In a recent report, NCI-funded researchers described a new artificial intelligence algorithm, called SPHINKS (Substrate PHosphosite-based Inference for Network of KinaseS). This algorithm may prove helpful in identifying and treating the underlying biological causes of glioblastoma multiforme (GBM).

Although not yet ready for clinical use, an algorithm that identifies kinase activity as a driver of tumor growth and progression could help assign patients for prospective clinical trials and target highly effective therapies. In addition, because kinase activity is common in many cancers, SPHINKS has the potential to work in a broad range of tumors, including certain breast, lung, and pediatric cancers.

Kinases are known targets in cancer therapy. These enzymes act along with phosphatases to form complex, interwoven networks that respond to internal and environmental factors to turn protein activity “on” or “off.” Most protein kinases promote cell growth and proliferation. When kinase activity is unchecked, it can lead to many types of cancer. Medications that inhibit protein kinases are effective but don’t work the same in every patient.

GBM has numerous subtypes, with unique omics profiles (i.e., transcriptomics, proteomics, phospho-proteomics, metabolomics, and more) and distinct clinical and imaging features. In this study, using SPHINKS, the team added kinase data to these earlier GBM profiles. The addition enabled them to see the master kinases involved in activating or deactivating the proteins that lead to cancer in four common subtypes of GBM.

“Until recently, we’ve had to rely on imprecise methods, like immunohistochemistry, to define cancer subtypes,” notes NCI Program Officer Dr. Kristine Willis, Division of Cancer Biology. “If we can improve our ability to identify different subtypes using modern methods, including predicting the underlying biological drivers, it could help us select the most effective treatments and better target patients to clinical trials.”

Dr. Anna Lasorella, the co-corresponding author of the study, added, “Identifying abnormal protein kinases linked to specific subtypes of glioblastoma enables us to quickly select highly precise treatments. This would be an important advance in treating GBM, which is an especially aggressive and deadly cancer that remains without effective therapies.”