Faculty Bibliography

Machine learning is becoming increasingly important in both research and clinical applications in radiology due to recent technological developments, particularly in deep learning. As these technologies are translated toward clinical practice, there is a need for radiologists and radiology trainees to understand the basic principles behind them. This primer provides an accessible introduction to the vocabulary and concepts that are central to machine learning and relevant to the radiologist.

BACKGROUND AND PURPOSE:No MR imaging measurement criteria are available for the diagnosis of optic nerve atrophy. We determined a threshold optic nerve area on MR imaging that predicts a clinical diagnosis of optic nerve atrophy and assessed the relationship between optic nerve area and retinal nerve fiber layer thickness measured by optical coherence tomography, an ancillary test used to evaluate optic nerve disorders. MATERIALS AND METHODS:We evaluated 26 patients with suspected optic nerve atrophy (8 with unilateral, 13 with bilateral and 5 with suspected but not demonstrable optic nerve atrophy) who had both orbital MR imaging and optical coherence tomography examinations. Forty-five patients without optic nerve atrophy served as controls. Coronal inversion recovery images were used to measure optic nerve area on MR imaging. Retinal nerve fiber layer thickness was determined by optical coherence tomography. Individual eyes were treated separately; however, bootstrapping was used to account for clustering when appropriate. Correlation coefficients were used to evaluate relationships; receiver operating characteristic curves, to investigate predictive accuracy. RESULTS:< .001). CONCLUSIONS:has moderately high sensitivity and specificity for predicting optic nerve atrophy, making it a potential diagnostic tool for radiologists.