How New Infrastructure Drives the Next Evolution of Medical AI: Industry Experts Reveal the Logic Behind the Transformation

Counting from the planning initiated in 1990, China’s path of large-scale infrastructure construction has spanned three decades. As time entered the second decade of the 21st century, the GDP growth driven by traditional infrastructure—namely railways, highways, and bridges—showed signs of fatigue. Redundant construction and rising marginal costs of infrastructure have not only hindered the efficient flow of resources but also risked fueling bubbles and exacerbating industrial risks during economic downturns. In response, the central government introduced the concept of “New Infrastructure” to guide the transformation of traditional manufacturing toward high-end sectors.

Yet industrial transformation is inherently slow, and few cities are willing to take the first bold step—until the COVID-19 pandemic emerged as a catalyst. At this juncture, “New Infrastructure,” symbolizing the advancement of high-end technologies and new drivers of economic growth, became a lifeline for changing the status quo.

The underlying reasons can be summarized into two factors: first, low production efficiency resulting from the lack of innovative technologies such as automation, necessitating an acceleration of industrial transformation; second, the visibly slowing economic downturn on a global scale.

So, why can “new infrastructure” solve the above two problems?

Ultimately, we can now foresee the technological development logic of the next era. Developing technologies such as 5G, AI, and IoT is an inevitable path to accelerating the construction of smart cities, smart industries, and smart healthcare. The first logic can be described in the words of Ben Einstein, founder of the U.S. industrial machine learning startup Instrumental: “The pandemic has made us realize the necessity of automation. In electronics manufacturing, the market is driving us to achieve five years’ worth of innovation within the next 18 months.”

Second is policy-driven momentum. On May 7, 2020, the Shanghai Municipal People’s Government released the Action Plan for Promoting the Construction of New Infrastructure in Shanghai (2020–2022). The document identified and assessed the first batch of 48 major projects and project packages to be implemented over the following three years under the new infrastructure initiative, with total investment estimated at RMB 270 billion. At that time, Shanghai became the first city worldwide to issue a concrete plan for new infrastructure development.

Amidst the current wave of infrastructure development, healthcare has undoubtedly emerged as a key focus area. The 5G networks, big data centers, and artificial intelligence (AI) highlighted in this transformative initiative have all been central to the advancement of medical technology in recent years. In particular, AI is increasingly integrating with 5G and big data, synergistically enhancing diagnostic capabilities at primary care levels through telemedicine and aiding in the discovery of hidden insights through advanced data mining.

Today, with policies and strategic preparations in place, medical AI companies stand at a crossroads of development.

The use of artificial intelligence to assist physicians in diagnostic decision-making is no longer novel; medical AI companies such as YITU Medical have long been

Deploying relevant applications in hundreds of Grade 3A hospitals or at the primary care level is not the norm for new infrastructure development.

In the words of Fang Cong, Vice President of YITU Medical, the “new” in new infrastructure lies in the cross-boundary technological integration and the closed-loop connectivity of point-specific applications.

Taking AI in medical imaging as an example, if one merely develops a standalone application for image recognition, it will remain confined to the laboratory. To break through spatial limitations, AI can be integrated with 5G technology to extend diagnostic capabilities to grassroots areas with communication infrastructure. To overcome procedural constraints, AI can be combined with big data to develop applications covering all diseases of a single organ. Cross-boundary integration gives rise to emerging capabilities.

Point-based application of closed-loop connectivity represents a state that transcends the current landscape of artificial intelligence. From a horizontal perspective, this requires integrating the entire patient journey by combining medical history, clinical symptoms, physical signs, and other relevant auxiliary examinations. From a vertical perspective, since a single disease may manifest across multiple organs, it is essential to adopt a systematic approach to assessing the patient’s condition.

Although no finished products have been developed, research in both directions is ongoing.

Taking the intelligent emergency and rehabilitation management system for stroke developed by Yitu Medical as an example, this system leverages the advantages of 5G technology and Yitu Medical’s full-stack product matrix. It applies 5G communication to the 120 cloud platform to enable AI-powered cloud-based intelligent consultations, thereby assessing the likelihood of stroke.

Covering the entire patient care journey, the process begins with the rational deployment of dedicated stroke ambulances for 120 emergency services. Upon patient pickup, appropriate receiving hospitals are selected based on a stroke map. Pre-hospital diagnosis and CT scans can be performed within the ambulance, with data synchronized in real time to the stroke center at the target hospital. The stroke center leverages AI to interpret CT images within seconds, determining subsequent clinical management plans. Meanwhile, the ambulance utilizes 5G technology for intelligent traffic route planning to expedite transport to the target hospital, enabling remote consultations and initiating essential patient family education en route. Post-discharge, human keypoint recognition technology facilitates out-of-hospital rehabilitation, thereby reducing the burden on patients.

YITU has also applied its integration with 5G technology to the field of early cancer screening. Due to a shortage of radiologists and other factors, primary care hospitals lag behind urban facilities in terms of image interpretation efficiency, accuracy, and consistency. However, limited network transmission capabilities have prevented radiological images generated at primary healthcare institutions from being transmitted promptly to regional or third-party reading centers, thereby significantly restricting the application of remote image interpretation. With its high bandwidth, low latency, and high reliability, 5G communication technology enables the massive, low-latency transmission of data. This capability promises to facilitate the real-time transmission of imaging information—previously confined to intra-hospital sharing—and support real-time diagnosis, thus unlocking the potential for real-time remote consultation and treatment.

“However, these applications still place high demands on infrastructure development,” said Fang Cong. “Currently, the practical application of AI technology remains focused on addressing isolated needs; only by accelerating the establishment of a robust foundational layer can we build more effective advanced-level applications.”

Stripping away the conceptual layers, “New Infrastructure” must still return to its core essence: “infrastructure.” In the realm of general health, the foundational pillars for the people are the systems for managing critical illnesses, chronic diseases, pediatrics, and responding to major public health emergencies. It is upon these areas that medical new infrastructure must be firmly grounded.

“For a long time, critical illnesses have continuously imposed payment pressures on the medical insurance system and placed significant life burdens on patients. Taking malignant tumors as an example, although technological advancements have improved survival rates for patients in advanced stages, the optimal approach remains to detect and treat them at the earliest possible stage. This is where early screening at the primary care level becomes crucial,” stated Fang Cong. “Such early screening must be implemented at the grassroots level. Given China’s large population, secondary hospitals and above cannot possibly shoulder such a massive volume of early screenings.”

To realize this vision, the greatest challenge stems from the scarcity of medical resources, which is why Yitu Healthcare has developed an AI-assisted diagnostic system for primary care institutions. During the development of new infrastructure, Yitu Healthcare needs to standardize this screening approach and integrate it with the public health defense system.

“Take our AI Cancer Prevention Map as an example. Early screening for critical illnesses typically stratifies the population by age; for instance, the national guideline recommends annual screening for men aged 55 and older and women aged 50 and older. Our objective is to assist the public health system in establishing a screening database while providing an AI-assisted diagnostic system. Through data accumulation over several years, we can calculate the prevalence of various tumors within specific urban districts. Against this backdrop, comparing data across different years enables us to understand trends in the incidence of different cancer types. On one hand, if there is a surge in the number of patients with a specific cancer, we can initiate investigations promptly. On the other hand, this will help the government formulate medical strategies and allocate healthcare resources more rationally, ensuring that more patients can access appropriate physicians when needed.”

The same principle can also be applied to the prevention of major public health emergencies. As long as the database collects data structures on highly infectious diseases, the system will automatically issue an alert, which will greatly enhance public health early warning capabilities. After all, data does not lie.

During the pandemic, YITU Healthcare officially launched the “Intelligent Prevention and Control Solution for Regional Infectious Diseases.” Built upon YITU Healthcare’s self-developed public health big data platform, this solution integrates AI and big data technologies—including natural language processing, speech recognition, computer vision, and knowledge graphs. By enabling intelligent collection, governance, and analysis of multi-dimensional, full-lifecycle health data for regional populations, it provides capabilities such as intelligent detection, reporting, situational forecasting, and decision support, thereby enhancing the three-tier prevention system for infectious diseases.

Furthermore, through similar approaches, Yitu Medical has established a chronic disease management platform and developed an AI-based bone age assessment system tailored for children. In summary, Yitu is striving to build a closed-loop application ecosystem leveraging artificial intelligence and big data to enhance overall healthcare capabilities.

Setting aside the “New Infrastructure” label, YITU Medical’s “Cancer Prevention Map” initiative had already been advancing toward a similar vision a year earlier. In 2019, the “AI Cancer Prevention Map” was launched in multiple provinces and municipalities, including Guangdong, Fujian, Henan, Zhejiang, Chongqing, Hubei, and Liaoning, cumulatively serving hundreds of thousands of individuals. It conducted over 5,000 AI-powered early screenings for lung cancer, identifying more than 50 suspected high-risk patients.

Subsequently, YITU Medical closely followed the “New Infrastructure” strategy by platformizing and standardizing these discrete AI-enabled achievements. By integrating AI-assisted diagnostic technologies with central hubs such as cloud imaging platforms and regional Centers for Disease Control and Prevention, it has established a systematic artificial intelligence solution for grassroots regions, thereby transforming the previous “breadth” into “depth.”

Leveraging Yitu Healthcare’s resource advantages from its presence in major Grade-A tertiary hospitals, along with Yitu Technology’s strengths in chips and data security, Yitu Healthcare holds a significant first-mover advantage in the AI track of new infrastructure development. Driven by this new momentum, Yitu Healthcare’s performance is worth anticipating.