Digital Construction of Pathology Department of Ruijin Hospital:
Reading 1000 Slide Images Per Second
If you observe the daily work of pathologists, you may find them not the same as traditional doctors who run routine diagnostic process, but more like lab scientists.
For most diseases, especially tumor diseases, pathological diagnosis is recognized as the gold standard of clinical medical diagnosis. For example, if a patient is diagnosed with a tumor, determining whether the tumor is benign or malignant is critical. If it turns out to be malignant, then the patient is likely to be diagnosed with cancer. Despite the availability of preliminary diagnosis through B-scan ultrasonography, CT, and MRI, pathology is the decisive factor for doctors when arriving at a final diagnosis. It enables doctors to determine the nature of the disease, which significantly affects the method of treatment, prescription and prognosis for the patient. The level of pathological diagnosis affects the misdiagnosis and missed diagnosis rate to a large extent. For cancer patients in particular, it is usually a life-and-death situation. Unreliable pathological reports can often result in either overtreatment or undertreatment. As Zhong Nanshan, Academician of Chinese Academy of Engineering said, the level of clinical pathology is an important sign of the quality of national medical care.
Ruijin Hospital Affiliated to Shanghai Jiao Tong University School of Medicine (referred to as Ruijin Hospital), a grade-A tertiary hospital (comprehensive hospital of the highest level) in China, integrates medical care, teaching, and scientific research. It has ranked No. 1 in Shanghai and East China hospitals for several consecutive years. As one of the key subjects of Ruijin Hospital, the pathology department has the state-of-the-art pathological examination platform and significant soft power. As a result, it receives massive pathological diagnosis from all over the country. The pathology department of Ruijin Hospital livestreams discussions on difficult pathology-based diagnosis every two weeks, which has become a major learning platform for pathologists in China.
Wang Chaofu, the chief of the pathology department of Ruijin Hospital and head of department of pathology of Shanghai Jiao Tong University School of Medicine, is a renowned pathology expert in China. In 2020, he led the first pandemic etiology research expert group dispatched by Shanghai Jiao Tong University School of Medicine to offer the assistance in Wuhan. There are many stories of patients he received that helped provide insights into the importance of an accurate pathological diagnosis.
In 2010, Doctor Wang attended a kidney tumor consultation for a 32-year-old male. The tumor was diagnosed as benign in the original hospital. However, the first intuition told him that it was not benign once he read the slide. The tricky part was that he had never seen such a tumor, and there was no corresponding kidney tumor in the World Health Organization (WHO) tumor classification. He had a gut feeling that it might be a new type of kidney cancer, which was confirmed later on. It made Doctor Wang the first doctor to discover this type of kidney cancer in China, when only 17 cases were reported in the world.
In early 2016, a 30-year-old female patient from Jiangsu had a tumor in her ankle. The local hospital organized several consultations for this difficult case and finally concluded that the tumor was bone cancer, requiring amputation. Following a consultation with Doctor Wang, it was diagnosed as ossifying myositis (benign lesion). In the end, the young mother's leg was saved.
Since 2018, a male patient in Jiangxi had experienced frequent lymphatic swelling in his thighs, head, and neck. The pathology department of multiple hospitals in China were unable to confirm the diagnosis after a consultation. In despair, the elderly patient even expressed his willingness to donate his remains for scientific research. In July 2020, after going through hardships and wandering around, his daughter finally got to Ruijin Hospital. After visiting Doctor Wang, the patient was diagnosed with a special form of peripheral T-cell lymphoma. After treatment, he has recovered and discharged from the hospital.
There are more difficult miscellaneous diseases alike. However, not every patient is that lucky to come across an expert like Wang Chaofu for in-person consultation in time. It means that many people are unfortunately misdiagnosed or missed diagnosed. The distribution of medical resources in China is also heavily unbalanced, the pathology field is no exception. Most pathologists work in grade-A tertiary hospitals. The number of registered pathologists is less than 20,000, with a gap of 140,000. Moreover, both secondary and tertiary hospitals are in desperate need of pathologists. China has the highest incidence of cancer in the world, and therefore it is high time to transfer the pathology capabilities of large tertiary hospitals such as Ruijin Hospital to grassroots hospitals.
Besides, well-known departments such as the pathology department of Ruijin Hospital undertake an influx of various pathological examination projects of clinical departments every day. There are more than 3500 pathological slides per day, amounting to a total annual slide volume exceeding 1 million. Therefore, pathologists are in dire need of a more manageable workload. On top of this, storing physical pathological slides poses significant challenges, as they tend to fade and become unusable after several years. The lack of convenient information management (such as handwritten labels) also restricts the full-process tracking and clinical research of samples.
All of these problems now point to a solution, that is, digitization.
The application of the digital pathology system began in 1985. In the 1990s, the digital pathology system was applied to the commercial field in the United States, and since 2000, it gradually replaced the traditional microscope in medical schools. Since then, 50% of medical schools in the United States and around the world have introduced or are preparing to bring in digital pathology systems.
Compared with the traditional pathology system, the digital system effectively addresses a range of pain points as just mentioned. For example, precious pathological slide data can be permanently stored in a digital library, thereby solving problems such as storage, fading, damage, loss, and retrieval of glass slides. Doctors can read any area of a digital slide image at any magnification ratio they need anytime and anywhere. The monitor can show both the entire view of a digital slide and the zoomed-in image of a specific area to assist high-precision diagnosis. At the same time, the digital pathology system supports side-by-side display of multiple images for comparison and analysis, which can be easily shared with other doctors. This greatly benefits consultations, telemedicine, online meetings where slide images are shared and examined, as well as online teaching to cultivate more pathology talent.
However, the digital pathology system also faces new challenges. For example, the generation, storage, and reading of massive digital slide images pose challenges on the data storage system which is the ICT foundation for digital slide image reading. The reading efficiency and accuracy of slide images are severely affected by long latency, frame freezing, and mosaics. The data volume of a single digital slide image is equivalent to that of an HD movie. Hundreds to tens of thousands of pathological slide image data can be generated every day depending on the size of the hospital. This data needs to be stored long-term to meet regulatory compliance requirements, which burdens the storage resources of hospitals. The requirements ranging from the quantity of equipment, the size of equipment rooms, and the backup of data, to the fast and effective slide image reading without mosaics are extremely complicated, requiring a high level of expertise.
Not to mention issues such as confidentiality, sharing, and defense against external hacker attacks in daily use. These potential risks have undoubtedly become stumbling blocks for many hospitals in their way to build digital pathology systems.
To overcome these problems and improve the analysis efficiency of massive digital slide images, the pathology department of Ruijin Hospital cooperates with Huawei data storage. It is a qualitative leap in high-performance digital pathological slide image reading and data reduction technologies based on Huawei OceanStor Pacific scale-out storage, resolving key pain points in digital pathology.
To name a few pain points, the slow reading and frequent mosaics and tens of seconds' frame freezing in traditional image reading severely hinder doctors' ability to read images and make diagnosis. In addition, the pathology department of Ruijin Hospital imposes higher requirements on storage systems. That is, clinical, scientific research, and teaching personnel in the hospital should be able to quickly access pathological slide images in real time, accelerating the value conversion of pathological data. After adopting Huawei OceanStor Pacific scale-out storage, the pathology department of Ruijin Hospital achieved reading 1000 slide images within a second, 80 times more than that of traditional storage.
To solve the issue of insufficient storage capabilities, Huawei's unique secondary compression algorithm for digital pathology implements intelligent identification of pathological data, achieving a data compression ratio of over 30%. The high-density hardware design of OceanStor Pacific scale-out storage enables less space to store more data for a longer time, saving 73% space of the equipment room for the hospital.
The unbalanced distribution of medical resources and development of medical level are ongoing issues in China, so providing high-quality medical services for as many people as possible under limited conditions is not only a vision of all doctors and patients, but also the vision of science and technology. Through long-term technological innovation, Huawei deeply integrates technologies with healthcare applications, making breakthroughs in fundamental digital pathology technologies and streamlining the last mile of hospital digital transformation.