— Investigational Multi-Cancer Blood Test Detects Strong Signal for 12 Deadly Cancer Types When They are Still Localized, with 99 Percent Specificity, and Identifies Tumor’s Tissue of Origin with High Accuracy —
— GRAIL Plans to Advance Development of Multi-Cancer Test Toward Commercialization —
— Data Presented at 2019 ASCO Annual Meeting Support Feasibility of GRAIL’s Multi-Cancer Approach —
MENLO PARK, Calif. –GRAIL, Inc., a healthcare company focused on the early detection of cancer, today announced new data from the Circulating Cell-free Genome Atlas (CCGA) study that demonstrates the ability of GRAIL’s technology to detect cancer early with a single blood test. Data showed GRAIL’s investigational multi-cancer blood test detected a strong signal for 12 deadly cancer types at early stages with a very high specificity of at least 99 percent (or a false positive rate of one percent or less). In addition, the test identified where the cancer originated in the body (the tissue of origin) with high accuracy.
Detection rates (sensitivity) for the 12 deadly cancer types ranged from 59 to 86 percent at early stages (stages I-III). A combined analysis of this group of cancers showed robust detection at early stages (34 percent, 77 percent, and 84 percent at stages I, II, and III, respectively). In addition, a tissue of origin result was provided for 94 percent of all cancers detected and, of these, the test correctly identified the tissue of origin in 90 percent of cases.
The 12 pre-specified cancer types included anorectal, colorectal, esophageal, gastric, head and neck, hormone receptor negative breast, liver, lung, ovarian, and pancreatic cancers, as well as multiple myeloma and lymphoid neoplasms. Together, these cancer types account for approximately 63 percent of all cancer deaths in the United States.
These data will be presented in a poster tomorrow, Saturday, June 1 at the 2019 American Society of Clinical Oncology (ASCO) Annual Meeting by Minetta Liu, MD, Research Chair and Professor, Department of Oncology, Mayo Clinic (Abstract 3049).
“These exciting results suggest we can achieve what we believe are the requirements for a cancer screening blood test, including detection of multiple deadly cancer types at early stages in a single test, high accuracy in determining where the cancer originated, and a very low false positive rate,” said Jennifer Cook, Chief Executive Officer at GRAIL. “Our improved methylation-based technology has the potential to address gaps that exist with today’s screening options, which are limited to a few cancer types and only screen for one cancer type at a time. Based on these positive data, we plan to advance development of our test toward commercialization.”
“These very promising data indicate that a highly specific blood test for early cancer detection is approaching reality,” said Dr. Liu, who is also an investigator on the CCGA study. “The exceptional accuracy in determining the tissue of origin across all stages for those malignancies with significant cancer-specific mortality suggests that, if a cancer is detected, the test will inform where the tumor originated in the majority of cases. This factor is critical to streamline the clinical workup.”
“When we set out on this journey, we knew that to be successful, a blood-based screening tool would need to detect the clinically important cancers and not contribute to overdiagnosis of indolent cancers at the earliest stages,” said Rick Klausner, MD, Former Director of the National Cancer Institute, and Founder and Director of GRAIL. “Data being presented at ASCO suggest GRAIL’s test preferentially detects the most lethal cancers and can detect tumors when they are still localized and amenable to successful treatment. The high detection rate of stage II cancers at 77 percent in the group of 12 deadly cancers is particularly compelling and supports the potential benefit of our multi-cancer approach.”
Additional Results
In the second pre-planned case-control sub-study of CCGA, GRAIL’s multi-cancer early detection blood test is being evaluated in approximately 4,500 participants for its ability to detect cancer and identify the tissue of origin when cancer is present. Data reported at ASCO are based on an initial analysis of 2,301 participants from the training phase of the sub-study, including 1,422 participants with more than 20 cancer types across all stages and 879 participants without diagnosed cancer. This sub-study is ongoing, and the company is planning to present additional results at future medical conferences.
All results are reported at 99 percent specificity, which equates to a false positive rate of one percent. The actual false positive rate may be less than one percent, since cancer incidence rates suggest some of the individuals enrolled in the non-cancer group may have had an undiagnosed cancer at enrollment. Follow-up of participants in CCGA is ongoing, and outcomes will be collected for five years.
The overall detection rate for more than 20 cancer types in the sub-study across all stages was 55 percent (n=784/1,422; 95% confidence interval: 52.5-57.7%). A tissue of origin result across more than 20 cancer types was provided for 94 percent of all cancers detected by GRAIL’s test (n=735/784), and of these, the test correctly identified the tissue of origin in 90 percent of cases (n=663/735).
The overall detection rate for the 12 pre-specified deadly cancer types across all stages was 76 percent (n=671/882; 95% confidence interval: 73-79%). The tissue of origin accuracy for this group of cancers was consistent regardless of stage ranging from 84 to 92 percent.
Additional results for the 12 pre-specified deadly cancer types are detailed below.
Detection Rates (Sensitivity) at Early Stages (Stages I-III) for 12 Deadly Cancer Types at 99 Percent Specificity
| Cancer Type (N) | Sensitivity (95% CI) |
| Lung (151) | 59% (51-67%) |
| Colorectal (72) | 74% (62-83%) |
| Hormone Receptor Negative Breast (67) | 64% (52-76%) |
| Lymphoma☨ (60) | 70% (57-81%) |
| Pancreatic (41) | 78% (62-89%) |
| Head and Neck (36) | 86% (71-95%) |
| Multiple Myeloma (34) | 71% (53-85%) |
| Ovarian (30) | 67% (47-83%) |
| Esophageal (25) | 76% (55-91%) |
| Liver (25) | 68% (46-85%) |
| Anorectal (14) | 79% (49-95%) |
| Gastric (9) | 78% (40-97%) |
☨Excludes leukemias, which are not staged
Detection Rates (Sensitivity) and Tissue of Origin Accuracy by Stage for 12 Deadly Cancer Types at 99 Percent Specificity☨
| Stage | Sensitivity (N) (95% CI) | Tissue of Origin Accuracy* (N) |
| Stage I | 34% (151) (27-43%) | 84% (44) |
| Stage II | 77% (171) (70-83%) | 91% (124) |
| Stage III | 84% (242) (79-89%) | 92% (192) |
| Stage IV | 92% (281) (88-95%) | 91% (249) |
☨Excludes leukemias, which are not staged
*Performance in 94% of cases where a tissue of origin result was provided
Tissue of Origin Prediction Across All Stages (Stages I-IV) for 12 Deadly Cancer Types at 99 Percent Specificity*
| Cancer Type (Number of Predictions) | Correct Tissue of Origin |
| Lung (165) | 92% |
| Colorectal (87) | 97% |
| Breast (78) | 96% |
| Lymphoid Neoplasms (107) | 89% |
| Pancreatic (68) | 79% |
| Head and Neck (53) | 81% |
| Multiple Myeloma (23) | 96% |
| Ovarian (25) | 96% |
| Upper Gastrointestinal (Esophageal and Gastric combined) (44) | 89% |
| Liver (29) | 72% |
| Anorectal (3) | 67% |
*Performance in 94% of cases where a tissue of origin result was provided
Poster Presentation Details
Posters will be available online at https://grail.com/science/publications/ at time of presentation.
Abstract 3049*
Minetta C. Liu, et al. Genome-wide cell-free DNA (cfDNA) methylation signatures and effect on tissue of origin (TOO) performance
Poster Session: June 1, 2019: 8:00-11:00AM CDT, Hall A, Poster Board #41
*Includes data from second CCGA sub-study
Abstract 3103
Darya Filippova, et al. The Circulating Cell-free Genome Atlas (CCGA) study: Size selection of cell-free DNA (cfDNA) fragments
Poster Session: June 1, 2019: 8:00-11:00AM CDT, Hall A, Poster Board #95
Abstract 5574
Allen Cohn, et al. The Circulating Cell-free Genome Atlas (CCGA) study: Follow-up (F/U) on non-cancer participants with cancer-like cell-free DNA signals
Poster Session: June 1, 2019: 1:15-4:15PM CDT, Hall A, Poster Board #397
Abstract 1545
Geoffrey R. Oxnard, et al. Prognostic significance of blood-based cancer detection in plasma cell-free DNA (cfDNA): Evaluating risk of overdiagnosis
Poster Session: June 3, 2019: 1:15-4:15PM CDT, Hall A, Poster Board #39
About CCGA
The Circulating Cell-free Genome Atlas (CCGA) study is a prospective, observational, longitudinal, case-control study that has completed enrollment of approximately 15,000 participants with and without cancer across 142 sites in the United States and Canada. CCGA is designed to characterize the landscape of genomic cancer signals in the blood and to discover, train, and validate GRAIL’s multi-cancer early detection blood test through three pre-planned sub-studies.
About GRAIL’s Investigational Multi-Cancer Early Detection Test
GRAIL is developing a next-generation sequencing (NGS) blood test for the early detection of multiple deadly cancer types. GRAIL’s high efficiency methylation-based technology preferentially targets the most informative regions of the genome and is designed to use its proprietary database and machine-learning algorithms to both detect the presence of cancer and identify the tumor’s tissue of origin. GRAIL’s sequencing database of cancer and non-cancer methylation signatures is believed to be the largest of its kind and covers approximately 30 million methylation sites across the genome. More than 20 cancer types across stages are represented within the database.
DNA methylation is a natural process used by cells to regulate gene expression. It is a chemical modification to DNA and a well-studied epigenomic feature of the genome. In cancer, abnormal methylation patterns and the resulting changes in gene expression can contribute to tumor growth. For example, hypermethylation can cause tumor-suppressor genes to be inactivated.
Contacts
GRAIL
Charlotte Arnold
pr@grail.com
