Chronic thromboembolic pulmonary hypertension (CTEPH) is often diagnosed late in acute pulmonary embolism survivors: more efficient testing to expedite diagnosis may considerably improve patient outcomes. The InShape II algorithm safely rules out CTEPH (failure rate 0.29%) while requiring echocardiography in only 19% of patients but may be improved by adding detailed reading of the computed tomography pulmonary angiography diagnosing the index pulmonary embolism.

We evaluated 12 new algorithms, incorporating the CTEPH prediction score, ECG reading, N-terminal pro-brain natriuretic peptide levels and dedicated computed tomography pulmonary angiography reading, in the international InShape II cohort (n=341) and part of the German FOCUS cohort (n=171). Evaluation criteria included failure rate, defined as the incidence of confirmed CTEPH in pulmonary embolism patients in whom echocardiography was deemed unnecessary by the algorithm, and the overall net reclassification index compared to the InShape II algorithm.

The algorithm starting with computed tomography pulmonary angiography reading of the index pulmonary embolism for six signs of CTEPH, followed by ECG/N-terminal pro-brain natriuretic peptide level assessment and echocardiography resulted in the most beneficial change compared to InShape II, with a need for echocardiography in 20% (+5%), a failure rate of 0% and a net reclassification index of +3.5%, reflecting improved performance over the InShape II algorithm. In the FOCUS cohort, this approach lowered echocardiography need to 24% (–6%) and missed no CTEPH cases, with a net reclassification index of +6.0%.

Dedicated computed tomography pulmonary angiography reading of the index pulmonary embolism improved the performance of the InShape II algorithm and may improve the selection of pulmonary embolism survivors who require echocardiography to rule out CTEPH.

慢性血栓栓塞性肺动脉高压 （CTEPH） 通常在急性肺栓塞幸存者中被诊断为晚期：更有效的检测以加快诊断可能会大大改善患者的预后。InShape II 算法可以安全地排除 CTEPH （失败率 0.29%），同时只有 19% 的患者需要超声心动图检查，但可以通过增加计算机断层扫描肺血管造影的详细读数来诊断指数肺栓塞来改善。

我们评估了 12 种新算法，包括国际 InShape II 队列 （n=341） 和部分德国 FOCUS 队列 （n=171），包括 CTEPH 预测评分、心电图读数、N 末端脑钠肽前体水平和专用计算机断层扫描肺血管造影读数。评估标准包括失败率，定义为算法认为不需要超声心动图的肺栓塞患者确诊 CTEPH 的发生率，以及与 InShape II 算法相比的总体净重分类指数。

与 InShape II 相比，该算法从计算机断层扫描肺血管造影读取 CTEPH 的 6 个体征的指标肺栓塞开始，然后是 ECG/N 末端脑钠肽前体水平评估和超声心动图，结果是最有益的变化，需要超声心动图的 20% （+5%），失败率为 0%，净重分类指数为 +3.5%， 反映了比 InShape II 算法更好的性能。在 FOCUS 队列中，这种方法将超声心动图需求降低到 24% （-6%），并且没有漏诊 CTEPH 病例，净重分类指数为 +6.0%。

专用计算机断层扫描指标肺栓塞的肺血管造影读数提高了 InShape II 算法的性能，并可能改善需要超声心动图排除 CTEPH 的肺栓塞幸存者的选择。