Never underestimate the power of the tide. "AI+ empowerment" is becoming an important engine to promote the expansion of new quality productivity in various industries, and the medical and health system is no exception. For the medical and health system, improving productivity and efficiency while reducing the cost of patients and healthcare systems is an eternal topic. This also means that the greatest potential of AI+ healthcare lies not only in the application of one or two scenarios, but in assisting and optimizing medical services in all aspects, and completely rewriting the order of all links of "diagnosis and treatment".

At this year's CIIE, through the display of Roche, the industry leader, we can see that this trend is already very obvious: AI is penetrating in multiple dimensions to meet the individualized diagnosis and treatment needs of patients and change the direction of the tide in multiple fields. Although many market participants believe that AI empowerment of the medical industry still needs to be explored, in fact, changes are happening all the time: not only in the drug research and development link, but also in dimensions such as clinical decision support and patient remote monitoring management, the order may be reshaped by AI. At this year's CIIE, the three practical cases brought by Roche gave us a good reference. The general trend of AI subverting the diagnosis and treatment process is fully reflected in the hepatocellular carcinoma treatment efficacy prediction model presented by Roche.

In recent years, the development of liver cancer treatment drugs has made significant progress, especially immunotherapy represented by PD-(L)1, which has greatly improved the overall survival of patients. However, in actual treatment, due to factors such as the heterogeneity of the disease itself and the differences in the ability of doctors, patients often find it difficult to quickly obtain the most effective treatment plan, resulting in extremely high trial and error costs and prices.

In order to solve this problem, Roche cooperated with the top scientific research hospitals in China to develop a digital decision-making assistance model in the field of liver cancer. This is not easy, because the realization of this model requires not only extensive but also in-depth and continuous data.

Fortunately, as one of the leading companies in smart healthcare, Roche relies on large-scale meaningful data and advanced analytical technology to apply to the development of digital decision-making assistance models, integrating complete longitudinal data from the beginning to the end of patient treatment, as well as multi-dimensional information of patients, such as basic information, imaging information, laboratory test results, treatment response and treatment endpoint, in order to provide more basis for improving patient prognosis.

At present, Roche's hepatocellular carcinoma treatment efficacy prediction model has demonstrated great potential to solve clinical pain points. According to the display at the CIIE, the model can quickly and accurately locate the target population and assist doctors in making decision support. For example, when facing a specific patient, the model can comprehensively analyze the patient's lesion location, number, size and other characteristics, and after analysis by the efficacy prediction model, recommend the treatment plan with the highest degree of benefit.

In actual application, the accuracy of the efficacy prediction model is higher than 80%. After the model prediction, the clinical benefit of the patient's treatment is expected to increase by 2.6 times, which can not only improve the patient's long-term prognosis, but also save medical costs.

The digital decision-making assistance model in the field of liver cancer that Roche cooperated with Zhongshan Hospital is currently in the clinical verification stage, and the goal of providing Chinese patients with precise treatment plans is getting closer.

In fact, the efficacy prediction model of hepatocellular carcinoma treatment is only a small part of Roche's "clinical decision system". From Roche's exploration, the "clinical decision system" can empower multiple links such as diagnosis, treatment and prognosis management.

For example, relapse is an unavoidable problem in the field of lymphoma. Data show that 40% of diffuse large B-cell lymphoma (DLBCL) patients cannot be cured after standard R-CHOP treatment. Although more and more advanced therapies are on the market, doctors have more treatment options for relapsed patients, but how to identify relapsed patients early and match them with appropriate therapies as soon as possible has become one of the current clinical problems.

Roche is expected to push this field to a new height. On the one hand, Roche's AI imaging algorithm "lymphoma PET/CT automatic delineation and evaluation model" has the ability to respond quickly and accurately: it can output quantitative and visual results for imaging doctors within 3 minutes.

On the other hand, Roche is also ready to cooperate with domestic institutions based on its advantages and experience to develop a DLBCL relapse prediction model that meets China's clinical needs, and further help improve the cure rate of Chinese patients.

In addition, in the long-term follow-up and monitoring of the disease, digitalization can also reflect its core value.

Roche's digital patient remote management tool for lymphoma has given us a good reference. As a real-time communication platform for doctors and patients, this tool is simple and easy to use, but it can achieve several core points:

First, the refined management of diseases. The tool can provide real-time and accurate key nodes in the long-term follow-up process of patients, including follow-up visits, to avoid patients forgetting and improve the quality of management.

Second, improve the productivity and efficiency of the overall medical and health industry. On this platform, patients upload their physical index data, doctors pay attention in real time, the communication chain is streamlined and efficient, and the diagnosis and treatment paths inside and outside the hospital are successfully opened up.

Based on this tool, doctors' standardized treatment can be achieved, while optimizing the patient's course of disease management, making the goal of improving the cure rate of lymphoma patients a big step forward.

The "three practices" mentioned above are just Roche's initial attempts in the field of smart healthcare. If we expand to the entire medical industry, we should see the trends behind these "three practices" that cannot be ignored.

In the past few years, there have been many discussions about AI reshaping medicine, but it is also a reality that the scenarios are difficult to implement. The three practical cases brought by Roche have subverted this cognition to some extent.

However, more importantly, this allows us to see the three key elements of smart medical practice applications: large-scale meaningful data, advanced analytical technology and valuable application scenarios.

Many companies may only realize the importance of data and analytical technology, but ignore the element of "valuable application scenarios". It is not difficult to understand that "valuable application scenarios" are not easy to create.

On the one hand, valuable application scenarios are not easy to be discovered. The scenario is not imagined out of thin air, but discovered by the doctor group in front-line clinical practice. Doctors may have many ideas, but only a few may be truly valuable. How companies grasp valuable "ideas" is the key.

On the other hand, it is extremely difficult to implement the scenario. The implementation of the scenario requires a series of processes such as "discovering problems, defining problems, optimizing concepts, verifying concepts, product development, optimizing products, and landing transformation". Ideally, doctors discover many valuable scenarios, and companies provide reasonable support and judgment, efficiently screen out truly valuable ideas, and then quickly transform them into practical applications. But the extremely long chain is destined to be a challenge. Any flaw in any link may cause valuable ideas to be difficult to implement.

In-depth analysis shows that the core of Roche's ability to continuously create value in the field of smart healthcare lies in its successful creation of a pyramid structure:

The base is meaningful data and advanced analytical technology on a large scale; the upper layer is its strong "ecological" empowerment strength, which enables the implementation of many valuable scenarios.

Roche's secret weapon to strengthen its "ecosystem" is the "digital healthcare incubator". The core advantages of the "digital healthcare incubator" are:

It has a strong ability to support the entire industrial chain, and ultimately achieves efficient and seamless links from concept verification to implementation. More importantly, this is a two-way process, and it will also help clinical experts upgrade their knowledge and capabilities. Therefore, it is more like a vibrant innovation community, which continues to stimulate the participation of clinical experts and create more breakthrough solutions.

In view of past successes, Roche continues to promote the iteration of the "digital healthcare incubator" and launched the "digital healthcare incubator 2.0" at this year's CIIE. Compared with the first version, version 2.0 significantly enhances the capabilities of two dimensions:

First, it enhances the ability to verify concepts.

The implementation of valuable medical scenarios is a very critical factor in concept verification. That is, after the doctor proposes a preliminary application scenario, the enterprise can identify and judge the feasibility and value of the scenario.

As mentioned above, only a small number of the many needs discovered by doctors are valuable. The key lies in how to discover the ability of the "small number". To this end, Roche's "Digital Medical Incubator 2.0" has strengthened the concept verification capabilities of the enterprise, enabling doctors to truly identify the problems of rigid needs and transform them into greater impacts, thereby changing the problems of patient treatment.

Second, it is the ability to scale, systematize, and be sustainable.

The implementation of the scenario requires a series of tedious links. To achieve scale, systematization, and sustainability, the core is that professional people do professional things. Only when there are the top talents in the industry in each link can the role of the "incubator" be maximized.

Previously, Roche already had this foundation and set up a product incubation team. However, Roche was not satisfied with this. In the "Digital Medical Incubator 2.0", Roche made further upgrades and set up an elite college to attract more capable people to join it, making the innovation of digital medicine more scaled, systematic, and sustainable.

At the same time, Roche is also focusing on improving conversion efficiency.

At present, Roche is actively exploring the use of AI tools to improve scientific research efficiency, allowing scientific researchers to devote more time and energy to the source of innovation.

For example, in response to the actual needs of doctors in scientific research, clinical and patient education work scenarios, Roche has developed functions such as literature interpretation, intelligent writing, intelligent question and answer, and content generation in combination with the latest generative AI technology, which will help experts significantly improve their work efficiency.

These tools will first be applied to colleagues within Roche's medical department, and the quality of tool generation will be continuously improved through actual use, and then gradually promoted to external users.

Based on the three core elements, Roche comprehensively arranges around the patient's clinical diagnosis and treatment path, thereby achieving a triple improvement in medical quality, efficiency, and experience, which is in line with the direction of the era of the medical industry. The launch of Roche's "Digital Medical Incubator 2.0" seems to indicate that this competition is accelerating.

Roche's attack not only demonstrates its ambition in the medical field, but also reflects the transformation of the underlying logic of pharmaceutical companies.

A clear point of view is that AI empowers the medical industry, not limited to a certain link, nor to certain specific fields, but a comprehensive trend, from today's tumors to next year's autoimmune and other fields, everything is possible.

The success of leaders such as Roche indicates that platform companies have begun to emerge. In the future, these companies will inevitably continue to iterate and apply AI technology to a wider range of fields.

As a company focused on innovation, Roche has also demonstrated such ambitions.

Its product pipeline covers five core therapeutic areas, including oncology/hematology, neurology, ophthalmology, cardiovascular renal metabolism and immunology, and is deployed in 11 disease areas including breast cancer, lung cancer, malignant hemoglobin, hemophilia, multiple sclerosis, Alzheimer's disease, inflammatory bowel inflammation, chronic obstructive pulmonary disease, retinal vascular disease, macular degeneration, obesity, etc. At the CIIE, Roche demonstrated its innovative achievements and future layout in more therapeutic areas. Among them, the field of chronic obstructive pulmonary (COPD) disease is a top priority.

In this field, Roche is showing off the muscle of ST2 monoclonal antibody Astegolimab. Astegolimab blocks the IL-33 signaling pathway by targeting the IL-33 receptor ST2.

Currently, a Phase III clinical trial is underway to treat patients with moderate to severe COPD. Regardless of the patient's blood EOS level (referring to the count and percentage of eosinophils (EOS) in the blood) and smoking status, its unique mechanism of action may benefit a wider range of patients and is expected to become the first ST2 monoclonal antibody for the treatment of chronic obstructive pulmonary disease.

For COPD, a single therapeutic drug is not enough. COPD is characterized by "four highs and three lows", which has brought a heavy burden to the society and health care system. The key to solving this burden lies in solving the "four highs and three lows" problem. Roche, a leader in the field of smart healthcare, may bring comprehensive solutions in the future.

At the CIIE, Roche mentioned that it will continue to focus on identifying and meeting key clinical needs, while building a vibrant ecosystem in the field of smart healthcare through close cooperation with medical experts and technology pioneers.

At present, the demand for healthcare is increasing and upgrading. Aminojun believes that leaders like Roche are needed to guide the direction. Under the tide of the times, all pharmaceutical companies need to continue to work hard to adapt to the widespread application and continuous iteration of AI technology in the medical field to meet the needs of Chinese patients. Only by following the general trend and insisting on innovation can we achieve twice the result with half the effort and move forward with a promising future.

https://mp.weixin.qq.com/