A simple method for semi-automatic readjustment for positioning in post-mortem head computed tomography imaging. (March 2019)

• Record Type: Journal Article
• Authors: Kawazoe, Yusuke; Morishita, Junji; Matsunobu, Yusuke; Okumura, Miki; Shin, Seitaro; Usumoto, Yosuke; Ikeda, Noriaki
• Is Part Of: Journal of forensic radiology and imaging. Volume 16(2019)
• Page Start: 57
• Page End: 64
• Publication Date: 2019-03
• Subjects: Forensic radiology -- Head computed tomographic imaging -- Orbitomeatal line -- Post-mortem imaging; Forensic radiography -- Periodicals; Magnetic resonance imaging -- Periodicals; Diagnostic imaging -- Periodicals; Diagnostic imaging; Forensic radiography; Magnetic resonance imaging; Periodicals; 614.1
• DOI: 10.1016/j.jofri.2019.01.004
• Languages: English
• ISSNs: 2212-4780

Abstract

Highlights: Only 5 landmarks on post-mortem head CT image were used for positional adjustment. Our method resulted in successful adjustment in 96.4% (217/225) of the cases. Readjustment of post-mortem head CT images can be conducted within a short time. Semi-automated positional adjustment would be useful for personal identification. Abstract: The purpose of this study is to evaluate the usefulness of a simple method for readjustment of positioning in post-mortem (PM) head computed tomography (CT) imaging. Two hundred and twenty-five autopsied cases were used to test our proposed method. The coordinates of 5 landmarks on axial planes in PM head CT images were determined manually to calculate Yaw, Pitch, and Roll angles for the sake of readjustment. A radiologist and a forensic pathologist evaluated the readjustment by mutual agreement. Intraclass correlation coefficients (ICCs) for Yaw, Pitch, and Roll angles were measured for 3 operators. Similarity in AM and PM images after the readjustment was tested for 18 cases to evaluate the usefulness of the readjustment in personal identification for forensic cases. If similarity of the same person showed the highest value among all combinations in 18 cases, it was regarded as the correct identification. The processing time of our readjustment was limited to within 2 min per case. The orbitomeatal line (OML) was included in a single axial plane in 96.4% (217/225) of the cases after the readjustment. Reproducibility for determining landmarks showed acceptable ICC. Moreover, rank-one identification rate before and after readjustment were 16.7% (3/18) and 88.9% (16/18), respectively. Thus, our method for semi-automated readjustment exhibited potential usefulness in forensic cases. A simple method for semi-automatic readjustment would be a useful tool for post-mortem CT imaging with high reproducibility and a short processing time.