OR WAIT null SECS
In 323 cases, IBM's artificial intelligence identified additional treatment options that a panel of experts overlooked.
Time and again, healthcare experts have doubted the ability of Watson to help fight cancer. But the findings of a new study pose an argument in favor of IBM’s flagship artificial intelligence (AI) technology.
Led by the University of North Carolina (UNC) Lineberger Comprehensive Cancer Center, researchers employed Watson for Genomics to dig into “large volumes of data” from studies and databases, along with genetic information, to target treatment options and clinical trials for 1018 patients with “tumors with specific genetic abnormalities.” The team then compared the AI’s choices regarding to those made by a molecular tumor board consisting of cancer experts, according to an announcement.
The findings boded well for Watson. Its cognitive computing confirmed 703 cases in which the expert panel identified “actionable genetic alterations,” according to the university. And the AI unearthed additional “potential therapeutic options” for 323 patients—or a third of the participant base. Humans had not recognized actionable mutations in 96 of these patients, according to the study.
“Our findings, while preliminary, demonstrate that cognitive computing might have a role in identifying more therapeutic options for cancer patients,” the report’s corresponding author, William Kim, MD, an associate professor at UNC’s medical school, told the university. “I can tell you that as a practicing oncologist, it’s very reassuring for patients to know that we’re able to explore all possible options for them in a very systematic manner.”
The research squad, which also included IBM Watson Health investigators, concluded that the analysis and usefulness of next-generation sequencing are “evolving too rapidly to rely solely on human curation,” according to the study, which was published yesterday in the journal The Oncologist. The authors advocated for molecular tumor boards to take advantage of cognitive computing when it comes to examining data and seeking clinical trials.
In the extra 323 cases targeted by Watson for Genomics, just 8 genes made the difference, Kim, who has served as an adviser to IBM, noted. The group of experts had not considered those genes. Yet Watson matched most of the patients with a new clinical trial, including one that kicked off within only a week of the AI analysis, according to the researchers.
Kim cautioned that his team did not design the study to examine how cognitive computing can improve outcomes, including “prolonged survival or treatment response.”
Further, most of the patients—who had undergone exon sequencing at UNC—no longer had active cancer or had already died. Researchers, however, reported findings for 47 active cancer patients to their physicians, according to the study.
“To my knowledge,” Kim said, “this is the first published examination of the utility of cognitive computing in precision cancer care.”
As the body of genomic sequencing, treatment, and therapy response data grows, AI like Watson for Genomics stands to become more powerful, Kim said. He noted that researchers must continue to ask and answer questions surrounding the technology.
The study is titled “Enhancing Next-Generation Sequencing-Guided Cancer Care Through Cognitive Computing.” Grants from the University Cancer Research Fund and the National Cancer Institute funded the effort.