Abstract:

The Systems Management and Production Center at the University of Alabama in Huntsville has been working closely with the College of Nursing in developing 3D printed models of the body for teaching human anatomy. This paper presents the development of three 3D printed models of a full size human skull from a CT (computed tomography) scan with no IV (intravenous) contrast agent and a slice thickness of 3.75mm. Model1 of the full skull could not be printed full size. Model2 was two halves of the skull that resulted from the segmentation in the transverse, or axial plane. The inside view of the transverse plane looking down (inside the 3D printed model) clearly showed the depressions in the floor of the cranial base of the anterior cranial fossa, the middle cranial fossa and the posterior cranial fossa. The skull cap (top half) of Model2 was printed and connected with magnets to the bottom half resulting in a full size model of the skull (Model1). Model3 was a half skull of Model1 that resulted from the segmentation in the longitudinal (sagittal) plane. The inside view of the longitudinal plane (inside the 3D printed model) clearly showed the many cavities and bones below the cranial base. 3DSlicer was used to segment the model from the CT scan using a threshold level of 114. The scissors option in 3DSlicer Segment Editor was used to remove unwanted regions such as the “starburst” resulting from dental fillings. 3DBuilder software was used to visualize the STL (stereolithography) files of the 3D printed models and helped in identifying extraneous objects that needed to be removed before 3D printing. A CT scanner collects x-ray data at many different angles around the patient. The x-ray data are then reconstructed into slice thicknesses depending on the use of the data.

The steps in 3D printing the model from a CT scan in DICOM (Digital Imaging and Communications in Medicine) format were:
1) load CT scan in DICOM format into 3DSlicer,
2) use 3DSlicer Segmentation Editor to anonymize the DICOM data set (remove any identification to specific patient) and to segment the skull, remove noise and obtrusions and convert the segmented data set to an STL format,
3) use 3DBuilder to view the 3D computer simulated model (model can be rotated and enlarged),
4) use PrusaSlicer software to convert STL to G-code, and
5) print the STL file on the Prusa i3MK1 3D printer

The skull Model1 was easily and quickly segmented from the CT scan with no IV contrast agent. The model had very little noise, except for the “starburst” extrusions resulting which had to be carefully removed using 3DSlicer. Consequently, very little cleaning was required. 3DSlicer smoothing was unnecessary to improve model detail. Segmenting the skull Models2 and 3 was easily and quickly done, requiring no cleaning or smoothing. Fifteen of the twenty-one foramina in the skull were identifiable in Model2 which included all foramina greater than four millimeters. The models are being introduced into the Simulation Based Learning Experiences in the College of Nursing for training nursing students in human anatomy. The 3D printed models gave students the ability to touch and feel full size 3D models of the skull. The 3DBuilder software allowed nursing students to view the skull models in 3D and to rotate and enlarge the models.

Keywords:

3D printed model, simulation, human skull, CT scan, 3DSlicer, nursing student training.