Fuzhou Dache Launches Cancer Data Analytics Platform Powered by Machine Learning to Advance Precision Oncology

Current cancer treatments often adopt a “one-size-fits-all” approach that fails to stratify cancers within the same category, frequently resulting in suboptimal therapeutic outcomes. Precision medicine, which analyzes patients’ tumor genomic profiles and incorporates individual variations in genetics, environment, and lifestyle into disease prevention and management strategies, offers a promising solution to enhance the efficacy of diagnosis, treatment, and prevention, as well as to improve survival rates among cancer patients.

Nowadays, with the development of cross-disciplinary applications involving genomic sequencing technologies, bioinformatics, and big data science, precision medicine has gradually reached maturity. The upgrading of healthcare consumption has further promoted market expansion, leading to a continuous emergence of innovative solutions.

Fuzhou Dache Precision Medicine Technology Co., Ltd. (hereinafter referred to as “Fuzhou Dache”) specializes in research on the molecular subtyping of cancer.The company performs molecular subtyping of patients based on tumor gene expression profiles through AI-driven big data analysis, thereby achieving the goals of precision medicine.Meanwhile, by integrating appropriate treatment regimens, this solution can avoid the severe toxic side effects associated with inappropriate chemotherapy drugs, thereby reducing the suffering endured by patients during treatment.

It is reported that an invention patent for breast cancer subtyping has been granted. The patent for the four-subtype classification of gastric cancer was also published in February this year and has entered the stage of substantive examination. Clinical data indicate that the four-subtype classification method can significantly improve the survival rate of cancer patients. Furthermore, the company’s solutions have accumulated extensive clinical statistical data from cancer patients, and several software copyrights for subtyping algorithms have been approved. Among these, the company’s self-developed molecular subtyping technology for gastric cancer has matured and is now being promoted in the market.

“Proactively and precisely determining the optimal therapy, timing, and treating physician for patients represents a major challenge facing the healthcare industry today. Although we are accustomed to relying on experience to guide diagnostic and therapeutic decisions, machine learning approaches can perform in-depth analyses of disease progression, further evaluate deep phenotypes and risks, maximize clinical outcomes, and thereby realize precision medicine.”Lei Zhengdeng, Chairman and Chief Scientist of Fuzhou Dache, stated.

In his conversation with VCBeat, Lei Zhengdeng revealed the critical role of “machine learning” in Fuzhou Dache’s solutions, stating that it has always been his strategic direction.

Although he majored in Chemistry at Peking University during his undergraduate studies, he devoted more than half of his college years to learning computer programming. Since beginning his doctoral studies, he has conducted extensive research on the application of machine learning in biomedicine.

Driven by a strong interest in machine learning, Lei Zhengdeng pursued his Ph.D. in bioinformatics in the United States, with a focus on machine learning research. After graduation, he joined Memorial Sloan Kettering Cancer Center in New York as a High-Throughput Computational Analyst and Leader of the Bioinformatics Group, where he engaged in high-throughput cancer drug screening and signaling pathway analysis, collaborating with world-leading cancer experts including Nobel laureate and former NIH Director Harold Varmus.

To further advance the development of precision medicine, Lei Zhengdeng returned to academia to conduct research on molecular subtyping of cancer. He pursued postdoctoral research at the Duke-NUS Medical School in Singapore, focusing primarily on personalized treatment for gastric cancer. His supervisors, Steve Rozen and Patrick Tan, are chief scientists of the International Cancer Genome Consortium and recipients of Singapore’s highest scientific award.

After completing his postdoctoral fellowship, he returned to the University of Illinois Chicago in the United States to serve as a Bioinformatics Specialist, where he engaged in high-throughput data analysis and the development of novel algorithms.

To date, Lei Zhengdeng has published more than 50 academic papers in international journals such as Gastroenterology (impact factor: 22.7) and Gut (23.1), with his work cited over 2,500 times. He also serves as Associate Editor for the British academic journal BMC Research Notes and as a reviewer for more than ten international journals and conferences, including Gastroenterology (impact factor: 22.7) and Bioinformatics (6.9).

Industry exploration, the accumulation of academic experience, and long-term focus on a specific field have enabledLei Zhengdeng is eager to translate his research findings into concrete solutions, ultimately implementing them in clinical settings.In late 2016, Lei Zhengdeng founded Fuzhou Dache Precision Medicine Technology Co., Ltd., and served as its Chairman and Chief Scientist.Leveraging artificial intelligence technology, we have independently developed a comprehensive suite of data analysis systems and corresponding solutions.

Given the significant clinical value of the company’s proprietary cancer molecular subtyping method, it has established collaborations with multiple hospitals and large enterprises to jointly advance the research and application of precision medicine technologies in oncology treatment.

Precision medicine originates from genomics, which involves the application of genomics and its derived omics disciplines (such as proteomics, metabolomics, and epigenomics) at the genetic and molecular levels to diagnose etiologies or identify therapeutic targets. This approach enables the transformation of previously uniform treatment regimens into effective therapies tailored to individual etiologies.

Fuzhou Dache's currentIts core business area is the molecular subtyping of gastric and breast cancer patients, based on tumor gene expression profiles and leveraging the gastric and breast cancer data analysis system developed by Lei Zhengdeng.andRecommend corresponding treatment plans and medication guidance based on the typing results to overcome the drawbacks of traditional treatment models and achieve the effects of precision medicine.

Fuzhou Dache shared a set of data with VCBeat:

First, the proportion of patients currently using targeted therapies is small. For instance, patients with HER2-positive gastric cancer account for only about 10%. Meanwhile, the proportion of patients testing positive for gene mutations via DNA sequencing is also low.

Second, the use of targeted therapies in HER2-positive gastric cancer patients improves survival rates by only approximately 10%. However, guiding medication based on cancer genotyping can significantly enhance survival rates for cancer patients.

Third, the response rate to immunotherapy guided by PD-L1 is approximately 50%, whereas that guided by genomic profiling reaches 90%.

Compared with current pharmacological treatments, guiding medication based on cancer genotyping offers significant advantages. Based on the clinical data obtained, the company has foundMolecular subtyping derived from gene expression RNA sequencing reveals differential drug sensitivities. Specifically, patients can truly benefit and achieve improved survival rates only when treated with specific drugs tailored to their particular molecular subtype.

However, achieving true precision medicine based on cancer genotyping requires overcoming numerous challenges. Genes mutually regulate one another to form complex gene networks, and the increasing volume of research data makes it difficult to uncover the underlying mechanisms of pathogenesis.

Fuzhou Dache derives molecular subtypes through unsupervised clustering analysis, then leverages existing data as training sets to continuously refine its algorithms, establishing a positive feedback loop. Ultimately, it employs specialized algorithms to classify genomic data while integrating patient clinical profiles, survival outcomes, and medication usage records, utilizing machine learning techniques to identify disease patterns.

Challenges in Precision Therapy Through Molecular SubtypingThe challenges lie in the large volume of data and significant data noise. In-depth analysis of numerous details is required to identify the optimal model. Generally, in scenarios with high-dimensional data, machine learning models are prone to overfitting, a phenomenon where the model performs well on the current training data but yields inaccurate predictions on new data or demonstrates poor performance in practical applications.

Fuzhou Dache has taken note of this issue,Minimize the negative impact of overfitting by developing proprietary algorithms.According to clinical data,Fuzhou Dache’s solution not only significantly improves cancer patient survival rates, but also demonstrates highly reproducible results, with consistent patterns observed across patient cohorts in Singapore, South Korea, the United States, and other regions.

It is reported that,The company conducts research on breast cancer, lung cancer, colorectal cancer, pancreatic cancer, and other malignancies, and has identified common molecular subtypes across these diseases. In the future, leveraging its core technologies, the company will continue to collaborate with hospitals to explore additional molecular subtypes, thereby enabling its solutions to be applied in a broader range of cancer treatment scenarios.