Can a $300M, 50,000-Participant Study Crack the Code on Multi-Cancer Early Detection?

A single blood draw for multi-cancer screening has become a consensus among participants in early cancer detection programs.

As domestic companies such as Burning Rock Biotech gradually unveiled their pan-cancer research and development initiatives, New Horizon Health unexpectedly launched a large-scale cohort study for early pan-cancer screening. “Achieving early screening and diagnosis of multiple high-incidence cancers with a single blood draw is the ‘Mount Everest’ that researchers worldwide are striving to conquer,” stated Zhu Yeqing. Following the completion of internal validation of its technical platform’s performance, New Horizon Health swiftly joined this endeavor.

On November 1, the Peking University Health Science Center and New Horizon Health officially launched the PANDA study (Pan-cancer Early Detection in China), a cohort study focused on early screening and diagnosis of pan-cancers. Concurrently, the joint laboratory established by New Horizon Health and Peking University Health Science Center was unveiled. “We regard the development of pan-cancer products as part of our long-term strategy,” stated Zhu Yeqing at the ceremony. The PANDA project is scheduled to span six years, with New Horizon Health investing over RMB 200 million in R&D and enrolling 50,000 participants.

It is understood that this study selected more than 20 cancer types, including lung cancer, breast cancer, colorectal cancer, gastric cancer, and liver cancer, from the top 10 cancers with the highest incidence and mortality rates to construct algorithmic models. Over 50,000 participants will be enrolled in four phases. Specifically, the initial algorithm model development phase (PANDA-1) is a retrospective study enrolling 7,500 participants; the model optimization and finalization phase (PANDA-2) is a retrospective study enrolling 5,000 participants; the independent model validation phase (PANDA-3) is a prospective study enrolling 17,500 participants; and the real-world cohort study phase (PANDA-4) will enroll 20,000 participants.

In the race for early cancer screening, there has never been a shortage of bold and strategic pioneers. Can New Horizon Health, having entered the field midway, rise to the challenge of pan-cancer early detection and make multi-cancer screening from a single blood draw a widespread practice?

VCBeat analyzed in an article several months ago that pan-cancer early screening has become a fiercely contested battleground for companies in the tumor early screening industry.

For instance, Burning Rock Biotech launched PREDICT in 2020, the first prospective, multi-center pan-cancer early screening clinical study in China, with a planned enrollment of 14,000 participants. New Horizon Health, known for its ambitious strategies, entered the field relatively late but has presented a highly clear roadmap.

The fervent development of pan-cancer early screening products is driven, on one hand, by the fact that pan-cancer early screening addresses a genuine need. Data indicates that approximately one-third of people will develop cancer during their lifetime; this proportion rises to one-half for those who live beyond 80 years of age. Cancer prevention and treatment have long become unavoidable topics in daily life. However, in clinical practice, most cancers do not present specific physical symptoms in the early or even middle-to-late stages. The fear associated with cancer stems largely from the psychological and physical torment of being diagnosed at an advanced stage. Currently, there are no clinical screening tools with sufficient sensitivity for most early- to mid-stage cancers. If a test could highly specifically and sensitively screen for the likelihood of early-stage cancer in a single assay, while narrowing down the potential tumor location to just one or two sites, it is unlikely that anyone would truly refuse it.

On the other hand, pan-cancer early screening technologies are continuously evolving to a level that enables large-scale deployment in real-world settings. Prospective clinical data indicate that algorithmic models are projected to reduce cancer-related mortality by 39%, with the potential to further lower cancer incidence. This progress is partly attributable to the development and application of DNA methylation technology in early diagnosis and screening.

However, overall, multi-cancer early detection (MCED) has yet to find a solution that breaks through the technical bottlenecks. Taking the CCGA-3 data released by Grail in June 2021 as an example, after enrolling a cumulative total of 150,000 participants and spending over $400 million, Grail’s MCED product, Galleri, achieved an overall sensitivity of only 51.5% and a tissue-of-origin accuracy of 88.7%. When analyzed by clinical stage, the sensitivity for Stage I patients was merely 17%, and 40% for Stage II. Clearly, these results do not meet the requirements for large-scale commercial application, yet they represent nearly the best performance achievable by current products under development.

The underlying reason is that the technical challenges in developing pan-cancer early screening products are immense, making them the veritable “crown jewel” of the tumor early screening field. This is because pan-cancer early screening requires simultaneous and accurate determination of both cell nature and tissue origin, while the available biological signals are weaker than those in other types of tests.

As is well known, cancer signals originate from tumor cells. Specifically, a single human tumor cell contains one DNA molecule weighing approximately 6 pg, up to 10 million copies of RNA molecules with a cumulative weight of about 20 pg, and roughly 42 million protein molecules. These constitute all the potentially detectable cancer signals. In early-stage cancer, the number of tumor cells is low, resulting in very weak cancer signals. Taking lung cancer as an example, in patients with early-stage lung cancer, the tumor diameter is approximately 1 cm, containing a total of about 1 billion cells. If 15 mL of peripheral blood is drawn for testing, it would contain tumor-derived signals equivalent to only approximately 1.7 copies of genomic data, making detection extremely challenging.

Currently, biomarkers such as plasma free proteins, gene sequence mutations, genomic copy number variations (CNVs), DNA methylation, and DNA fragmentation are widely used in research related to tumor diagnosis and treatment. However, these markers either exhibit low signal abundance or limited signal dimensionality, making it difficult to achieve the dual functions of capturing tumor signals and tracing their tissue of origin.

In practice, relying solely on existing biomarkers makes it extremely difficult to further improve the accuracy of pan-cancer early screening, regardless of how model designs are optimized. Unlike most established companies that focus on integrating existing biomarkers, the development of novel biomarkers and the construction of multi-omics algorithmic models are becoming key innovation trends for new entrants in the field of pan-cancer early screening.

Based on the publicly available information from the PANDA research project, New Horizon Health’s pan-cancer early screening product has selected a highly sophisticated biomarker: cell-free RNA (cfRNA).

This is a biomarker that is easily overlooked, yet it exhibits relatively high signal abundance and dimensionality, making it a crucial source of tumor signals. Traditionally, RNA fragments were considered unstable, which led to relatively slow progress in their clinical application and development. However, practical experience with COVID-19 vaccines has demonstrated that RNA fragments possess high stability when stored under appropriate conditions.

Studies have shown that at 1 cm³In early-stage tumor tissues of varying sizes, fewer than 10% of tumor cells are undergoing apoptosis at different stages. Most apoptotic cells are phagocytosed by macrophages in the body and do not release DNA. A smaller subset of tumor cells escapes phagocytosis and passively releases DNA, resulting in small amounts of cell-free DNA. In contrast, every viable tumor cell secretes large quantities of RNA, particularly small non-coding RNA molecules, which circulate freely in the bloodstream.

Guided by the traditional central dogma, research has primarily focused on messenger RNA (mRNA) that facilitates DNA-encoded protein synthesis. In reality, coding mRNA constitutes less than 5% of total cellular RNA. In other words, non-coding RNA represents a highly abundant source of biological signals. These RNAs and their fragments are present in substantial quantities in bodily fluids such as blood, urine, saliva, and feces. A growing body of evidence indicates that they play critical roles in regulating vital physiological processes.

Dr. Chen Yiyou, Chief Scientist at New Horizon Health, stated that another key reason for selecting cell-free RNA as the biomarker to build a pan-cancer early screening algorithm model is that the core intellectual property patents for biomarkers such as DNA fragmentation and DNA methylation are largely held by overseas companies. To forge a path for pan-cancer early screening with independent international intellectual property rights, it is essential to explore new dimensions.

Currently, New Horizon Health’s proprietary STAR-seq technology platform enables mixed library preparation of cell-free DNA and RNA, allowing for the simultaneous detection of both DNA and RNA signals in samples through a single sequencing reaction. This serves as the foundational technology underpinning New Horizon Health’s strategy for pan-cancer early screening. While the technological innovation itself may not present overly striking features, the continuously accumulating research data are steadily validating the robustness and efficacy of the STAR-seq platform.

Following the proposal of this technological concept, New Horizon Health rapidly completed a series of small-scale preliminary validations, which laid the foundation for the internal approval of its pan-cancer project. Previously, New Horizon Health applied its STAR-seq technology platform to the development of Gan Zheng Qing, an early-screening product for liver cancer, and released the preliminary research data for Gan Zheng Qing at the Chinese Society of Clinical Oncology (CSCO) Annual Meeting in September 2021.

In this study, New Horizon Health enrolled a total of 645 participants, including 264 patients with confirmed liver cancer, aiming to closely mirror the high-risk target population for real-world screening rather than focusing on patients with intermediate or advanced-stage liver cancer. The positive cohort primarily included patients with stage A or B liver cancer who had alpha-fetoprotein (AFP) levels below 400 ng/mL, or even below 20 ng/mL. The control cohort mainly comprised individuals at high risk for liver cancer, such as those with liver cirrhosis or hepatitis, as well as healthy individuals. Data demonstrated that, through high-throughput sequencing combined with multiplex protein detection and proprietary artificial intelligence algorithms, both the sensitivity and specificity of GanZhengQing in determining positive and negative results reached 97.9%.

In terms of tissue-of-origin identification, New Horizon Health has also leveraged public databases to validate the performance of its STAR-seq technology platform in tracing the origins of 14 major types of cancer. According to Dr. Chen Yiyou, taking breast cancer as an example, the STAR-seq technology confirmed that more than 210 out of 215 breast cancer cases originated from breast tissue during model validation. Overall, the tissue-of-origin identification accuracy of the STAR-seq technology reached 96% in the training set and 95% in the validation set.

In terms of cancer type coverage, it is understood that New Horizon Health has selected 14 key cancers, encompassing 10 of the highest-mortality cancers and 10 of the highest-incidence cancers in China. These cancer types account for more than 80% of all cancer cases and cancer-related deaths. Furthermore, across different application scenarios, New Horizon Health plans to incorporate an additional eight cancer types, thereby increasing the covered population proportion to 90%. “The design concept of the PANDA project is to cover no fewer than 22 types of cancer; this is our goal,” stated Dr. Chen Yiyou.

Regarding the screening accuracy of the PANDA project, Dr. Chen Yiyou outlined New Horizon Health’s dual objectives. The first is to accurately distinguish between cancer and non-cancer cases. “We aim for a 98% accuracy rate when no cancer signals are detected, meaning that the specificity of New Horizon Health’s multi-cancer early detection (MCED) product should exceed 98%.” Secondly, based on positive results, New Horizon Health aims to achieve a tissue-of-origin identification accuracy of over 90%, as verified by clinical gold standards such as endoscopy and imaging. At the same event, Zhu Yeqing also stated that, unlike single-cancer early detection products, which focus more on maximizing negative predictive value, the development of MCED products should strive to maximize positive predictive value to minimize the likelihood of subjecting users to unnecessary clinical examinations.

Prior to the announcement of the PANDA research project, New Horizon Health was widely recognized as the most prominent driver in commercializing single-cancer early screening products. From its PP Tube and Colotect Plus to YouYou Tube, Cervical Clear, and GlycoClear, New Horizon Health’s single-cancer early screening portfolio covers colorectal cancer, gastric cancer, cervical cancer, and liver cancer—some of the most prevalent cancers in China—thereby establishing itself as an oncology early-screening enterprise with platform-level capabilities.

In a sense, New Horizon Health has galvanized the development and practice of early cancer screening. Now, its high-profile entry into pan-cancer early screening is not only a strategic move aligned with industry trends but may also usher in a new era of product commercialization for early cancer screening in China.

So, even with New Horizon Health entering the fray, will this six-year R&D marathon reach the finish line? And will the view at the finish line be what everyone hopes to see? From VCBeat’s perspective, at least looking forward from the present, New Horizon Health has made adequate preparations in terms of “personnel, funding, and resources” for the grueling battle to develop a pan-cancer early screening product.

First and foremost, the human element. Zhu Yeqing told VCBeat that New Horizon Health’s entry into pan-cancer early screening aims to achieve the productization of this highly challenging technology, which undoubtedly imposes higher requirements on early-stage research. To this end, New Horizon Health assembled a robust pool of clinical experts for the PANDA project study, with Academician Qiao Jie, Executive Vice President of Peking University and Director of its Health Science Center, serving as the principal investigator (PI). This initiative mobilized more than ten hospitals within the Peking University Health Science Center system to participate jointly, thereby ensuring the smooth progress of the research from the perspectives of expertise, scientific research, laboratory resources, talent development, and clinical trials.

In his address, Academician Qiao Jie expressed strong endorsement of the STAR-seq technology platform and its early validation studies, placing high hopes on the R&D collaboration between Peking University Health Science Center and New Horizon Health. “We believe that the PANDA study project—a nationwide cohort for early screening and diagnosis of pan-cancers in China—will certainly contribute to the Healthy China Initiative, accelerate the construction of a new development paradigm, and promote high-quality development.” In her view, “The PANDA project will establish a joint laboratory of international caliber, assemble a ‘national team’ for R&D, conduct rigorous clinical trials, and help China achieve preventable and treatable outcomes for high-incidence cancers at an earlier date.” This top-down endorsement undoubtedly serves as a key pillar supporting New Horizon Health’s advancement in the development of pan-cancer early screening products.

Next is cash flow. At every stage of developing pan-cancer early screening products, the financial strength of R&D enterprises is put to the test. Substantial investments are required, whether for processing massive amounts of bioinformatics data during algorithm optimization or for sustaining the human and material resources needed in large-scale prospective studies.

Since its listing in 2021, New Horizon Health has maintained triple-digit revenue growth and accumulated substantial cash reserves. According to the latest financial report, the company generated RMB 226 million in revenue in the first half of 2022 alone, a year-on-year increase of 414%, with a gross profit of RMB 185 million, up 650% year on year. Its gross margin rose from 56.2% in the first half of 2021 to 82.0%. The company is expected to reach break-even by 2024, providing New Horizon Health with the financial capacity to support the entire research endeavor.

Finally, there is technical prowess. Similar to how it directly benchmarked against Exact Sciences, the global leader in early cancer screening, during the development of its single-cancer early screening products, New Horizon Health anchored itself to Grail from the outset of its multi-cancer early screening product R&D. According to Zhu Yeqing, New Horizon Health chose biomarkers and algorithmic models entirely different from those of Grail to tackle the challenge of multi-cancer early screening, akin to attempting to summit Mount Everest via a different route. “We aim to overtake despite starting later.” The PANDA study project represents the world’s first breakthrough innovation covering the full dimensions of DNA, RNA, and proteins to achieve liquid biopsy-based multi-cancer screening and early diagnosis. It has also demonstrated potential advantages in cancer detection and tissue-of-origin tracing in preliminary studies.

Of course, we still cannot predict whether New Horizon Health’s determination will change the accessibility of pan-cancer early screening. However, this large-scale endeavor will inevitably drive such products toward commercialization; we simply hope that this transition arrives faster and more steadily.