Blood-Based Test for CRC Screening Hits Sensitivity, Specificity Endpoints

WASHINGTON — A blood-based test for the early detection of colorectal cancer (CRC) met its prespecified sensitivity and specificity endpoints, according to results from the PREEMPT CRC study.

Using genomics testing, the blood-based test had a sensitivity for CRC of 79.2% and a specificity for non-advanced colorectal neoplasia of 91.5%, while multiomics testing showed a sensitivity of 77.8% and a specificity of 93.7%, reported Aasma Shaukat, MD, MPH, of New York University Grossman School of Medicine, during the Digestive Disease Week annual meeting.

“PREEMPT CRC is the largest study evaluating a blood-based colorectal screening test and enrolled a diverse U.S. population,” Shaukat said in her late-breaker presentation.

The negative predictive value for non-advanced colorectal neoplasia was 90.8% using both testing methods, while the positive predictive value was 15.5% with genomics testing and 17.7% with multiomics testing.

The sensitivity for advanced adenoma was 12.5% with genomics testing, Shaukat said, indicating that this endpoint missed its cutoff by 0.2%, while the sensitivity for high-grade dysplasia or carcinoma in situ was 29.1%.

Shaukat explained that the blood-based test was developed using case-control samples of patients with CRC and advanced adenomas and went through multiple testing platforms, including genomics, multiomics, and proteomics. The data were then fed into a machine-learning and artificial intelligence algorithm to develop a signature of advanced CRC.

There were two classifiers that underwent clinical classification — one was based on the genomic signature from a genomic platform, and the second was a multiomics platform that combines genomics with a protein model. The results from the multiomic platform mirrored the genomics platform results.

Shaukat said the need for a more widely accepted CRC screening program — aside from colonoscopy and stool sampling — is obvious. “Current screening efforts for colorectal cancer covers 59% of the population, which gives us a lot of area for improvement,” she said.

Session moderator Phillip Ge, MD, of the University of Texas MD Anderson Cancer Center in Houston, called the test “very promising.”

“We always see a lot of these types of colonoscopy alternatives at these meetings, but it is just a matter of what gets approved by regulators,” he told MedPage Today. “We hope this will add to our testing armamentarium.”

“In the U.S., there is a considerable hesitancy to undergo colonoscopy. Everyone says that colonoscopy is the gold standard, and that everyone should get one,” he noted. “But the reality is that we have never been good at convincing everybody who needs a colonoscopy to get a colonoscopy.”

In addition, “not everyone who needs a colonoscopy has access to a colonoscopy,” he added. “So there is both a resistance and an access issue. Stool-based systems or this blood-based system — while maybe not as good as colonoscopy — will fill a vital gap and unmet medical need.”

For this study, Shaukat and colleagues included 27,010 evaluable patients ages 45-85 who were enrolled from 201 study centers from May 2020 through April 2022. Participants had no personal history of CRC, colorectal adenoma, or inflammatory bowel disease; no family history of CRC; and no diagnosis of hereditary gastrointestinal cancer syndromes.

Mean age was 58, 55.8% were women, 73% were white, 11.2% were Black, and 8.8% were Asian. About 12% of the population identified as Hispanic or Latino.

Patients had their blood drawn, then underwent a standard bowel preparation and colonoscopy. Specimens underwent histopathology and other testing, and then the samples were validated clinically. “This was all overseen by a data quality oversight committee,” Shaukat said.

CRC was diagnosed in 72 patients, while advanced adenoma was identified in 2,567 patients, and non-advanced colorectal neoplasias were identified in 7,270 patients.

Using genomics testing, the sensitivity for CRC by stage was 57.1% for stage I disease, 100% for stage II disease, 82.4% for stage III disease, and 100% for stage IV disease.

Shaukat noted that the study was not designed to assess the impacts of this blood-based test on CRC incidence and mortality, nor was it designed to compare adherence to and effectiveness of different CRC screening tests.

She said her research team is now working to tweak the algorithm in order to improve the test.

Please enable JavaScript to view the

comments powered by Disqus.