Photon-counting detector CT with an ultra-low-dose contrast media to diagnose a renal pseudoaneurysm: A case report.

A 75-year-old male, weighing 71 kg, was admitted to our institution with anemia related to a subcapsular hematoma after accidental extraction of a nephrostomy catheter. While the patient exhibited the progression of chronic kidney disease, he was not yet on dialysis. His serum creatinine level increased to 6.8 mg/dL, with an estimated glomerular filtration rate of 7.4 mL/min/1.73 m2. Radiologists planned contrast-enhanced photon-counting detector CT (PCD-CT) with an ultra-low-dose contrast media to mitigate the impact on renal function. The contrast media dosage was set at 7.4 gI, which was 82.6% lower that used in the standard protocol for a male weighing 71 kg. Non-contrast-enhanced PCD-CT identified a low-density nodular area within the renal subcapsular hematoma. Contrast-enhanced PCD-CT revealed contrast enhancement in both the early and late phases corresponding to the nodular area. On virtual monoenergetic images, the renal pseudoaneurysm was most clearly delineated at 40 keV. Following the diagnosis of a pseudoaneurysm, transcatheter arterial coil embolization was performed. No subsequent progression of anemia or the deterioration of renal function was observed, showcasing the potential of ultra-low-dose contrast-enhanced PCD-CT for the detection of small vascular abnormalities while minimizing adverse effects on renal function.

Utility of unsupervised deep learning using a 3D variational autoencoder in detecting inner ear abnormalities on CT images.

BACKGROUND AND PURPOSE: To examine the diagnostic performance of unsupervised deep learning using a 3D variational autoencoder (VAE) for detecting and localizing inner ear abnormalities on CT images. METHOD: Temporal bone CT images of 6663 normal inner ears and 113 malformations were analyzed. For unsupervised learning, 113 images from both the malformation and normal cases were used as test data. Other normal images were used for training. A colored difference map representing differences between input and output images of 3D-VAE and the ratio of colored to total pixel numbers were calculated. Supervised learning was also investigated using a 3D deep residual network and all data were classified as normal or malformation using 10-fold cross-validation. RESULTS: For unsupervised learning, a significant difference in the colored pixel ratio was seen between normal (0.00021 ± 0.00022) and malformation (0.00148 ± 0.00087) cases with an area under the curve of 0.99 (specificity = 92.0%, sensitivity = 99.1%). Upon evaluation of the difference map, abnormal regions were partially and not highlighted in 7% and 0% of the malformations, respectively. For supervised learning, which achieved 99.8% specificity and 90.3% sensitivity, a part of and no abnormal regions were highlighted on interpretation maps in 34% and 8% of the malformations, respectively. Abnormal regions were not highlighted in 4 malformation cases diagnosed as malformations and were highlighted in 6 cases misdiagnosed as normal. CONCLUSIONS: Unsupervised deep learning of 3D-VAE precisely detected inner ear malformations and localized abnormal regions. Supervised learning did not identify whole abnormal regions frequently and basis for diagnosis was sometimes unclear.