Phase 1 - Phantom development:
Commercially available imaging anthropomorphic phantoms, which represented adult head, pelvis and heels (Figure 1 ) were scanned using CT (Toshiba 16 Toshiba Medical Systems, Tokyo, Japan)[3].
After that the DICOM format CT image data was converted to NRRD file format using Slicer Software (https://www.slicer.org/) , then the NRRD file was converted to STL format using embodi3D.com. Prior to printing, STL data were processed using MeshLab (version 1.2.3-64; http://meshlab.sourceforge.net) to correct 3D surface anomalies. Four 3D printing sessions occurred – head, pelvis, left heel and right heel.
Once the four components had been printed, they were linked together to represent the human body (Figure 3). To join components together, a frame of 25 mm x 25 mm x 1.5 mm box aluminium at 1.75 meters long, was used. The box of aluminium was fixed by plastic connectors to maintain rigidity. For knee and elbow ‘positions’, flexible plastic connectors were used to allow the structure to bend in order to simulate potential human limb motion. Finally, urethane foam was used to fill the 4 printed components to give them adequate rigidity to make them strong enough to withstand weights that will be applied to them from above.
In order to simulate a range of human weights, XSensor (http://www.sumed.co.uk) [4] interface pressure data from a published study with 27 volunteers [5] was used to determine the weights to be applied to the 3D phantom. Volunteer characteristics were: average height=164.63cm (SD=7.64), BMI=28.18 (SD=6.75), weight 77.04kg (SD=22.18); range 50-148kg (SD=7.39); 24 females and 3 males. The 27 volunteers were grouped into 5 weight categories: maximum (148kg), third quartile (84kg), mean (76kg), first quartile (64kg), and minimum (50kg). To represent these 5 categories, average weights for each were calculated from XSensor data to determine the amount of sand to be added to the heels, pelvis and head phantoms (Table 1, final column – ‘weight for 3D phantom’).
Phase 2 - phantom validation and proposing the method for analysing and presenting data.
For the jeopardy areas, Peak Pressure Values (PPV) and pressure profiles were compared between the 3D phantom and that of the human study.
3D phantom interface pressure data and the previously published human interface pressure data from head, sacrum and heels were acquired using the same memory mattress. Phantom validation was achieved by comparing phantom interface pressure distribution characteristics, using the 5 phantom weights, against 27 sets of human interface pressure distribution data from a published study ([1]). Pressure distribution was assessed using Xsensor Technology.
To minimise measurement error in humans, data is acquired for 20 minutes, following a settling time of 6 minutes (Angmorterh et al., 2018 )[5]. However, the 3D phantom does not contain soft tissue materials and so it only required 3 minutes settling time to achieve stabilised pressure readings, and 15 minutes for data collection. Weight (sand), used for the phantom are indicated in Table 1.