"Describing life with numbers is the underlying logic of computational medicine," stated Zhao Yu, a researcher at the Institute of Computational Technology, Western Branch of the Chinese Academy of Sciences. "Many people don't understand the difference between genetic testing and computational medicine. I often use this analogy to explain: you can't walk into a wheat field and ask the person planting the wheat to make you a bowl of noodles; they can't do it. In other words, genetic testing only produces data, while computational medicine interprets and processes data using artificial intelligence or mathematical methods."
According to statistics, there are over 50,000 diseases worldwide, with over 14,000 clinically recognized in China alone. Many of these diseases lack effective treatments, and even when medications are available, drug pricing by pharmaceutical companies is often criticized.
Why is drug development so slow and expensive? Zhao Yu believes the reason lies in the outdated capacity of the existing drug development technology. Computational medicine not only reduces the cost of developing new drugs but also enables the repurposing of existing drugs.
"As a disruptive non-consensus technology, computational medicine often gives people a sense of the future or science fiction. But there is no doubt that it is already on the path of reality," Zhao Yu remarked.
China has nearly 10 million epilepsy patients, 40% of whom are women of childbearing age, with 80% not receiving standardized diagnosis and treatment. The infertility rate, stillbirth rate, and maternal mortality rate among female epilepsy patients are 11 times, 5 times, and 5 times higher, respectively, than those of ordinary women.
To address this, Dr. Chen Lei, Chief Neurosurgeon at West China Hospital, Sichuan University, and his team have established a precise health management system for female epilepsy patients throughout their entire lifecycle, from adolescence to postpartum. They have also promoted this system in nearly 200 medical centers across 34 provinces and regions nationwide.
"We have established an ASM teratogenic risk model based on FAERS data prediction. In addition, we have built a pregnancy ASM medication model based on population pharmacokinetics to propose recommended medication regimens for different clinical scenarios," Dr. Chen Lei said. They have also developed a remote epilepsy seizure monitoring and early warning system based on behavior recognition to ensure the safety of female epilepsy patients of childbearing age at home.
Through data-driven approaches, Dr. Chen Lei's team has also made innovative breakthroughs in epilepsy treatment methods, such as pioneering minimally invasive surgery for closure of patent foramen ovale internationally, with an effective rate of up to 70% in controlling drug-resistant epilepsy.
In breast cancer molecular subtyping, over 60% are HR+ subtype, with 20% diagnosed at an advanced stage, and a median survival period of only 2 to 3 years. CDK4/6 inhibitors are the preferred first-line treatment for HR+HER2- advanced breast cancer recommended by guidelines, but approximately 20% of HR+HER2- advanced breast cancer patients are intrinsically resistant to CDK4/6 inhibitors.
"This is very challenging because first-line resistance may delay the timing of second-line treatment for patients, or even lead to a loss of treatment confidence. So, we considered conducting a drug sensitivity prediction study to differentiate between sensitive and resistant populations, but we have not succeeded so far," said Dr. Yang Mei, Deputy Director of the Breast Cancer Department at Guangdong Provincial People's Hospital.
Due to the close relationship between breast cancer and genetics, Dr. Yang Mei's team turned to embryonic genomic pathogenesis research with computational medicine teams. "We integrated the germline genes of each individual to calculate and establish embryonic genomic pathogenesis. In a global research approach, we also combined germline gene variations with tumor variations to study the effectiveness and resistance of neoadjuvant therapy for triple-negative breast cancer. The results showed that this method can more effectively distinguish between resistant and sensitive populations."
Based on this, Dr. Yang Mei considered establishing a drug model from a new perspective. Since the action of drugs involves individuals, it is necessary to view the progression of tumors globally. Thus, patient digital twins were created, followed by tumor digital twins, and virtual clinical trials were conducted to validate whether they could produce outcomes similar to those of real-world data. The results showed that virtual clinical trials not only had comparable outcomes to real data but also explained possible mechanisms for differences.
"In fact, there are many deep learning and AI models in clinical practice, which can correlate images and features and show differences, but lack interpretability in efficacy and image correlation. Digital twins are different; they start from mechanisms and are interpretable, which helps further research into mechanisms and explore potential targets," Dr. Yang Mei said, adding that drug models based on this are a new way to understand diseases and cancer.
"Our team performs the most surgeries for skull base spinal cord tumors in the department, and possibly the most surgeries for skull base spinal cord tumors in the world," said Dr. Bai Jiwei, Chief Neurosurgeon at Beijing Tiantan Hospital, Capital Medical University. However, in terms of incidence, spinal cord tumors are rare diseases, with a prevalence rate of less than 1 in 100,000. In the realm of treatment, surgery for spinal cord tumors is acknowledged as a highly intricate and challenging procedure in neurosurgery. In recent years, surgery for skull base spinal cord tumors has gradually shifted from predominantly using traditional craniotomy approaches to primarily utilizing endoscopic transnasal routes, significantly increasing the rate of complete resection. As for radiation therapy, conventional radiation struggles to deliver high doses to lesions around critical structures such as the brainstem and optic nerves. Particle therapy devices, such as proton and heavy ion therapy, are limited in availability and costly, posing a considerable burden on patients. While radiation therapy can delay recurrence for some tumors, treating recurrent spinal cord tumors post-radiation presents even greater challenges, with a lack of effective salvage treatment options. Hence, Bai Jiwei believes that drug therapy represents the ultimate hope for spinal cord tumor patients, yet regrettably, there are currently no highly effective drugs available.
Recently, a collaborative study between Bai Jiwei's team and computational medicine researchers revealed that patients with high endoplasmic reticulum stress-associated cancer-associated fibroblasts (ERS-CAFs) have a shorter progression-free survival, higher stromal scores, higher immune scores, and increased expression of several immune checkpoints (PD1, PD-L2, CTLA-4, TIM-3, CD86), indicating relatively poor prognosis. However, the study also found that patients with high ERS-CAF scores in spinal cord tumors may be more sensitive to tyrosine kinase inhibitors (TKIs) and CDK4/6 inhibitors. Additionally, there have been preliminary indications of favorable treatment outcomes in recurrent and refractory spinal cord tumor patients using CDK4/6 inhibitors.
"Reflecting on this study, computational medicine has provided us with an efficient predictive method in drug development. Our next step involves conducting clinical trials based on CDK4/6 inhibitors, with hopes of achieving desirable outcomes," Bai Jiwei stated.
From a regulatory science perspective, Zhang Wei, Chairman of the China Association for Drug Regulatory Affairs, pointed out that digital twins and virtual clinical trials are pivotal technologies. However, the transition from models to tools is still in its early stages. "Regulatory authorities need to resonate with the development of new technologies, synchronously conducting research and development on regulatory science tools, evaluating and accepting digital evidence, to actively contribute to reducing the number of animal experiments and human trials, lowering the costs of drug development, and enhancing review efficiency in the future."
"The future world will undoubtedly be driven by computation, and the future of medicine will also continue to develop with the assistance of powerful computational tools," remarked Gu Chengming, Chairman of the China Conference on Innovative Drugs and Medical Devices (CAMC).
