A Simplex algorithm approach to determine the optimal keV and window setting for appendicitis visualization.

PURPOSE: Contrast-enhanced dual energy CT, with low-mono-energetic CT scanning (conventional 70 keV and low-mono energetic 40 keV) can improve visualization of acute inflammatory processes, but the high iodine attenuation often requires adjustment of window display settings to optimize interpretation. Here, we aim to determine optimal keV and window settings for the visualization of acute appendicitis from images derived from low-mono-energetic-CT.

METHODS

34 patients diagnosed with appendicitis, using dual energy CT, were retrospectively identified. An iterative four-dimensional Simplex optimization algorithm was used to determine optimal keV and windowing for assessment. 14 different keVs (40-70 range) and window parameters (center 5-290 HU; width 205-570 HU) were evaluated by 6 radiologists. Each reader individually graded the conspicuity of appendix wall enhancement using a 5-point scale and the overall image quality on a 3-point scale. Three-dimensional contour maps of the relationship between window settings and overall averages of reader grade at 40 keV were constructed.

RESULTS

Convergence to optimal windowing was achieved over ten parameter iterations following the initial evaluation of 4 different parameter sets. This amounted to a total of 2,856 reads (14 keVs and window settings x 6 readers x 34 cases). The highest average conspicuity grade (4.4 ± 0.0) was achieved at 40 keV with windows of 57-79 HU center and 357-418 HU width, with best background in this range at window 57 HU and center 418 HU.

CONCLUSIONS

Algorithmic optimization of dual energy contrast-enhanced CT scans may allow better visualization of acute inflammatory pathological processes such as mucosal enhancement in appendicitis by tailoring keV levels and window settings.