How AI is changing cancer care and research?
The potential for artificial intelligence to disrupt healthcare and health systems is becoming evident. In this article we explore the AI-powered future of cancer care and research.
Cancer. Tumors. Harsh, jarring words that often trigger painful memories and fears. Doctors worldwide are seeking to reinvent the way we treat this group of diseases. Oncology is the branch of medicine that specializes in the diagnosis and treatment of cancer. Interestingly, it also includes three different sub-specialties: medical oncology, radiation oncology and surgical oncology. Why am I going into this much detail? Artificial intelligence (AI) is playing an important role in each of these sub-specialties.
In our two-part series on the impact of artificial intelligence (AI) on Healthcare, we were exploring how artificial intelligence is affecting various parts of the healthcare spectrum. To briefly remind you, we’ve talked about the Google Deepmind Project at the University College London Hospital, where AI was used to analyze targeted radiotherapy based on anonymized scans of 700 patients. The process involved in the analysis, known as segmentation, usually takes four hours to complete, but with AI they have seen the potential to reduce that time to one hour. Convenience and more efficient planning of radiotherapy lead to better-served patients.
How is AI changing healthcare in terms of oncology?
Physicians from the University of Texas MD Anderson Cancer Center and the Palo Alto Medical Foundation in California have begun exploring potential uses of AI and Big Data in the fight against cancer (JAMA Oncology). They’ve suggested 14 scenarios in which cancer care and research can significantly benefit from such discoveries. To summarize these findings, AI researchers and clinicians will accelerate oncology research in three main ways:
By further developing and integrating existing cancer registries, from the local to the international level. These are analyzed and interpreted for a better understanding of cancer mechanisms (from common to rare cancers). Big Datasets provide a credible evidence base, while AI helps in the analysis.
By improving cancer treatment pathways around the world through analysis of best practices and trends.
By dramatically facilitating the implementation of cost-efficient trials.
AI in oncology – reinventing the tools we use to diagnose cancer
Traditionally, cancers are detected using clinical methods such as ultrasonography, X-ray, Computed Tomography (CT), and Magnetic Resonance Imaging (MRI). However, a number of cancers cannot be recognized by these techniques. An alternative approach is the analysis of microarray gene profiles. Sounds complex? Cancers can be detected using incredibly small amounts of genetic material to assess the degree to which certain genes are expressed. The data produced by this genetic material create enormous data sets that must be analyzed. This analysis may take hours to complete. Now imagine that this analysis could be performed quickly through the use of AI. Artificial intelligence, in fact, plays an important role here, as seen in studies from 2001 and 2003, and when we fast-forward to 2017, we see researchers using Cascaded Neural Networks to classify cancer through innovative techniques such as Gene Masking.
Case studies for AI in oncology
The ugly face of tumors shows up when scientists struggle to understand, or even better, predict, the way that tumors behave. Countless patients and their families who have been through the cancer journey struggle with the possibility of a relapse. Relapse occurs when a few of the original cancer cells survive the initial treatments or were too small to be detected during the follow-up immediately after treatment.
I have been inspired by the intriguing collaboration that took place between the Stanford Artificial Intelligence Laboratory and Stanford Medical School. This resulted in an incredible effort involving TensorFlow and a database of 130,000 skin disease images. They trained the TensorFlow algorithm to visually diagnose potential skin cancer. Best of all, they tested the diagnoses produced by this algorithm against the opinion of 21 board-certified dermatologists. The outcome? A ground-breaking research paper where the AI-enabled diagnostic algorithm matched the performance of dermatologists. Want to learn more? Check out this video focused on the technological aspect:
Startups using AI in the battle against cancer
The artificial intelligence scene within the field of oncology is growing, and there are five companies worth keeping an eye on. Enlitic is using deep learning to automatically detect lung cancer nodules in chest CT images with 50% greater accuracy than an expert panel of thoracic radiologists.
All these kinds of treatments are supported by insights and analytics, and Oncora Medical is bringing predictive insights and risk analytics to radiation oncology. In this way they’re helping radiation oncologists to make better decisions and make use of the diverse and valuable data they’re generating
How to improve Emergency Departments with AI?
Faith in the ER is being restored. Thanks to Artificial Intelligence predicting patient flow and avoiding unnecessary hospital visits.
Pathologists around the world diagnose cancer on a daily - if not hourly - basis, and their work includes analyzing thousands of slides. Now imagine if there was a simple way to help these specialists to filter through all the normal slides and to flag the ones that require further attention. Proscia has designed a digital pathology platform allowing pathologists and researchers to “leverage the pathology data in every slide”.
The future of AI in the field of oncology
In ESMO Open, an open-access oncology journal, Dr. Curioni-Fontecedro’s article “A new era of Oncology through Artificial Intelligence” succinctly explains the current situation. That is, that although this technology and research exist and are available for cancer care and research, they have not yet spread throughout the oncological community. The missing ingredients needed to take this to the next level in oncology are clinician buy-in, funding opportunities for implementation, and education.
“I’m looking forward to a future where cancer will be treated in a simple and seamless manner, providing evidence-based hope and opportunities to patients suffering from these diseases.”
The future is bright in the field of cancer research and treatment. I’m looking forward to a future where cancer will be treated in a simple and seamless manner, providing evidence-based hope and opportunities to patients suffering from these diseases. This is, after all, not a matter of replacing an entire profession with AI, but rather providing timely care to patients in a disease specialty where the time factor is critical and treatment needs to be swift and precise.