Researchers from the Johns Hopkins Kimmel Cancer Center and the Johns Hopkins University School of Medicine have developed a computational method to predict which patients with hepatocellular carcinoma, a primary liver cancer, may benefit most from combination immunotherapy and targeted therapy. The findings were published online July 14 in the Proceedings of the National Academy of Sciences.
The spatial QSP modeling platform was developed in the laboratory of Aleksander Popel, Ph.D., professor of biomedical engineering and oncology at Johns Hopkins. This platform combines quantitative systems pharmacology (QSP), which uses mathematical equations to simulate whole-body responses to treatment, with an agent-based model that tracks individual cell behavior within a tumor scenario. Together, these tools map both the quantity and location of cells, allowing predictions about tumor dynamics and its surrounding microenvironment.
In this study, researchers expanded their modeling platform to include fibroblasts—a cell type previously linked with resistance to immunotherapy in liver cancer—and created a machine-learning calibration workflow using real clinical trial data. This process generated virtual patients whose predicted responses could be compared against actual outcomes. “We generate a virtual tumor to see what happens in the microenvironment. Do the cancer cells resist? If you change the architecture of the tumor, does that help the cancer cells or the immune cells?” said Deshpande.
One advantage cited for this computational approach is its ability to scale quickly: simulations based on small phase I studies can generate virtual populations large enough for phase III trials without risk to real patients. When simulating treatment with cabozantinib (a targeted therapy) and nivolumab (an immunotherapy), both individually and together, predicted response rates closely matched those observed in actual clinical trials. The team also validated their model by comparing simulated tumor architectures against post-treatment tissue samples.
Researchers found that fibroblasts remodeled tumors’ microenvironments by creating physical barriers—"Even if immune cells were located near the tumor, the fibroblast would block [them] from reaching [it],” Deshpande said. The team suggested that architectural features identified before treatment could eventually help predict patient benefit from certain therapies.
