Optimization of postprocessing parameters for abdominal forensic CT
Optimization of postprocessing parameters for abdominal forensic CT scans Pernille A Nielsen 1, Dina M Bech 1, Julie B Nielsen 1 Pernille L Hansen 1, Dennis L Hansen 2, 3 Svea D Mørup 1, Peter M Leth 4, Helle Precht 1, 5 1 CONRAD 2 3 Research Program, University College, Odense, Denmark Department of Hematology, Odense University Hospital, Odense, Denmark Department of Clinical Research, University of Southern Denmark, Odense, Denmark 4 Institute 5 of Forensic Medicine, University of Southern Denmark, Odense, Denmark Cardiology Research department, Odense University hospital, Svendborg, Denmark
Introduction • Post-Mortem Computed Tomography is used as a supplement to autopsy • However, imaging of the deceased is different from imaging living persons • Decay, dismemberment, metal artefacts and dehydration affect image quality • Reconstruction methods is needed that aims post-mortem conditions • Reconstruction parameters has previously been suggested, but never evaluated Abdominal CT Newly deceased Early decay Late decay
Materials and methods • In forensics the aim is to determine between inflicted injuries, pathologies and accidents • We chose to focus on imaging of abdomen, and the visceral organs • In CT, even slight decay can impede distinction between visceral organs • Persons found dead, or persons dead under suspicious circumstances, has often been dead in hours or days
Materials and methods • We compares four reconstruction protocols • Three newly constructed and one routinely used • The protocols was used on scans of 20 deceased, in various states of decay • Image quality of the reconstructions was compared using visual grading analysis • The images was evaluated separately by four forensically trained radiologists Recon 1 Recon 2 Recon 3 Recon 4 Algorithm DETAIL STANDARD Hounsfield WW / WL 350/40 4000/80 350/40
Results Visual presentation of reconstructions R 1 - DETAIL R 2 – HU 3000 R 3 – HU 4000 R 4 - STND
Results Percentage of each reconstruction type deemed useful for diagnostic purpose Scans R 1 - DETAIL R 2 – HU 3000 R 3 – HU 4000 R 4 - STND 95, 2% 91, 7% 89, 3% 91, 7% Decomposed bodies 50% 62, 5% 50% Excluding decomposed bodies 100% 94, 7% 93, 4% 97, 4% All
Results Visual grading analysis of image quality criteria • Grading from 1 to 5 where 5 is the best • Results depicted as median, with interquartile range Spatial Resolution Low Contrast Resolution Contrast Artifacts/ Beam Hardening R 1 – DETAIL 4 R 2 – HU 3000 2 R 3 – HU 4000 2 R 4 - STND 4 [iqr: 3 -4] [iqr: 1 -3] [iqr: 3 -4] 4 2 3 4 [iqr: 3 -4] [iqr: 2 -3] [iqr: 3 -4] 2 3 3 2 [iqr: 2 -3] [iqr: 3 -4] [iqr: 2 -3] 4 4 [iqr: 3 -4] [iqr: 4 -5] [iqr: 4 -4. 5] [iqr: 3 -4]
Conclusions • Stage of decomposition influenced the evaluation • Above 90% of scans was rated usefull for non-decomposed in all reconstructions • The new reconstructions increased soft tissue details • However, several reconstructions are needed to visualize various tissues • Protocols must be targeted to different conditions of the deceased • The study provides a basis for further research which could potentially improve the quality and/or utility of forensic CT scans
Limitations • Observers was experienced in forensic imaging but not in evaluating reconstructions • Study population was limited to 20 scans • Of the 20 deceased included, two was severely decomposed which resulted in mixed evaluations
Ethics committee approval / Funding • This study was a quality improvement study and did as such not require ethics committee approval under Danish law • There has not been applied for funding for this study
- Slides: 10