Two identical black and white pictures of murky shapes sit side-by-side on a computer screen. On the left side, Ismail Baris Turkbey, MD, a radiologist with 15 years of experience, has outlined an area where the fuzzy shapes represent what he believes is a creeping, growing prostate cancer. On the other side of the screen, an artificial intelligence (AI) computer program has done the same--and the results are nearly identical. The black and white image is an MRI scan from someone with prostate cancer, and the AI program has analyzed thousands of them. "The [AI] model finds the prostate and outlines cancer-suspicious areas without any human supervision," Turkbey explains.

Computational medical imaging approaches can improve the analytical accuracy of interpretation in cancer identification and characterization, allowing for earlier disease detection and a better understanding of physiology and pathology. Machine Learning (ML) models have revolutionized many activities of medical imaging applications, such as novel imaging techniques, segmentation, registration, and synthesis, by analyzing large amounts of quantitative imaging biomarkers. While ML models outperform traditional methods on these tasks, they are still largely tacit in terms of explaining the data under consideration.This has reduced the interpretability of ML models, which is one of the major obstacles to ML-based pathology identification and generalized single- or multi-modal and multi-scale interpretation in medical imaging. Detailed examples of model behaviors are expected in current clinical practices to promote reliability and improve clinical decision making. Furthermore, the primary challenge for designing explainable models is to provide rationales while retaining high learning results as one of the most exciting areas of medical imaging science.We hope to attract novel, high-quality research and survey papers that represent the most recent developments in ML models in innovative medical imaging (MRI, CT, PET, SPECT, Ultrasound, histology and others) modalities or multi-modalities, by investigating novel methodologies either by interpreting algorithm components or by ex...

Despite technological and medical advances, the detection, interpretation, and treatment of cancer based on imaging data continue to pose significant challenges. These include high inter-observer variability, difficulty of small-sized lesion detection, nodule interpretation and malignancy determination, inter- and intra-tumour heterogeneity, class imbalance, segmentation inaccuracies, and treatment effect uncertainty. The recent advancements in Generative Adversarial Networks (GANs) in computer vision as well as in medical imaging may provide a basis for enhanced capabilities in cancer detection and analysis. In this review, we assess the potential of GANs to address a number of key challenges of cancer imaging, including data scarcity and imbalance, domain and dataset shifts, data access and privacy, data annotation and quantification, as well as cancer detection, tumour profiling and treatment planning. We provide a critical appraisal of the existing literature of GANs applied to cancer imagery, together with suggestions on future research directions to address these challenges.

Despite technological and medical advances, the detection, interpretation, and treatment of cancer based on imaging data continue to pose significant challenges. These include high inter-observer variability, difficulty of small-sized lesion detection, nodule interpretation and malignancy determination, inter- and intra-tumour heterogeneity, class imbalance, segmentation inaccuracies, and treatment effect uncertainty. The recent advancements in Generative Adversarial Networks (GANs) in computer vision as well as in medical imaging may provide a basis for enhanced capabilities in cancer detection and analysis. In this review, we assess the potential of GANs to address a number of key challenges of cancer imaging, including data scarcity and imbalance, domain and dataset shifts, data access and privacy, data annotation and quantification, as well as cancer detection, tumour profiling and treatment planning. We provide a critical appraisal of the existing literature of GANs applied to cancer imagery, together with suggestions on future research directions to address these challenges. We analyse and discuss 163 papers that apply adversarial training techniques in the context of cancer imaging and elaborate their methodologies, advantages and limitations. With this work, we strive to bridge the gap between the needs of the clinical cancer imaging community and the current and prospective research on GANs in the artificial intelligence community.

The blog post below is adapted from a longer article in VA Research Currents. When Facebook suggests a new friend for you, or Gmail shows you ads based on your email content, or Alexa or Siri understands your verbal command to do some chore in the house, that's artificial intelligence at work. Or, for a more dramatic example, think of driverless cars that read traffic and make lightning-fast decisions to stay on course and avoid accidents. Basically, artificial intelligence (AI) means using computers to simulate human thinking. Computers will never be able to fully replicate the human mind in all its amazing nuance, speed, and complexity--at least most people hope not!--but scientists have made remarkable strides in teaching computers to handle tasks such as finding patterns in data, analyzing and weighing risk factors, choosing the best option from among many choices, predicting future events based on past ones, and solving problems.