Transbronchial needle aspiration(TBNA) is a common procedure for diagnosis of mediastinal and hilar lymphadenopathy. Currently, pulmonologists prepare for this procedure by examining CT image slices; then, they estimate the location of the lymph node or tumor during the bronchoscopy. The success rate for this approach is near 70%. Our study aims to increase this success rate. By utilizing electromagnetic tracking and a registration technique, corresponding virtual images that show the lymph node or mass behind the airway may be displayed simultaneously alongside the real bronchoscope images. This technique may increase the accuracy of TBNA.
A custom phantom was CT-scanned with fiducials on the surface. We also developed image guidance software that includes manual segmentation, point-based registration, tracking, 3D visualization, and virtual bronchoscopy components. Using this software, we manually segmented the mass and generated a centerline to the biopsy target. Before inserting the bronchoscope, we registered the CT space and the patient space by providing corresponding points in two spaces based on the fiducials. For bronochoscope tracking, we used the Aurora electromagnetic tracking device. Using a flexible bronchoscope, we inserted an electromagnetically tracked catheter into the phantom. Using the frame-grabber card, we could capture the video sequence from the bronchoscope.
The average root-mean-square (RMS) registration error was 2.5 mm. The phantom model showed good correlation between the virtual and video bronchoscopic images. The software displayed lymph nodes, great vessels, and adjacent structures by making the airways transparent. By providing real-time virtual display during the actual bronchoscopy, we expect to improve the accuracy of TBNA.
This study is the first step towards our final goal of providing real-time assistance for TBNA with virtual images. Next, we will develop a real-time registration algorithm to align virtual and video bronchoscope images. We will narrow the search space by using the electromagnetic tracking device, and estimate the virtual location that best matches the virtual and video images.
Virtual bronchoscopy with electromagnetic tracking may lead to an improved yield of TBNA.
Jae Choi, None.