From the October 2016 issue of HealthCare Business News magazine
In a time of increasing complexity, increasing data intensity and increasing demand for real-time health care delivery, we must drive the next digital revolution to radically transform medicine.
The current model is not sustainable. One hundred and twenty-one years after the first X-ray image, 2.5 exabytes of data and almost 2 million imaging exams are generated every day globally. Data produced by health care devices are expected to increase 50-fold by 2020. And the radiologist’s workload continues to swell as she is given more to analyze.
Providers are looking for more efficient, data-driven and outcomes-oriented solutions. The future of health care is critically intertwined with the next digital revolution. Traditional analysis of health care data will no longer be adequate. Leveraging the power of data analytics through deep learning and artificial intelligence (AI) is going to drastically change the future of medicine. Using deep learning to generate contextual algorithms will deliver more accurate clinical decision support, workflow and health care systems efficiencies and better outcomes. This is the way of the future.
In the not-so-distant future, computer-intensive approaches to biomedical analysis, in addition to deep learning, will start becoming practical with cloud computing:
Midmark Workstations are made to order with customization that can assist with the integration of telehealth and other technology at the point of care, wherever that may be. See more>>>
One could imagine a “digital twin” of the imaging system as well as of human anatomy, physiology and function becoming computationally feasible. These digital twins, along with real-time access to large patient databases, would enable “smart” imaging systems that could prescribe optimal patient specific imaging protocols and then organize and analyze the massive amounts of data generated to aid clinical interpretation.
Quantitative imaging data will be increasingly combined and co-analyzed with large sets of clinical, genomic, monitoring and pathology data to improve and personalize the diagnosis, prognosis and treatment selection for complex disorders.
AI-based systems will become an extension of the clinician. These systems, employing deep learning and neuro-cognitive AI, will operate in a continuous reinforcement-learning mode with a clinical expert in the loop to develop an unprecedented level of machine intelligence in medical decision-making.
The fields of radiomics and radiogenomics — the correlation of large image-derived feature sets with clinical outcomes and with the underlying biology and molecular phenotypes, respectively — have the potential to unlock “hidden” information from data-rich patient scans.