An AI-powered software to enhance stroke treatment in Pakistan

Low and middle-income countries (LMICs) face 40% more stroke incidences and 74% higher chances of death or disability compared to high-income nations.

In Pakistan, this challenge is compounded by a critical shortage of neurologists, with only one specialist per million people (Farooq et al., 2021b). The estimated stroke rate is 250 per 100,000, resulting in 350,000 new cases annually (Pakstrokesociety, 2019; Farooq et al., 2021b). About 80% of these strokes are ischemic, requiring urgent intervention within a 4.5-hour window to prevent fatal or disabling outcomes.

However, in these conditions, analyzing multiple data (imaging, clinical, demographic) within this short time frame is complex and requires expertise and human resources. Therefore, an automated system capable of performing this analysis could help clinicians make timely treatment decisions.

The project aims to develop AI-powered cloud-based software to enhance the accuracy and time burden of stroke treatment decisions. Indeed, AI models can quickly analyze multimodal data, enabling optimal decision-making. By improving decision-making, the program aims to increase patient outcomes and throughout. The AI is based on an input-output system. AI can be trained using imaging, demographics and clinical data from past patients, and after training can analyze the new patient data and produce a forecast.

The software is designed to integrate into existing hospital workflows without requiring additional hardware or complex IT setups. The solution initially targets hospitals with CT/CTA scan facilities and established stroke care protocols, high patient volumes, and readiness to integrate AI into clinical workflows. The system will be accessible via Web, allowing healthcare workers to use it on existing hospital computers, tablets, or mobile devices.

By introducing an AI-powered imaging diagnostic tool, the project aims to strengthen stroke care in Pakistan and support neurologists in making faster, more accurate treatment decisions. By the end of the implementation phase, the project expects to achieve the following:

• AI-generated imaging diagnostics available to neurologists and clinicians in stroke units, enabling quicker and more reliable decision-making.
• Shorter intervention times, reducing stroke-related mortality and long-term disability.
• Improved quality of care and reduced treatment delays through the integration of AI support in clinical workflows.
• Lower burden on hospitals and healthcare personnel, with fewer complications and follow-up treatments required.
• Demonstrated cost-effectiveness through time savings and better allocation of medical resources. Ultimately, the solution is aiming at contributing to democratize access to AI-driven stroke care, extending its benefits to remote and underserved areas, and ultimately, contribute to decreasing the prevalence of severe disability and mortality due to stroke in Pakistan.

_Photo credit: National cancer institut_e