Raji Cyrus A, Willeumier Kristen, Taylor Derek, Tarzwell Robert, Newberg Andrew, Henderson Theodore A, Amen Daniel G
Department of Radiology, UCLA Medical Center, 757 Westwood Blvd, Los Angeles, CA, 90095, USA.
Department of Research, Amen Clinics, Inc, Costa Mesa, CA, USA.
Brain Imaging Behav. 2015 Sep;9(3):527-34. doi: 10.1007/s11682-015-9385-5.
PTSD and TBI are two common conditions in veteran populations that can be difficult to distinguish clinically. The default mode network (DMN) is abnormal in a multitude of neurological and psychiatric disorders. We hypothesize that brain perfusion SPECT can be applied to diagnostically separate PTSD from TBI reliably in a veteran cohort using DMN regions. A group of 196 veterans (36 with PTSD, 115 with TBI, 45 with PTSD/TBI) were selected from a large multi-site population cohort of individuals with psychiatric disease. Inclusion criteria were peacetime or wartime veterans regardless of branch of service and included those for whom the traumatic brain injury was not service related. SPECT imaging was performed on this group both at rest and during a concentration task. These measures, as well as the baseline-concentration difference, were then inputted from DMN regions into separate binary logistic regression models controlling for age, gender, race, clinic site, co-morbid psychiatric diseases, TBI severity, whether or not the TBI was service related, and branch of armed service. Predicted probabilities were then inputted into a receiver operating characteristic analysis to compute sensitivity, specificity, and accuracy. Compared to PSTD, persons with TBI were older, male, and had higher rates of bipolar and major depressive disorder (p < 0.05). Baseline quantitative regions with SPECT separated PTSD from TBI in the veterans with 92 % sensitivity, 85 % specificity, and 94 % accuracy. With concentration scans, there was 85 % sensitivity, 83 % specificity and 89 % accuracy. Baseline-concentration (the difference metric between the two scans) scans were 85 % sensitivity, 80 % specificity, and 87 % accuracy. In separating TBI from PTSD/TBI visual readings of baseline scans had 85 % sensitivity, 81 % specificity, and 83 % accuracy. Concentration scans had 80 % sensitivity, 65 % specificity, and 79 % accuracy. Baseline-concentration scans had 82 % sensitivity, 69 % specificity, and 81 % accuracy. For separating PTSD from PTSD/TBI baseline scans had 87 % sensitivity, 83 % specificity, and 92 % accuracy. Concentration scans had 91 % sensitivity, 76 % specificity, and 88 % accuracy. Baseline-concentration scans had 84 % sensitivity, 64 % specificity, and 85 % accuracy. This study demonstrates the ability to separate PTSD and TBI from each other in a veteran population using functional neuroimaging.
创伤后应激障碍（PTSD）和创伤性脑损伤（TBI）是退伍军人中两种常见的病症，在临床上可能难以区分。默认模式网络（DMN）在多种神经和精神疾病中均表现异常。我们假设，脑灌注单光子发射计算机断层扫描（SPECT）可应用于在退伍军人队列中，利用DMN区域可靠地将PTSD与TBI进行诊断区分。从一个大型多地点的精神病患者群体队列中选取了196名退伍军人（36名患有PTSD，115名患有TBI，45名患有PTSD/TBI）。纳入标准为和平时期或战时的退伍军人，无论其军种如何，且包括那些创伤性脑损伤与服役无关的人。对该群体在静息状态和注意力集中任务期间均进行了SPECT成像。然后，将这些测量值以及基线 - 注意力集中差异，从DMN区域输入到单独的二元逻辑回归模型中，该模型控制了年龄、性别、种族、诊所地点、共病的精神疾病、TBI严重程度、TBI是否与服役相关以及军种。然后将预测概率输入到受试者工作特征分析中，以计算敏感性、特异性和准确性。与PTSD患者相比，TBI患者年龄更大，为男性，且双相情感障碍和重度抑郁症的发病率更高（p < 0.05）。SPECT的基线定量区域在退伍军人中以92%的敏感性、85%的特异性和94%的准确性将PTSD与TBI区分开来。进行注意力集中扫描时，敏感性为85%，特异性为83%，准确性为89%。基线 - 注意力集中（两次扫描之间的差异指标）扫描的敏感性为85%，特异性为80%，准确性为87%。在将TBI与PTSD/TBI区分开来时，基线扫描的视觉读数敏感性为85%，特异性为81%，准确性为83%。注意力集中扫描的敏感性为80%，特异性为65%，准确性为79%。基线 - 注意力集中扫描的敏感性为82%，特异性为69%，准确性为81%。在将PTSD与PTSD/TBI区分开来时，基线扫描的敏感性为87%，特异性为83%，准确性为92%。注意力集中扫描的敏感性为91%，特异性为76%，准确性为88%。基线 - 注意力集中扫描的敏感性为84%，特异性为64%，准确性为85%。这项研究证明了利用功能神经成像在退伍军人群体中将PTSD和TBI相互区分开来的能力。
