Artificial intelligence, or AI, programs contain algorithms designed to learn specific tasks. Once completed and ready for use, AI software performs assigned jobs faster and more accurately than people. The programs also have the ability to learn and improve themselves to perform more efficiently. The ability of AI to function rapidly is of great benefit to medicine by enabling physicians to better diagnose and devise treatment plans for individual patients. In this way, the programs are able to individualize therapy.
Medical facilities use computers to record and store a wealth of patient information. Along with contact information and diagnoses, data entered includes personal medical history, family medical history, current medications, current vital signs and status of the individual. When taught what to look for in terms of disease processes, AI has the ability to alert the medical team should a patient’s condition change. Some programs might also recommend a treatment plan appropriate to the situation.
AI might be used to develop intelligent medications. Once administered to the patient, the medication knows what area of the body and what cells are affected by the disease process. The formulation targets the cells and performs the action that it was designed to accomplish. The medication might alleviate local or systemic infections or perhaps eradicates tumors. But the programmed formula would only interact with the diseased cells and not healthy surrounding tissues.
In 2018, the scientists from the College of Medicine at the National University Hospital in Seoul developed an AI algorithm named DLAD, which stands for Deep Learning-based Automatic Detection. The program was designed to analyze chest images in order to detect possible abnormal cell development. DLAD was tested against the diagnostic skills of physicians using the same images. The algorithm was able to diagnose faster and more efficiently than 17 out of 18 physicians.
During the same year, researchers from Google AI Healthcare developed LYNA or Lymph Node Assistant. The program was tasked to analyze slides of tissue samples taken from lymph nodes to determine whether the cells contained breast cancer. The program was able to locate and identify suspicious areas that were not readily recognized by the eyes of human technicians. When tested on two separates sets of malignant and non-malignant sets of data, LYNA proved to be 99 percent accurate. The software also required half the time to analyze each slide before coming to a conclusion.