iStroke: Racing Against Time in Stroke Care – BioMind Empowers Doctors to Save More Lives

Acute Ischemic Stroke (hereinafter referred to as “AIS”)—a cerebrovascular disease with the highest lifetime incidence rate globally among the Chinese population—is claiming over one million lives annually, with at least one in every five deaths attributable to stroke. Characterized by sudden onset and rapid progression, AIS results in the death of millions of neurons per minute of delay, potentially leading to disability or even death, thereby imposing a substantial burden on both the nation and individual families. A pressing medical challenge that demands immediate attention is how to leverage medical artificial intelligence to enhance the diagnosis and treatment of AIS, improve the management of acute cerebral infarction, optimize prognoses for Chinese patients with acute stroke, reduce disability rates, and alleviate the national economic burden associated with healthcare.

At the 7th Annual Academic Conference of the Chinese Stroke Association and the Tiantan International Stroke Conference (hereinafter referred to as the “Tiantan Conference”), BioMind, a global leader in medical artificial intelligence, unveiled its flagship product, the “iStroke – Intelligent Decision-Making Platform for Reperfusion Therapy in Acute Stroke” (hereinafter referred to as “iStroke”), presenting an AI-driven healthcare solution for the diagnosis and treatment of acute ischemic stroke (AIS).

Timely and effective reperfusion therapy for acute cerebral infarction is the primary treatment modality for Acute Ischemic Stroke (AIS). The clinical objective of improving the quality of reperfusion therapy for AIS has been elevated to a national initiative. On February 20 this year, the National Health Commission (hereinafter referred to as the “NHC”) issued the Notice on the 2021 National Medical Quality and Safety Improvement Goals (hereinafter referred to as the “Goals”). The Goals explicitly set forth ten major objectives, including “increasing the reperfusion therapy rate for acute cerebral infarction,” and mandated that quality control organizations at all levels and across all specialties prioritize the implementation of these Goals as their core task, providing technical support to achieve them. It is reported that in March this year, the National Center for Neurological Disorders Medical Quality Control (Beijing Tiantan Hospital) launched the “National Action for Quality Improvement in Reperfusion Therapy for Acute Cerebral Infarction.”

As a high-tech tool for the “National Action for Quality Improvement in Reperfusion Therapy for Acute Cerebral Infarction,” iStroke, recently launched by Beijing Tiantan Hospital, centers on patients to efficiently coordinate medical resources and optimize emergency care processes. It establishes a highly efficient “one-click green channel” collaborative team for pre-hospital emergency care and in-hospital thrombolysis and thrombectomy, thereby enhancing overall treatment capabilities. This creates an integrated, coordinated care system covering pre-hospital emergency services, in-hospital imaging assessment, reperfusion therapy decision-making, in-hospital secondary prevention, and post-discharge follow-up management.

Professor Miao Zhongrong, Director of the Neurointerventional Center at Beijing Tiantan Hospital, stated, “Reperfusion therapy for acute cerebral infarction, as a primary treatment modality for acute ischemic stroke (AIS), involves administering intravenous thrombolysis and arterial thrombectomy to AIS patients within the therapeutic time window (typically within 6 hours). Identifying suitable candidates among AIS patients presenting beyond the standard time window for interventional therapy remains a significant clinical challenge.” iStroke can extend the effective therapeutic time window from the traditional 6 hours to 24 hours, holding substantial practical significance for “reperfusion therapy in acute cerebral infarction” and providing robust technical support to increase the rate of reperfusion therapy and improve the quality of care for AIS.

This is because iStroke is a one-stop solution for acute ischemic stroke, enabling hemorrhagic lesion identification and hematoma expansion prediction (i-ICH), infarct lesion identification and severity scoring (i-ASPECTS), automatic large vessel occlusion detection (i-LVO), and supporting automated multimodal brain tissue perfusion analysis (i-Mismatch).

Under normal conditions, cerebral blood flow is 70 mL per 100 grams of brain tissue per minute. Once vascular occlusion reduces blood flow to below 20 mL, symptoms such as hemiplegia and aphasia occur; if it falls below 8 mL, infarction ensues. The range between 8 and 20 mL is referred to as the “ischemic penumbra.” Tissue that has already undergone infarction following ischemia is irreversible and untreatable. Therefore, the therapeutic goal in ischemic stroke is to salvage brain tissue at risk of infarction but still viable, namely, the ischemic penumbra.

However, the reality is that formulating a treatment plan within the traditional therapeutic window after an acute ischemic stroke (AIS) onset is no easy task for a considerable number of physicians. Particularly after six hours from symptom onset, when most brain tissue regions have already suffered damage, primary-care physicians often lack sufficient diagnostic capabilities. Coupled with the clinical absence of reliable and accurate automated assessment tools, they are unable to precisely identify the core infarct zone or determine the amount of salvageable brain tissue (penumbra). Consequently, the rate of reperfusion therapy has remained stubbornly low. Since 2015, China’s reperfusion therapy rate for acute cerebral infarction has been far below that of European and American countries; the actual implementation rate of intravenous thrombolysis for acute cerebral infarction patients within 4.5 hours of onset and without contraindications stands at only 22.9%.

Wu Zhenzhou, Chief Technology Officer of Ande Medical Intelligence’s BioMind, stated that iStroke can accurately detect infarct lesions and perform automated, intelligent, and rapid segmentation of blood supply regions, watershed areas, and structural regions according to diverse clinical needs. This enables rapid qualitative and quantitative analysis of infarct lesions, with the multi-region simultaneous segmentation model achieving an accuracy rate of 97.5%.

Specifically, iStroke can identify multiple sequences in CT and MR scans, mine unique features of various lesions from different dimensions such as pixel-level and region-level, and rapidly complete precise segmentation, characterization, and quantitative assessment of the lesions within 5 seconds. It also provides accurate anatomical localization of the lesions, overall lesion volume, and other detailed quantitative assessments, while automatically generating structured reports.

“Leveraging the high-quality medical data from Beijing Tiantan Hospital, iStroke deeply learns the diagnostic and treatment expertise of senior clinical experts at Tiantan Hospital. It supports automated analysis of multimodal imaging, including CT angiography (CTA), CT perfusion (CTP), and MRI, thereby accommodating the equipment capabilities of hospitals at various levels and assisting clinicians in making optimal treatment decisions,” stated Dr. Xiong Yunyun, a researcher at the National Medical Quality Control Center for Neurological Diseases, during her introduction of the iStroke product.

“For patients with AIS, time is life,” said Professor Miao Zhongrong.

However, the reality is that for a significant proportion of patients with acute ischemic stroke (AIS), the interval from arrival at the emergency department to the neurologist’s emergency treatment decision alone can take as long as 100 minutes, not including prehospital transport time, thereby often causing them to miss the golden window for treatment.

“iStroke is expected to halve preoperative decision-making time through technological support, enabling more patients with acute ischemic stroke (AIS) to receive timely treatment,” stated Professor Miao Zhongrong. He emphasized that increasing the rate of reperfusion therapy for acute cerebral infarction helps reduce disability and mortality rates among these patients, improves their quality of life, and alleviates the burden on society and families.

Wu Zhenzhou further explained that iStroke can significantly shorten the time required for imaging diagnosis, providing a rapid and precise tool for the automated imaging assessment of acute ischemic stroke (AIS), with final evaluation and surgical intervention performed directly by neurologists. iStroke transforms the in-hospital emergency care process for AIS patients into a hub-and-spoke model: with iStroke as the central hub, imaging analysis results are instantly fed back to physicians across various departments, while simultaneously sharing patient intake information, AI-derived imaging data, and critical clinical laboratory results. By replacing the traditional linear workflow, this approach can substantially improve the efficiency of AIS treatment, enabling more patients to receive timely care.

Meanwhile, iStroke supports automated multimodal imaging analysis, including intelligent CTA (CT angiography)/CTP (CT perfusion) analysis and intelligent MRI analysis, meeting the equipment requirements of hospitals at various levels.

In the future, iStroke will comprise a PC platform, a mobile app, smart wristbands, and AI-powered medical image analysis servers, forming a comprehensive one-stop assistance platform for the treatment of stroke patients.

“For example, through its companion app, iStroke provides mobile smart medical services for stroke emergency care. Family members can promptly enter the patient’s basic information and time of symptom onset; based on this data, physicians assess the likelihood of stroke and activate the emergency green channel with a single click; the stroke map automatically recommends the nearest hospital; and the patient’s information along with the ambulance’s real-time location is pushed to the designated hospital, enabling the patient to proceed directly to the CT suite for imaging upon arrival,” said Wu Zhenzhou.

According to the “2020 Report on the Development of Artificial Intelligence in Medical Imaging in China,” the current hospital-side paid penetration rate for AI medical imaging stands at 4.5%–7%. As the value of AI medical imaging products gains greater recognition and hospitals’ willingness to pay increases, the penetration rate is projected to reach 15%–20% by 2023.

China’s stroke treatment market has already exceeded RMB 100 billion, but its future potential extends far beyond this scale. By the end of 2020, 562 Advanced Stroke Centers had been established nationwide, and 607 district- and county-level hospitals were routinely providing thrombolytic therapy. In addition, the Stroke Emergency Map system has been implemented in more than 170 cities across 26 provinces, covering over 2,000 medical institutions and emergency centers. Building on this foundation, the Stroke Prevention and Treatment Engineering Committee of the National Health Commission launched the “Thousand Counties and Ten Thousand Towns Stroke Recognition Initiative” and the “Hundred Cities, Hundred Maps Construction Campaign” for the China Stroke Emergency Map. This means that stroke centers will proliferate throughout China, extending their reach deep into grassroots communities.

Ande Yizhi BioMind has launched major industry products at the Tiantan Conference for four consecutive years. Last October, Ande Yizhi BioMind released an auxiliary diagnostic product for cerebral aneurysms at the Tiantan Conference.

Medical AI has entered a new era of integrated diagnosis and treatment, creating significant market opportunities for multimodal, whole-body AI products. BioMind adopts a dual-pronged approach combining “Imaging-Assisted Diagnosis” and “Clinical Treatment Decision-Making,” enabling AI empowerment across the entire continuum from diagnosis, decision-making, and treatment to prognosis. Compatible with imaging modalities such as CT, MRI, and ultrasound, it comprehensively supports AI analysis of multimodal imaging data, whole-body-level AI application deployment, and AI-driven responses to comprehensive healthcare needs, promising substantial future market potential.

Wu Zhenzhou believes that under the tiered diagnosis and treatment policy, AI products can only gain market recognition and demonstrate their true value by genuinely assisting radiologists and clinicians in managing acute, critical, and severe diseases, and by providing grassroots hospitals with comprehensive, integrated diagnostic and therapeutic solutions.