Liquid Biopsy Assay Detects 50+ Types of Cancer and Identifies Cancer Origin in Tissue

April 6, 2020
Danielle Ternyila

The first liquid biopsy assay, a blood test, to detect over 50 types of cancer has been developed and is able to identify in which part of the body that the cancer originated in, based on findings from a prospective multicenter case-control observational trial, the CCGA study published in Annals of Oncology. The test also identified cancer prior to symptoms in most patients, according to a press release.

The first liquid biopsy assay, a blood test, to detect over 50 types of cancer has been developed and is able to identify in which part of the body that the cancer originated in, based on findings from a prospective multicenter case-control observational trial, the CCGA study (NCT02889978) published inAnnals of Oncology.The test also identified cancer prior to symptoms in most patients, according to a press release.1

The test had a 0.7% false-positive rate (95% CI, 0.1-2.6) for cancer detection. It was also able to predict in which tissue the cancer originated in 96% of samples with 93% accuracy, according to the study.2

“Considering the burden of cancer in our society, it is important that we continue to explore the possibility that this test might intercept cancers at an earlier stage and, by extension, potentially reduce deaths from cancers for which screening is either not available or has poor adherence,” Michael Seiden, MD, PhD, president of US Oncology, Texas, said in a statement.1“To our knowledge, this is the largest clinical genomics study, in participants with and without cancer, to develop and validate a blood test for early detection of multiple cancers.”

Samples from 6,689 patients either with a previously untreated cancer (n =2,482) or without cancer (n = 4,207) were divided into a training set and a validation set for the study. Overall, 4,216 patients were evaluable for analysis, which included 1,531 samples with cancer and 1,521 without cancer in the training set (n = 3,052) and 654 samples with cancer and 610 without in the validation set (n = 1,264). More than 50 types of cancer were included in the study.2

Blood samples were analyzed with the machine learning classifier to identify methylation changes as well as classify the samples as cancer or non-cancer. The tissue of origin was also identified.

Sensitivity was high in both the training set (99.8%; 95% CI, 99.4-99.9) and the validation set (99.3%; 95% CI, 98.3-99.8;P=.095).

Its ability to identify cancer was consistent between the 2 sets of patients. The true positive rate was 67.3% in 12 of the deadliest cancer types across clinical stages I, II, and III. These 12 types of cancer, including anal, bladder, bowel, esophageal, stomach, head and neck, liver and bile duct, lung, ovarian, and pancreatic cancers, as well as lymphomas and cancers of white blood cells (like multiple myeloma), account for about 63% of all cancer deaths in the United States annually; the majority of these cancer types have no way of screening prior to the development of symptoms.

The true positive rate was 43.9% across all cancer types for clinical stages I, II, and III. Investigators noted that detection improved with each stage of disease in both the training and validation sets.

Among the 12 pre-specified cancers, the sensitivity for stages I through III was 69.8% (95% CI, 65.5-73.7) in the training set versus 67.3% (95% CI, 60.7-73.3) in the validation set (P=0.988). Stage I-IV sensitivity was 77.9% (95% CI, 75.0-80.7) in the training set versus 76.4% (95% CI, 71.6-80.7) in the validation set (P=.573). The true positive rate was 39%, 69%, 83%, and 92% in stages I, II, III, and IV, respectively.

Sensitivity was consistent in both the training and validation sets for all cancer stages I to III. The sensitivity was 44.2% (95% CI, 41.3-47.2) in the training set versus 43.9% (95% CI, 39.4-48.5) in the validation set (P=1.000). Stage I-IV sensitivity was 55.2% (95% CI, 52.7-57.7) in the training set versus 54.9% (95% CI, 51.0-58.8) in the validation set (P=.897). The true positive rate for all cancer types in the study was 18% for stage I, 43% for stage II, 81% for stage III, and 93% for stage IV.

Cell-free DNA (cfDNA) consists of DNA shed from tumors into the blood, but cfDNA can come from other types of cells as well, making it difficult to identify what cfDNA comes from tumors. However, this assay analyzes chemical changes to the DNA called “methylation,” which generally controls gene expression. Abnormal patterns of methylation and resulting changes in gene expression can lead to tumor growth, which can cause signals in the cfDNA that could detect and localize cancer.1

Investigators continue to validate the test and the feasibility for screening populations in large, prospective studies, including STRIVE (NCT03085888) and PATHFINDER (NCT04016740) in the United States as well as in the SUMMIT study (NCT03934866) in the United Kingdom. These ongoing trials will evaluate the test in broader patient populations.

“This is a landmark study and a first step toward the development of easy-to-perform screening tools. Earlier detection of more than 50% of cancers could save millions of lives every year worldwide and could dramatically reduce morbidity induced by aggressive treatments,” Fabrice André, MD, PhD, director of Research at the Institut Gustave Roussy, Villejuif, France, said in a statement. “While numbers are still small, the performance of this new technology is particularly intriguing in pancreatic cancer, for which mortality rates are very high because it is usually diagnosed when it’s at an advanced stage.”

References:

  1. Blood Test Accurately Detects Over 50 Types of Cancer, Often Before Symptoms Show [news release]. European Society of Medical Oncology; March 31, 2020. https://bit.ly/2RalRGv. Accessed April 6, 2020.
  2. Liu MC, Oxnard GR, Klein EA, et al. Sensitive and specific multi-cancer detection and localization using methylation signatures in cell-free DNA [Published Online March 30, 2020].Annals of Oncology. doi.org/10.1016/j.annonc.2020.0011.