Infervision Deploys COVID-19 AI System in Japan with Doctor Net to Address Radiologist Shortage

In March, snow and ice still blanketed Hokkaido. It should have been the peak tourist season, bustling with visitors, but the sudden outbreak of the novel coronavirus pandemic left the entire region desolate and gripped by panic. Zhou Xiaoyan, head of Infervision’s Asia-Pacific division, who has been stationed in Japan for three years, told VCBeat that the past two weeks had been a critical period for epidemic prevention. Japan was currently experiencing cluster infections and outbreaks with unidentified sources of transmission. Shortages of masks and toilet paper were also reported in Japanese supermarkets.

According to a report by Japan’s NHK on the 12th, citing data released by Japan’s Ministry of Health, Labour and Welfare and local governments, Japan recorded 53 new confirmed cases of COVID-19 on the 11th, bringing the cumulative total to 620. In addition, there were 696 confirmed cases among passengers and crew members aboard the Diamond Princess cruise ship, and 14 individuals who returned to Japan on Japanese government-chartered flights tested positive. The cumulative number of confirmed COVID-19 cases in Japan thus reached 1,330, with 475 patients having recovered. The Japanese government has placed on its agenda plans to enact legislation to address the epidemic and implement a nationwide state of emergency.

On the 10th, the World Health Organization released its latest data showing that the number of confirmed COVID-19 cases outside China reached 32,778. In addition to Japan, outbreaks in Italy, Iran, South Korea, and other regions are also causing significant concern. As the pandemic ravages the globe, combating COVID-19 is no longer solely the responsibility of a single country or government, but a collective effort required from the entire international community.

In the cold February, the phrase “Though our lands are divided by mountains and rivers, we share the same wind and moon” warmed the hearts of the Chinese people. Now is the time to reciprocate kindness with generosity. As the global pandemic intensifies, China, having entered the final stage of its domestic battle against the epidemic, has begun donating supplies such as face masks and protective suits to Japan. In addition to material aid, China’s anti-epidemic measures have proven effective through practice, prompting the country to share its experience in combating the virus with other nations, including the latest diagnostic technologies.

As a new force in the technological fight against the epidemic, Infervision joined forces with Tongji Hospital affiliated to Huazhong University of Science and Technology, Zhongnan Hospital of Wuhan University, and Shenzhen Third People’s Hospital (National Clinical Research Center for Infectious Diseases) to rapidly develop an intelligent auxiliary screening and epidemic monitoring system for COVID-19, which was put into clinical use in Wuhan, Hubei Province at the earliest opportunity.

In Japan, leveraging the solid foundation it had previously established in the Japanese market, Infervision promptly partnered with Doctor Net, the country’s largest teleradiology company, to rapidly deploy its “Intelligent AI-Assisted Screening and Epidemic Monitoring System for Pneumonia.” In Japan, where CT scanner availability is high but radiologists are in short supply, this COVID-19 AI system—trained on thousands of cases during the pandemic and acting as an “expert”—has created a safety net to reduce missed diagnoses and enable precise diagnosis, establishing the first line of defense in epidemic screening through chest CT imaging.

Given the differences in healthcare systems between countries, bringing Chinese tech-driven anti-pandemic products to the Japanese market cannot achieve recognition by simply replicating domestic experiences and copying products. Facing the vastly different healthcare systems and clinical scenarios in the two countries, how can the AI system for COVID-19 play its role? How did Infervision complete the localization of its products in a short period? VCBeat interviewed Zhou Xiaoyan, Head of Infervision’s Asia-Pacific Division.

When launching the intelligent AI-assisted screening and epidemic monitoring system for pneumonia in China, Infervision tackled challenges such as tight deadlines, heavy workloads, and disrupted transportation due to regional lockdowns, amidst a surge of unconfirmed suspected cases and a severe shortage of medical personnel in Hubei Province.

In Japan, although the number of confirmed and suspected cases is far lower than that in China, new cases with unknown sources of infection have been reported across various regions since mid-February, posing significant challenges to COVID-19 diagnosis.

Zhou Xiaoyan stated, “Japan has the largest installed base of CT scanners in the world. When we entered the Japanese market, we were optimistic about the vast amount of imaging data available there. Japan ranks first globally in the per capita prevalence of medical imaging equipment. However, the proportion of radiologists specializing in diagnostic radiology within its population is relatively low among developed countries, with only slightly over 6,000 registered radiologists nationwide. Consequently, there is a problem of excessive workload for radiologists.”

The highly contagious COVID-19 pandemic, with its aggressive onset, will continue to strain the existing radiologists’ capacity as case numbers keep rising. Meanwhile, due to the limited number of COVID-19 cases in Japan, there are few radiologists with experience in imaging-based diagnosis of the disease.

In the face of the ongoing spread of the COVID-19 pandemic, Japanese medical institutions are also under immense pressure. As the number of confirmed cases rises, PCR testing is plagued by issues such as prolonged turnaround times and insufficient capacity. Under these circumstances, it is imperative to establish chest CT imaging diagnosis as the first line of defense for epidemic screening.

As a global health-tech company rooted in China, Infervision has been making every effort to deliver its latest anti-epidemic products to the places in most urgent need. At the earliest opportunity, Infervision partnered with Doctor Net, Japan’s largest remote imaging diagnostics enterprise, to deploy Infervision’s “Intelligent Assistant Screening and Epidemic Monitoring System for Pneumonia.”

Upon implementation of this system, when CT images suggestive of COVID-19 are encountered, the AI system immediately alerts physicians to the potential presence of COVID-19 infection, prompting timely clinical intervention. This system helps accelerate diagnostic workflows, enables rapid feedback of diagnostic results to client hospitals, and reduces the risk of epidemic spread. Furthermore, it assists physicians in accurately assessing patient conditions, providing a basis for diagnosis and treatment, formulating precise therapeutic plans tailored to individual patient presentations, and enhancing the standardization and quality of pneumonia diagnosis and management.

The introduction of the AI engine can be described as bringing to Japan a “battle-hardened expert” with extensive experience in Wuhan. The Infervision Pneumonia AI Special Edition was developed in Wuhan, the epicenter of the outbreak. In January 2020, after being trained on thousands of COVID-19 cases provided by multiple medical institutions, including Tongji Hospital in Wuhan, this special edition was first deployed at Tongji Hospital affiliated with Tongji Medical College of Huazhong University of Science and Technology, a designated facility for fever clinics specializing in novel coronavirus infections. During the pandemic, Tongji Hospital served as the primary designated hospital in Wuhan for the centralized treatment of the largest number of severe cases.

During the rapid promotion of its AI pneumonia solution in Japan within a short timeframe, Infervision encountered significant challenges. The first hurdle was product localization to align with the clinical workflow habits of physicians in both countries.

Zhou Xiaoyan stated that Infervision has been exploring the Japanese market for three years. During this period, we took many detours and spent considerable time communicating with physicians to understand their product requirements. Simply replicating our experience from China would inevitably lead to challenges in local adaptation. For instance, regarding the nomenclature of pulmonary lobes and segments, Chinese physicians are accustomed to textual descriptions such as “lower segment of the left lung lobe.” When we introduced our product to Japan, some physicians questioned this labeling convention. In Japan, physicians prefer using alphanumeric codes like S1 and S2 to denote the nine anatomical regions of the lungs. The same principle applies to our AI system for COVID-19; we must first adapt to the operational habits of Japanese physicians and achieve thorough localization.

In addition to adapting to physicians’ operational habits at the micro level, AI must also accommodate the fundamentally different healthcare systems of the two countries. China’s healthcare system is centered around tertiary Grade A hospitals, with an imperfect tiered diagnosis and treatment framework. In contrast, Japan’s healthcare system comprises community hospitals, specialized hospitals, and university hospitals, with third-party imaging centers serving as a coordinating link among these three tiers.

Zhou Xiaoyan explained, “Taking imaging examinations for patients as an example, patients first seek treatment at community hospitals rather than going directly to tertiary hospitals, as is common in China. If patients visit large hospitals without a referral letter, these hospitals may decline to admit them. Community hospitals are equipped with adequate facilities to perform basic examinations. In cases where community hospitals lack radiologists, doctors from university-affiliated hospitals may hold consultations on scheduled days during the week, or a large volume of imaging data may be transmitted to third-party imaging centers for interpretation.”

In terms of epidemic control, due to differences in Japan’s healthcare system, patients have not been centrally hospitalized as in China. Therefore, the Japanese government’s current approach is to contain patient care within the scope of community hospitals. In this context, Tuixiang Technology selected third-party imaging centers as its entry point and carried out its initial deployment.

Since third-party imaging centers receive data from various CT scanner models and images acquired by different technologists, this places higher demands on AI systems for COVID-19.

Leveraging Infervision’s robust expertise in AI-powered medical imaging and its deep understanding of the Japanese healthcare system, Infervision’s AI system for COVID-19 has also received high recognition from Doctor Net. As Japan’s largest professional platform for radiological diagnosis, the Doctor Net platform supports remote image-based diagnoses for approximately 6,000 cases daily, handles around 1.8 million cases annually, and serves approximately 700 contracted radiologists across Japan who utilize its image interpretation services.

In its statement, Doctor Net expressed its hope that by empowering contracted radiologists on its platform with Infervision’s AI system for COVID-19, the accuracy of imaging diagnoses for potentially increasing cases of suspected novel pneumonia could be improved.

The crisis brought by this epidemic has caused the economy to go into "shock." In this tough battle, the epidemic is not a pause button that allows us to return to normalcy once restarted. During this outbreak, racing against the virus has directly tested the R&D capabilities, organizational capabilities, and globalization capabilities of medical AI companies. In fact, these capabilities all rely on early strategic layout and long-term accumulation.

Doctor Net, a leading company in Japan’s remote radiology diagnostics sector and the partner of Infervision in this collaboration, is celebrating its 25th anniversary this year and collaborates with approximately 900 medical institutions across Japan. In Japan’s hierarchy-conscious society, Infervision has made significant efforts to gain recognition.

Zhou Xiaoyan stated, “We began positioning ourselves in the Japanese market in 2017, introducing our products into Japan’s relatively mature healthcare system for refinement, thereby making them better aligned with local usage habits and application scenarios. If our products can meet the rigorous standards of such a mature healthcare system as Japan’s, we will be even more confident in promoting them to other countries and regions.”

Over the past three years, Infervision’s product applications have expanded to cover community hospitals, university hospitals, and third-party imaging centers in Japan. Moving forward, Infervision will continue to deepen its presence in the Japanese market, support the country’s tiered diagnosis and treatment system, and leverage AI technology to control and even reduce healthcare costs.

As Infervision’s Asia-Pacific headquarters outside of China, Japan will see the company collaborate with Doctor Net to develop AI-RAD, a platform that integrates remote imaging diagnostic support services with an artificial intelligence engine.

In China, although CT has ceased to serve as the clinical diagnostic standard, Infervision’s AI system for COVID-19 continues to engage in research collaborations on imaging-based diagnosis of the novel coronavirus with multiple frontline hospitals involved in the pandemic response, including Zhongnan Hospital of Wuhan University. Acting as an additional pair of eyes for physicians, the system quantifies disease progression and assists in establishing precise diagnosis and treatment plans.

In this life-and-death race against the virus, Infervision rapidly developed a product within less than a month, leveraging real-world cases from the front lines of the Wuhan outbreak, and subsequently deployed it to global pandemic front lines.

Zhou Xiaoyan said, “This is a process of achieving success through sustained accumulation. In this pandemic, no one can remain unaffected. Now is not the time to dwell on what we lack; rather, we should consider what we can accomplish with the resources currently at our disposal.”