Is it worth spending money?
Artificial intelligence determines the effectiveness of immunotherapy
Researchers are using AI with routine computer scans to predict how well lung cancer patients will respond to expensive treatment based on changes in the texture structure inside and outside the tumor, according to a press release from Case Western Reserve University Using Artificial Intelligence to determine whether immunotherapy is working.
Article by Khorrami et al. Changes in CT radiomic features associated with lymphocyte distribution predict overall survival and response to immunotherapy in non-small cell lung cancer published in the journal Cancer Immunology Research.
Scientists from the Case Western Digital Imaging Laboratory, which has already used artificial intelligence (AI) to predict the success of chemotherapy, can now determine which lung cancer patients will benefit from expensive immunotherapy.
The researchers train the computer to find previously invisible changes in CT scan patterns made during the first diagnosis of lung cancer, compared with images taken after the first 2-3 cycles of immunotherapy. And, as in the previous work, these changes were found both inside and outside the tumor.
Currently, only about 20% of all cancer patients actually benefit from immunotherapy, which differs from chemotherapy in that it uses drugs to help your immune system fight cancer, while chemotherapy uses drugs to directly destroy cancer cells.
Anant Madabhushi from the Center for Computer Visualization and Personal Diagnostics said that the recent work of his laboratory will help oncologists find out which patients will really benefit from therapy and which will not.
"Despite the fact that immunotherapy has provided new opportunities for cancer treatment, it remains extremely expensive – about $ 200,000 per patient per year," Madabhushi said. "This leads to the fact that about 42% of all those diagnosed with cancer lose their savings within a year after diagnosis."
Having a tool based on the research currently being conducted by his lab will go a long way to "better determine whether patients will respond to immunotherapy or throw away $800,000," he added, citing that four out of five patients will not benefit from the treatment.
"This is important because when a doctor makes a conclusion based on a CT scan alone whether a patient has responded to therapy, it often depends on the size of the lesion," Khorrami said. – We found that changing the texture is the best indicator of whether the therapy is working. Sometimes the nodule may look bigger after therapy for another reason, for example, because of a broken vessel inside the tumor - but therapy really works. Now we have a way to find out."
Prasanna, a postdoctoral fellow in Madabhushi's lab, said the study also showed that the results were consistent when scanning patients treated at two different locations and with three different types of immunotherapy agents.
"This demonstrates the fundamental value of the program in that our machine learning model can predict the response in patients receiving various immune checkpoint inhibitors," he said. "We are dealing with a fundamental biological principle."
Prasanna said the initial study used CT scans of 50 patients to train a computer and create a mathematical algorithm to detect changes in the lesion. He said that the next step would be to test the program on cases received from other places and using various means of immunotherapy.
In addition, according to Madabhushi, the researchers were able to show that the CT patterns that were most associated with a positive response to treatment and with the overall survival of patients were closely related to the location of immune cells on the initial diagnostic biopsies of patients.
This suggests that CT scans seem to record the tumor's immune response to exposure, and that those with the strongest immune response show the most significant structural changes and respond best to immunotherapy," he said.
Portal "Eternal youth" http://vechnayamolodost.ru