Keywords:
Lung, CT, Computer Applications-3D, Image registration
Authors:
I. Dudurych1, A. Garcia-Uceda Juarez2, R. Vliegenthart1, M. de Bruijne2; 1Groningen/NL, 2Rotterdam/NL
DOI:
10.26044/ecr2021/C-10975
Methods and materials
The initial airway segmentation 3D-Unet [2] was trained on datasets from the Danish Lung Cancer Screening Trial [3] and from Erasmus MC, Rotterdam, NL. The resulting incomplete airway segmentations of 13 ImaLife scans [4] were imported into Slicer 3D 4.0 [5] and viewed with a window setting of Width: 800 and Level: -625.
The corrections are performed using the segment editor tool in Slicer3D. First, the incomplete airway segmentation is obtained from the segmentation volume. An airway is followed with the cursor in the 3D view of the segmentation. In this manner it is possible to quickly identify and complete airways as they are identified on axial, sagittal and coronal views simultaneously, with the results instantly visible on the 3D view.
Next, in a separate colour, the airways are completed by using the paint tool.
The completed segmentation is exported as a set of DICOM slices and can be used for the training of AI tools.
We utilised an airway analysis tool [6] to automatically calculate the number of airway branches per mathematical generation, the total length of the branches per generation and the lumen diameter per generation. Then, we compared the initial incomplete segmentation to the manually corrected segmentation. Comparison and calculation of median and interquartile ranges was performed using Python (Python Software Foundation, https://www.python.org/).