Professor Bu Hong of West China Hospital, Sichuan University: Intelligence and Cloud-Based Solutions Are Inevitable Trends for the Future of Pathology
From December 20 to 22, the 2019 Future Healthcare 100 Conference, hosted by VCBeat and themed “New Growth in Life,” was grandly held at Jiuhua Resort in Beijing.
At the Digital Pathology and Precision Diagnosis Forum, Professor Bu Hong from the Department of Pathology at West China Hospital of Sichuan University delivered a presentation titled “The Future of Pathology: Intelligence and Cloud Integration.” This article provides a comprehensive compilation of the highlights from his talk.
The following is an edited transcript of the speech:
Intelligent and cloud-based pathology is an inevitable trend for the future.Intelligentization represents a further advancement in the integration of artificial intelligence into pathology, while cloud-based transformation signifies a shift in pathological practice from individual work to collaborative, group-based workflows.
Pathological diagnosis is evolving toward specialty pathology grounded in clinical treatment, with molecular pathology and digital pathology representing key developmental directions. Today, we emphasize that pathology is not an individual empirical endeavor; its standardization is even more critical than the recognition of tissue morphology.It was previously believed that pathologists required a greater degree of morphological intuition; in reality, pathological diagnosis is a field that can be standardized, replicated, quantified, and normalized, which forms the foundation for its development toward intelligence.
Pathology currently faces numerous issues and shortcomings, with suboptimal "user experience" being a significant aspect. This includes factors such as report turnaround time, report format, report content, and content standardization.
In my view, the development of intelligent and cloud-based pathology can be broadly divided into the following stages:
Phase I: Technological Innovation and Telepathology. Whole-slide digital scanning marks the beginning of digital pathology.
Phase II: Widespread Adoption and Digital Pathology Departments. Comprehensive digitalization of pathology departments and widespread adoption of digital pathology applications.
Phase III: Future Aspirations and Intelligent Pathology. The integration of AI-driven, data-based precision pathology.
Smart pathology must steadily progress through these three stages. In fact, practices in recent years have demonstrated that many hospitals are still at the first stage. Currently, many entry-level artificial intelligence products cannot be applied in the pathology industry. The primary reason is that the application of AI-based pathology products requires appropriate workflow scenarios and a digital foundation within the pathology department. As we have not yet completed this foundational stage, the practical implementation of smart pathology remains challenging.
Clearly, without full digitalization of pathology data, it is impossible to advance to the third stage of intelligent pathology. Many pathologists have noted that while there are indeed many excellent software solutions available now, rescanning slides for each diagnosis is time-consuming and labor-intensive for physicians. Experienced pathologists can typically resolve cases in just a few dozen seconds based on their expertise, leaving little room for intelligent pathology to demonstrate its value.
Without motivation, application scenarios, or a foundation in digitalization, the goal of intelligent pathology will be difficult to achieve.
Therefore, it is crucial to accelerate the digital transformation of pathology departments.
As a pathologist, I believe there are at least five reasons to substantiate that this is an inevitable trend. They are as follows:
I. Solutions to Staff Shortages
Staff shortages constitute a major challenge in the pathology industry, one that is unlikely to be effectively resolved in the foreseeable future. Big data statistics on the pathology sector this year indicate that personnel deficits will persist in the years ahead.
Statistical data shows that there are over 4,900 pathology departments nationwide, with more than 17,000 pathologists and over 13,000 pathology technicians.
Based on the statistics of professional titles, 9.8% of pathologists hold senior professional titles (Zhenggao), 19.5% hold associate senior titles (Fugao), and 70.7% hold intermediate or junior titles, with the vast majority being entry-level junior staff. Among pathology technicians, 4.5% hold senior professional titles, while 95.5% hold intermediate or junior titles.
Based on educational attainment data, among pathologists, 11% hold doctoral degrees, 33% hold master’s degrees, and 56% have bachelor’s degrees or lower. In contrast, among pathology technicians, only 1% hold doctoral degrees, 8% hold master’s degrees, and 91% have bachelor’s degrees or lower.
The number of personnel in pathology departments who truly possess the talents required for the new era is far from ideal. Coupled with the ever-increasing workload, the future outlook remains highly challenging.
II. Specialized Solutions
Currently, the average staffing level in the vast majority of primary-care pathology departments is around three personnel, making it extremely difficult to promote subspecialization. Without implementing subspecialization, pathology—operating as a second-level discipline—will remain in a persistently passive position when confronted with the clinical reality that has already progressed to fine-grained subdivisions into fourth-level disciplines and subspecialty directions, thereby failing to adequately meet clinical needs.
How can the challenges of specialization be addressed? I believe that in the future, this must rely on collaboration and resource sharing among all stakeholders—achievable only through the “cloud concept.”
III. Inevitable Requirements for the Operational Framework
I am accustomed to dividing the operational framework into roughly three tiers: national-level teams and national regional medical centers, tertiary hospitals, and secondary hospitals. This framework was originally designed to balance tasks across all levels; however, the current reality is that those at the top are “overworked to exhaustion,” while those at the bottom are “starved for resources.” Of course, such a dilemma can only be resolved through a cloud-based model.
What should the future landscape of pathology departments in China, or the development model for primary-care pathology departments, look like? We envision a model led by pathology departments at tertiary Grade A hospitals, with the baseline requirement of one pathology department per county, and collaborative development involving third-party laboratories, thereby forming a cross-regional collaborative framework.
IV. Continuous Advancements in Technical Methods
With the continuous advancement of technology, including 5G and other innovations, the enabling technologies for intelligence and cloud-based solutions will develop faster than we imagine.
V. Ample Evidence from Prior Practices
Practice over the past decade has demonstrated that the initial applications of intelligence and cloud-based solutions are successful and effective. What early traditional pathologists deemed impossible is gradually becoming a reality. China has completed more than one million cases of remote pathological diagnosis, successfully implemented remote initial pathological consultations and intraoperative frozen section diagnoses, and subspecialty remote pathological diagnosis has also begun to play a positive role.
Without dreams, there is no development and no future.The future vision of intelligence and cloudification should pursue: smart pathology, happy pathology; specialization and quality improvement; continuous enhancement of reporting speed and efficiency; self-correction and error-prevention capabilities; automated integration of pre-analysis functions; and balanced and equitable development of tiered diagnostic structures.
At this year’s National Pathology Annual Conference, I mentioned that pathology should be “intelligent pathology” and “fulfilling pathology.” While many felt this vision was still distant, it is, in fact, not so far off.
We can envision a future where pathology, driven by cloud computing and artificial intelligence, transforms the workflow of pathologists. Instead of rushing through congested city traffic to reach the hospital, a pathologist could simply start their day at home, logging into their computer to review data processed overnight by intelligent systems along with preliminary reports, and finalize the diagnosis after further verification. Once transmitted via the cloud, smart tools assist in error checking, report integration, and providing clinically standardized recommendations, thereby generating formatted reports. The day will surely come when we can serve patients while relaxing by the seaside under the sun.
The reason many technologies fail to perform effectively is that we have overlooked the need for final implementation in real-world work settings, which require the participation of pathologists.While engineers can handle the technical aspects, the scenarios must be refined by pathologists.
Future pathology should not be limited to interpreting and applying only self-generated data, but should also encompass the interpretation and integration of data generated across the entire disease domain.This is the vision of future pathology. Perhaps the pathological diagnosis system in the future will be a more complete and systematic integrated system, rather than following the traditional mindset and framework of today's pathology.
The field of pathology is undergoing significant changes, and every aspect of these transformations requires the support of intelligence and cloud-based solutions. Only by achieving these two milestones can future developments be successfully realized.
Currently, the intelligentization and cloud-based transformation of pathology can be achieved by focusing on the following six key areas: 1. Proper preservation of specimens; 2. Standardization of tissue sampling and slide preparation; 3. Innovation in staining and non-staining techniques; 4. Continuous improvement of intelligent and cloud-based technologies; 5. Ongoing advancement of diagnostic protocols and theoretical frameworks; 6. Continuous enhancement of the experience for both physicians and patients.
In fact, today we still adhere to the mindset of traditional pathology and continue to implement the biomedical model, having not yet transitioned to a truly modern medical paradigm.
I think,Addressing these challenges in pathology cannot be achieved by the pathology sector alone; rather, it requires collaboration with enterprises and cross-industry partners, which represents the correct path to finding solutions.
These are my basic thoughts on the future trend toward intelligence and cloud integration.