gms | German Medical Science

Introduction: Stenosis of the trachea represents a frequent complication after tracheotomy or long time artificial respiration. Symptomatic stenosis needs surgical resection and reconstruction. Nowadays resection is planned by slice-by-slice inspection of 2-D radiological data whereas diagnostic endoscopy gives additional information. 3-D reconstructions generated by radiologists are sometimes rather simple and frequently not sufficient. This study investigates different possibilities of segmentation and visualization-methods of CT-data and their benefit in trachea resection planning.

Methods: We designed a software (NeckVision) to segment and visualize CT-data. The integrated semiautomatic image analysis algorithms based on the ILab-library (MeVis, Bremen, Germany) were summarized in an user-friendly interface with fixed segmentation-order. Manual segmentation was made possible for difficult cases. Additionally we designed a surgical planning software (NeckSurgeryPlanner) to improve individual planning. Segmentation was done by surgeons themselves and additionally by professional image processors. The trachea and adjacent structures were segmented and visualized in 3-D. Diameter, length and position of stenosis were measured. The results were compared to intraoperative and histopathologic findings. 8 patients were evaluated, where 3 patients were undergoing resection of the stenosis.

Results: The average time requirement was 35 min., indeed professional image processors needed less time for post-processing. The colours were standardised for rapid recognition. Different levels of transparency were used to prevent cover up structures in the back. The planning software allowed individual movement, choice of structures and measurements as needed. The results of surgical evaluation of trachea stenosis based on 2-D-CT-data or diagnostic tracheoscopy differed from histopathologic findings, where stenosis was estimated to long. Measurements based on 3-D-reconstruction corresponded exactly to histopathologic findings. A secure classification of possibility and extent of trachea resection was enabled. The advantage was made clear for individual and detailed surgical planning and common discussions with other medical specialists.

Conclusion: This kind of 3-D visualisation was characterized by its compressed content of information. The possibility of image-post-processing by surgeons itself was proved. Surgical planning was simplified and patients’ safety during surgical intervention was enlarged. Besides the received results can be used for patients` information, students` teaching and interdisciplinary discussions.