Philips partners with medical AI company Parsley Medical. Is platformization the only way for medical device manufacturers?

Robert, a member of the US National Academy of Sciences and a renowned expert in behavioral analysis and game theory·Axelrod, in his book "The Evolution of Cooperation," stated:“True power doesn't come from conquering others, but from fully guiding cooperation.”

This statement is equally applicable to Philips and ParsView Medical.

2016Years later, Philips China officially implemented a digital transformation strategy, focusing on the“Health Technology”field. To this end, Philips adopted an approach of addition and subtraction: internally, it abandoned businesses it wasn't proficient in and invested heavily in its areas of expertise; externally, it abandoned areas it didn't understand, leveraging its existing experience to build a medicalAIecosystem.

Just a year later, at2017theRSNAconference, Philips launched the AI platformIntelliSpace Discovery 2.0(ISD2.0). This is a platform and research tool for radiologists; it can be considered the medicalAIfield's"App Store"。

It is noteworthy that the first product to enter this“app store”came from a startup company—ParsView Medical. Its nasopharyngeal cancer radiotherapy clinical target delineation system based onCT/MRIbecame the only mature product at the time that had passed the strict review of a top medical equipment manufacturer and entered commercial use.AIA collaboration completed in one month

一个月完成的一次合作

2017year5month, Lu Yao and his university classmate Shen Shuo founded ParsView Medical.

Lu Yao previously served as a researcher at the University of Michigan Medical School.2013year, Lu Yao returned to China as a selected candidate for the Organization Department of the CPC Central Committee's“Thousands of Talents Plan”Youth Program, and served as a professor at the School of Data Science and Computer Engineering and director of the Institute of Data Science at Sun Yat-sen University.

Lei Feng.com learned that it took onlyISD，仅用时1a little over a month from the initial contact to the embedding of the application software into

According to Lu Yao's recollection, ParsView Medical was founded4months later, Zhou Zhenyu, senior director of the Clinical Science Department of Philips' Integrated Solutions Center, was on a business trip to Shenzhen and visited ParsView Medical in Guangzhou afterwards. Before this, Philips had signed a comprehensive strategic cooperation agreement with Sun Yat-sen University Cancer Center to jointly conduct cutting-edge research and clinical practice from image diagnosis to treatment planning.

Philips globalCEO Frans Van Houton，Greater ChinaCEOHe Guowei, and Zhou Zhenyu, after listening to Lu Yao's introduction of the nasopharyngeal cancer radiotherapy clinical target delineation system, fully affirmed the prospect of integrating this product into the PhilipsISDplatform. During this visit, both sides sat down and talked for two hours, with Lu Yao and Zhou Zhenyu sharing ParsView Medical's products and future plans.

Looking back at the cooperation process, Zhou Zhenyu said that at that time, there were100manyAImedical start-ups in China. The reason why ParsView Medical was chosen first was:

First, China's top radiotherapy department is located at Sun Yat-sen University Cancer Center, and ParsView Medical's technology was developed in deep cooperation with Sun Yat-sen University Cancer Center;

Second, ParsView Medical did not choose lung nodules, diabetic retinopathy screening, etc., like mostAImedical startups, but chose to tackle the more difficult nasopharyngeal cancer radiotherapy target delineation first, which is quite unique;

Third, the application scenario is excellent. It takes professional doctors3-5hours to complete the delineation, but ParsView Medical's intelligent radiotherapy delineation solution (IRCS) can complete it in tens of seconds.

Based on these considerations, after the National Day holiday, Philips sent personnel to ParsView Medical to cooperate with the Guangzhou-based startup team to embed theAIproduct into theISDplatform.

The most time-consuming part was communication. The team responsible for the PhilipsISDplatform interface was located in Germany, and the two teams had a six-hour time difference. In addition, the pre-releaseISDversion was still quite rough, and the platform's interface friendliness was not good. To connect the interfaces between the two systems, both teams had to spend more time coordinating the details.

Lu Yao said that this was a tough task, but it was something that had to be done in the early stages of a startup,“after all, only by being more proactive and assertive is it possible to survive in this track.”

A month later, at the end of2017year11month, Philips officially released itsRSNAat the North American Radiological Society annual meeting (RSNA)ISDPlatform, and ParsView Medical successfully becameISDthe first resident on the platform.

Is platformization the only way for medical device manufacturers?

Why did Philips choose to cooperate with medicalAIcompanies? In fact, this is a microcosm of the entire medical device industry's pursuit of transformation.

In recent years,“software and hardware integration”has become a keyword for medical device manufacturers. AtCMEF's exhibition booth, in addition to showcasing its own hardware“nuclear weapons”，GE，Philips, Siemens and other industry giants have tacitly followed“platformization”direction.

For example,GEMedical's"Edison"platform can seamlessly integrate with existing artificial intelligence partners and their related products. Clinical partners are based on"Edison"platform to develop algorithms, and technology partners can apply the latest data processing results to applications and smart devices based on"Edison"the platform.

As Siemens Healthineers, a subsidiary of Siemens, a German electronics and electrical engineering company, has more than170years of history. At this year'sCMEFexhibition booth, Siemens Healthineers also reserved a space for its digital medical ecosystemTeamplaytwo ChineseAIpartners.

For example, its collaboration with Yason Technology to jointly promote its AI-assisted diagnosis in Alzheimer's disease (AD) research and implementation, as well as image recognition related products targeting lung nodules, stroke, breast cancer and other diseases with Deepwise Medical were also unveiled at this exhibition.AI图像识别相关产品也在此次展会上进行了亮相。

Lu Yao said that the transformation ideas of medical device manufacturers are actually easy to understand.“Medical device manufacturers essentially need to provide services through platformization to meet users' soft needs, such as clinical training and scientific research, which will be the key development direction for medical device manufacturers in the future.”

For those who started with“hardware”giants, their revenue mostly comes from this, and the capabilities of their software departments are relatively weak. Under the trend of precision and personalization, the service needs of hospitals are very high, and the software department of a medical device manufacturer cannot meet all the needs.

TakeISDplatform as an example.ISDThe origin of stems from the long-standing pain points in the field of medical imaging research. For example, patient prognosis analysis and adaptive radiotherapy in radiotherapy departments are long-standing problems for doctors in clinical practice, but it is difficult for doctors to transform these problems into their own research results.

The reason is that doctors are often not good at programming and development, and their valuable experience and massive data cannot be quickly transformed into new results and new applications.

And medicalAIcompanies bring value is to allow medical device manufacturers to solve clinical needs more flexibly, quickly, and professionally.“MedicalAIstart-up companies will use the company's full strength to take root in some very specific technical fields, which is something that medical device manufacturers are unlikely to do. Secondly, using the platform form can enrich the product lines of medical device manufacturers at a lower cost, and support the promotion of medical device manufacturers' products.”

According to Leifeng.com, ParsView Medical mainly has two product lines: intelligent radiotherapy target delineation solution (IRCS) and intelligent medical image processing solution, which integrates a series of radiotherapy and radiotherapy tools, including automatic segmentation, target delineation, data annotation, radiomics analysis and quantitative analysis.

For example, ParsView Medical's AI delineation system provides automatic delineation and convenient modification tools for radiotherapy targets and organs at risk of various tumors. Target delineation includesGTV(tumor target area)、CTV(clinical tumor target area) delineation, ParsView Medical can achieve accuracy≧80%automatic delineation, only a small amount of modification by doctors is needed for clinical use;

in the time-consuming and laborious delineation of organs at risk, ParsView Medical can complete the automatic delineation of up to20organs at risk in nasopharyngeal cancer radiotherapy within43seconds, reducing the delineation time from the original1hours to10minutes (including the doctor's modification time).

These tools are a very beneficial supplement to the software departments of traditional medical device manufacturers, and also accelerate the landing application of medical device manufacturers' products.

2018year8month,ISDfirst landed in Jilin University First Hospital, and then entered Union Hospital, Tiantan Hospital, Xuanwu Hospital, Huaxi Hospital, etc. In the next two years, that is, by2021year, the planned installed capacity of hospitals in China will reach50units.

In the long run, theISDplatform carrying different applications provides greater imagination space for Philips' development: it can not only incubate new technologies and new products, but also sell new technologies and new products. This will undoubtedly add more possibilities to the thin hardware sales model of medical device manufacturers.

And all this requires Philips and numerous medical enterprises to work together.AIAfter all, if there is only"App Store"and no large and small"App"，this“app store”would also be difficult to maintain.

According to Lei Feng Wang, within this year, application software derived from two solutions of Bosch Medical will be released onISDplatform.

Of course, besides Philips’ISDplatform, Bosch will not reject cooperation with other instrument manufacturers, and will provide imagingAItechnology empowerment services to instrument manufacturers and medical service providers.

Lu Yao said in an interview,“We hope that the other party has good technical integration capabilities and excellent channel expansion capabilities. In addition, the agreements we sign with our partners are non-exclusive, so we can cooperate with many manufacturers, with the aim of maintaining a mutually beneficial ecological balance.”

Conclusion

In a previous interview with Lei Feng Wang, Philips Greater ChinaCEOHe Guowei once said that Philips' openness is multi-layered, that is, seeking cooperation with institutions in all aspects of the medical process.

The medical industry has its particularities, withAIas the center of the health care ecosystem needs to cover every link in the entire industrial chain, including a large number of sub-fields and application scenarios.

In other words, instrument manufacturers cannot be involved in all scenarios, and platformization and seeking open cooperation are a necessary path for these instrument giants in the future.

And with Bosch Medical as such a medicalAIenterprise, will take high-quality data and constantly evolving algorithms as the core, and join hands with traditional medical equipment and medical service providers to form a new ecological structure in medical health.