Evaluation of frontal sinus and skull measurements using spiral CT scanning: an aid in unknown person identification.

The present study was undertaken to test a simple system for the identification of unknown bodies using spiral CT images of frontal sinus and other skull measurements among selected Iraqi sample. Ninety patients (45 males and 45 females) with age range from 20 to 49 years were selected in this study. Three features and two groups of measurements of frontal sinus and three skull dimensions were obtained from the CT images. Three basic features were F (presence or absence of frontal sinus), S (septum) and S (scalloping). Measurements selected for the study were frontal sinus width, height and anteroposterior length. In addition to measurements of total width, the distance between the highest points of the two sinuses, the distance between the highest points of each sinus to its maximum lateral limit. Skull measurements included; maximum skull length, prostio-bregmatic height and maximum skull width. All data were subjected to a descriptive and discriminative analysis using the SPSS (Version 17.0). The pre-post comparison (number of discordant items) resulted in 95% accurately predicted perfect match for intra-examiner calibration and 90% accurately predicted perfect match for inter-examiners calibration and the result for one discordant item was 5% for intra-examiner calibration and 10% for inter-examiners calibration. The discriminative analysis showed that the ability of the frontal sinus to identify gender was 76.9%, adding the skull measurements to the frontal sinus measurements gave a higher overall classification accuracy for gender (85.9%). Frontal sinus measurements are valuable method in differentiating gender. Adding skull measurements to the frontal sinus measurements can significantly improve accuracy of gender determination using discriminant analysis. CT based films can provide valuable and precise measurements not only for frontal sinus but even for the whole skull that cannot be approached by other means.