Andersen L, Wenzel A
Department of Oral Pathology, Medicine and Toxicology, Royal Dental College, Aarhus University, Denmark.
Forensic Sci Int. 1995 Mar 21;72(1):55-64. doi: 10.1016/0379-0738(94)01676-v.
An analysis of simulated post- and antemortem (p.m. and a.m.) victim radiographs was performed. Existing pairs of bitewing radiographs randomly sampled from a large population of adolescents were used. Two categories of individuals were included based on dental therapy conditions (none or few simple fillings). Subtraction images were performed of pairs of bitewings originating from the same individual (identical images) and different individuals (non-identical images) within each group. Four observers were asked to match the bitewings and to assess the subtraction images with respect to identity based on the score system: 1 = 'eliminated', 2 = 'possible', 3 = 'likely', or 4 = 'certain'. Three observers each mismatched the film radiographs among two out of 12 individuals without fillings. True positive identification (identified by scores 2 + 3 + 4) of subtraction images were made in 10-12 individuals out of a possible 12. The range of false positive (FP) and false negative (FN) scorings was 3-15 and 0-2, respectively. All observers were capable of correctly matching bitewings originating in the 12 individuals with amalgam fillings. By use of subtraction radiography, 12-14 identical images out of a possible 15 were assessed correctly. The range of FP and FN scorings of subtraction images in this group was 1-4 and 1, respectively. Defining only score 4 as positive identification, the sensitivity of the subtraction technique decreased from approximately 0.90 to 0.70 in both categories of individuals, while specificity increased to almost 1.00 from 0.82 vs. 0.96 in the categories with and without dental restorations. Bitewing radiographs from single individuals within a group of individuals were sufficiently identical to allow for valid identification by a strict criterion. The subtraction technique may add to the subjective matching of radiographs as a screening test in victim identification.
对模拟的死后和生前受害者X光片进行了分析。使用了从大量青少年人群中随机抽取的成对咬合翼片X光片。根据牙科治疗情况（无或仅有少量简单补牙）纳入了两类个体。对每组中来自同一人（相同图像）和不同人（不同图像）的成对咬合翼片进行了减影成像。四名观察者被要求匹配咬合翼片，并根据评分系统评估减影图像的一致性：1 =“排除”，2 =“可能”，3 =“很可能”，或4 =“确定”。在12名无补牙的个体中，有三名观察者各自在其中两人身上将胶片X光片匹配错误。在可能的12人中，有10 - 12人对减影图像做出了真阳性识别（评分为2 + 3 + 4）。假阳性（FP）和假阴性（FN）评分范围分别为3 - 15和0 - 2。所有观察者都能够正确匹配12名有汞合金补牙个体的咬合翼片。通过使用减影X射线摄影，在可能的15张相同图像中，有12 - 14张被正确评估。该组中减影图像的FP和FN评分范围分别为1 - 4和1。仅将评分为4定义为阳性识别时，减影技术的敏感性在两类个体中均从约0.90降至0.70，而特异性从有和无牙齿修复的类别中的0.82对0.96增加到几乎1.00。一组个体中单个个体的咬合翼片X光片足够相似，以便通过严格标准进行有效识别。减影技术可以作为受害者识别中的一种筛查测试，辅助X光片的主观匹配。
