Author: Izabella L. Barreto, Benjamin Taylor Heggie, Stephanie M. Leon 👨🔬
Affiliation: University of Florida College of Medicine, University of Florida 🌍
Purpose: Filament deposition modeling (FDM) 3D printers may utilize proprietary calibration methods inadequate for medical imaging. Proper techniques are necessary to achieve imaging uniformity standards that are reproducible for manufactured filaments.
Methods: Four FDM printers were evaluated using polylactic acid (PLA) spools by printing five test rods. Each rod consisted of five layers with differing settings; infill pattern (rectilinear, grid, triangles, cubic, & gyroid), infill speed (25, 75, 125, 175, & 225 mm/s), flow percentage (90%, 95%, 100%, 105%, and 110%) and infill density (10-50% and 60-100%). These rods were placed in a water equivalent CT phantom and scanned with dual-energy CT. An ROI was used to measure the CT number (HU) at 70 keV, effective atomic number (Zeff) and electron density (ρe-) of each rod. Infill pattern and speed settings were optimized based on water background measurements, where standard deviation was used as a measure of uniformity. Flow percentages were plotted linearly as a function of CT number to determine flow settings that will produce identical infill density measurements (HU, Zeff, and ρe-) between printers. Lastly, a one-way ANOVA test was used to determine whether PLA spool batches produce rods with statistically significant HU measurements.
Results: An increase in standard deviation (HU, Zeff, and ρe-) was observed for higher infill speeds and intersecting infill patterns. One of four printers used a Bowden extruder which showed higher standard deviation measurements compared to the remaining three direct drive printers. Identical infill density measurements were observed between printers using calibrated flow settings. Spool batches showed statistically significant mass differences with respect to HU measurements (p-value < 0.5).
Conclusion: The optimization techniques introduced in this work proved successful in improving 3D print uniformity for medical imaging applications; however, the process is printer-specific, and CT number calibration must be repeated for each spool batch.