Why Did DiAn Diagnostics Build a 'Pathology Cloud'? Insights from Jiang Tang at the 2016 Big Data Expo

The entire evolution of medicine is, in essence, a journey of human self-exploration and self-understanding. From observing ants to detect sugar in urine, which led to the discovery of diabetes, to the introduction of various advanced Western technologies—including current innovations such as 3D printing and gene sequencing—each step represents our gradual deepening understanding of the human body.

During the era when the ancients compiled the *Huangdi Neijing* (The Yellow Emperor’s Inner Canon), our understanding relied primarily on empirical data gathered through the four diagnostic methods: inspection, auscultation and olfaction, inquiry, and palpation. At that time, the human body was conceptualized in terms of twelve meridians. Subsequently, with the advent of various physicochemical analytical instruments, we gained the ability to analyze diverse metabolic parameters, leading to the development of detailed anatomical maps of the human body.

Gradually, our technologies have become increasingly advanced, enabling us to acquire various types of multidimensional data, including genomics, metabolomics, genetics, and transcriptomics, while striving to integrate these datasets. This progress has, of course, been facilitated by the availability of the first human genome map.

Therefore, medical progress lies in the relentless effort to quantify the human body and acquire diverse physiological data. The advent of new tools—such as computers, computing technologies, the Internet of Things (IoT), and mobile health—has brought about profound transformations to society and human life, while also marking a significant turning point in the field of medicine.

If there is Industry 4.0, is there also Medicine 4.0? To this day, there is still no complete definition. I attempt to provide a rough classification: The 1.0 era was characterized by traditional medical practices of observation, auscultation and olfaction, inquiry, and pulse-taking; the 2.0 era saw the introduction of Western medicine, enabling diagnosis through inspection, palpation, percussion, and auscultation; with the development of various advanced technologies, we have arrived at the current era, and with cutting-edge technologies such as imaging, the Internet of Things (IoT), and big data, we are on the verge of entering the Medicine 4.0 era.

We recognize that, for patients, the entire healthcare experience is private, subjective, direct, and one-on-one. In this context, many medical decisions are based on population averages derived from statistics. We tend to prioritize safety over efficacy, often exhibiting a more conservative and authoritative stance. As medical specialties become increasingly fragmented, many decisions risk resembling the parable of the blind men and the elephant—each perceiving only a part of the whole. This is the scenario in which we acquire data and learn from experience.

In medical decision-making, data collection is crucial; it determines what kind of data we should gather to make informed clinical decisions. Diagnosis precedes treatment. Without access to valid data, it is difficult to make appropriate therapeutic choices.

In clinical diagnosis, the three pillars are laboratory testing, medical imaging, and medical records. In the era of precision medicine, this concept has been increasingly narrowed in scope, with many mistakenly equating precision medicine solely with genomics and believing that genetic analysis can resolve all medical issues.

In fact, at least 96% of today’s medical practices rely on these “three pillars.” Every day, we analyze human blood and urine samples to generate a wide range of data that help clinicians make accurate diagnoses and treatment decisions.

The proteins, sugars, and fats we consume undergo highly complex metabolic pathways. Any obstruction in these pathways can lead to the development of metabolic diseases, a problem that cannot be resolved solely through genetic interventions. For instance, in the case of diabetes, although we have deciphered the genetic codes labeled A, B, C, and D, there remains 97% of “dark matter” whose genes, functions, significance, and clinical relevance are still unknown to us.

Over the past two decades, China has established 22 provincial-level central laboratories. We have conducted more than 2,000 projects, serving over 12,000 medical institutions. Building on this foundation, we have created an ecosystem spanning the entire value chain—from upstream to downstream—to collect patient data and provide diagnostic information to clinicians. In the current period of downturn in the pharmaceutical industry, diagnostics stands in contrast, as it must continue to grow. Only third-party testing can deliver more precise data and achieve a better cost-effectiveness ratio.

Imaging is a relatively familiar technology and is considered at the forefront in the medical field, as it can acquire information to aid diagnosis. Pathology, however, is the least digitized component among the “three pillars” of clinical diagnostics. Today, medical practice still relies on tissue sectioning, microscopic examination, and physical storage of slides.

We have introduced the most advanced technology, and on this basis, we have established a digital pathology cloud system across China. Collaborating with 12,000 hospitals, we provide all pathological information to consultation experts both domestically and internationally via the cloud platform, thereby enabling accurate diagnoses for end-user hospitals.

For our physician expert platform, we have established a three-tiered system comprising 22 basic diagnostic centers covering China, advanced pathology diagnostic centers, and premium advisory centers in collaboration with leading U.S. pathology institutions, fostering cooperation across various subspecialties.

In this context, we can assist in teaching, scientific research, and clinical care. On top of that, the number of pathologists in China actually meets less than one-quarter of the demand. Pathology is a specialty in urgent need of revitalization, with a severe shortage of pathologists. Under these circumstances, we leverage big data, cloud-based pathology platforms, and internet technologies to enhance the overall efficiency of pathological diagnosis, promote the integration of high-quality resources, and help improve the quality of pathological diagnosis in underdeveloped regions.

Our objective is not merely to generate a set of data, but rather to engage in critical analysis of this data. In the context of big data for clinical laboratory testing, the volume is immense and exhibits strong longitudinal continuity, characterized by the coexistence of continuous and discontinuous data streams. The characteristics of imaging big data are even more pronounced, displaying multidimensionality with both structured and unstructured pathological features. Regarding pathological features and electronic health record (EHR) data, the characteristics become even more complex. Such data is not only massive in volume but also demands high storage and query efficiency. It is challenging to integrate, with constantly evolving data schemas. Furthermore, issues such as data collection and acquisition standards, accuracy, validity, as well as ethics, privacy, and security, remain subjects requiring further exploration.

Therefore, our ultimate goal is to determine how we collect data, how we build such a platform, and how we conduct data mining and apply data algorithms to better serve medical diagnosis and treatment.

Applications of Medical Big Data: It can assist hospitals and physicians in conducting disease risk screening, facilitating early diagnosis, and supporting patients’ post-treatment rehabilitation.

Therefore, we can help pharmaceutical companies identify drug candidates with greater precision, assist insurance providers, and facilitate early health management, disease management, and chronic disease management for patients. Furthermore, we can support governments in managing public epidemics and enhancing the overall efficiency of healthcare systems.

Therefore, if our healthcare system can truly integrate the aforementioned elements, we will have genuinely entered the era of smart healthcare. Consequently, the development of the internet, big data, and 3D printing will inevitably disrupt the current healthcare system.