Multicancer early detection test could detect cancer with an AUC of 0.94 for all cancer types and stages.
A liquid biopsy-based multicancer early detection (MCED) test could detect 12 types of cancers, including low DNA-shedding cancers and early-stage cancers, according to research presented at the American Association for Cancer Research Annual Meeting 2023.
“Detection of cancer in early stages through screening programs has been demonstrated to save lives,” said presenter Ben Ho Park, MD, PhD, director of the Vanderbilt-Ingram Cancer Center, in a press release. “However, only a few cancer types have screenings available today, and the majority of cancer deaths are from cancers for which there are no screening methods.”
Blood-based MCED tests are being developed for early detection in multiple cancer types and are able to test for multiple cancers with a single blood sample. However, Park noted that early-stage disease detection and cancers that shed small amounts of cell-free DNA (cfDNA) is challenging.
Aberrant DNA methylation is a key driver of cancer development, with different methylation patterns between cancerous and non-cancerous cells. There are also shared methylation patterns across cancer types, as well as differentially methylated regions that are tissue-specific, Park said. Therefore, MCED approaches based on methylation could work to both detect cancer and predict the tissue of origin.
Park and his colleagues are investigating a novel genome-wide methylome enrichment platform that captures methylated cfDNA molecules in the bloodstream without chemical or enzymatic treatment. This enables sequencing of the whole methylome without the need to sequence the entire genome, making it cost-effective and preserving the quality of the cfDNA.
The investigators conducted a retrospective case-control study to assess the ability of this platform to detect 12 cancer types in a cohort of approximately 4000 participants, including individuals with newly diagnosed, treatment-naïve cancer and age- and gender-matched non-cancer controls. Approximately 50% of the cancer cases were stage 1 or 2.
For the initial analysis, the researchers reported findings from cross-validation within a subset of 1903 samples in the training cohort to assess the performance of a machine learning algorithm designed to distinguish cases from controls, using 80% of the samples for training and 20% for testing. Performance was assessed using area under the curve (AUC), which estimated how accurately the test was able to differentiate cancer from non-cancer, with a score of 1 representing perfect discrimination.
According to the study, the test could detect cancer with an AUC of 0.94 for all cancer types and stages. Importantly, the performance of the MCED platform to distinguish stage 1 and 2 cancers from controls remained high, with an AUC of 0.92 and 0.95, respectively.
The AUCs for detection of individual cancer types ranged from 0.89 to 0.99. Performance was also high for low-shedding cancers, including breast, bladder, renal, prostate, and endometrial cancers. Combined, these cancers were distinguished with an AUC of 0.91.
“This platform allowed for a higher signal-to-noise ratio and led to increased performance in the more challenging applications where the cfDNA burden is the lowest,” Park said in the press release. “Cancers that shed a high amount of cfDNA had the best performance, but even cancers that are typically difficult to detect with cfDNA assays were detected with high performance in this interim readout.”
The key limitation of the study was its use of samples obtained from multiple biobanks. Park noted that there could be confounding factors by biobank or population that limit the performance of the test. Additionally, this is an early development study and additional studies will be needed to confirm these findings before the test is ready for use in a clinical setting.
“At this early stage in development, the robust detection of early-stage and low-shedding cancers with this genome-wide methylome enrichment platform is very promising,” Park concluded.
A Liquid Biopsy-based Multicancer Early Detection Test May Detect Early-stage Disease and Low DNA-shedding Cancers. News release. American Association for Cancer Research. April 16, 2023. Accessed April 12, 2023.