CT MEASUREMENT OF APPENDIX DIAMETER USING A 3D PHANTOM: IMPACT OF IMAGE QUALITY ON RADIOLOGIST ACCURACY

Abstract

Background: Accurate measurement of appendix diameter on CT is important for diagnostic evaluation and for selecting the appropriate closure technique during laparoscopic appendectomy. Factors influencing measurement accuracy remain poorly understood, particularly under varying image quality conditions.

Purpose: To investigate how CT acquisition parameters affect radiologist and automated measurements of appendix diameter using a 3D printed abdominal phantom with a geometrically defined appendix insert.

Materials and Methods: A custom abdominal phantom containing a 7.5 mm appendix analogue was scanned using eleven CT protocols that varied tube voltage, tube current, slice thickness, and reconstruction kernel. Image quality was characterised using image noise and the spatial frequency at 10 percent of the modulation transfer function (f₁₀). Appendix width was measured using an automated full width at half maximum (FWHM) method in ImageJ and independently by two radiologists using a standard DICOM viewer. Pearson correlations and multiple linear regression were used to identify predictors of radiologist deviation (Δw) from the true width.

Results: Automated FWHM measurements closely matched the nominal diameter across all protocols, with deviations below 3 percent and no significant dependence on noise or resolution. Radiologist measurements showed larger variability and significant negative correlations with noise (r = −0.609, p = 0.047) and f₁₀ (r = −0.649, p = 0.031). Regression analysis identified f₁₀ (p < 0.001), noise (p = 0.001), and slice thickness (p = 0.006) as independent predictors of radiologist deviation. CTDIvol, tube voltage, tube current, and reconstruction kernel did not contribute additional explanatory value.

Conclusion: Radiologist measurement accuracy for small soft-tissue structures is influenced by image quality characteristics, particularly spatial resolution, noise, and slice thickness. Automated measurements remained stable across all conditions. These findings support careful consideration of protocol-specific image appearance when appendix diameter is used for clinical decision making.