Researchers have created a new machine learning software (artificial intelligence) that can predict survival and response rates in patient treatments with ovarian cancer.
The artificial intelligence software, which was created by researchers at Imperial College London and Melbourne of the University, was able to make predictions on the prognoses of patients with ovarian cancer with higher accuracy than other methods. It can also predict which treatment will be more effective for the patient after diagnosis.
The relevant research was published in Nature Communications, with the test taking place at Hammersmith Hospital. According to the researchers, this new technology can help doctors provide the best treatments for patients and pave the way for more individualized medicine. They also hope it could be used to separate ovarian cancer patients into classes, based on the small differences seen in their CT scans.
“Survival rates for long-term advanced ovarian cancer patients are advanced, despite the advances in anti-cancer therapies. There is an urgent need to find new ways to tackle the disease. Our technology can provide physicians with more accurate and detailed information on how patients are likely to respond to different treatments, paving the way for better and more targeted treatment decisions”,
said Professor Eric Amboagie, lead author and professor of Cancer Pharmacology and Molecular Imaging at Imperial College London.
“Artificial Intelligence has the potential to transform the way in which Health is working and to improve the developments for patients. Our software is such an example and we hope it can be used as a tool to help doctors better manage and treat patients with ovarian cancer”,
said Professor Andrea Rochol, of the survey participants.
Diagnosis is normally done in a number of ways, including blood tests to identify a substance called CA125, which indicates cancer, and CT scanning for detailed tumor imagery. This helps doctors to decide how much the disease has spread and what treatment the patient should take (surgery, chemotherapy, etc.).
The researchers used a mathematical software tool, TEXLab, for realizing how aggressive tumors are in the CT scans and tissue samples from 364 women with ovarian cancer between 2004 and 2015.
The software examined four tumor biological features for prognostic purposes and the patients were given a score known as RPV (Radiomic Prognostic Vector), indicating how advanced the disease is. The researchers compared the results with blood tests and “conventional” prognostic scores that doctors use. It was found that the software was up to four times more accurate in predicting death from ovarian cancer compared with conventional methods.
The team also found that 5% of patients with high RPV scores had survival rates below two years. The high RPV also seemed to be associated with resistance to chemotherapy and not good developments in surgery, indicating that the RPV can be used as a “biomarker” that provides how patients will respond to treatment. According to Professor Amboagie, this technology can be used to identify patients less likely to respond to normal therapies so that alternatives are sought.