Online Master's in Biomedical Engineering @
A Mighty Force - read full story

Experts Convene at the Artificial Intelligence in Oncology Symposium


Above: Pallavi Tiwari, Hugo Aerts, Brian Hobbs, Donna Plecha and Satish Viswanath.


On October 24, 2019, the Case Comprehensive Cancer Center and the Center for Computational Imaging and Personalized Diagnostics (CCIPD) hosted the inaugural Artificial Intelligence in Oncology Symposium. Held on the campus of Case Western Reserve University, the event attracted 270 attendees from 12 states and eight countries, with an additional 50 tuning in to a livestream. 

“The potential for artificial intelligence to improve cancer care is extraordinary,” said Pamela Davis, MD, Dean of the Case Western Reserve University School of Medicine in her welcome message (pictured left). “Cancer is incredibly complex, and that complexity really lends itself to the kind of analytics you can derive from AI. The gains that could be possible with AI could be life-transforming for patients.” A leader in the field, the Case Comprehensive Cancer Center has investigators on more than 10 projects in AI in oncology funded by the National Cancer Institute.  

The symposium featured an impressive lineup of speakers and panelists, organized into four sessions that focused on developments and opportunities for AI in radiology, pathology and genomics, immuno-oncology, and policy and ethics. The gathering of a variety of stakeholders at the intersection of oncology and artificial intelligence was timely given that the National Cancer Institute named AI one of the three pillars of its strategic plan for 2020 and beyond.

“Artificial intelligence could have a profound and transformative impact on the management and care of cancer patients, but it’s not just about AI for diagnosis and detection of the disease,” said Anant Madabhushi in his opening remarks. Madabhushi is director of the CCIPD and F. Alex Nason Professor II of Biomedical Engineering at Case Western Reserve University. “There is a huge opportunity for the role of AI in prognosticating disease outcome – for stratifying noninvasively with imaging data the more aggressive cancers from the less aggressive cancers. There is also a clinical role for AI in predicting therapeutic response to identify which patients will respond to specific therapies in advance of that treatment.” Since its inception in 2012, the CCIPD has developed more than 80 technologies utilizing computer analytics and big data in oncology and other medical specialties.

One of the highlights of the symposium was the keynote address by Sohrab Shah, the Nicholls-Biondi Endowed Chair in Computational Oncology at Memorial Sloan Kettering Cancer Center, who shared insight into the computational methodologies he has developed and applied in cancer genomics. “In the lab, we view cancer as a dynamic disease, and that means the cancer diagnosis is often very different than the cancer that achieves relapse or resistance to treatment or undergoes metastatic spread to different parts of the anatomy,” he said. “Ultimately, these processes are driven by changes in the genome, so we think of cancer as a disease of the genome.” Shah’s address focused on the co-evolution of malignant and immune cells in high-grade serous ovarian cancer.

Throughout the keynote address, speaker presentations and panel discussions, a few overarching themes emerged as experts discussed the road maps for successful translation and clinical adoption of AI tools:

  • AI in oncology requires vast, validated datasets. “Cancer is a computational problem,” said symposium speaker Kevin White, chief science officer at Tempus, a Chicago-based biotechnology company. “Over the next decade, the greatest progress in technology will be made by those who have access to the deepest and most comprehensive datasets.”
  • Datasets must be shared over multiple sites. “It is incumbent on us to push our health systems to develop data models and data marts that allow us to take all of the transactional data we generate on a day-to-day basis and mobilize it in shareable ways,” said symposium speaker Michael Feldman, MD, a professor of pathology and laboratory medicine and director of the Office of Pathology Informatics at the University of Pennsylvania. 
  • Experts must breakdown professional silos and work together. “Having a data dump where we place all the data and think something beautiful will drop out is not realistic,” said Shah. “But if we integrate the clinicians, oncologists and surgeons and ask them to articulate the pressing clinical problems that can be solved through integrating multi-modal data points, then we will likely get somewhere.”

Though the road ahead for AI in oncology is challenging, attendees at the Artificial Intelligence in Oncology symposium were upbeat about the field’s potential to ultimately make a difference in the lives of cancer patients. “I think in the next decade we will see a huge growth area in integrated multi-modal systems,” said Shah. “It’s an exciting time because data are ubiquitous.