In this study, we described a novel 3D CT quantification method for TEV in acute PE, and we found that it was independently associated with the Qanadli score, echocardiographic findings (sPAP and RV dysfunction), SpO2, and troponin I. Most of these measurements are commonly available and can be obtained from the emergency department. Furthermore, we noticed that RV dysfunction, as assessed using echocardiography, is strongly correlated with TEV but not the Qanadli score, and that CT 3D measurements of clot burden played an important role in predicting impending shock but not short-term mortality.
Historically, MDCT has been considered the gold standard for diagnosing patients with acute PE12. However, it has played a limited role in prognostication. In the current clinical guidelines issued by the AHA and ESC, a high RV/LV ratio is one of the CT factors associated with poor prognosis2,3. A meta-analysis showed that a RV/LV ratio > 1.0 on CT was associated with a 2.5-fold increased risk for all-cause mortality (OR: 2.5, 95% CI: 1.8–3.5)13. In our study, we also found that an increased RV/LV ratio was significantly associated with an increased risk for impending shock. Conversely, TEV may be an important predictor for stratifying acute PE patients who are referred to the emergency department. Increases in sPAP were correlated with increasing TEV quartiles (Table 3) and a similar pattern was seen with RV dysfunction and troponin I, while the reverse pattern was observed with SpO2. We also found that of the MDCT and echocardiogram findings in patients without cancer, TEV and RV/LV ratio were significantly correlated with impending shock (Table 6). Furthermore, in the backward stepwise logistic regression analysis, TEV, age, and respiratory rate were independently associated with impending shock (Table 5). The presence of shock is clinically relevant to the management and prognosis of acute PE. Approximately 5% of intermediate-risk PE patients who initially show hemodynamic stability will develop hemodynamic decompensation within the first 48 hours and up to as late as five days later14. Traditionally, for acute PE patients, the presence of RV dysfunction on echocardiography and the RV/LV ratio measured via CT are key indicators for shock. Our study suggests that precisely measured 3D embolic volume in the pulmonary arteries may be a more accurate predictor of impending shock than these two indicators. To date, there are limited data on the clinical implications of embolic burden. Herein, we have introduced the clinical feasibility of this novel imaging modality in PE patients referred to the emergency department. We propose that its use for the timely diagnosis of PE and stratification of risk for impending shock in such patients may save lives in acute settings by indicating when urgent thrombolysis or thrombectomy interventions are warranted.
For echo-based RV dysfunction (Table 3), Q2 (6.06%) is smaller than Q1 (27.8%). After excluding counts with only RV dilatation (diameter < 30 mm), which is weak evidence of RV dysfunction, we found that the percentages of Q1 and Q2 were equal (4%). This result supports other data (sPAP, RV diameter, RV/LV) that show that when the embolic burden is small (maximal TEV in Q2 is < 4 cm3), changes in cardiac morphology and function are minor. To our knowledge, no other studies prove the relationship between cardiac function and precise embolic burden, and we offer this perspective to clarify the relationship.
Compared to the Qanadli score, TEV not only more accurately assesses 3D emboli burden but is also better correlated with RV dilatation and echo-based RV dysfunction. Increasing TEV and Qanadli score can both indicate a larger clot burden in the pulmonary arterial circulation, which can lead to pressure overload and RV dilatation. In our study, TEV and the Qanadli score were well correlated, especially in the case of main pulmonary arterial involvement. However, the Qanadli score only assigns one point to each subsegmental embolus, irrespective of its length and the number of related subsegmental arteries. Therefore, this score cannot reflect the actual volume of the clot burden. Indeed, we showed that the Qanadli score is only well correlated with a smaller TEV (Fig. 2), due to this underestimation. Furthermore, as a semiquantitative method, the Qanadli score is difficult to calculate and has low reproducibility and high interobserver variability15. Here, we have demonstrated that TEV measurement, which is both semiautomatic and fully quantitative, has better reproducibility than the Qanadli score (ICC: 0.99 vs 0.74, respectively).
In this study, there was no significant association between TEV and 30- and 90-day mortality rates, but this was not unexpected. There are only two small studies published before 2010 that have shown a significant association between semiquantitative clot burden assessment and short-term survival9,16, whereas several more recent studies with a large number of subjects have failed to show a significant association between the Qanadli score and short-term survival17,18. This may be due to improvements in the early diagnosis and treatment of patients with shock and cardiac arrest from acute PE, which has increased their overall survival rate. Recently, Stein et al. investigated the mortality of PE patients in the United States and showed that, from 1999 to 2017, the mortality of all high-risk patients decreased from 72.7–49.8%19. A decreasing mortality rate for acute PE makes demonstrating a significant correlation between radiologic markers and mortality more difficult. This could explain why, despite 79.3% of our patients having a high-risk sPESI score, TEV was not significantly associated with short-term mortality. In contrast, cancer, which is a major risk factor for PE and is used in calculating the sPESI score, was the leading cause of death in our patients (Fig. 4). Similarly, other PE risk factors, such as advanced age and certain medical comorbidities20,21 can significantly affect morbidity and mortality rates, even in patients who are considered to have a low-risk PE because they are normotensive with normal biomarker levels and no RV dysfunction on imaging15. Therefore, to more comprehensively stratify the severity of acute PE and the risk of early (in-hospital or 30-day) death, additional factors such as hemodynamic status, clinical condition, RV dysfunction, and troponin I levels are important and need to be explored19.
Our study has some limitations. First, it is a retrospective study with an exclusively Asian population from a single medical center, which may represent a selection bias and limit generalizability. Second, our novel 3D CT volumetric method is only semiautomatic and still requires manual adjustments. Future studies using alternative technologies, such as artificial intelligence, for fully automatic quantification of clot burden may help to improve clinical efficiency by generating 3D volumetric results before radiologist interpretation and reducing possible manual operator errors.
In conclusion, our study demonstrated that a novel 3D CT method for quantifying acute pulmonary embolism provides results that are significantly linked to clinical condition, laboratory data, RV dysfunction, and impending shock. 3D CT measurements of clot burden may thus become a useful and feasible method for acute PE risk stratification and prognostication.