Oncoseek 2.0: An updated blood-based assay for multi-cancer early detection.

10551 Background: Early identification of cancer remains essential for improving survival and reducing downstream treatment costs, yet effective and scalable multi-cancer screening strategies are still limited. OncoSeek is a previously validated, AI-based multi-cancer early detection (MCED) assay integrating protein tumor markers (PTMs) with clinical data, demonstrating 58.4% sensitivity, 92.0% specificity, and 70.6% tissue-of-origin accuracy in > 15,000 participants. To improve detection of specific cancer types and expand cancer-type coverage, we developed OncoSeek 2.0 by incorporating three additional PTMs and evaluated its performance in validation cohort. Methods: OncoSeek 2.0 retained the original machine-learning framework of OncoSeek 1.0, with expanded PTM inputs (ProGRP, SCCA, and tPSA). Performance was evaluated in a retrospective validation cohort comprising 1,267 cancer and 355 non-cancer individuals. Overall performance, including cancer-type-specific sensitivity, was assessed and compared with OncoSeek 1.0. Results: OncoSeek 2.0 incorporated three additional PTMs (ProGRP, SCCA, and tPSA) to enhance cancer-type–specific detection. ProGRP levels was significantly elevated in small cell lung cancer compared with other cancers and healthy controls; SCCA concentrations were higher in squamous cell carcinomas (esophageal, cervical, and lung) than in non-squamous cancers and controls; and tPSA was selectively increased in prostate cancer (all P < 0.0001). Compared with OncoSeek 1.0, OncoSeek 2.0 achieved a higher AUC (0.917 vs 0.860) and improved sensitivity from 65.3% to 77.6% at 90.1% specificity across 15 pre-specified cancer types collectively accounting for 76.5% of global cancer mortality. Sensitivity gains were notable in lung (75.1% → 84.3%), prostate (58.8% → 88.2%), cervical (55.6% → 65.6%), and esophageal (39.6% → 62.4%) cancers. When stratified by tumor stage, OncoSeek 2.0 achieved sensitivities of 54.2% for stage I, 71.7% for stage II, 79.8% for stage III, and 88.4% for stage IV, consistently outperforming OncoSeek 1.0, which achieved sensitivities of 30.2%, 53.8%, 62.3%, and 85.7% for stages I through IV, respectively. Conclusions: By integrating three additional cancer-type–specific PTMs, OncoSeek 2.0 substantially improves sensitivity while maintaining high specificity and low reagent cost (~$30 per test). The upgraded assay enhances detection across multiple tumor types and disease stages, with particularly pronounced gains in lung, prostate, cervical, and esophageal cancers. These results support OncoSeek 2.0 as a scalable and clinically actionable MCED approach for population-level early detection, particularly for cancers lacking USPSTF-recommended screening and in low- and middle-income countries.

Authors

Institutions

Publication Details

Journal
Journal of Clinical Oncology
Published
2026-05-27
DOI
https://doi.org/10.1200/jco.2026.44.16_suppl.10551
Primary Topic
Cancer Genomics and Diagnostics
Type
article
Field-Weighted Citation Impact
0.00
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
article

Oncoseek 2.0: An updated blood-based assay for multi-cancer early detection.

Yinyin Chang, Wei Wu, Qingxia Xu, P. W. Xing et al.
Journal of Clinical Oncology
Cancer Genomics and Diagnostics
article

Oncoseek 2.0: An updated blood-based assay for multi-cancer early detection.

Yinyin Chang, Wei Wu, Qingxia Xu, P. W. Xing, Chenyu Ding, Wei Cui, Dandan Zhu, Mao Mao, Shiyong Li, Yong Shen
article en

Abstract

10551 Background: Early identification of cancer remains essential for improving survival and reducing downstream treatment costs, yet effective and scalable multi-cancer screening strategies are still limited. OncoSeek is a previously validated, AI-based multi-cancer early detection (MCED) assay integrating protein tumor markers (PTMs) with clinical data, demonstrating 58.4% sensitivity, 92.0% specificity, and 70.6% tissue-of-origin accuracy in > 15,000 participants. To improve detection of specific cancer types and expand cancer-type coverage, we developed OncoSeek 2.0 by incorporating three additional PTMs and evaluated its performance in validation cohort. Methods: OncoSeek 2.0 retained the original machine-learning framework of OncoSeek 1.0, with expanded PTM inputs (ProGRP, SCCA, and tPSA). Performance was evaluated in a retrospective validation cohort comprising 1,267 cancer and 355 non-cancer individuals. Overall performance, including cancer-type-specific sensitivity, was assessed and compared with OncoSeek 1.0. Results: OncoSeek 2.0 incorporated three additional PTMs (ProGRP, SCCA, and tPSA) to enhance cancer-type–specific detection. ProGRP levels was significantly elevated in small cell lung cancer compared with other cancers and healthy controls; SCCA concentrations were higher in squamous cell carcinomas (esophageal, cervical, and lung) than in non-squamous cancers and controls; and tPSA was selectively increased in prostate cancer (all P < 0.0001). Compared with OncoSeek 1.0, OncoSeek 2.0 achieved a higher AUC (0.917 vs 0.860) and improved sensitivity from 65.3% to 77.6% at 90.1% specificity across 15 pre-specified cancer types collectively accounting for 76.5% of global cancer mortality. Sensitivity gains were notable in lung (75.1% → 84.3%), prostate (58.8% → 88.2%), cervical (55.6% → 65.6%), and esophageal (39.6% → 62.4%) cancers. When stratified by tumor stage, OncoSeek 2.0 achieved sensitivities of 54.2% for stage I, 71.7% for stage II, 79.8% for stage III, and 88.4% for stage IV, consistently outperforming OncoSeek 1.0, which achieved sensitivities of 30.2%, 53.8%, 62.3%, and 85.7% for stages I through IV, respectively. Conclusions: By integrating three additional cancer-type–specific PTMs, OncoSeek 2.0 substantially improves sensitivity while maintaining high specificity and low reagent cost (~$30 per test). The upgraded assay enhances detection across multiple tumor types and disease stages, with particularly pronounced gains in lung, prostate, cervical, and esophageal cancers. These results support OncoSeek 2.0 as a scalable and clinically actionable MCED approach for population-level early detection, particularly for cancers lacking USPSTF-recommended screening and in low- and middle-income countries.

Journal of Clinical OncologyVol. 44(16_suppl)
Zhengzhou University (CN), Development Research Center (CN), Henan Cancer Hospital (CN)
Good health and well-being
Openalex Percentile: Top 11%
Cancer Genomics and Diagnostics
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.