Feasibility of low dose chest CT for virtual bronchoscopy navigation in a porcine model

BACKGROUND: Virtual bronchoscopy navigation (VBN) is widely used for assistance in the histological examination of lung nodules. However, little is known about the optimal CT radiation dose for VBN. Therefore, we performed an animal study to evaluate the feasibility of low dose CT (LDCT) for VBN. METHODS: Ten pigs underwent standard dose CT (as a reference) and four different LDCT protocols: LDCT 1, 120 kVp, 15 mAs; LDCT 2, 120 kVp, 8 mAs; LDCT 3, 100 kVp, 7 mAs; and LDCT 4, 100 kVp, 4 mAs. As targets for the VBN, 10 mm virtual lesions were created in the central and peripheral bronchi. To assess the performance of the VBN, the navigation direction (direction of reconstructed pathways to the target) and the number of branching’s (the number of peripheral bronchi to the target) were evaluated. RESULTS: The mean effective doses significantly differed across the four LDCTs (P < 0.001). For both central and peripheral virtual targets, there were significant differences in the accuracy of the navigation direction and the number of branching’s of the VBNs across the four LDCTs (P < 0.001 for all). Regarding the accuracy of the navigation direction and the number of branching’s, the areas under the curves of the ROCs were 0.9352 and 0.9324, respectively, for central virtual targets, and 0.8696 and 0.8783, respectively, for peripheral virtual targets. Youden’s index indicated that the optimal effective CT scan dose for both central and peripheral virtual targets was 0.238 mSv. CONCLUSIONS: LDCT is feasible for VBN.