Key Points
Question
What is the clinical performance of a circulating tumor DNA–based blood test for colorectal cancer detection in average-risk individuals?
Findings
This prospective, population-based, observational study enrolled 48 995 participants aged 45 to 85 years at average risk of colorectal cancer. In a cohort of 27 010 evaluable participants, the blood test demonstrated 79.2% sensitivity for colorectal cancer and 91.5% specificity for advanced colorectal neoplasia using colonoscopy as the reference standard.
Meaning
In this large diagnostic accuracy study evaluating a colorectal cancer screening test, the blood-based test demonstrated acceptable accuracy for colorectal cancer detection in an average-risk population and provides an additional option for colorectal cancer screening.
Abstract
Importance
Colorectal cancer screening is widely recommended but underused. Blood-based screening offers the potential for higher adherence compared with endoscopy or stool-based testing but must first be clinically validated in a screening population.
Objective
To evaluate the clinical performance of an investigational blood-based circulating tumor DNA test for colorectal cancer detection in an average-risk population using colonoscopy with histopathology as the reference method.
Design, Setting, and Participants
Prospective, multicenter, cross-sectional observational study enrolling participants between May 2020 and April 2022 who were asymptomatic adults aged 45 to 85 years, at average risk of colorectal cancer, and willing to undergo a standard-of-care screening colonoscopy. Participants, staff, and pathologists were blinded to blood test results, and laboratory testing was performed blinded to colonoscopy findings. The study was conducted at 201 centers across 49 US states and the United Arab Emirates. Site-based and mobile phlebotomy were used for blood collection.
Exposures
Participants were required to complete a screening colonoscopy after blood collection.
Main Outcomes and Measures
The primary end points were sensitivity for colorectal cancer, specificity for advanced colorectal neoplasia (colorectal cancer or advanced precancerous lesions), negative predictive value for advanced colorectal neoplasia, and positive predictive value for advanced colorectal neoplasia. The secondary end point was sensitivity for advanced precancerous lesions.
Results
The median age of participants in the evaluable cohort (n = 27 010) was 57.0 years, and 55.8% were women. Sensitivity for colorectal cancer was 79.2% (57/72; 95% CI, 68.4%-86.9%) and specificity for advanced colorectal neoplasia was 91.5% (22 306/24 371; 95% CI, 91.2%-91.9%). The negative predictive value for advanced colorectal neoplasia was 90.8% (22 306/24 567; 95% CI, 90.7%-90.9%) and the positive predictive value for advanced colorectal neoplasia was 15.5% (378/2443; 95% CI, 14.2%-16.8%). All primary end points met prespecified acceptance criteria. The sensitivity for advanced precancerous lesions was 12.5% (321/2567; 95% CI, 11.3%-13.8%), which did not meet the prespecified acceptance criterion.
Conclusions and Relevance
In an average-risk colorectal cancer screening population, a blood-based test demonstrated acceptable accuracy for colorectal cancer detection, but detection of advanced precancerous lesions remains a challenge, and ongoing efforts are needed to improve test sensitivity.
Trial Registration
ClinicalTrials.gov Identifier: NCT04369053
This diagnostic study assesses the accuracy of an investigational blood-based circulating tumor DNA test for detection of colorectal cancer in an asymptomatic average-risk population.
Introduction
Colorectal cancer remains the second most common cause of cancer death in the US, despite the decline in incidence and mortality in recent decades.1 Screening can prevent cancer by detecting and prompting removal of premalignant polyps and can improve cancer survival through early detection.2 Evidence-based guidelines3,4,5,6 recommend colorectal cancer screening for average-risk adults starting at age 45 years.
The most widely used modality for colorectal cancer screening in the US is colonoscopy, followed by noninvasive stool tests.7,8 Despite the availability of multiple screening modalities, only an estimated 59% of eligible individuals aged 45 years or older were up to date with guideline-recommended screening in 2021,9 well below the 80% national goal.10 Screening rates vary greatly among populations according to age, race, and socioeconomic status, with rates for some demographic groups as low as 20%.9 Several factors contribute to suboptimal uptake, including (1) the invasive nature, potential discomfort and risks, and resource utilization associated with colonoscopy6,7,11,12; (2) racial, socioeconomic, and geographic disparities in access to different screening methods9; (3) the impact of missed workdays to prepare for and undergo colonoscopy, as well as the need for transportation assistance6,13; and (4) aversion to handling stool.14 Thus, noninvasive and more convenient testing approaches are needed to improve population-level access to colorectal cancer screening.
The noninvasive, blood-based circulating tumor DNA test for colorectal cancer screening evaluated in the PREEMPT CRC study is an investigational, qualitative, next-generation sequencing in vitro diagnostic test for the detection of patterns of CpG (cytosine followed by guanine) dinucleotide methylation associated with advanced colorectal neoplasia, including colorectal cancer and advanced precancerous lesions, in plasma derived from whole blood samples. The aim of the study was to validate the diagnostic accuracy of the blood-based test for early detection of colorectal cancer in average-risk individuals aged 45 to 85 years.
Methods
Study Design and Oversight
The PREEMPT CRC study protocol (Supplement 1; see also Supplement 2 for the statistical analysis plan) was developed by the authors and sponsor and approved by a central institutional review board (https://www.wcgclinical.com/) as well as individual site-based institutional review boards and ethics committees. From May 2020 to April 2022, the study enrolled eligible participants across 49 US states and the United Arab Emirates (at a facility affiliated with a US health care system). Site-based and mobile phlebotomy were used for blood collection. This study followed the Standards for Reporting of Diagnostic Accuracy (STARD) reporting guidelines for diagnostic studies.
Eligible participants provided written informed consent. A data monitoring committee reviewed adverse events, colonic neoplasia event rates, and colonoscopy quality and provided stopping guidance for the study. Data collection and management were conducted by IQVIA, a clinical research organization. Data monitoring was conducted by the sponsor and clinical research organizations, including IQVIA and LabCorp. Data were analyzed by study statisticians (A.P. and C.-K.S.).
Study Population
Eligible study participants were aged 45 to 85 years, asymptomatic, at average risk of colorectal cancer, and willing to undergo a standard-of-care screening colonoscopy. Key exclusion criteria included at least 1 first-degree relative diagnosed with colorectal cancer before age 60 years; at least 2 first-degree relatives diagnosed with colorectal cancer at any age; family history of hereditary gastrointestinal cancer syndromes; personal history of colorectal cancer or colorectal adenomas, inflammatory bowel disease, or any malignancy (except for nonmelanoma skin cancer) at the time of enrollment or within the past 5 years; colonoscopy within the past 9 years; sigmoidoscopy or computed tomography colonography within the past 4 years; stool DNA testing within the past 2 years; and fecal occult blood testing or fecal immunochemical testing (FIT) within the past 6 months. Full inclusion and exclusion criteria are described in the eAppendix in Supplement 3. Race and ethnicity were self-reported by participants; fixed categories were provided for selection.
Clinical Procedures
Participants were initially required to complete a screening colonoscopy within 90 days of blood collection; this window was increased to 120 days to account for decreased access to care due to the COVID-19 pandemic. When multiple lesions were found, the most clinically significant finding was defined as the index lesion (eTable 1 and eTable 2 in Supplement 3). All colonoscopy reports and applicable histopathology reports were reviewed by a central pathologist to assign a histopathological category (eTable 1 in Supplement 3). Cases that were categorized as colorectal cancer or certain subtypes of advanced precancerous lesions and a subset of negative findings (eTable 1 in Supplement 3) were additionally assessed by another independent central pathologist, who reviewed the tissue samples. Any inconsistency between site-reported and central pathology assessments were adjudicated by a third independent pathologist.
Laboratory Procedures
Up to 50 mL of whole blood from each participant was collected in Streck cell-free DNA blood collection tubes. Plasma was isolated from deidentified blood samples and stored at −80 °C for blinded analytical processing. Plasma was assessed for CpG methylation associated with advanced colorectal neoplasia (eAppendix in Supplement 3). The classification model, which was prespecified and locked after model training prior to processing clinical samples, was used to generate a score that was compared with a predefined threshold to yield either a positive or a negative result. Laboratory testing was performed independent of the colonoscopy findings. All participants, site staff, and central pathologists remained blinded to the blood test results.
Primary and Secondary End Points
The 4 prespecified primary end points were sensitivity for colorectal cancer, specificity for advanced colorectal neoplasia, negative predictive value for advanced colorectal neoplasia, and positive predictive value for advanced colorectal neoplasia (eTable 2 in Supplement 3). A secondary end point was sensitivity for advanced precancerous lesions. Advanced precancerous lesions included carcinoma in situ or high-grade dysplasia, adenoma with villous growth pattern (≥25%), adenoma of 1.0 cm or greater, sessile serrated lesion of 1.0 cm or greater, and traditional serrated adenoma. The end points were assessed using screening colonoscopy with histopathology as the reference method.
Statistical Analysis
The study was designed to have at least 90% power to assess all primary and secondary end points. The primary analysis was to determine whether the lower bounds of the 2-sided 95% CIs for sensitivity for colorectal cancer, specificity for advanced colorectal neoplasia, negative predictive value for advanced colorectal neoplasia, and positive predictive value for advanced colorectal neoplasia exceeded 65.0%, 85.0%, 90.4%, and 10.9%, respectively. A secondary analysis was to determine whether the lower bound of the 2-sided 95% CI for sensitivity for advanced precancerous lesions exceeded 11.5% (eAppendix in Supplement 3). The Centers for Medicare & Medicaid Services will cover a blood-based biomarker colorectal cancer screening test once every 3 years for Medicare beneficiaries if the test has both sensitivity greater than or equal to 74% and specificity greater than or equal to 90% in the detection of colorectal cancer.
Based on US Food and Drug Administration (FDA) recommendations, the P values and 95% CIs for sensitivity and specificity were calculated using the Wilson score method, and the P values and 95% CIs for positive and negative predictive values were computed using the 95% CIs of the corresponding likelihood ratio (positive likelihood ratio and negative likelihood ratio, respectively) using the noniterative score method,15 in which prevalence is treated as a constant. Based on an expected colorectal cancer prevalence of 0.3%, a minimum of 56 evaluable colorectal cancer cases among 19 000 evaluable participants were required to power the colorectal cancer sensitivity primary end point. Assuming an unevaluable participant rate of 20%, an enrollment of 24 000 participants was required. Enrollment was closed after the data monitoring committee assessed that the necessary number of colorectal cancers had been reached.
The clinical validation cohort consisted of participants consecutively enrolled on or after a cutoff date of April 5, 2021, to mitigate the then-unknown impact of the COVID-19 pandemic on the representativeness of the study population.16 Participants enrolled before April 5, 2021, were not included in developing or assessing the clinical performance of the classification model, and their blood samples were not tested. The cutoff date specifically coincided with a study protocol amendment in which further COVID-19 mitigations were implemented, generally coincided with COVID-19 vaccine expansion to all adults in the US, and is further discussed in the eAppendix in Supplement 3.17 Participants who had confirmed clinical findings and valid blood test results were considered evaluable.
Univariate and multivariate subgroup analyses were performed to assess test performance based on lesion and demographic characteristics. Control of overall type I error for primary and secondary end points was achieved through a prespecified hierarchical testing procedure (eAppendix and eFigure in Supplement 3). Any other reported 95% CIs were not adjusted for multiple comparisons, and any inferences drawn from them may require further confirmation. A multiple imputation analysis was performed to evaluate potential bias from missing data (eAppendix in Supplement 3).
Details of the analysis methods are provided in the eAppendix in Supplement 3 and the statistical analysis plan in Supplement 2. The analyses were conducted using R software, version 4.2.1.18
Results
Study Participants
The study was conducted across 201 sites and enrolled 48 995 participants. A total of 32 731 participants sequentially enrolled on and after April 5, 2021, were included in the clinical validation cohort, and 27 010 (82.5%) had results that were evaluable (Figure). Twelve of the evaluable participants were from the United Arab Emirates site. A total of 72 evaluable participants (0.3%) were found to have colorectal cancer on colonoscopy, of whom 60 (83.3%) had stage I to III disease. A total of 2567 evaluable participants (9.5%) had advanced precancerous lesions, 7270 (26.9%) had nonadvanced precancerous lesions, and 17 101 (63.3%) had negative findings.
Figure. Participant Enrollment and Disposition in the PREEMPT CRC Study.
The listed order of the exclusion criteria reflects a prespecified hierarchy used to exclude participants.
The median age of participants in the evaluable cohort was 57.0 years, and 55.8% were women. The evaluable cohort consisted of 8.8% Asian, 11.2% Black or African American, and 73.0% White participants; 11.8% were Hispanic or Latino (Table 1).
Table 1. PREEMPT CRC Cohort Demographic Characteristics.
| Characteristics | Cohorta | ||
|---|---|---|---|
| Enrolled (n = 48 995) | Clinical validation (n = 32 731) | Evaluable (n = 27 010) | |
| Age, yb | |||
| Mean (SD) | 57.9 (8.0) | 58.1 (8.2) | 58.1 (8.2) |
| Median (range) | 56.0 (45.0-86.0) | 57.0 (45.0-86.0) | 57.0 (45.0-86.0) |
| No. (%) | |||
| 45-49 | 4750 (9.7) | 3717 (11.4) | 2968 (11.0) |
| 50-54 | 16 986 (34.7) | 10 612 (32.4) | 8899 (32.9) |
| 55-64 | 16 352 (33.4) | 10 602 (32.4) | 8725 (32.3) |
| 65-74 | 9486 (19.4) | 6730 (20.6) | 5604 (20.7) |
| ≥75 | 1367 (2.8) | 1043 (3.2) | 814 (3.0) |
| Unknown | 54 (0.1) | 27 (0.1) | 0 |
| Sex, No. (%) | |||
| Female | 26 967 (55.0) | 18 149 (55.5) | 15 076 (55.8) |
| Male | 21 974 (44.8) | 14 555 (44.5) | 11 934 (44.2) |
| Unknown | 54 (0.1) | 27 (0.1) | 0 |
| Race, No. (%) | |||
| American Indian or Alaska Native | 159 (0.3) | 96 (0.3) | 78 (0.3) |
| Asian | 3336 (6.8) | 2698 (8.2) | 2381 (8.8) |
| Black or African American | 5484 (11.2) | 3890 (11.9) | 3038 (11.2) |
| Native Hawaiian or Other Pacific Islander | 131 (0.3) | 107 (0.3) | 72 (0.3) |
| White | 34 142 (69.7) | 23 416 (71.5) | 19 707 (73.0) |
| More than 1 race reported | 302 (0.6) | 157 (0.5) | 136 (0.5) |
| Other or unknownc | 5441 (11.1) | 2367 (7.2) | 1598 (5.9) |
| Ethnicity, No. (%) | |||
| Hispanic or Latino | 5498 (11.2) | 4115 (12.6) | 3189 (11.8) |
| Not Hispanic or Latino | 38 121 (77.8) | 26 526 (81.0) | 22 421 (83.0) |
| Unknown | 5376 (11.0) | 2090 (6.4) | 1400 (5.2) |
See the Figure for descriptions of the 3 cohorts.
Age was recorded based on year of birth; fractional age was not calculated. Per eligibility, the study population included participants aged 45 to 85 years within 1 month of enrollment. Participants outside this age range were confirmed by the respective sites to meet the age eligibility at time of enrollment.
No further specific information available.
Test Performance
All primary end points met prespecified acceptance criteria (eTable 3 in Supplement 3). The blood-based test identified 57 of 72 participants with colorectal cancer (sensitivity, 79.2%; 95% CI, 68.4%-86.9%; P = .006) (Table 2), with 16 of 28 stage I (sensitivity, 57.1%; 95% CI, 39.1%-73.5%), 15 of 15 stage II (sensitivity, 100%; 95% CI, 79.6%-100%), 14 of 17 stage III (sensitivity, 82.4%; 95% CI, 59.0%-93.8%), and 11 of 11 stage IV (sensitivity, 100%; 95% CI, 74.1%-100%). There was 1 case of unstaged cancer, which was detected by the blood-based test (Table 2). Based on the positive likelihood ratio for colorectal cancer and observed colorectal cancer prevalence, a positive blood test result would increase the probability of having colorectal cancer from 0.27% to 2.33%. There was no substantial difference in sensitivity for colorectal cancer by demographic characteristics (eTable 4 in Supplement 3). Sensitivity for colorectal cancer increased with larger lesion size (eTable 5 in Supplement 3). There was no statistically significant difference in test sensitivity by lesion location when adjusting for lesion size and demographic characteristics in a prespecified multivariate analysis.
Table 2. Sensitivity and Specificity of the Blood-Based Test.
| No./total | Sensitivity or specificity, % (95% CI) | |
|---|---|---|
| Positive findings (sensitivity) | ||
| Colorectal cancer, by stage | 57/72 | 79.2 (68.4-86.9) |
| IV | 11/11 | 100.0 (74.1-100.0) |
| III | 14/17 | 82.4 (59.0-93.8) |
| II | 15/15 | 100.0 (79.6-100.0) |
| I | 16/28 | 57.1 (39.1-73.5) |
| Unknown | 1/1 | 100.0 (20.7-100.0) |
| Advanced precancerous lesions | 321/2567 | 12.5 (11.3-13.8) |
| Adenoma with carcinoma in situ or high-grade dysplasia, any size | 32/110 | 29.1 (21.4-38.2) |
| Adenoma, villous growth pattern (≥25%), any size | 86/564 | 15.2 (12.5-18.4) |
| Adenoma ≥1.0 cm | 169/1470 | 11.5 (10.0-13.2) |
| Sessile serrated lesion with or without cytological dysplasia ≥1.0 cm | 30/394 | 7.6 (5.4-10.7) |
| Traditional serrated adenoma, any size | 4/29 | 13.8 (5.5-30.6) |
| Negative findings (specificity) | ||
| No advanced colorectal neoplasiaa | 22 306/24 371 | 91.5 (91.2-91.9) |
| Nonadvanced precancerous lesions | 6655/7270 | 91.5 (90.9-92.2) |
| Negative findings (category 4.1-4.4)b | 15 651/17 101 | 91.5 (91.1-91.9) |
| No findings (category 4.4)b | 11 383/12 446 | 91.5 (91.0-91.9) |
Advanced colorectal neoplasia comprised colorectal cancer and advanced precancerous lesions.
See definitions in eTable 1 in Supplement 3.
Among 24 371 participants without colorectal cancer or advanced precancerous lesions, specificity was 91.5% (22 306/24 371; 95% CI, 91.2%-91.9%; P < .001) (Table 2). For those with no findings on colonoscopy (category 4.4 in eTable 1 in Supplement 3), specificity was 91.5% (11 383/12 446; 95% CI, 91.0%-91.9%) (Table 2). Specificity correlated inversely with age (P < .001) (eTable 4 in Supplement 3).
The negative predictive value for advanced colorectal neoplasia was 90.8% (22 306/24 567; 95% CI, 90.7%-90.9%; P < .001) and the positive predictive value for advanced colorectal neoplasia was 15.5% (378/2443; 95% CI, 14.2%-16.8%; P < .001) (eTable 3 and eTable 6 in Supplement 3).
Of 2567 participants with advanced precancerous lesions, 321 tested positive, resulting in a sensitivity of 12.5% (95% CI, 11.3%-13.8%; P = .055) (Table 2), which did not meet the prespecified acceptance criterion (eTable 3 in Supplement 3). The sensitivity for advanced precancerous lesions correlated with age (P < .001) (eTable 4 in Supplement 3). The sensitivity for carcinoma in situ or high-grade dysplasia was 29.1% (32/110; 95% CI, 21.4%-38.2%) (Table 2).
There were no reported serious adverse events related to the blood test or blood collection procedure (eTable 7 in Supplement 3).
Multiple imputation of missing data was performed for all 32 731 participants in the clinical validation cohort. Results were almost identical to those for the evaluable cohort (eTable 6 in Supplement 3). The sensitivity for colorectal cancer was 79.3% (95% CI, 68.0%-87.3%), specificity for advanced colorectal neoplasia was 91.5% (95% CI, 90.6%-92.3%), and sensitivity for advanced precancerous lesions was 12.5% (95% CI, 10.9%-14.2%).
Discussion
This study was a large-scale diagnostic accuracy evaluation of a colorectal cancer screening test using colonoscopy as the reference standard. The study enrolled a diverse population, approximating the racial and ethnic distribution of the US. All primary end points met the prespecified acceptance criteria, demonstrating 79.2% sensitivity for colorectal cancer and 91.5% specificity for advanced colorectal neoplasia. These findings were robust in a prespecified direct standardization adjustment to the sex and age distribution of the US population (eAppendix and eTable 8 in Supplement 3), an adjustment method used by the FDA for other colorectal cancer screening products.19,20
The study population may have included a more challenging cohort of colorectal cancers, given the higher proportion of stage I cancers compared with recent studies.21,22 Despite this, the test demonstrated detection accuracy comparable with a recently approved cell-free blood DNA test.21 Notably, the performance of the approved test decreased after age-standardized adjustment,20 and the sensitivity for colorectal cancer and specificity for advanced neoplasia decreased after multiple imputation analysis for missing data.21
The sensitivities for colorectal cancer of both the investigational test in PREEMPT CRC and the approved cell-free blood DNA test are similar to that of FIT at its commonly used 20-μg hemoglobin/g feces cutoff, higher than that of high-sensitivity guaiac fecal occult blood tests, and lower than those of multitarget stool DNA and multitarget stool RNA tests.23,24,25 Sensitivity for stage I cancer was lower than for stages II to IV cancer, and sensitivity for colorectal cancer increased with larger lesion size. Of note, this study population included a higher representation of colorectal cancers smaller than 5 mm than similar screening studies.20,26 The specificity of the blood-based test for advanced colorectal neoplasia is lower than FIT but higher than multitarget stool DNA or RNA tests.22,23,24,27 The inverse correlation between age and specificity may be attributable to age-specific alterations in DNA methylation signatures.28,29,30,31
The sensitivity for advanced precancerous lesions of the blood-based test did not meet the prespecified criterion and is lower than FIT, multitarget stool DNA, and multitarget stool RNA.23,24 However, the sensitivity for advanced precancerous lesions was comparable with that of the approved cell-free blood DNA test, and both are within the range of the advanced precancerous lesion sensitivity of high-sensitivity guaiac fecal occult blood tests.21,23 High-grade dysplasia or carcinoma in situ was better detected in the present analysis than that reported for the approved cell-free blood DNA test.21 Such polyps are an important target for screening tests and are associated with 4.3 and 13.2 greater odds of developing subsequent advanced adenomas and colorectal cancer, respectively.32 Detection of advanced precancerous lesions using blood-based tests remains a challenge,21,30,33 despite promising assay development.34 This may be due to reduced shedding of DNA from precancerous lesions into the bloodstream.35 These findings underscore the importance of further work to improve advanced precancerous lesion detection by blood-based tests.
Given that adherence to screening remains a challenge in the US, blood-based tests may fill an unmet need, especially among those unscreened.6,7,9 Recent studies suggest that patient choice plays an important role in increasing adherence to screening36 and that including a blood-based option alongside or after offering FIT and colonoscopy increases screening participation.37,38,39 There is also evidence of strong patient preference for noninvasive or minimally invasive modalities, which may lower the initial barrier to screening for unscreened individuals who either refuse or have limited access to colonoscopy.40,41 Even within the category of noninvasive or minimally invasive screening modalities, blood-based testing may have additional benefits, including collection at the time of a medical visit and not requiring users to overcome an aversion to fecal collection. Coupling blood-based screening with other routine blood tests is conceptually appealing to many screening-eligible individuals7 and is also attractive for programmatic screening, as blood draws are often widely accessible, are easy to complete, and have minimal complications. In a randomized clinical trial comparing FIT with a blood-based test in individuals who were overdue for screening, adherence to the blood-based test was 11 percentage points higher than that of the stool test.42 Recent modeling work suggests that relative to stool-based tests, higher screening adherence for a blood-based test might compensate for lower precancerous lesion sensitivity when estimating the long-term benefits of colorectal cancer screening.43 Similar to other noninvasive screening tests, an abnormal result from any noninvasive or minimally invasive test must be investigated by follow-up colonoscopy,3,6 for which completion rates can still be low.44,45
The study enrolled a diverse population across the US, including urban and rural locations as well as a range of practice settings. A decentralized clinical trial platform, combining both direct-to-participant and site-based channels, was used to increase enrollment from underrepresented geographical regions and overcome challenges of enrollment during the COVID-19 epidemic. Other notable strengths are the blinded assessment of the blood-based test and colonoscopy results, central pathology review and tiered adjudication for select histological diagnoses and all colorectal cancers, inclusion of consecutive participants in the validation cohort, and complete staging information for all but 1 case of colorectal cancer.
Limitations
This study has some limitations. First, it focused on the 1-time diagnostic accuracy of the blood-based test and did not assess the comparative acceptance and effectiveness of different colorectal cancer screening approaches. Second, the downstream effect of this novel testing approach on colorectal cancer incidence and mortality will require further evaluation. Third, the frequency of testing intervals for the blood-based test remains uncertain. Future studies, including modeling analyses, that consider test performance, colorectal cancer progression, adherence, and cost will help determine optimal screening frequency. Expanding screening options to include blood-based tests in the future will provide opportunities to study comparative effectiveness and adherence.
Conclusions
In an average-risk colorectal cancer screening population, a blood-based test demonstrated acceptable accuracy for colorectal cancer detection, but detection of advanced precancerous lesions remains a challenge, and ongoing efforts are needed to improve test sensitivity.
Study Protocol
Statistical Analysis Plan
eAppendix. Supplementary Methods
eFigure. Hierarchical Testing Sequence for Endpoint Evaluation
eTable 1. Histopathology Classification
eTable 2. Key Terms and Definitions
eTable 3. Hierarchical Testing Strategy and Results
eTable 4. Sensitivity and Specificity Based on Demographic Characteristics
eTable 5. Sensitivity and Specificity Based on Lesion Characteristics
eTable 6. Multiple Imputation
eTable 7. Adverse Events
eTable 8. US Census Sex-Age Weighted Test Performance
eReferences
Data Sharing Statement
References
- 1.Siegel RL, Giaquinto AN, Jemal A. Cancer statistics, 2024. CA Cancer J Clin. 2024;74(1):12-49. doi: 10.3322/caac.21820 [DOI] [PubMed] [Google Scholar]
- 2.Gupta S. Screening for colorectal cancer. Hematol Oncol Clin North Am. 2022;36(3):393-414. doi: 10.1016/j.hoc.2022.02.001 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Davidson KW, Barry MJ, Mangione CM, et al. ; US Preventive Services Task Force . Screening for colorectal cancer: US Preventive Services Task Force recommendation statement. JAMA. 2021;325(19):1965-1977. doi: 10.1001/jama.2021.6238 [DOI] [PubMed] [Google Scholar]
- 4.Shaukat A, Kahi CJ, Burke CA, Rabeneck L, Sauer BG, Rex DK. ACG clinical guidelines: colorectal cancer screening 2021. Am J Gastroenterol. 2021;116(3):458-479. doi: 10.14309/ajg.0000000000001122 [DOI] [PubMed] [Google Scholar]
- 5.Patel SG, May FP, Anderson JC, et al. Updates on age to start and stop colorectal cancer screening: recommendations from the U.S. Multi-Society Task Force on Colorectal Cancer. Gastroenterology. 2022;162(1):285-299. doi: 10.1053/j.gastro.2021.10.007 [DOI] [PubMed] [Google Scholar]
- 6.Wolf AMD, Fontham ETH, Church TR, et al. Colorectal cancer screening for average-risk adults: 2018 guideline update from the American Cancer Society. CA Cancer J Clin. 2018;68(4):250-281. doi: 10.3322/caac.21457 [DOI] [PubMed] [Google Scholar]
- 7.Shaukat A, Levin TR. Current and future colorectal cancer screening strategies. Nat Rev Gastroenterol Hepatol. 2022;19(8):521-531. doi: 10.1038/s41575-022-00612-y [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Shapiro JA, Soman AV, Berkowitz Z, et al. Screening for colorectal cancer in the United States: correlates and time trends by type of test. Cancer Epidemiol Biomarkers Prev. 2021;30(8):1554-1565. doi: 10.1158/1055-9965.EPI-20-1809 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Siegel RL, Wagle NS, Cercek A, Smith RA, Jemal A. Colorectal cancer statistics, 2023. CA Cancer J Clin. 2023;73(3):233-254. doi: 10.3322/caac.21772 [DOI] [PubMed] [Google Scholar]
- 10.National Colorectal Cancer Roundtable . The 80% in Every Community campaign. Accessed September 16, 2024. https://nccrt.org/our-impact/80-in-every-community
- 11.Lin JS, Perdue LA, Henrikson NB, Bean SI, Blasi PR. Screening for colorectal cancer: updated evidence report and systematic review for the US Preventive Services Task Force. JAMA. 2021;325(19):1978-1998. doi: 10.1001/jama.2021.4417 [DOI] [PubMed] [Google Scholar]
- 12.Subramanian S, Tangka FKL, Hoover S, Cole-Beebe M, Joseph D, DeGroff A. Comparison of program resources required for colonoscopy and fecal screening: findings from 5 years of the Colorectal Cancer Control Program. Prev Chronic Dis. 2019;16:E50. doi: 10.5888/pcd16.180338 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Dong MH, Kalmaz D, Savides TJ. Missed work related to mid-week screening colonoscopy. Dig Dis Sci. 2011;56(7):2114-2119. doi: 10.1007/s10620-010-1545-2 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Kotzur M, Macdonald S, O’Carroll RE, et al. What are common barriers and helpful solutions to colorectal cancer screening? a cross-sectional survey to develop intervention content for a planning support tool. BMJ Open. 2022;12(9):e062738. doi: 10.1136/bmjopen-2022-062738 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Nam JM. Confidence limits for the ratio of two binomial proportions based on likelihood scores: non-iterative method. Biom J. 1995;37(3):375-379. doi: 10.1002/bimj.4710370311 [DOI] [Google Scholar]
- 16.Fleming TR, Labriola D, Wittes J. Conducting clinical research during the COVID-19 pandemic: protecting scientific integrity. JAMA. 2020;324(1):33-34. doi: 10.1001/jama.2020.9286 [DOI] [PubMed] [Google Scholar]
- 17.AJMC Staff . A timeline of COVID-19 vaccine developments in 2021. Am J Manag Care. Accessed August 13, 2024. https://www.ajmc.com/view/a-timeline-of-covid-19-vaccine-developments-in-2021
- 18.R Core Team . R: A Language and Environment for Statistical Computing. http://www.R-project.org/
- 19.US Food and Drug Administration . FDA executive summary: Exact Sciences Corporation Cologuard. Published March 27, 2014. Accessed September 17, 2024. https://wayback.archive-it.org/7993/20170405192818/https:/www.fda.gov/AdvisoryCommittees/CommitteesMeetingMaterials/MedicalDevices/MedicalDevicesAdvisoryCommittee/MolecularandClinicalGeneticsPanel/ucm390219.htm
- 20.US Food and Drug Administration . Summary of safety and effectiveness data [Guardant Shield]. Published July 26, 2024. Accessed September 16, 2024. https://www.accessdata.fda.gov/cdrh_docs/pdf23/P230009B.pdf
- 21.Chung DC, Gray DM II, Singh H, et al. A cell-free DNA blood-based test for colorectal cancer screening. N Engl J Med. 2024;390(11):973-983. doi: 10.1056/NEJMoa2304714 [DOI] [PubMed] [Google Scholar]
- 22.Imperiale TF, Porter K, Zella J, et al. ; BLUE-C Study Investigators . Next-generation multitarget stool DNA test for colorectal cancer screening. N Engl J Med. 2024;390(11):984-993. doi: 10.1056/NEJMoa2310336 [DOI] [PubMed] [Google Scholar]
- 23.Lin JS, Perdue LA, Henrikson NB, Bean SI, Blasi PR. Screening for Colorectal Cancer: An Evidence Update for the US Preventive Services Task Force. Agency for Healthcare Research and Quality; 2021. Accessed January 31, 2022. https://www.ncbi.nlm.nih.gov/books/NBK570913/ [PubMed]
- 24.US Food and Drug Administration . Summary of safety and effectiveness data [Geneoscopy ColoSense]. Published May 3, 2024. Accessed September 17, 2024. https://www.accessdata.fda.gov/cdrh_docs/pdf23/P230001B.pdf
- 25.Barnell EK, Wurtzler EM, La Rocca J, et al. Multitarget stool RNA test for colorectal cancer screening. JAMA. 2023;330(18):1760-1768. doi: 10.1001/jama.2023.22231 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.US Food and Drug Administration . Summary of safety and effectiveness data [Cologuard]. Published August 11, 2014. Accessed September 27, 2024. https://www.accessdata.fda.gov/cdrh_docs/pdf13/p130017b.pdf
- 27.Imperiale TF, Ransohoff DF, Itzkowitz SH, et al. Multitarget stool DNA testing for colorectal-cancer screening. N Engl J Med. 2014;370(14):1287-1297. doi: 10.1056/NEJMoa1311194 [DOI] [PubMed] [Google Scholar]
- 28.Ahlquist DA. Aberrantly methylated gene marker levels in stool: effects of demographic, exposure, body mass, and other patient characteristics. J Mol Biomark Diagn. 2012;3(5):1000133. doi: 10.4172/2155-9929.1000133 [DOI] [Google Scholar]
- 29.Horvath S, Raj K. DNA methylation-based biomarkers and the epigenetic clock theory of ageing. Nat Rev Genet. 2018;19(6):371-384. doi: 10.1038/s41576-018-0004-3 [DOI] [PubMed] [Google Scholar]
- 30.Potter NT, Hurban P, White MN, et al. Validation of a real-time PCR-based qualitative assay for the detection of methylated SEPT9 DNA in human plasma. Clin Chem. 2014;60(9):1183-1191. doi: 10.1373/clinchem.2013.221044 [DOI] [PubMed] [Google Scholar]
- 31.Imperiale TF, Kisiel JB, Itzkowitz SH, et al. Specificity of the multi-target stool DNA test for colorectal cancer screening in average-risk 45-49 year-olds: a cross-sectional study. Cancer Prev Res (Phila). 2021;14(4):489-496. doi: 10.1158/1940-6207.CAPR-20-0294 [DOI] [PubMed] [Google Scholar]
- 32.Fairley KJ, Li J, Komar M, Steigerwalt N, Erlich P. Predicting the risk of recurrent adenoma and incident colorectal cancer based on findings of the baseline colonoscopy. Clin Transl Gastroenterol. 2014;5(12):e64. doi: 10.1038/ctg.2014.11 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Church TR, Wandell M, Lofton-Day C, et al. ; PRESEPT Clinical Study Steering Committee, Investigators and Study Team . Prospective evaluation of methylated SEPT9 in plasma for detection of asymptomatic colorectal cancer. Gut. 2014;63(2):317-325. doi: 10.1136/gutjnl-2012-304149 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Lin J, Ariazi E, Dzamba M, et al. Evaluation of a sensitive blood test for the detection of colorectal advanced adenomas in a prospective cohort using a multiomics approach. J Clin Oncol. 2021;39(3)(suppl):43. doi: 10.1200/JCO.2021.39.3_suppl.43 [DOI] [Google Scholar]
- 35.Myint NNM, Verma AM, Fernandez-Garcia D, et al. Circulating tumor DNA in patients with colorectal adenomas: assessment of detectability and genetic heterogeneity. Cell Death Dis. 2018;9(9):894. doi: 10.1038/s41419-018-0934-x [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.Wong MC, Ching JY, Chan VC, et al. Informed choice vs. no choice in colorectal cancer screening tests: a prospective cohort study in real-life screening practice. Am J Gastroenterol. 2014;109(7):1072-1079. doi: 10.1038/ajg.2014.136 [DOI] [PubMed] [Google Scholar]
- 37.Liang PS, Zaman A, Kaminsky A, et al. Blood test increases colorectal cancer screening in persons who declined colonoscopy and fecal immunochemical test: a randomized controlled trial. Clin Gastroenterol Hepatol. 2023;21(11):2951-2957. doi: 10.1016/j.cgh.2023.03.036 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38.Coronado GD, Jenkins CL, Shuster E, et al. Blood-based colorectal cancer screening in an integrated health system: a randomised trial of patient adherence. Gut. 2024;73(4):622-628. doi: 10.1136/gutjnl-2023-330980 [DOI] [PubMed] [Google Scholar]
- 39.Ioannou S, Sutherland K, Sussman DA, Deshpande AR. Increasing uptake of colon cancer screening in a medically underserved population with the addition of blood-based testing. BMC Cancer. 2021;21(1):966. doi: 10.1186/s12885-021-08678-8 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 40.Makaroff KE, Shergill J, Lauzon M, et al. Patient preferences for colorectal cancer screening tests in light of lowering the screening age to 45 years. Clin Gastroenterol Hepatol. 2023;21(2):520-531. doi: 10.1016/j.cgh.2022.07.012 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 41.Ong C, Cook AR, Tan KK, Wang Y. Advancing colorectal cancer detection with blood-based tests: qualitative study and discrete choice experiment to elicit population preferences. JMIR Public Health Surveill. 2024;10:e53200. doi: 10.2196/53200 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 42.Liles EG, Coronado GD, Perrin N, et al. Uptake of a colorectal cancer screening blood test is higher than of a fecal test offered in clinic: a randomized trial. Cancer Treat Res Commun. 2017;10:27-31. doi: 10.1016/j.ctarc.2016.12.004 [DOI] [Google Scholar]
- 43.Meester RG, Piscitello A, Baldo L, Liang PS. 539 Higher assumed adherence to blood-based vs stool-based colorectal cancer screening compensates for potential lower advanced adenoma sensitivity. Gastroenterology. 2024;166(5)(suppl):S122. doi: 10.1016/S0016-5085(24)00772-8 [DOI] [Google Scholar]
- 44.May FP, Yano EM, Provenzale D, et al. Barriers to follow-up colonoscopies for patients with positive results from fecal immunochemical tests during colorectal cancer screening. Clin Gastroenterol Hepatol. 2019;17(3):469-476. doi: 10.1016/j.cgh.2018.05.022 [DOI] [PubMed] [Google Scholar]
- 45.Alsayid M, Singh MH, Issaka R, et al. Yield of colonoscopy after a positive result from a fecal immunochemical test OC-Light. Clin Gastroenterol Hepatol. 2018;16(10):1593-1597. doi: 10.1016/j.cgh.2018.04.014 [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Study Protocol
Statistical Analysis Plan
eAppendix. Supplementary Methods
eFigure. Hierarchical Testing Sequence for Endpoint Evaluation
eTable 1. Histopathology Classification
eTable 2. Key Terms and Definitions
eTable 3. Hierarchical Testing Strategy and Results
eTable 4. Sensitivity and Specificity Based on Demographic Characteristics
eTable 5. Sensitivity and Specificity Based on Lesion Characteristics
eTable 6. Multiple Imputation
eTable 7. Adverse Events
eTable 8. US Census Sex-Age Weighted Test Performance
eReferences
Data Sharing Statement
