A new study suggests that an artificial intelligence system may be able to perform tasks as accurately as a highly trained radiologist. The paper published in the Journal of the National Cancer Institute outlines how an AI system can accurately detect evaluate digital mammography in breast cancer screenings. Breast cancer screenings are an important tool in the early detection of breast cancer and the reduction of breast cancer-related mortality. Screenings currently are very labor intensive due to the high volume of women needing scans. In some parts of the world, including the U.S. there is a scarcity in the number of highly trained breast screening radiologists which has led to the development of AI systems that can do some of the tasks related to evaluating mammograms.
One in three women over 50 has delayed or not attended their cervical screening test, which should take place every five years, according to a survey from a cervical cancer charity. The average delay was more than two years, but one in 10 put off the test for more than five years. Jo's Cervical Cancer Trust surveyed 1,000 women over 50. It said not attending cervical screening was the biggest risk factor to developing cervical cancer. The survey found a lack of understanding of cervical cancer and cancer screening among women in that age group.
Kheiron Medical Technologies (Kheiron), a machine learning startup that's setting out to help radiologists detect early signs of cancer, has raised $22 million in a series A round of funding led by European VC firm Atomico, with participation from Greycroft, Connect Ventures, Hoxton Ventures, and Exor Seeds. Founded out of London in 2016, Kheiron offers a breast-screening product called Mia, which serves as a "second reader" to help radiologists decide whether to recall a patient for further evaluation. It's designed as a supportive tool rather than to replace medical professionals -- an automated second opinion, if you like. Mia's machine learning and data-processing smarts integrate directly into existing radiology workflows and software and look at areas of interest in full-field digital mammography (FFDM) images from breast cancer screenings, which can be difficult to read with the naked human eye if the tumors are small. This difficulty is often compounded by other distracting "noise" in a scan.
Artificial intelligence is better than specialist doctors at diagnosing lung cancer, a US study suggests. The researchers at Northwestern University in Illinois and Google hope the technology could boost the effectiveness of cancer screening. Finding tumours at an earlier stage should make them easier to treat. The team said AI would have a "huge" role in the future of medicine, but the current software is not yet ready for clinical use. The study focused on lung cancer, which kills more people - 1.8 million a year - than any other type of cancer.
In the year to come, an estimated 1,688,780 in the United States are expected to get a cancer diagnosis, and cancer will claim the lives of a projected 600,920. That death toll, however grim, represents a death rate from cancer that is 25% lower than it was a quarter-century ago -- a drop driven by steady reductions in smoking rates and advances in early detection and treatment. Between 1991 and 2014, that boost in cancer survivorship translates to approximately 2,143,200 fewer cancer deaths than might have been expected if death rates had remained at their peak. Better treatment protocols and more targeted therapies have driven the most dramatic improvements in the survival of patients with malignancies of the blood and lymph system, says the American Cancer Society's annual report card on cancer. In the mid-1970s, patients diagnosed with acute lymphocytic leukemia, for instance, stood a 41% probability of being alive five years later.