PURPOSE: Virtual bronchoscopy (VB) allows for more peripheral assessment of the airways (up to the 9th generation bronchioles) than fiberoptic bronchoscopy (FOB). Although VB which is non invasive has been shown to correlate with FOB, airway injury scoring based on VB alone has not yet been proposed. We explored the use of VB for scoring of airway injury severity in swine.
METHODS: Using a Toshiba 64-slice scanner, swine chest CT scans were acquired before and after 1) smoke inhalation; 2) chlorine inhalation; or 3) right-sided pulmonary contusion. Image reconstruction for VB was performed using commercial software. In our VB lung injury scoring system, the carina, left secondary carina, and right secondary carina on each image were scored. A 1 cm square at the carina and 5 mm squares at the secondary carinas were drawn and the number of ulcerations within each square were counted with 1–2 ulcerations equalling one point, 3–4 ulcerations equalling 2 points, 5 or more ulcerations equallling three points. Debris, edema, blistering, carbonaceous material/soot, hemorrhage, and rough mucosa also contribute one point each to the score. A total score of 1–3 points equals mild injury, 4–6 moderate injury and a score greater than 7 equals severe injury. Airway diameters were also measured. To assess interobserver variability, the images were graded by four different observers.
RESULTS: Twelve scans have been scored thus far. At baseline mean main carina score was 1, and the secondary carinas scores were zero. Post-injury mean score at the main carina was 5.6, at the secondary carinas 1.3 (left) and 0.83 (right), with interobserver variability of zero between two observers thus far.
CONCLUSION: Our VB lung injury scoring system provides a standardized method to assess lung injury.
CLINICAL IMPLICATIONS: In lung injury, particularly inhalation injury, there is a need for a standardized and accurate system for diagnosis and grading severity of injury. Our method provides a standardized method for grading airway injury. We are further investigating correlation with clinical outcomes.
DISCLOSURE: Dara Regn, No Financial Disclosure Information; No Product/Research Disclosure Information