Artificial intelligence in nuclear medicine: Historical milestones and core principles

Czech version

Authors: M. Bejtic ^1,2; V. Kamírová ¹; O. Lang ¹; M. Lang ¹; M. Darsa ²
Authors‘ workplace: Prague Medical Care Department, s. r. o., Praha ¹; Fakulta biomedicínského inženýrství, České vysoké učení technické v Praze, Kladno, ČR ²
Published in: NuklMed 2025;14:60-64
Category: Review Article

Overview

Artificial intelligence (AI) represents a rapidly evolving technology with a significant impact on modern medicine, including nuclear medicine, where it offers new possibilities for improving diagnostics, quantitative assessment, and prognostic evaluation. This article provides a comprehensive overview of the historical development of AI –⁠ from its theoretical foundations in the 1950s, through the emergence of machine learning and deep neural networks, to contemporary architectures based on convolutional networks, generative adversarial models, and large language models.

The text highlights the first practical and clinically relevant applications of AI in nuclear medicine, particularly in cardiology, neurology, and oncology. These include automated interpretation of myocardial perfusion SPECT, PET brain analysis in neurodegenerative diseases, and quantitative evaluation of bone metastases. Special attention is given to models that have undergone independent validation and demonstrated real clinical benefit, such as automated bone scan index calculation and algorithms for detecting Alzheimer’s disease from FDG PET.

In conclusion, the article summarizes the key technological developments that have shaped today’s AI capabilities in nuclear medicine and illustrates how the progression from simple algorithms to advanced deep and generative models has influenced current clinical practice.

Keywords:

artificial intelligence; nuclear medicine; history; machine learning; deep learning; convolutional network

Sources

McCarthy J, Minsky ML, Rochester N, et al. A proposal for the Dartmouth Summer Research Project on Artificial Intelligence (1955). AI Magazine. 2006;27 : 12–14
Nensa F, Demircioglu A, Rischpler C. Artificial intelligence in nuclear medicine. J Nucl Med. 2019;60(S2):29S–37S
Visvikis D, Lambin P, Beuschau Mauridsen K, et al. Application of artificial intelligence in nuclear medicine and molecular imaging: a review of current status and future perspectives for clinical translation. Eur J Nucl Med Mol Imaging 2022;49 : 4452–4463
Rosenblatt F. The perceptron: a probabilistic model for information storage and organization in the brain. Psychol Rev 1958;65 : 386–408
Krizhevsky A, Sutskever I, Hinton GE. ImageNet classification with deep convolutional neural networks. Adv Neural Inf Process Syst 2012;25 : 1097–1105
Nakajima K, Kudo T, Nakata T, et al. Diagnostic accuracy of an artificial neural network compared with statistical quantitation of myocardial perfusion images: a Japanese multicenter study. Eur J Nucl Med. 2017;44 : 2280–2289
Betancur J, Commandeur F, Motlagh M, et al. Deep learning for prediction of obstructive disease from fast myocardial perfusion SPECT: a multicenter study. JACC Cardiovasc Imaging. 2018;11 : 1654–1663
Ding Y, Sohn JH, Kawczynski MG, et al. A deep learning model to predict a diagnosis of Alzheimer disease by using 18F-FDG PET of the brain. Radiology 2019;290 : 456–464
Armstrong AJ, et al. Phase 3 assessment of the automated bone scan index (aBSI) as a prognostic imaging biomarker of overall survival in men with metastatic castration-resistant prostate cancer: a secondary analysis of a randomized clinical trial. JAMA Oncol 2018;4 : 944–951
Park C, Na KJ, Choi H, et al. Tumor immune profiles estimated by FDG PET with deep learning correlate with immunotherapy response in lung adenocarcinoma. Theranostics 2020;10 : 10838–10848
Mota JM, Armstrong AJ, Larson SM, et al. Measuring the unmeasurable: automated bone scan index as a quantitative endpoint in prostate cancer clinical trials. Prostate Cancer Prostatic Dis 2019;22 : 522–530
Hwang D, Kang SK, Kim KY, et al. Generation of PET Attenuation Map for Whole-Body Time-of-Flight 18F-FDG PET/MRI Using a Deep Neural Network Trained with Simultaneously Reconstructed Activity and Attenuation Maps. J Nucl Med 2019;60 : 1183-1189

Labels

Nuclear medicine Radiodiagnostics Radiotherapy

Parameters of kits for the preparation of 99mTc-labeled particles and colloids authorised in the European Union and the European Economic Area

MUDr. Jan Urbánek, CSc. †

doc. Ing. František Melichar, DrSc. †

MUDr. Lubomír Mrhač †

61. Dny nukleární medicíny, Liberec

Editorial

Article was published in

Nuclear Medicine

2025 Issue 4