Don’t be fooled by the news you’re hearing that artificial intelligence is about to transform how medicine is practiced. While artificial intelligence is promising, it’s years from coming close to changing how you’re treated.
As reported by CNBC, Alphabet CFO Ruth Porat said during a panel on Tuesday for the World Economic Forum that “We needed only hundreds of biopsies from breast cancer patients hundreds in order to have a breakthrough in metastatic breast cancer because of the developments in AI.” Claims like these should be taken with a heavy grain of salt. Sure, Google (which Alphabet owns) may well have been able to diagnose metastatic breast cancer with this little data, but if you ask clinicians, its going to take a lot more data to impact the point of care. We spoke to Dr. Elizabeth Comen, oncologist at Memorial Sloan Kettering Cancer Center, who said of Porat’s comments, “Obviously technology will improve in years to come and AI will be able to increasingly do interesting things, but I think this is years away from transforming how we treat patients. It’s going to take significantly more evidence and vastly more data to really change standards of care and clinical practice.”
Read More Health tech innovation has undoubtedly garnered some big time attention in recent years.
Flatiron Health, a healthcare company that is improving electronic health records for cancer patients, was purchased by
Roche, a multinational healthcare company, for $2 billion last year.
Tempus, another startup trying to use clinical data to improve cancer care, is currently valued at over $2 billion as well. However, just because this much money is being thrown at companies like this, doesn’t mean we’re anywhere near the amount–nor
the kind–of data necessary for artificial intelligence to actually make diagnoses for us. A recent perspective piece in the
Journal of the American Medical Association comments about the type of data that is necessary for AI to make accurate clinical decisions, and where we currently stand on collecting it. In the article, the authors cite “inconsistent data quality, limited evidence supporting the clinical efficacy of AI, and lack of clarity about the effective integration of AI into clinical workflow” as just a few of the challenges that plague the advancement of artificial intelligence. The authors close by warning about the “hype of AI.” They write, “Without the difficult work needed to address these issues, the medical community risks falling prey to the hype of AI and missing the realization of its potential.” That’s not to say that there isn’t exciting work being done to utilize technology to improve cancer treatments. SurvivorNet spoke to
Dr. Eliezer Van Allen, medical oncologist at the Dana Farber Cancer Institute, about how his research group is
using genetic data to tailor treatments to late stage prostate cancer patients. He told us, “We think there are actually a diverse set of drugs that we can apply to men in those settings, and that’s actually paving the way for a lot of innovative new clinical trials to match patients to those drugs based off of their genetics, which again sort of emphasizes our goal of actually bringing precision medicine to prostate cancer.” Work like Van Allen’s will further personalize cancer treatments to individual patients, but actually using artificial intelligence to help diagnose cancer? We’re not quite there.
For now, listen to your physician–they know best. A machine won’t be replacing your doctor anytime soon.
Learn more about SurvivorNet's rigorous medical review process.
