胸部X线识别肺炎的临床效用:考虑影像报告诊断不确定性
本研究以胸部CT为参考标准,回顾性分析2411例患者,通过将胸部X线和CT报告中的不确定印象分别归为阴性或阳性形成5种情景,计算胸部X线诊断肺炎的敏感度、特异度、PPV和NPV及95%CI。结果显示不确定报告常见(胸片31.8%,CT 21.7%),胸片阳性印象有助于在影像学上确立肺炎诊断,但阴性结果排除疾病的能力差。
要点
- 胸部X线报告中不确定肺炎印象常见,占31.8%,CT报告不确定印象占21.7%。
- 胸片印象采用最强支持肺炎诊断的措辞时,有助于在影像学上确立肺炎诊断。
- 胸片阴性结果排除肺炎的能力差,敏感度在不同情景下仅13.4%至59.6%。
核心结论
- 胸部X线评估肺炎时不确定结果常见,胸片印象采用最强支持肺炎诊断的措辞有助于在影像学上确立肺炎,但阴性结果排除疾病的能力差。
研究设计与样本
- 回顾性分析2411例患者就诊,以胸部CT为参考标准,计算胸部X线诊断肺炎的统计性能指标及95%CI。
- 通过将胸部X线和CT的不确定报告印象分别归为全阴性或全阳性,形成5种情景进行分析。
诊断标准与关键数值
- 情景1(胸片和CT不确定印象均视为阳性):敏感度51.9%,特异度71.3%,PPV 59.4%,NPV 64.5%。
- 情景2(胸片不确定为阳性、CT不确定为阴性):敏感度59.6%,特异度67.1%,PPV 59.6%,NPV 67.1%。
- 情景3(胸片不确定为阴性、CT不确定为阳性):敏感度13.4%,特异度97.7%,PPV 82.6%,NPV 58.1%。
- 情景4(胸片和CT不确定印象均视为阴性):敏感度19.6%,特异度96.4%,PPV 81.6%,NPV 59.5%。
- 情景5(排除胸片和CT不确定印象):敏感度32.7%,特异度96.8%,PPV 89.2%,NPV 63.8%。
临床应用与阅片要点
- 胸片报告采用最强支持肺炎诊断的措辞时,有助于在影像学上确立肺炎诊断。
- 胸片阴性结果排除肺炎的能力差,阅片和报告时需注意这一局限。
局限性
- 原文未报告局限性。
以上解读依据论文摘要(全文未开放获取或许可不允许翻译转载)。
英文原文(摘要)· 点击展开
OBJECTIVE. Currently, chest radiography is the first-line imaging test for identifying pneumonia; chest CT is considered the reference standard. The purpose of this study was to calculate the statistical measures of performance of chest radiography for identifying pneumonia when taking into account uncertain results of both chest radiography and CT examinations. MATERIALS AND METHODS. Statistical measures of performance of chest radiography, using CT as the reference standard, were calculated with 95% CIs by varying uncertain radiology report impressions of both chest radiography and CT to all negative or all positive. The resulting scenarios were as follows: scenario 1, uncertain chest radiography and CT impressions are considered positive for pneumonia; scenario 2, uncertain chest radiography impressions are positive but uncertain CT impressions are negative; scenario 3, uncertain chest radiography impressions are negative and uncertain CT impressions are positive; scenario 4, uncertain chest radiography and CT impressions are negative; and scenario 5, uncertain chest radiography and CT impressions are excluded. RESULTS. A retrospective analysis of 2411 patient visits revealed the prevalence of uncertain radiology report impressions to be 31.8% for chest radiography and 21.7% for CT. Scenario 1 yielded the following performance values: sensitivity, 51.9%; specificity, 71.3%; PPV, 59.4%; and NPV, 64.5%. Scenario 2 produced the following performance values: sensitivity, 59.6%; specificity, 67.1%; PPV, 59.6%; and NPV, 67.1%. Scenario 3 showed the following performance values: sensitivity, 13.4%; specificity, 97.7%; PPV, 82.6%; and NPV, 58.1%. Scenario 4 yielded the following performance values: sensitivity, 19.6%; specificity, 96.4%; PPV, 81.6%; and NPV, 59.5%. Scenario 5 produced the following performance values: sensitivity, 32.7%; specificity, 96.8%; PPV, 89.2%; and NPV, 63.8%. CONCLUSION. Uncertain chest radiography results for the evaluation of pneumonia are prevalent. A chest radiography impression using the strongest language in support of a pneumonia diagnosis is useful to rule in pneumonia radiographically, but a negative result performs poorly at ruling out disease.
引用
- Makhnevich A, Sinvani L, Cohen SL, et al. The Clinical Utility of Chest Radiography for Identifying Pneumonia: Accounting for Diagnostic Uncertainty in Radiology Reports. AJR Am J Roentgenol. 2019. PMID: 31509449. doi:10.2214/AJR.19.21521
供执业医师工作参考,不替代临床判断;数值与分级以指南原文为准。