New research demonstrates the potential of AI-driven cancer risk assessment, low-cost multi-cancer early detection, and economic integration with existing screening programs
SHENZHEN, China and SAN DIEGO, Calif. — September 21, 2026 — SeekIn Inc., a developer of blood-based technologies for cancer detection and monitoring, today announced that three studies from its research program have been selected for Rapid Fire presentations at the World Cancer Congress 2026, taking place September 24–26 in Hong Kong.
Organized by the Union for International Cancer Control (UICC), the World Cancer Congress brings together researchers, clinicians, policymakers, advocates and other cancer-control professionals from around the world to advance more accessible, innovative and effective approaches to cancer prevention, diagnosis, treatment and care.
The three SeekIn studies address challenges in implementing multi-cancer early detection (MCED) at scale: identifying individuals at elevated cancer risk, improving the performance and affordability of blood-based MCED testing, and integrating MCED into existing national cancer-screening programs without creating unsustainable healthcare costs.
Three studies to be presented:
1. AI-Empowered Cancer Risk Assessment Model Based on Routine Laboratory Test Data with Age for Enriching Individuals at High Risk of Cancer
This study evaluated OncoGate, SeekIn's AI-based cancer risk assessment model, which generates a Personal Cancer Risk Score (PCRS) combining routine laboratory data with age to identify individuals at higher risk of cancer and improve the efficiency of downstream MCED screening.
The retrospective analysis included 7,672 individuals, including 1,245 cancer cases, with 5,392 participants also evaluated using OncoSeek. In a simulated population of one million adults, OncoGate-based risk stratification reduced the population receiving MCED testing from more than 536,000 individuals under age-based screening to approximately 179,000, while detecting 2,402 cancers compared with 2,327 using age-based screening.
Compared with age-based screening, OncoGate increased modeled positive predictive value from 7.6% to 15.9%, reduced false positives by 55.2%, and reduced the modeled cost per cancer detected from $18,436 to $5,950.
The findings suggest that combining readily available routine laboratory data with AI-based risk assessment could provide a practical strategy for enriching individuals for subsequent MCED testing, potentially improving the scalability and cost-effectiveness of population-level screening.
2. OncoSeek: An AI-Empowered Blood-Based Multi-Cancer Early Detection Test with Large-Scale Validation
This study reports large-scale validation and further development of OncoSeek, SeekIn's AI-powered blood-based MCED assay. OncoSeek integrates seven protein tumor markers—including AFP, CA125, CA15-3, CA19-9, CA72-4, CEA and CYFRA21-1—with age and sex using a machine-learning framework to generate a probability-of-cancer index and predict tissue of origin. The assay was trained and validated in 15,122 participants, including 3,029 cancer cases and 12,093 non-cancer participants, across seven independent cohorts, seven centers, three countries, four analytical platforms and two sample types. The study found that the AI-based approach substantially reduced the accumulation of false positives associated with conventional interpretation of individual tumor markers. Specificity increased from 58.5% to 92.0%, with an AUC of 0.829 and 58.4% sensitivity. The model predicted tissue of origin with 70.6% accuracy and detected 14 cancer types representing approximately 72% of global cancer mortality.
An upgraded version, OncoSeek 2.0, incorporated three additional cancer-type-specific protein tumor markers and further improved sensitivity from 65.3% to 75.9% at 90.1% specificity, with an AUC of up to 0.926. Significant sensitivity improvements were observed across multiple cancer types, including prostate, head and neck, esophageal, cervical and lung cancers. Stage I–III sensitivity reached 67.1%, while stage IV sensitivity remained 86.5%.
With an estimated reagent cost of approximately $15 per test, the findings support the potential of OncoSeek as a scalable and lower-cost MCED approach, particularly in healthcare systems where affordability and accessibility are major barriers to advanced cancer screening.
3. Integrating Multi-Cancer Early Detection into US Cancer Screening Programs: A National Economic and Budget-Impact Modeling Study
This study evaluated how MCED could be incorporated into existing U.S. cancer-screening programs from a national economic and budget-impact perspective.
The analysis modeled screening among 75.6 million individuals using U.S. screening volumes, costs, population data, cancer incidence and published screening-test performance parameters.
Under the model, existing standard-of-care screening detected 197,773 cancers at a total cost of $43.2 billion. Adding OncoSeek Duet, a two-step strategy combining OncoSeek with SeekInCare at a modeled cost of $143 per person, increased the number of cancers detected to 398,586—approximately twice the standard-of-care yield—while increasing total screening expenditure by 25% and reducing modeled cost per cancer detected from $218,276 to $135,429.
The study also evaluated an optimized colorectal cancer screening pathway replacing colonoscopy with fecal immunochemical testing (FIT). The resulting savings could potentially be reinvested in OncoSeek Duet, producing a modeled two-fold increase in cancers detected while generating a $12.3 billion budget surplus. Together, the findings suggest that the economic feasibility of MCED may depend not only on the price and performance of the MCED test itself, but also on how MCED is integrated with and potentially complements existing single-cancer screening pathways.
The three studies collectively support a strategy in which AI-based risk assessment identifies people most likely to benefit from testing, an affordable blood-based MCED assay expands the reach of cancer detection, and health-economic optimization enables integration with existing screening infrastructure.
“We believe the future of cancer screening is not simply about developing another test, but about building a screening strategy that can be deployed at scale,” said Dr. Mao Mao, Founder and CEO of SeekIn. “Our three studies address three critical pieces of that challenge: identifying people at higher risk, making multi-cancer detection more affordable, and demonstrating how MCED could be integrated into existing screening programs in a financially sustainable way. Ultimately, our goal is to make earlier cancer detection accessible to far more people, including populations where cost and healthcare resources remain major barriers.”
The studies will be presented during the Rapid Fire sessions at the World Cancer Congress 2026 in Hong Kong, September 24–26. The Congress will be held at the Hong Kong Convention and Exhibition Centre.