Since immunotherapy began to take off in the first decade of the 2000s, it has provided hope, strength and healing to patients with cancer. Although scientists have made great strides, the technology is still in its infancy—a fact supported by two important considerations: 1) it only works for about 20% of cancer patients (the why still alludes scientists), and 2) it is prohibitively expensive. At about $200,000 per patient per year, combined with the small success rate sample size, immunotherapy can not be considered a large-scale cancer cure/treatment—yet.
Researchers at the Center for Computational Imaging and Personalized Diagnostics (CCIPD) at Case Western Reserve University are dedicated to changing that. Since it’s inception in 2012, the center has become a global leader in the detection, diagnosis and characterization of various cancers and other diseases by meshing medical imaging, machine learning and AI. The lab’s latest discovery is just another step in the right direction.
Building on their already-pioneering work of using artificial intelligence (AI) to predict whether chemotherapy will be successful, the researchers can now determine which lung cancer patients will benefit from immunotherapy based on CT images.
Study co-authors Anant Madabhushi, Prateek Prasanna and Mohammad Khorrami used CT images from 50 patients to train a computer to see the differences in lung lesions pre-treatment compared with post-immunotherapy. From there, they created a mathematical algorithm/computer program that could note the changes in texture, volume and shape of a given lesion, not just its size.
“This is important because when a doctor decides based on CT images alone whether a patient has responded to therapy, it is often based on the size of the lesion," Khorrami, a graduate student at CCIPD said. "We have found that textural change is a better predictor of whether the therapy is working. Sometimes, for example, the nodule may appear larger after therapy because of another reason, say a broken vessel inside the tumor--but the therapy is actually working. Now, we have a way of knowing that.”
After consistent results across patients treated at two sites with three different immunotherapy agents, researchers noticed a pattern. The tumors with the most significant textural changes were the same ones showing the strongest positive immune response. In fact, overall patient survival was closely associated with the arrangement of immune cells in the original diagnostic biopsies of those patients.
Remarkably, the differences between pre- and post-treatment CT scans show both inside and outside of the tumor—an imaging characteristic common to CCIPD, but not many other digital imaging labs.
Prasanna said the next step for the lab and researchers is to test the computer algorithm on cases obtained from other sites, as well as across additional immunotherapy agents. Currently, the most successful immunotherapy targets include PD-L1 and T-cells.
Photo: Differences in CT radiomic patterns before and after initiation of checkpoint inhibitor therapy. Courtesy of Case Western Reserve University.