Algorithm-aided blood test could help locate cancer early
Patients have better odds in the fight against cancer if it's caught early, but diagnosis often involves invasive biopsies that aren't usually undertaken unless there's already reason to suspect the presence of cancer. But soon it could be as simple as a routine blood test, thanks to a new computer program from UCLA researchers that can spot biomarkers in a patient's blood sample and identify where in the body a tumor might be hiding.
The idea for a cancer-detecting blood test isn't a new one, with research teams tackling the problem by searching for different biomarkers, such as the RNA profiles of platelets, elevated levels of a certain protein, or the telltale battle scars that tumors have left on white blood cells.
The UCLA team's new program, called CancerLocator, works in a similar fashion, but is searching for a different footprint. Cells express genes through a process called DNA methylation, but when cancer starts to take hold, this process is interrupted, with the disease silencing certain genes to give itself the best chance at surviving. Analyzing blood samples with their program, the UCLA team can spot the molecular patterns this interference creates, and by consulting a database of methylation profiles, it can identify the presence, type and location of a tumor.
"Non-invasive diagnosis of cancer is important, as it allows the early diagnosis of cancer, and the earlier the cancer is caught, the higher chance a patient has of beating the disease," says Jasmine Zhou, co-lead author of the study. "We have developed a computer-driven test that can detect cancer, and also identify the type of cancer, from a single blood sample. The technology is in its infancy and requires further validation, but the potential benefits to patients are huge."
The researchers compiled a database of methylation profiles for a variety of cancer types, using the markers to indicate in which organs a tumor originated. The database also contains "molecular footprints" of healthy samples, so the system can separate DNA found in the blood into tumor DNA and non-tumor DNA.
To test the new program, the researchers pitted it against two other algorithms, called Random Forest and Support Vector Machine. All three were tasked with analyzing blood samples from 46 cancer patients: 29 of which had liver cancer, 12 afflicted with lung cancer and five suffering from breast cancer. Each individual test was run 10 times to ensure the consistency of the outcome.
The tumors were at the early stages in 25 of the 29 liver cancer patients, and five out of 12 with lung cancer. In 80 percent of those cases, CancerLocator was able to spot the biomarkers, which the researchers say is a good indication of its abilities to detect relatively low levels of tumor DNA in the blood. Unfortunately, it fared less well against breast cancer, though overall it did better than both other algorithms.
"Owing to the limited number of blood samples, the results of this study are evaluated only on three cancer types (breast, liver and lung)," says Zhou. "In general, the higher the fraction of tumor DNAs in blood, the more accurate the program was at producing a diagnostic result. Therefore, tumors in well-circulated organs, such as the liver or lungs are easier to diagnose early using this approach, than in less-circulated organs such as the breast."
The research was published in the journal Genome Biology.
Source: UCLA via Eurekalert