Philips Advances into Medical AI Market, Collaborating with Clinicians and Third-Party Partners to Build an Integrated Ecosystem

Recently, at the Global Artificial Intelligence Product Application Expo held in Suzhou, Philips showcased for the first time to the industry the achievements of AI technology applied in its existing products and solutions. VCBeat, as one of the witnesses, participated in this conference.

At this conference, Philips unveiled its “AI Healthcare” strategy for the first time—Upholding the seriousness of medicine and grounding all efforts in clinical reality, we are committed to becoming an innovative “practitioner” in the AI healthcare industry and a “driver” of the local ecosystem.。

For over a century, Philips has undergone multiple transformations, evolving from its origins in lighting to diversification and global expansion, until today’s precise focus on the “Health Technology” sector. Each transformation has been driven by the aim to better align with social development trends, seize market opportunities, and explore and meet people’s unmet needs.

In 2015, Philips spun off its lighting business as an independent publicly listed company and leveraged the proceeds to accelerate its comprehensive footprint in the “health technology” sector through more than ten acquisitions, ushering in the “Philips 6.0” era. Committed to leveraging extensive clinical expertise and advanced digital and AI technologies, Philips continues to roll out integrated solutions across the entire continuum of care, delivering better health outcomes and medical care at lower costs.

Philips stated that the value of artificial intelligence in the healthcare sector lies in leveraging technologies such as natural language processing, deep learning, and big data mining and analysis to “transform data derived from humans into clinically actionable insights,” ultimately achieving precision medicine and improving quality of life.

The vision for AI in healthcare is to capture and derive insights from comprehensive longitudinal patient data, including demographics, family history, genomic profiles, clinical records, imaging reports, medication usage, and even personal health metrics. This is akin to creating a “digital twin” for each individual. When complete longitudinal patient data is compared with cross-sectional data derived from big data analytics, it will have a profound impact on the entire healthcare system.

An AI-Centric Healthcare EcosystemAn AI-centric healthcare ecosystem encompasses every moment and scenario related to health across the “full life cycle,” covering every link in the entire health industry chain. It includes a vast array of specialized fields and application scenarios, such as: genomic analysis (disease prediction), medical imaging recognition, clinical decision support, drug R&D, hospital management, health management, virtual assistants, medical robots, and medical research platforms.

To realize this vision and address the various medical challenges facing humanity, no single individual, company, or solution can solve all problems alone. Therefore, in the field of AI-driven healthcare, breaking down barriers and boundaries is imperative.

According to VCBeat, Philips incorporated AI into its development strategy as early as five years ago, investing €1.7 billion annually in research and development, with 60% allocated to software development. The resulting achievements have already been integrated into its existing products and solutions. Currently, 25% of Philips’ scientists are engaged in approximately 250 research projects related to AI and big data, closely aligned with clinical scenarios and workflows. These initiatives encompass natural language processing, big data mining and analysis, construction of structured clinical databases, image recognition, imaging-assisted diagnosis, interventional therapy, genomics, chronic disease management, home care, and cloud platform solutions.

Philips’ philosophy is to uphold the seriousness of medicine, grounding all efforts in clinical reality. For decades, Philips has collaborated closely with more than 4,000 leading hospitals, research institutions, and innovation platforms worldwide. From data sourcing, modeling, and training to result testing and evaluation, every step strictly adheres to clinical guidelines and medical pathways. Based on scientific assessment standards and systems, rigorous and repeated validation ensures the provision of safe and reliable solutions for clinical practice.

As an active “driver” in building the AI healthcare ecosystem, Philips has created the industry’s first AI healthcare sharing platform—“HealthSuite Insights”—providing data scientists, software developers, clinicians, and healthcare institutions with a range of shared tools and technologies to develop, maintain, deploy, and scale artificial intelligence solutions. The Insights Marketplace aggregates AI assets from Philips and other companies, available for shared use through licensing agreements, helping to save the time, resources, and costs required for developing and deploying medical AI solutions.

In China, Philips leverages co-creation through integration with local smart ecosystems to apply digitalization, artificial intelligence, and other technologies to real-world clinical scenarios. It develops intelligent solutions tailored to the actual needs of the Chinese market that tangibly enhance user capabilities, and strives to accelerate the market deployment of AI healthcare solutions through innovative business models. This approach aims to make “high-threshold” AI medical technologies accessible to households at an affordable level, thereby achieving inclusive healthcare.

“Shenfeiyun,” co-created with Shenzhou Medical, is one of the key outcomes of Philips’ strategic implementation in China. Powered by “Shenfeiyun,” the “Philips Nebula System” has enabled a “cloud platform” service model with flexible payment mechanisms. By leveraging remote technologies, it allows primary-care hospitals to deliver high-quality healthcare services based on “precision diagnosis and treatment” to the general public, thereby supporting the development of medical consortia,” said He Guowei, CEO of Philips Greater China.

Ninety percent of medical data originates from medical imaging. In China, medical imaging data is growing at an annual rate of 30%, while the number, experience, and efficiency of radiologists are wholly inadequate to cope with this growth trend. Currently, the majority of medical imaging data still relies on manual analysis, the most obvious drawback of which is imprecision.

According to Philips’ 2018 “China Health Views” survey, China faces a shortage of professional healthcare personnel and an uneven distribution of high-quality medical resources. The lack of “precise diagnosis and treatment at the first visit” has become a bottleneck affecting the efficiency and effectiveness of managing major diseases, resulting in a poverty rate due to surgical treatment exceeding 50%. As deep learning methods in the field of computer vision mature, they offer new opportunities for automated analysis of medical imaging and assist physicians in achieving intelligent, precise diagnosis and treatment. Consequently, “AI + Medical Imaging” has become the most prominent segment within AI-driven healthcare.

Although China possesses a vast volume of medical imaging data, the industry remains in a transitional phase from traditional film to electronic data. A significant portion of imaging records has yet to be digitized, and the level of data sharing and interoperability among hospitals is low, making the acquisition of large-scale datasets a major bottleneck.

Furthermore, based on data acquisition, deep learning requires the integration of prior knowledge to train models. The training sets primarily rely on precise annotations of imaging data by highly skilled physicians, thereby enabling machine learning algorithms to identify lesions. In other words, only through collaboration with experts and clinicians possessing extensive clinical experience can “intelligent artificial intelligence” be effectively trained. Moreover, given that there are thousands of imaging-related disease types, favorable clinical outcomes must first be achieved in common, single-disease categories with relatively large datasets before gradually expanding to other disease areas.

Medicine is a rigorous science that emphasizes practice, relies on evidence-based approaches, and directly impacts quality of life. Therefore, AI-based medical imaging requires scientific evaluation standards and systems covering every stage—from data sourcing and model development to training and result assessment—so as to gain recognition from clinicians and authoritative bodies, ensure safe and reliable clinical application, and serve as an extension of the human mind and hands. As an interdisciplinary field, AI in healthcare can only produce “artificial intelligence” with practical clinical relevance when professionals skilled in artificial intelligence algorithms collaborate closely with experts who possess deep insights into the medical domain.

As a long-established professional medical company, Philips covers a wide range of clinical pathways, with its diagnostic imaging products serving as devices that generate and collect medical imaging data. Beyond embedding artificial intelligence (AI) technology into its equipment, Philips has developed a series of AI-driven intelligent tools based on actual clinical needs. These tools aim to optimize radiology workflows, assist physicians in achieving accurate initial diagnoses, and formulate personalized treatment plans.

For example, the “Philips Nebula AI Medical Imaging Platform,” which won the gold award at this product expo, comprises two components: the “Philips Nebula 3D Imaging Data Center” (IntelliSpace Portal) and the “Philips Nebula Exploration Platform.”

“Philips Nebula 3D Imaging Data Center” is an integrated, intelligent clinical imaging diagnosis platform that enables image processing across different brands and types of imaging equipment. It provides advanced visualization and post-processing for multi-modal imaging, assisting radiologists and clinicians in better lesion identification, as well as in monitoring, diagnosing, and following up on disease treatment.

It covers multiple clinical specialties in radiology, including cardiology, oncology (liver, lung, breast, prostate, etc.), neurology, emergency medicine, orthopedics, and dentistry. With more than 70 clinical application modules, it leverages advanced 3D image post-processing techniques based on clinical guidelines to present anatomical structures with greater clarity, thereby assisting clinicians in making rapid and precise diagnostic decisions and formulating personalized treatment plans.

Furthermore, the “Nebula Platform” enables the establishment of remote medical imaging centers that span across departments, hospital campuses, and geographic regions, thereby optimizing clinical workflows, facilitating cross-departmental sharing of imaging resources, promoting the equitable distribution of high-quality medical resources, and advancing tiered diagnosis and treatment.

“Philips IntelliSpace Discovery” supports clinical research, featuring an open-source architecture, multi-modal imaging support, compatibility with various programming languages, automated batch processing, and integrated AI components along with the XNAT database management system.

It covers three major modules—oncology, cardiology, and neurology—and provides a variety of tools for image segmentation, rigid and elastic as well as multimodal registration, along with computational utilities for Radiomics, DTI, IVIM, and various pharmacokinetic models, thereby offering researchers extensive image development tools and a robust development environment.

The newly released ISD 2.0 artificial intelligence platform enables multimodal image integration and mining of disease-specific imaging features. Its open software community allows Philips, clinicians, and third-party vendors to participate collaboratively in building a medical AI ecosystem.