Utility of Zero-Echo time sequence as an adjunct for MR evaluation of degenerative disease in the cervical spine.

OBJECTIVE: Determine the utility of ZTE as an adjunct to routine MR for assessing degenerative disease in the cervical spine. METHODS: Retrospective study on 42 patients with cervical MR performed with ZTE from 1/1/2022 to 4/30/22. Fellowship trained radiologists evaluated each cervical disc level for neural foraminal (NF) narrowing, canal stenosis (CS), facet arthritis (FA), and presence of ossification of the posterior longitudinal ligament (OPLL). When NF narrowing and CS were present, the relative contributions of bone and soft disc were determined and a confidence level for doing so was assigned. Comparisons were made between assessments on routine MR without and with ZTE. RESULTS: With ZTE added, bone contribution as a cause of NF narrowing increased in 47% (n = 110) of neural foramina and decreased in 12% (n = 29) (p = < 0.001). Bone contribution as a cause of CS increased in 25% (n = 33) of disc levels and decreased in 10% (n = 13) (p = 0.013). Confidence increased in identifying the cause of NF narrowing (p = < 0.001)) and CS (p = 0.009) with ZTE. The cause of NF narrowing (p = 0.007) and CS (p = 0.041) changed more frequently after ZTE was added when initial confidence in making the determination was low. There was no change in detection of FA or presence of OPLL with ZTE. CONCLUSION: Addition of ZTE to a routine cervical spine MR changes the assessment of the degree of bone involvement in degenerative cervical spine pathology.

Detecting pediatric wrist fractures using deep-learning-based object detection.

BACKGROUND: Missed fractures are the leading cause of diagnostic error in the emergency department, and fractures of pediatric bones, particularly subtle wrist fractures, can be misidentified because of their varying characteristics and responses to injury. OBJECTIVE: This study evaluated the utility of an object detection deep learning framework for classifying pediatric wrist fractures as positive or negative for fracture, including subtle buckle fractures of the distal radius, and evaluated the performance of this algorithm as augmentation to trainee radiograph interpretation. MATERIALS AND METHODS: We obtained 395 posteroanterior wrist radiographs from unique pediatric patients (65% positive for fracture, 30% positive for distal radial buckle fracture) and divided them into train (n = 229), tune (n = 41) and test (n = 125) sets. We trained a Faster R-CNN (region-based convolutional neural network) deep learning object-detection model. Two pediatric and two radiology residents evaluated radiographs initially without the artificial intelligence (AI) assistance, and then subsequently with access to the bounding box generated by the Faster R-CNN model. RESULTS: The Faster R-CNN model demonstrated an area under the curve (AUC) of 0.92 (95% confidence interval [CI] 0.87-0.97), accuracy of 88% (n = 110/125; 95% CI 81-93%), sensitivity of 88% (n = 70/80; 95% CI 78-94%) and specificity of 89% (n = 40/45, 95% CI 76-96%) in identifying any fracture and identified 90% of buckle fractures (n = 35/39, 95% CI 76-97%). Access to Faster R-CNN model predictions significantly improved average resident accuracy from 80 to 93% in detecting any fracture (P < 0.001) and from 69 to 92% in detecting buckle fracture (P < 0.001). After accessing AI predictions, residents significantly outperformed AI in cases of disagreement (73% resident correct vs. 27% AI, P = 0.002). CONCLUSION: An object-detection-based deep learning approach trained with only a few hundred examples identified radiographs containing pediatric wrist fractures with high accuracy. Access to model predictions significantly improved resident accuracy in diagnosing these fractures.