To enhance diagnostic accuracy of the RiSAT by integrating a weighted scoring system with real-time statistical calculations and likelihood ratios enhancing stroke diagnostic accuracy.

Prompt identification of ischemic stroke is critical in guiding management. However, access to CT angiography is limited in many rural/constrained settings delaying intervention and disproportionately affecting outcomes; imaging utility may also be limited due to renal impairment or smaller/lacunar strokes. To address these limitations, we developed RiSAT—a questionnaire software that leverages clinical findings to localize ischemic strokes. In this study, we refined RiSAT by introducing a dynamic, evidence-weighted scoring algorithm.

The initial binary algorithm was enhanced using a comprehensive literature review to assign weighted values to each symptom according to its published likelihood ratio and statistical significance for a given vascular territory. Data from over 2000 stroke cases were used to calibrate weights. Real-time statistical calculations were incorporated for enhanced diagnostic confidence.

For validation, over 150 external stroke cases were then chosen and ongoing testing is performed via a triple blind system:

• First researcher extracted clinical data.

• Second researcher entered data into RiSAT.

• Third researcher compared predicted versus confirmed stroke locations.

Preliminary results from the first 76 patients in our registry demonstrated the algorithm had an overall sensitivity of 94%, specificity of 88%. Interestingly, isolated unilateral leg weakness was most strongly associated with an ACA infarction with an LR+ >10 and LR- 0.2, and RRR of >50 within a 95% CI. Similar analyses have been performed for remaining focal neurological symptoms, and these results will be presented in the full dataset.

The incorporation of a weighted scoring and real-time statistical feedback substantially improved RiSAT’s diagnostic performance relative to prior studies. This approach may further enhance confidence in stroke localization using bedside findings, particularly valuable when imaging is delayed, unavailable, or inconclusive.