Reaching new heights by using AI in diagnosing coronavirus. AI into medicine!!
The year of 2020 which began with the word COVID-19 being spoken out of every mouth of almost everyone, brought destruction to the world by bringing many deaths caused due to the coronavirus. The virus is deadly when not detected soon enough for the treatment to begin, causes death. Now about 3,47,110 deaths around the world. The normal standard way of testing a patient if he is positive for the virus or not would take up to two days, says the experts. An AI tool can quickly spot or detect COVID-19 using the chest CT scans and the patient’s clinical data.
The chest CT scans are useful in finding out and analyzing the symptomatic patients suspected in having the virus, but there are limitations to these tools as they have predictive value only to a certain point. Thus, in pointing the need to combine both the clinical data and the process of diagnosis using Artificial Intelligence in doing so.
According to AI experts, Artificial Intelligence has the huge potential, that could analyze large chunks of data quickly and accurately, having such an attribute and putting it into use would be the best thing to do during pandemics such as the one mother earth is currently facing.
Read and learn more: Machine Learning- in fighting against COVID-19 virus
In the research conducted, the team tuned the algorithm on the data that was received from 626 patients out of 905 patients initially and then for 279 patients that were remaining to check the sensitivity of the tool. The results were satisfactory as it had 84% i.e. a higher sensitivity compared to the radiology results which had about 75% sensitivity.
The test focused on CT scans and blood tests, then the data from the scans was integrated along with the clinical data like blood test showing any abnormalities in white blood counts, age, sex, and the symptoms to build an AI algorithm. The tool then does what a human physician would do while analyzing and then gives the final prediction.
The combining of both the CT scans along with the clinical data, provided many accurate results as it improved the detection of COVID-19 positive patients who had negative CT scans alone but were positive for coronavirus when combined with clinical data. It recognized 68% of positive cases while all those cases were said to be negative by the radiologists as they all had negative CT scans.