Many people, myself included, look at the artificial intelligence, as an ultimate answer for oncology treatment improvement, through machine learning and automation. Something that will lead to a more streamlined and effective work process, while providing better care for the patients. Some people, however, argue, that despite the promises, the industry is still yet to deliver any formidable results - many of the “detailed diagnostics” lack both consistency and solid evidence of effectiveness. So, is the technology really not working, or is it just not working the way we want it to work?
The recent surge in AI development (more on AI development here) has highlighted the possibility of data to overcome some of the biggest challenges in healthcare, and the market sure reacted appropriately. Recently, the European Commission announced a €20Bn AI strategy for Europe with France launching their own program for €1.5Bn. And with expert estimations, predicting that medical knowledge will be growing in double every 73 days by the year 2020, everything seems to look rather positive. All of this, however, does not answer what role the oncology field will play in this development process. Oncologists have been struggling for decades in order to define small subsets of patients that may benefit from a concrete treatment, as seen with immunotherapies. The medical staff is overcrowded with all sorts of data, from genomics and co-morbidities to previous treatments and imaging. As such, developers would require to come up with better tools if they want to succeed in this combat.
Let us look at this example: over 80,000 new oncology papers were published during the year 2017 alone. With that much data on their hands, it is humanly impossible for doctors to stay up to date with the latest advances in the field. This is where the clinical decision support systems come to a place, ensuring that all the latest knowledge is taken into consideration, and the highest level of care is given as a result. Nearly a hundred new startups have begun to use artificial intelligence as a drug discovery tool in order to uncover synergistic combinations of possible drug targets. For breast cancer alone, there are 69 different approved standalone treatment drugs. Now take this number and add it to the number of working combination treatments out there, and oncologists face a catastrophe of a memory test. AI focus is then split into two directions - it needs to be increasing the efficiency of drug discovery as well as improving speed and accuracy of diagnosis.
Keep this example in mind, because now I want to talk about something that will really show you that the technology is only going forward. I am talking about something that is only possible with real dedication and passion, with thoughts of greater goods and an extension of people's lives. I am talking about breakthroughs. Let me start with mentioning an amazing this year’s American Society of Clinical Oncology (ASCO) conference. With the main trend being “less is more” which inlines perfectly with the question I have raised at the beginning of the article, two new clinical best practices made a big success story, by recommending areas where cancer patients could avoid all the risks of unnecessary costly treatments. The first one is about the people with an advanced kidney cancer, and that they are not actually required to have a surgery. The second practice tells us that most of the women in the early stages of breast cancer could skip chemotherapy with a diagnostic test guidance. These two studies were both convicted with the use of clinical trials, but this can only show how bigger and faster results we will be able to get, using the artificial intelligence technology.
Remember the drug discovery startups and clinical support systems we talked earlier? It is time to mention some of the formidable heroes in the field. And it is impossible not to mention IBM’s Watson Health, an AI clinical decision support system, which is able to analyze various data across different health sectors, like cardiology and diabetes. Watson for Oncology (WFO) focuses on the quick diagnosis and optimum treatment proposals for patients suffering from cancer. The recent study made by Indian researchers showed that WFO’s breast cancer treatment solutions were on part with the ones made by the expert oncologists. Out of 638 cases, the panel agreed with a 93% of them, which is an unbelievable result. The same panel of experts also agreed on treatment in colon, lung and rectal cancer cases.
I would also like to mention a company named Roche, which is one of the key leaders in the oncology field, and their recent partnership with GNS Healthcare. Together they developed a Flatiron Health, an Alphabet-backed oncology-driven digital health analytic startup. Flatiron's interface gives physicians access to different datasets that can be used decipher insight from inconsequential statistical noise. The company has the ability to isolate and validate both tumor biomarkers and therapeutic agents by using its lexicon of genomic and patients data. This is specifically focused on oncology and is intended to showcase the intrinsic drivers behind some of the most problematic tumor types.
And last, but not least, let us hear an insight from someone representing the industry and her view regarding AI’s place in the oncology field. During a recent panel at MedCity Converge conference in Philadelphia Dr. Tufia Haddad, chair of breast cancer medical oncology and of information technology at the Mayo Clinic in Rochester, gave her perspective on the matter: “Right now, for every hour I spend talking to patients about their health care plans, I spend two hours behind a computer. That time behind a computer screen involves a lot of data mining. Meanwhile, the doubling of knowledge in oncology can make it hard to keep up even for academic oncologists focused on specific cancer types, let alone community oncologists who must deal with solid tumors and hematological cancers across the spectrum. That is where AI can come in handy.”
In my opinion, we are slowly approaching a new era of oncology and the current generation of medical workers must be ready to approach this evolution with the curiosity that characterizes them in their profession. The incredible potential of these new technologies is that they can further learn and improve something that today we can only imagine to perfect.