The purpose of this research is to examine the reliability of posteroanterior (PA) left knee radiographs for positive identification using comparative radiography. PA radiographs of the knee were selected for their clinical significance in diagnosing knee osteoarthritis (OA), the most common skeletal pathology today, and their likely appearance in the antemortem record of a medicolegal death investigation. Therefore, this research examines the reliability of using a clinically relevant type of radiograph, PA left knee radiographs, for positive identification. All radiographs used in this research were downloaded from the Osteoarthritis Imitative (OAI), a public database, and compiled into a virtual survey which asked participants to identify matches between radiographs collected at various time periods throughout the OAI study. Images were collected from the 12, 48, 72, and 96-month collection periods of the OAI for 89 individuals and divided into three series. Each of the three series has different amounts of time between image acquisition, with three years (Series 1), five years (Series 2), and seven years (Series3), and were created by comparing images from the 12-month collection period to images from 48, 72, and 96-month collection period, respectively. Radiographs were edited into single side-by-side images for ease of comparison and were included in a Qualtrics XM survey that was distributed to those with and without comparative radiography experience. A total of 66 participants were asked to answer several demographic questions as well as nine radiographic comparisons questions. Each of the radiographic comparisons, comprised of OAI images of living individuals, asked if a specified antemortem “A” image and was a match or no match to a series of “postmortem” images. Accuracies, sensitivities, specificities, positive predictive values (PPV), and negative predictive values (NPV) were calculated for the survey respondents (n=66), series, and respondent demographics. The overall results (accuracy=94.2%, sensitivity=54.7%, specificity=97.5%, PPV=65.6%, and NPV=96.2%) indicated that respondents performed only moderately well in matching correct radiographs. However, the performance of the participants in the overall survey and in all three series was statistically significantly lower (α<0.05) for those with no comparative radiography experience and those with little-to-no forensic casework experience. The performance of all participants became progressively lower with each series, with Series 1 demonstrating the best performance and Series 3 demonstrating the worst performance among participants. Decreasing levels of performance suggests that bony changes associated with OA and the osseous and radiographic features used in making matches (e.g., morphology of the femur and tibia, joint space width) preclude radiographic matching of the knee over around three years between antemortem and postmortem image acquisitions. Therefore, the results of this research demonstrate that radiographic comparison using OA left PA knee radiographs is not recommended after more than five years between antemortem and postmortem images, due to the degraded performance of trained survey respondents. / 2024-05-02T00:00:00Z
| Identifer | oai:union.ndltd.org:bu.edu/oai:open.bu.edu:2144/47448 |
| Date | 03 November 2023 |
| Creators | Choat, Tara A. |
| Contributors | Tallman, Sean D., Moore, Tara L. |
| Source Sets | Boston University |
| Language | en_US |
| Detected Language | English |
| Type | Thesis/Dissertation |
| Rights | Attribution-NonCommercial-NoDerivatives 4.0 International, http://creativecommons.org/licenses/by-nc-nd/4.0/ |
Page generated in 0.0021 seconds