Colorectal cancer (CRC) is highly preventable and curable—when lesions are found early. The CRC field has been able to achieve marked reductions in mortality through improvements in screening, surveillance, and disease management. However, CRC constitutes to be the second most common cause of cancer mortality; clearly, more work must be done. Toward that end, two studies leveraging highly novel approaches show considerable merit for enhancing CRC screening and improving the efficiency of surveillance.
Improving Early Detection Using Liquid Biopsy
Researchers in Taiwan have developed a screening test for early CRC detection that requires a simple blood draw to assess for circulating tumor cells (CTCs) in the blood (Abstract 556). The test demonstrates 88% accuracy to detect all stages of colorectal illness, including precancerous lesions. If validated and made commercially available, this test could be readily integrated into a patient’s routine physical exam, thereby increasing CRC screening compliance.
The current screening methods for CRC work, but they have drawbacks. The bowel prep required for imaging tests and the steps involved in stool-based testing are inconvenient, and the procedures themselves make some individuals squeamish, which results in low compliance. Moreover, stool-based tests have limited sensitivity for the detection of precancerous lesions.
Dr. Wen-Sy Tsai presents Abstract 556 during Oral Abstract Session C.
Dr. Tsai and colleagues conducted a single-center, prospective, blinded clinical study during a period of more than 3 years to assess the ability of a novel assay, the CellMax (CMx) biomimetic platform, to detect and enumerate CTCs in the blood for early CRC detection. The CMx test is a proprietary microfluidic biochip that captures and catalogues CTCs, in this case on the basis of CK20+ and CD45− status. Only 2 mL of patient blood is needed for the test.
The study included 620 individuals, including 111 with adenomas or polyps, 327 with stage I to IV CRC, and 182 healthy controls. Each individual underwent a routine blood draw that was analyzed with the CMx platform. The CTC counts and patient ages were entered into a regression algorithm to determine a risk score for each individual, but the results were kept blinded. Thereafter, individuals were screened according to standard protocol with colonoscopy and biopsy (if indicated); again, the results were blinded. The CMx CTC results and colonoscopy findings then were compared to determine the predictive accuracy of the CMx platform.
The CMx platform and algorithm performed well, with an overall accuracy of 88%. The false-positive rate among healthy individuals was 3.3%, and the false-negative rate among individuals with precancers or cancers was 15.8%. This translates to 97.3% specificity and 84.0% specificity overall. The specificity and sensitivity rates among individuals with precancerous lesions were 97.3% and 76.6%, respectively. For individuals with CRC, these rates were 97.3% and 86.9%, respectively.
The sensitivity of the CMx test for CRC lesions is on par with sensitivities reported for other screening tests, including fecal occult blood testing, fecal immunochemical test, stool DNA, and colonoscopy. For precancerous lesions, the CMx sensitivity far exceeds all other screening methods except for colonoscopy, which has a sensitivity of 76% to 94%.
Discussant Douglas Corley, MD, PhD, of Kaiser Permanente, Northern California Division of Research, was impressed by the sensitivity and specificity achieved with the CTC test. “Unlike something such as fecal immunochemical testing, which is not testing specifically for cancer, or colonoscopy, which is quite invasive and detects a lot of things that may never progress to cancer, [the CTC test] offers this potential,” he said. It also provides the opportunity to markedly reduce the need for invasive follow-up tests, such as colonoscopy, after a positive screen.
Dr. Tsai noted that he and his colleagues are currently selecting collaborators to extend evaluation of the CMx platform and algorithm in U.S. populations. They are also evaluating the platform in other diseases, such as prostate and breast cancers.
Improving Postscreening Management With Polyp Genetics
Dr. Douglas Corley discusses Abstracts 555 and 556 during Oral Abstract Session C.
To gain better insight into the ideal frequency of surveillance procedures, Rodrigo Jover, MD, PhD, of the Alicante University General Hospital, in Spain, and colleagues studied the relationship between the genetic profile of polyps and the risk and timing of development of advanced metachronous lesions (AMLs; Abstract 555). AMLs were defined as polyps 1 cm or greater in size that showed high-grade dysplasia or a villous component.
The investigators consecutively enrolled 308 individuals with colonic polyps between 2007 and 2009 and then observed them until 2014 (median follow-up time, 26 months). These individuals contributed 997 colonic polyps, which were analyzed for BRAF and KRAS mutations as well as for high-level methylation on CpG islands (CIMP-H).
At the time of the first colonoscopy, 17.2%, 25.0%, and 17.9% of individuals had polyps with BRAF mutations, KRAS mutations, and CIMP-H, respectively, and 19.8% had existing AMLs.
The presence of CIMP-H in polyps correlated with a higher risk of subsequent AML and a shorter interval for their development. For example, 38.9% of patients with CIMP-H polyps identified at the first colonoscopy went on to develop AMLs, whereas only 8.5% did not (p < 0.001). The time to AML development was also shorter if CIMP-H was present versus absent in polyps at the first colonoscopy (31.0 vs. 54.0 months; p = 0.002).
These associations held true in multivariable analysis. After analysis was adjusted for several patient and disease characteristics, polyps with CIMP-H at the first colonoscopy portended an 11-fold increase in the likelihood of AML occurrence (OR 11.41, 95% CI [2.04, 63.70]; p = 0.006) and a nearly 4-fold increase in the rate of AML recurrence (OR 3.77, 95% CI [1.36,10.46;] p = 0.01).
Neither BRAF nor KRAS mutations emerged as independent predictors of AML risk or timing.
Dr. Jover suggested that determination of the CIMP profile of polyps at the time of the first colonoscopy could enhance the prognostic value of classic risk factors to predict for future AMLs. He estimated that an additional 15% of individuals at risk for AMLs would be captured by folding in the CIMP profile, in addition to the roughly 50% already identified by considering adenoma number and size.
Dr. Corley found these results intriguing and suggested an examination of whether CIMP-H prognostication could be extended to lower-risk adenomas to increase the potential impact of this biomarker.
To confirm the value of both the CTC screening test and the CIMP-H biomarker, Dr. Corley indicated that these tools require validation in other populations. However, he also offered praise. “Congratulations to the presenters of these abstracts,” Dr. Corley said. “This is tough work. You’re taking new technologies, and you’re applying them to clinical populations.”
–Kara Nyberg, PhD