It’s no secret that the medical field is technologically advancing, especially in the last several decades. In an era where pen and paper were mainly used to keep track of patient information, and treatment & prevention of illness were not nearly as effective as they are today, it is clear to see that modern medicine moved forward in tremendous ways. But why is this? One of the biggest factors driving these advancements in the healthcare industry is technology, and technology integration to create smarter health tools & devices. Some examples of these technological advancements include electronic health records, 3D reconstructions in trauma patients, DNA sequencing, laser technology, wearable devices, and the list goes on! Because of these new technologies, physicians have been able to improve their practice, and patients have been able to receive better care. Although, a specific subset of tech is currently taking over the healthcare world with bright and exciting promises; artificial intelligence. Artificial intelligence (AI) is a branch of computer science concerned with smart machines which have the ability to mimic human thinking and actions. Concepts of AI can already be seen implemented in our daily lives, such as opening our iPhone with face ID, personalized feeds on social media, digital voice assistants such as Siri & Alexa, and so many more! So the question comes down to, how is AI being used in healthcare?
Currently, AI has been found to be able to streamline the process of diagnosing a patient, classifying a disease, and making a decision on how to treat the disease. Although most AI diagnosis and classification of diseases is still under process, several researchers have conducted studies upon the implementation of AI healthcare technology and the effects they have had. For example, in May of 2019, Google researchers developed a deep learning tool which examines CT scans of a pair of lungs, and determines if there is any sign of lung cancer. They found that the model reduced false positives by 11 percent and false negatives by five percent, and had a 94.4 percent AUC (a type of performance measurement). Not only did the tool help in improving the accuracy in the detection of lung cancer, but when compared to the performance of six board-certified radiologists, the model performed in equation to or better than the human radiologists. Another study published by Stanford University developed an AI to classify skin cancer, conveying whether it is benign or malignant. The study showed that the researchers used a convolutional neural network (CNN) trained by a dataset of 129,450 clinical images in order to classify the skin cancer. The study found that the model achieved performance on par to that of tested medical experts/dermatologists, and that the model even has the capability to be used on a mobile device, increasing the access to medical tools in a low-cost manner. Not only has AI been able to diagnose and classify diseases in a patient, but AI is also being used by doctors in order to ease the decision-making process on which treatment to proceed with. Clinical Decision Support (CDS) has the ability to “analyze large volumes of data and suggest next steps for treatment, flagging potential problems and enhancing care team efficiency.” Although CDS systems are still being developed, many at the Healthcare Analytics Summit proposed plans on the integration of CDS, and the future of its implementation.
As we see the implementations of AI in healthcare start to rise, the positive effects start to come into place. According to the American Cancer Society, they estimate that in 2021 there are 235,760 new cases of lung cancer and about 131,880 deaths. Meanwhile, the AI tool Google developed would be able to significantly decrease this number, saving thousands of lives. And this is just one example! Implementing AI in all facets of healthcare can not only save lives and prevent patients from dying because of diagnosing unseen issues, but can also help doctors and physicians in the process. Physicians already deal with frequent stress and pressure, from the countless patients they help to apply their medical knowledge on a daily basis. By having a helping hand, doctors are able to perform better and in a more efficient & productive manner, all while the accuracy and reliability of their work increases in helping their patients!
Authored by Surbhi Kumar
Ardila, D., Kiraly, A.P., Bharadwaj, S. et al. End-to-end lung cancer screening with three-dimensional deep learning on low-dose chest computed tomography. Nat Med 25, 1319 (2019). https://doi.org/10.1038/s41591-019-0536-x
Esteva, A., Kuprel, B., Novoa, R. et al. Dermatologist-level classification of skin cancer with deep neural networks. Nature 542, 115–118 (2017). https://doi.org/10.1038/nature21056
Kent, Jessica. “How Machine Learning Is Transforming Clinical Decision Support Tools.” HealthITAnalytics, HealthITAnalytics, 26 Mar. 2020, healthitanalytics.com/features/how-machine-learning-is-transforming-clinical-decision-support-tools.
“American Cancer Society: Cancer Facts & Statistics.” American Cancer Society | Cancer Facts & Statistics, Jan. 2021, cancerstatisticscenter.cancer.org/#!/cancer-site/Lung%20and%20bronchus.