AGFA PACS software supported CT, MRI and PET windows impact on object appearances and size: phantom study

The aim of the present work is to assess the effect of window width and window level selection on feature size in a 2D displayed multimodality images. DICOM images of CT, MR and PET were retrieved and retrospectively analyzed using a local PACS system running an image manipulation software AGFA NX v2.0 that supports several windows for each image modality. The images analyzed were phantom based using standardized NEMA ICE body phantom that consist of six fillable variable spheres sizes. A NOVA t-test was used for statistical difference studies. Using bone and abdomen/ pelvis CT windows resulted in a sphere underestimation of up to 3 mm. While using T1 and T2 MRI windows yielded an underestimation of up to 3.5 mm especially in the small spheres. Significant spheres size differences were seen with PET windows shown an overestimation of up to 7 mm using CE equal and thermal windows. Improper selection of windowing while viewing 2D displayed images can yield considerable inconsistency leading to unreliable measurement.