A post-processing technique for cranial CT image identification

Objective: A major challenge in radiographic identification is the inconsistent orientation between clinical (ante-mortem, AM) and post-mortem (PM) radiographs. The objectivity and accuracy of radiological identification would be greatly enhanced by post-processing techniques that allow quantitative comparison of PM CT data in the same orientation as the AM CT data. Methods: We applied a post-processing technique to reposition a multislice computed tomography (MSCT) scan for spatial registration with a CT radiograph from the same patient. A second set of MSCT images from different individuals served as the non-matched control group. The consistency in radiographic positioning eliminated subjectivity in the comparison and identification process because the radiograph superposition provided objective evidence that confirmed the identification with fine detail. Results: A quantitative comparison with statistical validation was achieved by measuring a set of 14 landmarks from the images. Discrimination of identity based on logistic regression analysis of the earlier CT patient scans (the AM group) versus subsequent MSCT scans (the PM group) was objective and reliable. Conclusion: This quantitative comparison depends less on subjective judgment and the experience of the examiner, and so may meet legal standards.