Colorectal cancer (CRC) is the second leading cause of cancer-related mortality in the United States, causing nearly 53,000 deaths in 2022. This statistic may seem bleak, yet an estimated 68% of CRC deaths occurred in patients who failed to screen at appropriate intervals or to screen at all.1 In contrast to most cancers, which are only detectable by imaging tests, CRC can be detected through direct visualization during a colonoscopy. Early detection of CRC is known to improve survival, but still, many eligible people do not adhere to screening guidelines, in part because colonoscopies have a significant overhead, requiring specialized equipment and highly trained personnel. Thus, as a potentially more accessible option, a new blood-based CRC screening test was evaluated in people undergoing routine colonoscopies to determine its performance in CRC detection.2
How does the test work?
When cells (including tumor cells) die, small fragments of DNA, called cell-free DNA (cfDNA) are released into surrounding tissue and eventually make their way into circulation. The new blood-based test relies on this cfDNA for CRC detection. Specifically, it estimates two scores – one based on the patient’s fraction of tumor-derived cfDNA and a second based on an integrated score of common CRC somatic mutations (e.g., mutations in the KRAS and APC genes), aberrant methylation status, and aberrant fragmentation patterns. A training set of over 4,000 samples of both healthy and CRC patients was used to develop algorithms that determine these two scores. A positive result is based on either of these scores exceeding the thresholds determined during algorithm development.
What did the study find?
The study recruited 7,861 participants between 45 and 84 years of age who were of average risk for CRC and were undergoing a routine screening colonoscopy. A blood sample was taken from the participants ideally within 60 days leading up to their colonoscopy. Based on colonoscopy results, 65 participants were determined to have cancer, 1,116 had advanced precancerous lesions (advanced adenomas and sessile serrated lesions at least 10 mm in the largest dimension), 2,166 had non-advanced adenoma, and 4,514 had negative colonoscopies. The total study population size was determined based on an expected prevalence of colon cancer using a much larger cohort of 22,877 participants undergoing a screening colonoscopy and the randomly selected 7,796 participants without CRC racially mirrored recent U.S. census data. Even though the inclusion of participants without CRC should be random, the prevalence of both advanced precancerous lesions was 14.1% higher in this population than the 8-9% has been previously reported, when defined as either advanced adenomas or serrated sessile lesions at least 10 mm in the longest dimension.3 Even though the prevalence of CRC in this population matched current national trends, the increased prevalence of advanced precancerous lesions indicates that this screening population may not be fully representative of the general population.
Of the 65 participants with cancer, 54 were identified by the cfDNA test, which translates to an 83.1% sensitivity (95% CI: 72.2-90.3), the rate of a positive test in the presence of cancer. The test identified all Stage II (n=14), III (n=17), and IV (n=10) cancers; however, six of the 17 Stage I cancers detected by colonoscopy registered as negative by the cfDNA test, as did five of the seven confirmed cancers without complete cancer staging. Of these seven cancers without staging, five were malignant polyps (these patients are typically managed as Stage I) and two participants were lost to follow-up. The missed early-stage cancers tended to be smaller and were not detected by the cfDNA test when the largest lesion size was 9 mm or less. The detection rate of Stage I cancer was 55% when including the incompletely staged polyps, and as expected, there was an even lower detection rate for less advanced lesions. Only 147 of the 1,116 advanced precancerous lesions were identified as positive results, a 13.2% sensitivity (95% CI: 11.3-15.3). Although the test is more reliable for detecting higher-stage malignancy (indeed it seems that all stages beyond Stage I were identified), the number of patients with CRC for each stage was small enough that this conclusion cannot be definitively drawn.
Furthermore, out of 6,680 participants, 698 received a positive cfDNA test when they’d had a negative colonoscopy or the subsequent histological findings were considered non-advanced. This translates to an 89.6% specificity (95% CI: 88.8-90.3), the probability of not having CRC when receiving a negative cfDNA test result. While this was the level of specificity for the cohort as a whole, sub-analyses indicated that specificity decreased with age. In participants ages 45-59, specificity was 93.4% (95% CI: 92.3-94.3), but in participants ages 70 and older, it was only 81.3% (95% CI: 78.6-83.7). This is likely because methylation increases with age, increasing the likelihood of a false positive.
Due to the high rate of false positives, the positive predictive value (PPV) was 7.2%, meaning that of all people who received positive results, only about 7% had CRC diagnosed by colonoscopy. However, the prevalence of CRC in this population (0.83%) is higher than the expected prevalence of 0.5-0.7%. With a lower prevalence, the PPV of the test will be lower, as demonstrated in the Table below. The negative predictive value (NPV) of the cfDNA test was 99.8%, but since the prevalence of CRC in the pool of participants was 0.83%, the NPV improves the confidence that a negative result means an absence of CRC by 0.6%.
A comparison of non-invasive CRC tests
The 95% confidence intervals for the cfDNA test sensitivity and specificity are above the minimum thresholds specified for test approval by the FDA, so in all likelihood, this test will become an available option. Given the ever-growing number of non-invasive CRC screening tests (see the Table below), how does this new test stack up against others that already exist?
In a population with a CRC prevalence of 0.5% (an estimate of the U.S. prevalence), there is nothing striking about the performance of this cfDNA test, and its detection of precancerous advanced adenomas (AAs) is one of the lowest rates of all the tests. The potential utility of the cfDNA test is that it is a convenient blood test for someone already at a doctor’s office.
CT-colonography is the most intensive of the listed tests because it uses imaging with contrast after completing the same required bowel prep as a colonoscopy; however, it can detect CRC and AA lesions smaller than 10 mm (whereas the cfDNA test has not demonstrated this to date). Stool-based tests (e.g., gFOBT, FIT, DNA-FIT, and DNA-FIT 2.0) are convenient since they can be done at home, but most require collecting and shipping a stool sample. The benefit of the DNA-FIT tests (brand name Cologuard) is that they are more sensitive to advanced precancerous lesions and Stage I cancers than the cfDNA test. DNA-FIT tests had >85% sensitivity for Stage I CRC in both the first and second-generation tests compared to the 55% sensitivity of the cfDNA test. Although DNA-FIT tests have been approved for CRC screening for a decade, one of the barriers to stool-based CRC screening is a lack of knowledge that these options exist.4 More education and awareness about existing CRC screening options might be more effective than using tests that are less sensitive to early stage cancers.
In comparison to other blood-based tests, the new cfDNA test outperforms the Septin9 blood test, which determines the methylation status of the promoter region of the septin 9 gene in cell-free DNA. The GRAIL multi-cancer early detection is the best-performing blood test for CRC, but this does not test for CRC alone, and its overall sensitivity is 14%. So while a GRAIL test might not have many false positives for CRC, you have to be willing to do follow-up cancer diagnostics for CRC and many other types of cancer.
The superiority of colonoscopy
All non-invasive screening options require more frequent repeat tests than a colonoscopy; some require testing as frequently as every year. Despite the convenience of at-home tests, colonoscopies are superior to other tests because they can directly visualize the colon and have the benefit of being able to identify and remove lesions during the procedure. An additional preventative benefit of colonoscopy is that the removal of precancerous lesions prevents these polyps from developing into CRC. Noninvasive screening tests like the cfDNA might be helpful in expanding the pool of individuals who get some form of screening, but positive results will still need to be followed up with the gold standard – a colonoscopy.
The bottom line
The ideal cancer screening test is both highly sensitive and highly specific. The newest blood-based cfDNA test has comparable overall performance to existing tests, making it better than having no screening test at all. However, the sensitivity of this test is skewed towards detecting more advanced cancers compared to the latest stool-based CRC screening methods, whereas the best chance of survival comes from catching cancer as early as possible. The low sensitivity of this cfDNA test for AAs and Stage I cancers means that, at least for now, this test won’t be an improvement over other existing options.
For a list of all previous weekly emails, click here.
References
- Doubeni CA, Fedewa SA, Levin TR, et al. Modifiable failures in the colorectal cancer screening process and their association with risk of death. Gastroenterology. 2019;156(1):63-74.e6. doi:10.1053/j.gastro.2018.09.040
- Chung DC, Gray DM 2nd, 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
- Penz D, Waldmann E, Hackl M, et al. Colorectal cancer and precursor lesion prevalence in adults younger than 50 years without symptoms. JAMA Netw Open. 2023;6(12):e2334757. doi:10.1001/jamanetworkopen.2023.34757
- Kruse-Diehr AJ, Cegelka D, Holtsclaw E, et al. Barriers and facilitators to stool-based screening for colorectal cancer among Black Louisville residents. J Cancer Educ. 2023;38(3):1050-1058. doi:10.1007/s13187-022-02231-2
- 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
- Meklin J, SyrjÄnen K, Eskelinen M. Fecal occult blood tests in colorectal cancer screening: Systematic review and meta-analysis of traditional and new-generation fecal immunochemical tests. Anticancer Res. 2020;40(7):3591-3604. doi:10.21873/anticanres.14349
- 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
- Imperiale TF, Porter K, Zella J, et al. Next-generation multitarget stool DNA test for colorectal cancer screening. N Engl J Med. 2024;390(11):984-993. doi:10.1056/NEJMoa2310336
- 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
- Shao SH, Allen B, Clement J, et al. Multi-cancer early detection test sensitivity for cancers with and without current population-level screening options. Tumori. 2023;109(3):335-341. doi:10.1177/03008916221133136
