Abstract 97P
Background
Early detection of cancer is believed to lead to improved clinical outcome. Measuring cancer-related alterations in the tumour-derived portion of the cell-free DNA in plasma could offer an accurate, non-invasive approach for early cancer detection, leading to decreased cancer mortality. We report here a plasma targeted methylation marker panel performance for detection of 4 cancer types – colorectal, lung, pancreatic and breast.
Methods
Candidate methylation marker regions were evaluated in the plasma samples of 101 patients with colorectal cancer (N=20), breast cancer (N=29), lung cancer (N=37), pancreatic cancer (N=15) and 71 age/gender- matching non-cancer controls using methylation-sensitive restriction enzyme qPCR approach. Cancer patients who had received curative treatment prior to blood collection were excluded from the study. The non-cancer patients included, showed no clinical symptoms of the cancer at the time of recruitment. Performance of the methylation marker panel was evaluated for overall cancer detection and localization of tissue of origin. Accuracy was defined as the fraction of correct calls.
Results
10 methylation marker panel showed good pan-cancer detection potential with area under curve (AUC) of 89%, where sensitivity of detecting cancer of any origin was 79% (80/101) at 90% (64/71) specificity, with 100% of the CRC (20/20), 80% of the pancreatic cancer (12/15), and 75% of the breast cancer (21/28) and 73% of lung cancer (27/37) correctly identified as cancer patients. Notably, the sensitivity for stage I cancers was 75% (21/28). Cases that were correctly separated from control group were further evaluated for their tissue of origin. 16 methylation marker panel allowed us to correctly assign the tissue of origin to 80% of colorectal cancer (16/20), 78% of lung cancer (21/27), 75% of pancreatic cancer (9/12) and 62% of breast cancer (13/21) cases.
Conclusions
We show that targeted methylation marker panels have potential for early blood-based detection of multiple cancers with high sensitivity and specificity and good tissue of origin localization. This method could serve as the basis for further development of a highly accurate and minimally invasive blood-based multi-cancer screening test.
Clinical trial identification
Editorial acknowledgement
Legal entity responsible for the study
Universal Diagnostics S.L.
Funding
Universal Diagnostics S.L.
Disclosure
J. Kinross: Research grant/Funding (institution): NIHR: II-OL-1116-10027, NIH: R01-CA204403-01A1, Horizon H2020: ITN GROWTH. Imperial Biomedical Research Centre, SAGES research grant. Bowel and Cancer Research, CRUK; Advisory/Consultancy: Verb robotics; Advisory/Consultancy: Safeheal; Advisory/Consultancy: LNC therapeutics; Advisory/Consultancy: Medical iSight; Shareholder/Stockholder/Stock options: 1 Welbeck Day Surgery (day surgery unit); Officer/Board of Directors: Mangetoo.com (teledietetics); Officer/Board of Directors: 1Worldmedical; Officer/Board of Directors: Cerulean health; Advisory/Consultancy: Ethicon (J+J); Advisory/Consultancy: Medtronic; Research grant/Funding (self): Intuitive; Honoraria (self): Yakult. K. Kruusmaa: Leadership role, Shareholder/Stockholder/Stock options, Full/Part-time employment: Universal Diagnostics S.L.. M. Bitenc: Leadership role, Shareholder/Stockholder/Stock options, Full/Part-time employment, Officer/Board of Directors: Universal Diagnostics S.L.; Leadership role, Shareholder/Stockholder/Stock options, Licensing/Royalties, Full/Part-time employment, Officer/Board of Directors: Geneplanet d.o.o.. M. Chersicola: Full/Part-time employment: Geneplanet d.o.o.. All other authors have declared no conflicts of interest.