MedicalResearch.com Interview with:
Raymond H Mak, MD
Brigham and Women’s Hospital
MedicalResearch.com: What is the background for this study?
- Lung cancer remains the most common cancer, and leading cause of cancer mortality, in the world and ~40-50% of lung cancer patients will need radiation therapy as part of their care
- The accuracy and precision of lung tumor targeting by radiation oncologists can directly impact outcomes, since this key targeting task is critical for successful therapeutic radiation delivery.
- An incorrectly delineated tumor may lead to inadequate dose at tumor margins during radiation therapy, which in turn decreases the likelihood of tumor control.
- Multiple studies have shown significant inter-observer variation in tumor target design, even among expert radiation oncologists
- Expertise in targeting lung tumors for radiation therapy may not be available to under-resourced health care settings
- Some more information on the problem of lung cancer and the radiation therapy targeting task here:https://www.youtube.com/watch?v=An-YDBjFDV8&feature=youtu.be
MedicalResearch.com: What did you do to solve this problem in oncology?
- We aimed to develop artificial intelligence techniques to develop an automated lung tumor segmentation solution that can match the performance of an expert radiation oncologist in order to increase the efficiency and quality of this key step in radiation therapy planning.
- Recognizing that conventional academic approaches may not meet the need for more rapid cycles of innovation to develop a solution, we adopted a prize-based crowd innovation approach to crowd-source AI solutions that could replicate the human expert’s lung tumor targeting abilities.
MedicalResearch.com: What are the main findings?
- Over 10 weeks of contest time and with $55,000 in prize money for the contest, we challenged hundreds of data scientists to produce AI algorithms that match the performance of an expert radiation oncologists in targeting lung tumors
- The contestants generated 45 different solutions, and the winning AI solutions were able to target lung tumors with an accuracy comparable to the inter-observer variation seen between expert radiation oncologists
- These algorithms could perform the task in seconds (compared to average time of 8 minutes for an expert)
MedicalResearch.com: What should readers take away from your report?
- Response: Artificial intelligence is poised to revolutionize the way we deliver radiation therapy through increasing automation and can improve the quality of and access to this critical cancer therapy
MedicalResearch.com: What recommendations do you have for future research as a result of this work?
- Applying crowd innovation to problems in healthcare can lead to rapid prototyping of AI solutions and catalyze transformations in healthcare delivery.
- Clinical implementation of AI algorithms in therapeutic medicine will require careful validation by human experts and robust quality assurance methods are required.
MedicalResearch.com: Is there anything else you would like to add?
- The work represents a multi-disciplinary collaboration between Dana-Farber Cancer Institute, Brigham and Women’s Hospital, Harvard Catalyst, Harvard Business School and the Laboratory for Innovation Science at Harvard, in collaboration with the online crowdsourcing platform Topcoder
- The work was partially funded by The Laura and John Arnold Foundation, Harvard Catalyst, The Harvard Clinical and Translational Science Center NIH UL1 TR001102, and the Division of Research and Faculty Development at Harvard Business School.
Mak RH, Endres MG, Paik JH, et al. Use of Crowd Innovation to Develop an Artificial Intelligence–Based Solution for Radiation Therapy Targeting. JAMA Oncol. Published online April 18, 2019. doi:10.1001/jamaoncol.2019.0159
The information on MedicalResearch.com is provided for educational purposes only, and is in no way intended to diagnose, cure, or treat any medical or other condition. Always seek the advice of your physician or other qualified health and ask your doctor any questions you may have regarding a medical condition. In addition to all other limitations and disclaimers in this agreement, service provider and its third party providers disclaim any liability or loss in connection with the content provided on this website.