Diagnostic Accuracy and Failure Mode Analysis of a Deep Learning Algorithm for the Detection of Cervical Spine Fractures.

BACKGROUND AND PURPOSE: Artificial intelligence decision support systems are a rapidly growing class of tools to help manage ever-increasing imaging volumes. The aim of this study was to evaluate the performance of an artificial intelligence decision support system, Aidoc, for the detection of cervical spinal fractures on noncontrast cervical spine CT scans and to conduct a failure mode analysis to identify areas of poor performance. MATERIALS AND METHODS: This retrospective study included 1904 emergent noncontrast cervical spine CT scans of adult patients (60 [SD, 22] years, 50.3% men). The presence of cervical spinal fracture was determined by Aidoc and an attending neuroradiologist; discrepancies were independently adjudicated. Algorithm performance was assessed by calculation of the diagnostic accuracy, and a failure mode analysis was performed. RESULTS: Aidoc and the neuroradiologist's interpretation were concordant in 91.5% of cases. Aidoc correctly identified 67 of 122 fractures (54.9%) with 106 false-positive flagged studies. Diagnostic performance was calculated as the following: sensitivity, 54.9% (95% CI, 45.7%-63.9%); specificity, 94.1% (95% CI, 92.9%-95.1%); positive predictive value, 38.7% (95% CI, 33.1%-44.7%); and negative predictive value, 96.8% (95% CI, 96.2%-97.4%). Worsened performance was observed in the detection of chronic fractures; differences in diagnostic performance were not altered by study indication or patient characteristics. CONCLUSIONS: We observed poor diagnostic accuracy of an artificial intelligence decision support system for the detection of cervical spine fractures. Many similar algorithms have also received little or no external validation, and this study raises concerns about their generalizability, utility, and rapid pace of deployment. Further rigorous evaluations are needed to understand the weaknesses of these tools before widespread implementation.

Contrast-Enhanced 3D-FLAIR Imaging of the Optic Nerve and Optic Nerve Head: Novel Neuroimaging Findings of Idiopathic Intracranial Hypertension.

BACKGROUND AND PURPOSE: The sensitivity of contrast-enhanced 3D-FLAIR has not been assessed in patients with idiopathic intracranial hypertension. The purpose of this study was to evaluate whether hyperintensity of the optic nerve/optic nerve head on contrast-enhanced 3D-FLAIR imaging is associated with papilledema in patients with idiopathic intracranial hypertension. MATERIALS AND METHODS: A retrospective review was conducted from 2012 to 2015 of patients with clinically diagnosed idiopathic intracranial hypertension and age- and sex-matched controls who had MR imaging with contrast-enhanced 3D-FLAIR. Two neuroradiologists graded each optic nerve/optic nerve head on a scale of 0-3. This grade was then correlated with the Frisén Scale, an ophthalmologic scale used for grading papilledema from 0 (normal) to 5 (severe edema). To estimate the correlation between the MR imaging and Frisén scores, we calculated the Kendall τ coefficient. RESULTS: Forty-six patients (3 men, 43 women) with idiopathic intracranial hypertension and 61 controls (5 men, 56 women) with normal findings on MR imaging were included in this study. For both eyes, there was moderate correlation between the 2 scales (right eye: τ = 0.47; 95% CI, 0.31-0.57; left eye: τ = 0.38; 95% CI, 0.24-0.49). Interreader reliability for MR imaging scores showed high interreader reliability (right eye: κ = 0.76; 95% CI, 0.55-0.88; left eye: κ = 0.87; 95% CI, 0.78-0.94). Contrast-enhanced 3D-FLAIR imaging correlates with the Frisén Scale for moderate-to-severe papilledema and less so for mild papilledema. CONCLUSIONS: Hyperintensity of the optic nerve/optic nerve head on contrast-enhanced 3D-FLAIR is sensitive for the detection of papilledema in patients with idiopathic intracranial hypertension, which may be useful when prompt diagnosis is crucial.