis, including the likelihood of various arteriopathy subtypes.
Methods:
The Vascular Effects of Infection in Pediatric Stroke study prospectively enrolled 355 children with arterial ischemic stroke (2010-2014). Previously, a central panel of experts classified the stroke etiology. To create this tool, we used the 174 patients with definite arteriopathy and spontaneous cardioembolic stroke as the "derivation cohort" and the 34 with "possible" arteriopathy as the "test cohort." Using logistic regression models of clinical and imaging characteristics associated with each arteriopathy subtype in the derivation cohort, we built a decision framework that we integrated into a Web interface specifically designed to create a probabilistic differential diagnosis. We applied the Web-based tool to the "test cohort."
Results:
The differential diagnosis returned by our tool was in complete agreement with the experts' opinions in 20.6% of patients. We observed a partial agreement in 41.2% of patients and an overlap in 29.4% of patients. The tool disagreed with the experts on the diagnoses of 3 patients (8.8%).
Conclusions:
Our tool yielded an overlapping differential diagnosis in most patients that defied definitive classification by experts. Although it needs to be validated in an independent cohort, it helps facilitate high-quality, and timely diagnoses of arteriopathy in pediatric patients.
