Computed tomography in two recumbencies aides in the identification of pulmonary bullae in dogs with spontaneous pneumothorax.
Rachel DicksonValery F ScharfNathan C NelsonNicholas PetrovitchErin K KeenihanKyle G MathewsPublished in: Veterinary radiology & ultrasound : the official journal of the American College of Veterinary Radiology and the International Veterinary Radiology Association (2020)
Spontaneous pneumothorax presents a unique diagnostic and therapeutic challenge in veterinary medicine, specifically with regard to accurate identification of bullous lesions. Positioning of dogs with spontaneous pneumothorax during CT has not previously been evaluated. This retrospective, diagnostic accuracy study was performed to evaluate the sensitivity, positive predictive value (PPV), and interobserver variability for detection of pulmonary bullae with dogs positioned in multiple recumbencies. Dogs underwent CT in sternal and dorsal recumbency followed by thoracic exploration via median sternotomy. Three American College of Veterinary Radiology-certified veterinary radiologists blinded to surgical findings reviewed dorsal and sternal images simultaneously. Severity of pneumothorax, degree of atelectasis, lesion location and size, and view in which lesions were most confidently identified were compared to surgical and histologic findings. Sensitivities and PPVs for bulla detection ranged from 57.7% to 69.2% and 62.1% to 78.9%, respectively. For two of the 3 radiologists, the location of bullae in the thorax was significantly associated with the recumbency in which the lesion was best identified. Degree of atelectasis was found to be associated with the ability to identify lesions (P ≤ .02). The interobserver variability for identification was good (κ = 0.670). The sensitivity of CT when performed in both sternal and dorsal recumbency is similar to that previously reported. Because the distribution of bullae is unknown prior to advanced imaging and bulla location affects which recumbency is most useful for identification, acquisition of CT images in both sternal and dorsal recumbency may improve detection of bullous lesions and aid surgical planning.
Keyphrases
- computed tomography
- spinal cord
- dual energy
- image quality
- contrast enhanced
- neuropathic pain
- positron emission tomography
- artificial intelligence
- bioinformatics analysis
- loop mediated isothermal amplification
- deep learning
- high resolution
- pulmonary hypertension
- real time pcr
- spinal cord injury
- label free
- optical coherence tomography
- heart failure
- machine learning
- convolutional neural network
- magnetic resonance
- clinical trial
- study protocol
- photodynamic therapy
- pet ct
- aortic valve replacement
- double blind
- sensitive detection