AIの歴史と腎移植における活用—Society 5.0における臓器移植の在り方を求めて

<p>Three artificial intelligence (AI) booms have occurred since Alan Turing, the creator of the world's first computer, dreamed of reproducing human brain in electric circuits. In the first AI Boom Rosenblatt devised Perceptron (Neural Network: NN) that electrically represented nervous networks. The second AI boom started when Fukushima proposed Neocognitron in 1979, which imitated the visual nervous system of cats in the electronic circuit board. LeCun embedded it in computer program as a convolutional NN, LeNet in 1989. In 2006, Hinton proposed Autoencoder, a nonlinearization of Principal Component Analysis, and in 2012, AKrizhevsky multilayered it as AlexNet, the dramatic improvement of which in classification accuracy led to the third AI boom. Although there are not many applications in the organ transplantation yet, iBox which uses Cox regression has been proposed by Loupy to predict the prognosis of renal transplant patients. We proposed a model for predicting the difficulty of laparoscopic donor nephrectomy in living donor kidney transplantation, and a model for optimizing the trough concentration of immunosuppressive drug in renal transplant patients. AI in the field of organ transplantation would enhance the safety of surgery and improve the long-term prognosis of transplanted kidneys.</p>