Blood test reveals more than 20 types of cancer

Blood test reveals more than 20 types of cancer
A new blood test for cancer has shown great promise in a new study, with the ability to detect more than 20 types of cancer
A new blood test for cancer has shown great promise in a new study, with the ability to detect more than 20 types of cancer
View 1 Image
A new blood test for cancer has shown great promise in a new study, with the ability to detect more than 20 types of cancer
A new blood test for cancer has shown great promise in a new study, with the ability to detect more than 20 types of cancer

New technologies that enable doctors to detect key biomarkers at earlier stages promise the change the face of cancer treatment, and blood testing is one area where some exciting progress is being made. The latest advance comes from scientists at the Dana-Farber Cancer Institute, who have published a study describing a new type of blood test that can detect more than 20 types of cancer, and even trace them back to their source.

There are a number of cancer-detecting blood tests in development around the world, which are designed to detect different kinds of biomarkers as early indicators of the disease. Some search for elevated levels of certain proteins, some for damaged DNA in white blood cells and others for irregular platelet RNA profiles, with differing degrees of success.

The new test was developed by scientists at private company Grail Inc. and was investigated by scientists from Harvard University’s Dana-Farber Cancer Institute. The technology centers on the detection of methyl groups, which are small units of chemicals that can attach themselves to DNA and influence which genes are switched on or off.

When these patterns of on-and-off gene expression differ from the norm, they can be indicative of cancer, as we have seen demonstrated through previous studies on cancer-detection blood tests. The newly published study, however, does offer some particularly impressive and wide-ranging results.

The researchers applied the sequencing technology that scans for these irregularities to almost 3,600 blood samples, drawn from both healthy subjects and patients suffering from more than 20 types of cancer. It indicated the presence of the cancers with 99.4 percent accuracy, meaning that just 0.6 percent were incorrect diagnoses of healthy subjects.

But the test has the potential to go further than that. According to the investigators, the technology was able to reveal high-mortality cancers with a 76 percent accuracy, and among those, reveal stage one cancers with 32 percent accuracy, stage two cancers with 76 percent accuracy and stage three cancers with 85 percent accuracy.

Stage four cancers were detected with 93 percent accuracy, while 89 percent of the time, the test was able to correctly identify the organ or tissue where the cancer originated. The cancer types the test was able to reveal include breast, gall bladder, head, neck, lymphoid, lung, pancreatic and leukemia.

“Our previous work indicated that methylation-based assays outperform traditional DNA-sequencing approaches to detecting multiple forms of cancer in blood samples,” says the study’s lead author, Geoffrey Oxnard, MD, of Dana-Farber. “The results of the new study demonstrate that such assays are a feasible way of screening people for cancer.”

The team presented its results at the European Society for Medical Oncology (ESMO) 2019 Congress over the weekend.

Source: Dana-Farber Cancer Institute

Brian M
Pretty good with only a 0.6% incorrect diagnosis.
Unfortunately with the high-mortality cancers a 32% accuracy with stage one is perhaps not so good where it really matters for early treatment

Unfortunately no break down on the error rates as false positives or false negatives - both bad from a patients perspective.
I have to ask. How can it be 99.4% accurate for 20 types of cancer , if it was only 32% to 85% accurate for cancer at different stages? There is a little too much hype here. Hopefully, they develop a dependable test.
I'm excited for a day when cancer can be detected early and killed easily.
I was at the Dermatologist the other day when she cleared up a mistake in my understanding of cancer. I had been noting pain and discomfort with some skin anomalies when she stated simply that cancer does not cause pain or irritation. Whoa! I've been reacting to simple keratoses. That said, I think this blood test for cancer is super important and hope it comes to fruition soon, so everyone can possibly discover any cancer in their bodies before it gets to the point it would kill them. Now, why don't we get smart and pull some folks from minor cancer works, put them on this, and get this one prepped and delivered to the public soonest! // Nobody, it probably discovered cancer but misdiagnosed the actual type, leading to the discrepancy in accuracy reported.
99.4% sounds great, but if you're doing large population screenings it could still be a problem. For every million healthy people, 6,000 will be wrongly told they have cancer. They'll be subjected to additional procedures and biopsies, likely put lives and careers on hold etc. Some fraction of that 6,000 will end up dead as a result. (Overall, the public health tradeoff may be positive, especially if treatment saves or extends lives, but it's still rough on those false positives. And since cancer is still pretty rare on, say, a yearly basis, frequent screening could have false positive numbers comparable to accurate diagnoses.)