Cologuard is designed to detect altered genetic material in cells that are shed from the lining of the colon into stool from pre-cancerous and cancerous lesions. The Exact Sciences test can also detect blood in stool, which is another possible signal of colon cancer.
The more important and positive recent development was the release of new Cologuard validation data last month. Exact tweaked the biomarkers used in Cologuard resulting in higher sensitivities, particularly with pre-cancerous lesions and early-stage colon cancer. This is great news and bodes well for the ongoing pivotal validation study and the test's commercial potential.
The new validation data demonstrated 98% sensitivity for cancer and 59% sensitivity for advanced adenomas (pre-cancers) greater than or equal to 1 cm, at a specificity cut-off of 91%.
Sensitivity is the ability of a test to correctly identify a positive sample, i.e. if a person has a disease, how often does the test come back positive? Highly sensitive tests are good because they result in few false positives.
Exact wants Cologuard to be highly specific because that will cut down on the number of patients referred unnecessarily to colonoscopy. In other words, if Cologuard says a patient doesn't have colon cancer, it's very highly likely that colon cancer is not present. Fewer false positive results means fewer colonoscopies performed, which saves insurance companies money. In turn, insurers will actively encourage use of Cologuard and Exact can charge more for it.
Validation data presented in 2010 showed Cologuard to have 85% sensitivity for colon cancer and 64% sensitivity for pre-cancers. The latter number is higher than the new 59% sensitivity rate because lesions of 1 cm weren't included. When Exact applied the new, more rigorous pre-cancer standard to its 2010 samples, the sensitivity rate was 54%, so net, the newly tweaked Cologuard gained 5% improvement in sensitivity for pre-cancers.
A couple words of caution about the new Cologuard results: The data were obtained from 355 samples, so a smaller number than the 1,178 samples used to analyze the test in 2010. Also, the new samples come from patients referred after they underwent colonoscopies, which means the sensitivities could be artificially high. "Real world" samples would come from a pre-colonoscopy patient.