Conventional versus ultra-low dose computed tomogram in Crohn's disease: Do morphomics correlate with clinical data?

Abstract

BackgroundCross-sectional study to assess the body composition of patients with Crohn's disease (CD) on standard (SDCT) and low dose CT (LDCT) protocols for the abdomen and pelvis (CTAP). We aimed to assess if a low dose CT protocol reconstructed with model-based iterative reconstruction (IR) could evaluate body morphometric data comparable to standard dose examination. MethodsThe CTAP images of 49 patients who underwent a low dose CT scan (20% of standard dose) and a second at standard dose minus 20% were assessed retrospectively. Images were collected from the PACS system, deidentified and analysed using a web-based semi-automated threshold-based segmentation tool (CoreSlicer), capable of identifying tissue type based on differences in attenuation co-efficient. The cross-sectional area (CSA) and Hounsfield units (HU) of each tissue was recorded. ResultsMuscle and fat CSA is well preserved on comparing these derived metrics from low dose and standard dose CT scans of abdomen and pelvic in CD ((LDCT:SDCT mean CSA (cm2); Psoas muscle - 29.00:28.67, total lumbar muscle - 127.45:125.55, visceral fat- 110.44:114.16, subcutaneous fat - 250.88:255.05)). A fixed difference exists when assessing the attenuation of muscle, with higher attenuation on the low dose protocol (LDCT:SDCT mean attenuation (HU); Psoas muscle - 61.67:52.25, total lumbar muscle - 49.29:41.20). ConclusionWe found comparable CSA across all tissues (muscle and fat) on both protocols with a strong positive correlation. A marginally lower muscle attenuation suggestive of less dense muscle was highlighted on SDCT. This study augments previous studies suggesting that comparable and reliable morphomic data may be generated from low dose and standard dose CT images. Implications for practiceThreshold-based segmental tools can be used to quantify body morphomics on standard and low dose computed tomogram protocols.