A New Blood Test for Colorectal Cancer Screening: What the ECLIPSE Study Found

Table of Contents

• Key Points
• Background: Why This Research Matters
• How the Study Was Conducted
• Who Took Part in the Study
• Key Findings: How Well Did the Blood Test Work?
• Results by Cancer Stage
• Detection of Advanced Precancerous Lesions
• How Often the Test Was Correctly Negative
• Safety and Side Effects
• What These Findings Mean for Patients
• Limitations of the Study
• Recommendations for Patients
• Frequently Asked Questions
• Source Information

Key Points

• In a study of 7,861 average-risk adults, a blood test detected 83.1% of colorectal cancers found by colonoscopy.
• The test found 100% of stage II, III, and IV cancers but only 65% of stage I cancers.
• It detected only 13.2% of advanced precancerous lesions, so it cannot replace colonoscopy for removing growths.
• About 10% of people without advanced neoplasia had a positive result, which would lead to a follow-up colonoscopy.
• No serious adverse events related to blood collection were reported among 22,877 enrolled participants.

Background: Why This Research Matters

Colorectal cancer is the third most commonly diagnosed cancer and the second leading cause of cancer death in adults in the United States. The lifetime risk of developing colorectal cancer in the United States is about 4%. In 2024, approximately 53,000 people were expected to die from the disease.

Finding colorectal cancer early makes a major difference in survival. The 5-year survival rate is 91% for people whose cancer is found while it is still localized (confined to the colon or rectum). That rate drops to 14% for people whose cancer has spread to other parts of the body (metastatic disease).

Screening in people without symptoms reduces both the number of new colorectal cancer cases and the number of deaths from the disease. Leading professional organizations uniformly recommend screening. These organizations include the U.S. Preventive Services Task Force (USPSTF), the U.S. Multi-Society Task Force on Colorectal Cancer, and the American Cancer Society (ACS).

Despite these recommendations, screening rates remain low. Only about 59% of eligible people aged 45 years or older are up to date with screening guidelines. This is well below the target of 80% set by the National Colorectal Cancer Roundtable, which was established by the Centers for Disease Control and Prevention and the American Cancer Society.

The consequences of missed screening are serious. About 76% of colorectal cancer deaths occur in people who are not up to date with screening. There is a clear and pressing need for screening tests that are easier to complete and that increase the number of people who get screened.

Several factors contribute to low screening adherence. These include:

• The time required to perform screening
• Scheduling challenges
• Concern about how invasive or painful a test may be
• Fear of the test itself
• Discomfort or embarrassment associated with endoscopic examinations
• Lack of insurance coverage
• Distance from the test provider
• Lack of a physician recommendation for screening

A blood-based test that could be performed as part of a routine health care visit would add another screening option. Such a test would be relatively simple to complete, which could improve screening adherence. This study reports how well a cell-free DNA (cfDNA) blood-based screening test performed in an average-risk population.

How the Study Was Conducted

The study is called ECLIPSE, which stands for Evaluation of the ctDNA LUNAR Test in an Average Patient Screening Episode. The study was designed to evaluate how well the cfDNA blood-based test (Shield, made by Guardant Health) detects colorectal cancer that causes no symptoms and is at an early stage.

This was a prospective, observational, multicenter study. "Prospective" means participants were enrolled and followed forward in time. "Observational" means participants received standard care rather than being assigned to different treatments. "Multicenter" means the study took place at many locations.

Eligible people were enrolled at 265 U.S. sites. These sites included primary care and endoscopy centers in both academic and community-based institutions. The study protocol was approved by a central institutional review board and by site-specific review boards as required. All participants or their legal representatives provided written informed consent.

Guardant Health funded the study, and the authors designed it. A contract research organization called Premier Research gathered and monitored the data. The study statistician analyzed the data and confirmed the completeness and accuracy of the data and the fidelity of the study to the protocol. The first and last authors and two authors employed by Guardant Health also confirmed these points.

The first draft of the manuscript was written by an employee of Guardant Health and an employee of a contract research organization, funded by Guardant Health. All authors reviewed and edited the manuscript and agreed to submit it for publication. Authors who were not employed by Guardant Health signed clinical advisory agreements with Guardant Health related to the ECLIPSE study design, execution, data review, and analysis.

Who Was Eligible to Join

Eligible participants were 45 to 84 years of age at the time of consent. They were at average risk for colorectal cancer and were undergoing routine screening with colonoscopy.

The study excluded people with certain characteristics. Key exclusion criteria were:

• A history of cancer
• A known diagnosis of inflammatory bowel disease
• A hereditary predisposition to colorectal cancer
• A history of colorectal cancer in a first-degree relative
• Recent colorectal cancer screening includes colonoscopy within the preceding 9 years. Recent colorectal cancer screening also includes a positive fecal immunohistochemical test [FIT] or fecal occult blood test within the preceding 6 months. Recent colorectal cancer screening also includes completion of the multitarget stool DNA test or the methylated Septin9 blood test within the preceding 3 years.

Clinical Procedures

Participants provided written informed consent and a study blood sample before any medical preparation for colonoscopy. The standard-care screening colonoscopy was preferably performed within 60 days after enrollment. The date of enrollment was the date the study blood sample was obtained. An extended timeline for colonoscopy was allowed because of procedural delays caused by the coronavirus disease 2019 (COVID-19) global pandemic.

Colonoscopy bowel preparation was prescribed according to the standard of care. The endoscopist assessed the quality of bowel preparation for each participant. A completed colonoscopy was defined as visualization of the appendiceal orifice or ileocecal valve with photographic documentation. The exception was when a large lesion or mass prevented completion.

A repeat colonoscopy was permitted if it was done for clinical reasons and completed within the study window. The size and location of lesions found during colonoscopy were recorded. Resected lesions were sent for histopathological review (examination of tissue under a microscope). When multiple lesions were reviewed, the most advanced lesion was considered the primary lesion for final data analysis.

Endoscopic and histopathological reports for lesions assessed locally as advanced neoplasia were centrally reviewed. Advanced neoplasia means advanced precancerous lesions or colorectal cancer.

Laboratory Procedures

Whole-blood samples of 30 to 80 ml were collected in Streck cfDNA blood-collection tubes. The samples were shipped at ambient temperatures to the central biorepository. There, the samples were processed to plasma and stored at −80°C until shipment to the central laboratory for analysis at Guardant Health.

All samples were received in the central biorepository and central laboratory masked to clinical findings. This means the laboratory staff did not know the clinical details of the participants. The central laboratory remained unaware of the clinical attributes of the participants throughout the entire study.

The test under assessment is a cfDNA blood-based assay for the detection of colorectal cancer. The panel examines cfDNA genomic alterations (changes in DNA). The panel examines aberrant methylation status (abnormal chemical tags on DNA that affect gene activity). The panel examines fragmentomic patterns (the size and shape of DNA fragments). Results are integrated into a binary result: "abnormal signal detected" (positive test) or "normal signal detected" (negative test). Result thresholds were locked before analyses of the study samples. Binary results were reported to the contract research organization, where they were linked to clinical outcomes for analysis.

Primary and Secondary Outcomes

The coprimary outcomes were sensitivity of the cfDNA blood-based test for colorectal cancer and specificity of the test for advanced neoplasia. Both were compared with reference-standard screening colonoscopy with histopathological diagnosis.

Sensitivity means how often the test correctly identifies people who have the disease. Specificity means how often the test correctly identifies people who do not have the disease. The secondary outcome was sensitivity for detection of advanced precancerous lesions.

Advanced precancerous lesions were defined as advanced adenoma or sessile serrated lesions at least 10 mm in the largest dimension. Advanced adenoma means a tubular adenoma at least 10 mm in the largest dimension, or an adenoma of any size with villous features, high-grade dysplasia, or carcinoma in situ. Sessile serrated lesions counted if they were at least 10 mm in the largest dimension.

Any advanced colorectal neoplasia was defined as colorectal cancer or advanced precancerous lesions identified on colonoscopy. Any colorectal neoplasia was defined as colorectal cancer, advanced precancerous lesions, or nonadvanced precancerous lesions identified on colonoscopy.

Statistical Analysis

For previously approved screening tests for colorectal cancer, the Food and Drug Administration (FDA) has established acceptance criteria. Sensitivity for colorectal cancer is considered acceptable if the lower boundary of the two-sided 95% Wilson confidence interval exceeds 65%. Specificity for advanced neoplasia is considered acceptable if the lower boundary of the two-sided 95% Wilson confidence interval exceeds 85%. These were the coprimary outcome measures for this study.

A confidence interval is a range of values that likely contains the true result. A 95% confidence interval means that if the study were repeated many times, about 95% of the calculated intervals would contain the true value.

For the secondary outcome, sensitivity of the cfDNA blood-based test for advanced precancerous lesions was calculated and reported with the corresponding two-sided 95% Wilson confidence interval. Exploratory outcome measures included the prevalence-adjusted positive predictive value for advanced neoplasia, the negative predictive value for colorectal cancer, and a multiple imputation sensitivity analysis that accounted for missing data.

The study was powered for the coprimary outcomes. Other analyses were insufficiently powered, and the reported 95% confidence intervals are descriptive. The study sample size was calculated on the basis of a prevalence of colorectal cancer of 0.5 to 0.7%. Target enrollment was 68 evaluable participants with colorectal cancer and 7,000 evaluable participants who were negative for advanced neoplasia on colonoscopy. Enrollment continued until the target number of colorectal cancers was reached.

Final enrollment as of the data-cutoff date for the primary analysis was 65 evaluable participants with colorectal cancer. This provided the study with 85% power to establish that the sensitivity of the cfDNA blood-based test for colorectal cancer is greater than 65% at a true sensitivity of 82%.

The coprimary specificity outcome was sufficiently powered with 7,000 participants who were negative for advanced neoplasia on colonoscopy. The population without a diagnosis of colorectal cancer was randomly sampled to the target sample size of approximately 7,000 evaluable participants. Sampling used a stratified random approach, so that the age distribution of the selected participants without colorectal cancer followed the 2020 U.S. age distribution. Cohort sampling was completed before sample testing, with age being the only clinical variable considered.

The sample of 7,000 participants without advanced neoplasia corresponds to a power of at least 80% to establish the coprimary specificity of greater than 85%. This assumes that the true specificity for advanced neoplasia is at least 86.3%. There was a single interim analysis for futility of the outcome of specificity for advanced neoplasia, which was not met. A separate cfDNA assay combined with analysis of tumor-specific plasma proteins was independently evaluated in this study population. Statistical-power calculations and analyses were conducted with SAS software, version 9.8.

Who Took Part in the Study

Between October 2019 and September 2022, a total of 22,877 participants were enrolled. This included 65 evaluable participants with colonoscopy-identified colorectal cancer. Of these 65 participants, 48 (74%) had confirmed stage I, II, or III disease.

A total of 10,193 participants without colorectal cancer were randomly selected from the enrolled participants. This created a clinical validation cohort of 10,258 participants. Of these, 7,861 participants (76.6%) met all inclusion and exclusion criteria, had completed and valid colonoscopy results, had valid cfDNA blood-based test results, and were evaluable for final analysis.

The mean age of the participants in the evaluable cohort was 60 years, with a range of 45 to 84 years. Women made up 53.7% of the evaluable cohort.

With respect to race, 7.1% of participants were Asian, 11.8% were Black or African American, and 78.5% were White. With respect to ethnic group, 13.3% were Hispanic or Latino. These demographic characteristics closely mirror the racial and ethnic distribution in the 2020 U.S. Census.

The age distribution of the evaluable participants was as follows:

• 45–49 years: 640 participants (8.1%)
• 50–59 years: 3,055 participants (38.9%)
• 60–69 years: 2,440 participants (31.0%)
• 70–79 years: 1,670 participants (21.2%)
• 80 years or older: 56 participants (0.7%)

Of the 7,861 evaluable participants, 65 had colorectal cancer, 1,116 had advanced precancerous lesions, 2,166 had nonadvanced adenoma, and 4,514 had a negative test for colorectal neoplasia.

Key Findings: How Well Did the Blood Test Work?

The cfDNA blood-based test detected colorectal cancer in 54 of 65 participants (83.1%) who had colonoscopy-detected colorectal cancer. The remaining 11 participants (16.9%) had a negative cfDNA test. This means the overall sensitivity of the cfDNA blood-based test for colorectal cancer was 83.1% (95% confidence interval [CI], 72.2 to 90.3).

The lower boundary of the 95% confidence interval exceeded the acceptance criterion of 65%. This criterion was established for other FDA-approved screening tests for colorectal cancer.

The test performed even better for cancers that screening is most likely to catch. The cfDNA blood-based test identified 42 of 48 screening-relevant (stage I, II, or III) colorectal cancers. This gives a sensitivity of 87.5% (95% CI, 75.3 to 94.1) for these stages.

There were no substantial differences in sensitivity for colorectal cancer according to the primary tumor location. There were no substantial differences according to the tumor histologic grade (how abnormal the cancer cells look under a microscope). There were no substantial differences according to the demographic characteristics of the participants.

Seven of the 65 histopathologically confirmed colorectal cancers had insufficient clinical follow-up to confirm the cancer stage. Five of these were malignant polyps for which full cancer staging was not completed. Two were in participants who were lost to clinical follow-up. The sensitivity of the cfDNA blood-based test in this subgroup was 29% (95% CI, 8 to 64), meaning 2 of 7 cancers were detected.

Results by Cancer Stage

The test detected cancers at every stage, and detection improved as cancer stage advanced. The specific results by stage were:

• Stage I: 11 of 17 cancers detected (sensitivity, 65%; 95% CI, 41 to 83)
• Stage II: 14 of 14 cancers detected (sensitivity, 100%; 95% CI, 78 to 100)
• Stage III: 17 of 17 cancers detected (sensitivity, 100%; 95% CI, 82 to 100)
• Stage IV: 10 of 10 cancers detected (sensitivity, 100%; 95% CI, 72 to 100)

Sensitivity according to stage trended higher with more advanced cancer. However, the small sample did not allow for a formal comparison.

Detection of Advanced Precancerous Lesions

Among 1,116 participants with advanced precancerous lesions identified as the most advanced finding on colonoscopy, the cfDNA blood-based test was positive in 147 participants. This means the sensitivity for detecting advanced precancerous lesions was 13.2% (95% CI, 11.3 to 15.3).

This is an important limitation of the blood test. Advanced precancerous lesions are growths that are not yet cancer but can develop into cancer over time. A test that detects these lesions can prevent cancer by allowing them to be removed before they become malignant. The cfDNA blood test detected only about 1 in 8 of these lesions.

How Often the Test Was Correctly Negative

A total of 89.6% of participants without any advanced colorectal neoplasia identified on colonoscopy had a negative cfDNA blood-based test. The remaining 10.4% had a positive cfDNA blood-based test. This means the specificity for advanced neoplasia was 89.6% (95% CI, 88.8 to 90.3).

The lower boundary of the 95% confidence interval exceeded the prespecified acceptance criterion of 85%.

Specificity was also measured for people who had a completely negative colonoscopy. A negative colonoscopy was defined as no colorectal cancer, no advanced precancerous lesions, and no nonadvanced precancerous lesions identified on colonoscopy. Among these participants, 89.9% had a negative cfDNA blood-based test, while 10.1% had a positive cfDNA test. This indicates a false positive rate of 10.1% and a specificity for no neoplasia of 89.9% (95% CI, 89.0 to 90.7).

A false positive result means the test says something abnormal was found when, in fact, no cancer or precancerous lesion was present. Specificity for advanced neoplasia was inversely correlated with age, meaning specificity was lower in older participants.

Safety and Side Effects

There were no reported serious adverse events related to the blood-collection procedure. There were also no reported unanticipated cfDNA assay-related adverse events across the 22,877 enrolled participants. This indicates no serious or unexpected side effects were identified with the blood test in this study.

What These Findings Mean for Patients

In an average-risk screening population, this cfDNA blood-based test had 83% sensitivity for colorectal cancer, 90% specificity for advanced neoplasia, and 13% sensitivity for advanced precancerous lesions.

These results suggest the blood test could be a useful addition to the existing screening options. A blood test is simpler to complete than a colonoscopy or a stool-based test. It can be performed during a routine health care visit. This convenience could improve screening adherence among people who currently avoid or delay screening.

However, the test has an important limitation. It detects only 13.2% of advanced precancerous lesions. This means the blood test is much better at finding cancer that is already present than at finding growths that could be removed before they become cancer. Colonoscopy remains the only screening method that can both detect and remove precancerous lesions during the same procedure.

The test was especially reliable for later-stage cancers. It detected 100% of stage II, III, and IV cancers. It detected 65% of stage I cancers. Since earlier detection improves survival, the lower detection rate for stage I cancers is a consideration for patients and their doctors.

The false positive rate was about 10%. This means about 1 in 10 people without advanced neoplasia would receive a positive blood test result. A positive result would typically lead to a follow-up colonoscopy to determine whether cancer or a precancerous lesion is actually present.

Limitations of the Study

The study has several limitations that patients and clinicians should consider.

First, the sensitivity for advanced precancerous lesions was low, at 13.2%. The blood test cannot replace colonoscopy for the purpose of removing precancerous growths before they become cancer.

Second, the study was powered for the coprimary outcomes of sensitivity for colorectal cancer and specificity for advanced neoplasia. Other analyses, including the stage-specific sensitivity results and the sensitivity for advanced precancerous lesions, were insufficiently powered. The reported 95% confidence intervals for these analyses are descriptive rather than definitive.

Third, the number of participants with colorectal cancer was relatively small, at 65 evaluable participants. This limited the ability to make formal comparisons between groups, such as comparing sensitivity by cancer stage.

Fourth, seven of the 65 confirmed colorectal cancers had insufficient clinical follow-up to confirm the cancer stage. Five were malignant polyps for which full staging was not completed, and two were in participants lost to follow-up.

Fifth, the study enrolled participants who were already undergoing routine screening colonoscopy. This means the study population may differ from people who avoid screening altogether, who are the group most likely to benefit from a blood-based test.

Sixth, the study was funded by Guardant Health, the maker of the test. Several authors were employed by Guardant Health. The first draft of the manuscript was written by a Guardant Health employee and a contract research organization employee funded by Guardant Health.

Recommendations for Patients

If you are 45 years of age or older and at average risk for colorectal cancer, talk with your doctor about which screening option is right for you. Several options are available, including colonoscopy, stool-based tests, and now a blood-based test.

Consider the following points when making your decision:

1. Colonoscopy remains the most thorough option. It can detect cancer and remove precancerous growths during the same procedure. The blood test detected only 13.2% of advanced precancerous lesions.
2. The blood test is convenient. It can be done during a routine health care visit and requires no bowel preparation or sedation.
3. The blood test is good at finding cancer, especially later-stage cancer. It detected 83.1% of all colorectal cancers and 100% of stage II, III, and IV cancers.
4. A positive blood test requires follow-up. About 10% of people without advanced neoplasia received a positive result. If your blood test is positive, your doctor will likely recommend a colonoscopy.
5. Do not skip screening. About 76% of colorectal cancer deaths occur in people who are not up to date with screening. Any screening is better than no screening.

Discuss your personal risk factors, preferences, and any barriers to screening with your health care provider. Together, you can choose the screening approach that you are most likely to complete.

Frequently Asked Questions

Who was eligible to join this colorectal cancer blood test study?

Participants were 45 to 84 years old, at average risk for colorectal cancer, and already scheduled for routine screening colonoscopy. People were excluded if they had a personal history of cancer, inflammatory bowel disease, a hereditary colorectal cancer predisposition, a first-degree relative with colorectal cancer, or recent colorectal screening. The study enrolled 22,877 people at 265 U.S. sites.

How accurate was the blood test at finding colorectal cancer?

In 7,861 evaluable participants, the blood test detected 83.1% of colorectal cancers found by colonoscopy. It found 87.5% of stage I, II, or III cancers and 100% of stage II, III, and IV cancers. It detected 65% of stage I cancers. The study was funded by the test maker and included 65 people with colorectal cancer.

What does a 13.2% sensitivity for advanced precancerous lesions mean?

It means the blood test found only about 1 in 8 advanced precancerous lesions, which are growths that are not yet cancer but can become cancer over time. Because it misses most of these growths, the test is much better at finding cancer that is already present than at preventing cancer by catching growths early.

Were there any side effects or risks from the blood test?

Across 22,877 enrolled participants, there were no reported serious adverse events related to blood collection and no unanticipated adverse events related to the test itself. No serious side effects related to the blood draw were reported, and no unexpected side effects related to the test occurred, among the 22,877 enrolled participants. It requires no bowel preparation or sedation, unlike colonoscopy, and can be done during a routine health care visit.

Can this blood test replace colonoscopy for colorectal cancer screening?

No. Colonoscopy remains the only screening method that can both detect and remove precancerous growths during the same procedure. The blood test detected only 13.2% of advanced precancerous lesions, so it cannot replace colonoscopy for removing growths before they become cancer. It may be an added option for people who avoid or delay screening.

Why is colorectal cancer screening important if I feel fine?

Screening in people without symptoms reduces both new colorectal cancer cases and deaths from the disease. The 5-year survival rate is 91% when cancer is found while still localized, but drops to 14% after it has spread. About 76% of colorectal cancer deaths occur in people who are not up to date with screening.

If I get a positive result on the colorectal cancer blood test, should I get a second opinion before having a colonoscopy?

A positive blood test is not a diagnosis. About 10% of people without advanced neoplasia receive a positive result, so follow-up colonoscopy is typically recommended to see whether cancer or a precancerous lesion is actually present. Colonoscopy remains the only screening method that can both detect and remove precancerous growths in the same procedure. If you are deciding between screening options, or want your result and records reviewed before proceeding, an independent second opinion can help clarify your choices. Diagnostic Detectives Network provides independent expert second opinions.

Source Information

Original article title: A Cell-free DNA Blood-Based Test for Colorectal Cancer Screening NEJM 2024

Authors: Daniel C. Chung, M.D., Darrell M. Gray II, M.D., M.P.H., Harminder Singh, M.D., Rachel B. Issaka, M.D., M.A.S., Victoria M. Raymond, M.S., Craig Eagle, M.D., Sylvia Hu, Ph.D., Darya I. Chudova, Ph.D., AmirAli Talasaz, Ph.D., Joel K. Greenson, M.D., Frank A. Sinicrope, M.D., Samir Gupta, M.D., M.S.C.S., and William M. Grady, M.D.

Publication: The New England Journal of Medicine, March 14, 2024, Volume 390, Number 11, pages 973–983. DOI: 10.1056/NEJMoa2304714

Funding: Guardant Health

Clinical trial registration: ECLIPSE, ClinicalTrials.gov number NCT04136002

This patient-friendly article is based on peer-reviewed research.