Artificial Intelligence in medicine
Artificial intelligence is front line technology in computer science research. Basically AI is human intelligence simulated in the form of software, machine or robot. AI systems are cognitive in nature and are playing a major role in health care departments for the detection of various diseases. The use of artificial intelligence in medicines is significant and promises accurate diagnosis and treatment of acute diseases. Artificial intelligence is yielding medical screening and testing on a larger scale. Since diabetic retinopathy is increasing day-by-day and sufferers are expecting more from technology to provide a solution to their pain. Past year, a tool IDX-DR was launched to analyze this disease without help of doctor. The estimated number of patients is above million in every state which is facing complexities in curing this pain.
Diabetic Retinopathy Diagnosis
Diabetic retinopathy is a cause of blindness in adults who are suffering from diabetes. This disease gets complex when blood has high sugar levels and it damages the retina. People are scared to death by getting blind for a life-time from this disease and the situations are going worst and complex in medical centers.
AI solution: IDX DR FDA
IDX-DR is developed by using AI and is a first-ever approved algorithm for the detection of diabetic retinopathy by FDA. The device has an integrated retinal camera (Topcon, TRC-NW400) which captures eye images that are forwarded to cloud server. The server has enabled IDX DR software with a “deep learning” algorithm; the server performs calculations to uncover retinal findings in accordance with DR and using a large training dataset of representative images. The results obtained from this software are of two types
- If DR is detected, it recommends eye care professional
- If DR is not detected, it recommends rescreening after one year.
IDX DR FDA Approval
Before approval, the FDA tested IDX-DR on 900 subjects by performing an automated image examination. The device is trained to capture two images for one eye with a 45-degree angle and perform a comparison with images provided by FPRC (Wisconsin Fundus Photograph Reading Centre). After the acquisition of retinal images, the AI system detects disease in just 20 seconds.
Therefore, there is no full stop in technology. On the basis of IDX-DR, a new entity is evolved which is known as mtmDR (minimal DR). The scale used for diagnosis Early Treatment Diabetic Retinopathy Study (ETDRS) scale which is capable of detecting the presence of diabetic retinopathy on a scale of 35 or higher than this at the minimum in one eye. The algorithmic assessment of this technology is done with 96 percent of sufficient quality images in primary health care settings. This algorithm is worth of appreciation as it performs 100% successful detection of ETDRS on a scale of 43 or higher than this and its decisions are without the help of any eye specialist; that’s where its importance lies and it can also aid non-ophthalmic professionals if they want to treat any patient.
The IDX DR can be installed anywhere in the world and is easily used for eye-examination. A mild problem for patients is access to the system, but once they make their way towards it, they lead themselves towards successful detection of DR. Doctor involvement is omitted in such detection and analysis and hence it is proven from case studies that AI system is more accurate and precise than human in detection of DR. Operation of this system is easy-to-handle; it only requires a computer literate person to operate it. Patients having examination through this AI machine are quite happy as it is quick and smooth and it does not require regular visits just like ophthalmologists. However, detection gets troublesome with people having small pupil. If the percentage of such patients is reduced, the usefulness of the machine can be increased proportionally.
Everyone is aware of AI and its tools. FDA has approved an AI tool to assist medical professionals in the diagnosis of diabetic retinopathy and most patients are happy from this innovation but concerns of safety, reliability, efficiency, and accessibility are also there. On the same side, ophthalmologists are worrying about their jobs. In the upcoming year use of AI in medical sciences have to be transparent and benefits and risks have to be openly discussed for people believing and using AI in their life decisions. Artificial intelligence is making its way towards the diagnosis of macular degeneration and glaucoma beyond diabetic retinopathy and it is of no doubt that it will encourage physicians and patients to trust technology on the matter of life-saving decisions.