Cancer is one deadly disease that researchers around the world are trying to find a solution to. It is no news that cancer, of any kind, can be dealt with in the best way when detected at a very early stage.?
Lung cancer, for example, is normally detected after a CT scan. However, now a team of scientists at Stanford University have developed a blood test that uses machine learning to hunt for notorious cancer.?
The test looks for tumour DNA that is moving in a person¡¯s blood. While this has been known since quite some time, human technicians can¡¯t really hunt for these tiny specs of cancer in the blood, as something like this could take hours. But a carefully trained AI can do this in a matter of minutes.?
While the test is still in its early stages and the success rate for detecting stage-1 lung cancer is around 63 percent of patients who had it, which means it still needs some work to be applied.
However, the test will be a boon for people who cannot afford CT scans, as that¡¯s what prevents most people from getting scanned in time. Moreover, CT scans aren¡¯t the most accurate as the case of false positives are much higher.
Max Diehn lead researcher on this study from Stanford University said, ¡°We think our research is very exciting because we know that the best way to cure lung cancer is to detect it early before it has spread outside the lungs.¡±
He further added, ¡°Our method achieves sensitivities of approximately 40 per cent for stage I, 55 per cent for stage II, and 65 per cent for stage III. In aggregate, this means that the test could potentially help to detect over half of early-stage lung cancers. If confirmed in future studies, this could help to detect the majority of lung cancer before they have spread.¡±
Scientists also feel that once perfected, the system can be tweaked to look for other kinds of cancer to nip it in the bud.