Clinical trial end points are outcome measures used to assess the efficacy and safety of an intervention. The selection of end points is based on their perceived clinical importance and accuracy of measurement. If correctly defined and collected, these outcome variables enable the translation of clinical trial data into evidence-based recommendations. Yet as technology changes and clinical practices evolve, emerging clinical events that are unrelated to the study intervention may be missed when conventional end points are used. In the Scottish Computed Tomography of the Heart (SCOT-HEART) trial,1 coronary computed tomographic angiography,whenaddedtostandardofcare,wasassociated with a 41% lower incidence of death from coronary heart diseaseornonfatalmyocardial infarctionatamedianfollowup of 4.8 years. Examination of the incidence of unanticipated radiographic findings and their subsequent diagnostic outcomes was prespecified in the protocol and described in a subsequent report.2 This approach for clinicaltrialreporting—includingthenature,frequency,andoutcomes of incidental imaging findings and new unrelated diagnoses—is the exception to the contemporary norm in reporting the results of trials. However, developing better understanding of the clinical implications of unanticipated imaging findings represents an important opportunity as the use of diagnostic imaging in clinical trials increases.3 What is the prevalence of incidental imaging findings? Outside the trial context, their occurrence (referred to as radiographic“incidentalomas”)varieswidely,dependingon the imaging test, ranging from less than 5% for chest CT for pulmonaryembolismto34%withcardiacMRI.4 Dataabout the frequency of incidental imaging findings in clinical trials are limited. It is likely that the magnitude of the phenomenon in clinical research studies is not substantially different from that in clinical practice, although analyses of clinical trial data are needed to answer that definitively. What are the effects of incidental imaging findings on clinical outcomes of trial participants? Incidental findings are associated with a potential malignancy diagnosis in a sizable proportion of cases. In addition, noncancerous conditions are frequently encountered in the course of an evaluationofincidental imagingfindings,includingpituitary macroadenomas(30%ofpituitaryincidentalomas)andcerebral aneurysms (3.9% of brain incidentalomas).4 Financially, the evaluation of an incidentally detected adrenal nodule, among the most prevalent unanticipated finding, costs $746 to $1745, depending on the investigation modality selected. These unexpected findings and their subsequent evaluation and management have an important bearing on the long-term health of trial participants. If incidental findings are common and may influence patient outcomes, should they be integrated into trial outcome reporting? It seems that efficacy and safety assessments are incomplete without inclusion of a meticulous synthesis of incidental findings into outcome metrics. Unanticipated imaging findings reflect an unintended byproduct of the ability of modern technology to scan multiple organs with high-level precision and therefore are a form of safety evaluation. An individual found to have an incidental pulmonary nodule may have to undergo invasive and risky procedures and often submit to longlasting surveillance imaging (and sometimes active treatment).Theseinterventionsthereforehaveinnaterisksthat may partly counteract the benefit gained with a study intervention. However, incidental imaging findings are a heterogeneous entity. Certain incidentalomas may lead to an objective health benefit with an earlier diagnosis and treatment (eg, nodule representing curable early-stage cancer). Other findings may provide important information about cardiovascular health that could influence future risk-reduction interventions (eg, coronary calcification on chest computed tomography) and thus represent an opportunity to improve cardiopreventive care. In contrast, some incidental findings have no relationship with survival and may pose risks because of invasive evaluation, overdiagnosis, or unnecessary financial expenses. Investigators have an ethical duty to collect data on and report incidental findings in clinical research, given the important ramifications of such findings on patient health. Inaddition,researchershavearesponsibilitytoensurethat participants are notified about the possibility of an incidental finding, as part of the informed consent process, as well as to create a mechanism by which information about a finding is disclosed. The latter should occur as soon as a findingisdetectedtoallowtimelyfollow-upand, ifneeded, treatment. Individuals with certain findings may decide to withdraw from the study, and, although potentially affecting patient retention and outcome data, such a decision should be respected by researchers and the scientific community. Addressing incidental imaging findings in human researchbyestablishingapathwayforcommunicationand discussion with patients has paramount value because it fulfills the principles of patient autonomy and nonmaleficence and embodies respect for participants’ health and well-being—the highest priorities in any clinical trial. Incorporatingincidental imagingfindingsintotrialoutcomereportingwillrequireaconcertedeffortbyacademia, industry,andregulatoryagencies.Definingthescopeofthe problem will require quantification of the proportion of clinicaltrialsthatuseimagingtests, includingstructuraland functional modalities. Relevant parameters should include the distribution of patient ages, racial/ethnic groups, and sex in imaging-based clinical trials and the frequency and severity of unanticipated findings in each subgroup. Patientandtrialcharacteristicsassociatedwithhigherodds of incidental findings should be identified. In clinical trials that involve imaging studies, investigators and their sponsoring institutions should routinely collect information on unexpected radiographic findings. This includes the index imaging study, the primary organ VIEWPOINT

中文翻译：

---

临床试验中的偶然影像学发现

临床试验终点是用于评估干预措施的有效性和安全性的结果指标。终点的选择基于其感知的临床重要性和测量的准确性。如果正确定义和收集，这些结果变量可以将临床试验数据转化为基于证据的建议。然而，随着技术的变化和临床实践的发展，当使用传统终点时，可能会遗漏与研究干预无关的新出现的临床事件。在苏格兰心脏计算机断层扫描 (SCOT-HEART) 试验中，1 冠状动脉计算机断层扫描血管造影添加到标准护理中后，中位随访 4.8 年，冠心病或非致命性心肌梗死的死亡率降低了 41%。方案中预先指定并在随后的报告中描述了对意外影像学发现的发生率及其后续诊断结果的检查。 2 这种临床试验报告方法——包括偶然成像发现和新的无关诊断的性质、频率和结果——是报告试验结果的当代规范。然而，随着诊断成像在临床试验中的使用增加，更好地理解意外成像发现的临床意义代表了一个重要的机会。3 偶然成像发现的普遍性如何？在试验环境之外，它们的发生（称为放射学“偶发瘤”）因影像学检查而异，范围从肺栓塞的胸部 CT 小于 5% 到心脏 MRI 的 34%。4 关于临床试验中偶然成像发现频率的数据是有限的。尽管需要对临床试验数据进行分析才能明确回答这一问题，但临床研究中这种现象的严重程度可能与临床实践中的现象没有本质上的不同。偶然的影像学发现对试验参与者的临床结果有何影响？在相当大比例的病例中，偶然发现与潜在的恶性肿瘤诊断相关。此外，在评估偶然成像结果的过程中经常会遇到非癌性疾病，包括垂体大腺瘤（30% 的垂体偶发瘤）和脑动脉瘤（3.9% 的脑偶发瘤）4。对偶然检测到的肾上腺结节的评估是最常见的意外发现之一，其费用为 746 至 1745 美元，具体取决于所选的调查方式。这些意外发现及其随后的评估和管理对试验参与者的长期健康具有重要影响。如果偶然发现很常见并且可能影响患者结果，是否应该将它们纳入试验结果报告中？如果不将偶然发现的仔细综合到结果指标中，有效性和安全性评估似乎是不完整的。意外的成像结果反映了现代技术以高精度扫描多个器官的能力的意外副产品，因此是一种安全评估形式。发现偶发肺结节的个体可能不得不接受侵入性和危险的手术，并经常接受长期监测成像（有时是积极治疗）。因此，这些干预措施存在固有风险，可能会部分抵消研究干预措施所带来的好处。然而，偶然的影像发现是一个异质的实体。某些偶发瘤可能会通过早期诊断和治疗（例如，代表可治愈的早期癌症的结节）带来客观的健康益处。其他发现可能提供有关心血管健康的重要信息，这些信息可能会影响未来降低风险的干预措施（例如，胸部计算机断层扫描的冠状动脉钙化），从而代表改善心脏预防护理的机会。相比之下，一些偶然发现与生存无关，可能会因为侵入性评估、过度诊断或不必要的财务费用而带来风险。鉴于此类发现对患者健康的重要影响，研究人员有道德义务收集数据并报告临床研究中的偶然发现。此外，作为知情同意过程的一部分，研究人员有责任确保将偶然发现的可能性通知参与者，并创建一种机制来披露有关发现的信息。后者应在检测到发现后立即发生，以便及时跟进和治疗（如果需要）。有某些发现的个人可能会决定退出研究，尽管可能会影响患者保留率和结果数据，研究人员和科学界应尊重这样的决定。通过建立与患者交流和讨论的途径来解决人类研究中的偶然成像发现具有重要价值，因为它满足了患者自主和无恶意的原则，并体现了对参与者健康和福祉的尊重——这是任何临床试验中的最高优先事项。将偶然的影像学发现纳入试验结果报告需要学术界、工业界和监管机构的共同努力。确定问题的范围将需要量化使用影像学测试的临床试验的比例，包括结构和功能模式。相关参数应包括患者年龄分布、种族/民族、和基于影像的临床试验中的性别以及每个亚组中意外发现的频率和严重程度。应确定与偶然发现的高风险相关的患者和试验特征。在涉及影像学研究的临床试验中，研究人员及其赞助机构应定期收集有关意外影像学发现的信息。这包括索引成像研究、主要器官 VIEWPOINT