Efficacy of Computed Tomography Pulmonary Angiography as Non-invasive Imaging Biomarker for Risk Stratification of Acute Pulmonary Embolism

INTRODUCTION

Massive and submassive pulmonary embolism (PE) is associated with high mortality and morbidity. Rapid diagnosis and treatment are, therefore, essential to save patients’ lives. Risk stratification scores are used to determine the risk of complications and associated mortality. High-risk and intermediate-risk PE can be divided by a hemodynamic presentation.^[1]

Massive PE occurs in about 5% of patients with PE and carries a mortality rate of 18–65%. It is presented with sustained hypotension (systolic blood pressure [BP] <90 for at least 15 min), with signs or symptoms of shock.^[2] Submassive PE is seen in about 40% of patients with PE and carries a 5–25% mortality rate. Patients with submassive PE are systemically normotensive with evidence of myocardial ischemia and/or right ventricular dysfunction (RVD).^[2] The remaining 55% of patients with PE present with non-massive or low-risk PE, also called uncomplicated PE, with a mortality rate of only 1%. They are systemically normotensive, without RVD or myocardial ischemia.^[3]

The mainstay of treatment for all patients presenting with PE is anticoagulation.^[4] However, as the spectrum of PE severity varies widely, high-risk patients may require additional treatment measures to reduce the clot burden. Systemic intravenous thrombolysis or catheter-directed thrombolysis has been shown to improve mortality and hemodynamic stability when used for patients with high risk (massive) and high-intermediate risk (submassive) PE.^[5,6] Meanwhile, systemic thrombolysis is associated with a high risk of hemorrhage. Hence, emergency physicians must be able to risk stratify patients to ensure rapid access to advanced therapy for the appropriate candidate patients.^[7,8]

Diagnostic strategies for evaluating PE have undergone important changes over the past decades.^[9] Computed tomography pulmonary angiography (CTPA) has become the imaging method of choice in the diagnosis of acute PE (APE), being widely available, fast, and non-invasive technique, it can directly visualize emboli and may provide alternative diagnoses.^[10] CTPA may replace echocardiography, as it can identify pulmonary thrombus down to at least segmental level, to quantify the degree of PE, and at the same time, it allows evaluation of RV enlargement as an indicator of RVD.^[11] In this study, we attempt to evaluate the efficacy of CTPA as a non-invasive imaging biomarker for risk stratification of APE patients.

MATERIAL AND METHODS

This is a prospective study conducted on 150 patients who had been proven diagnosis of PE by CTPA. They were recruited from the emergency department, respiratory/ coronary, or general ICU presented by signs or symptoms of suspected APE. The ethics committee approval was obtained and all enrolled patients completed a written consent form.

RESULTS

The current study included 150 patients with CTPA proved APE. The final diagnosis regarding the presence or absence of PE was based on the result of MDCTPA. They were 98 (65%) females and 57 (35%) males. The average age of the patients was 58 ± 17 years. All the patients were above 18 years old. Regarding risk factors and comorbidities, we found that 42 patients (28%) were diabetic, 24 patients (16%) had chronic obstructive pulmonary disease, 12 patients (8%) were postpartum, 15 patients (10%) had a history of deep vein thrombosis, and 15 patients (10%) had a surgical history, 6 patients (4%) had a history of malignancy, 33 patients (22%) had more than one comorbidity, while no comorbidity found in 24 patients (16%). Concerning the distribution of the patients according to their sPESI score, 42/150 (28%) were low risk, while 108/150 (72%) were high-risk patients.

Regarding the distribution of patients according to the site of the clot, most patients had a central level of involvement with PE which was found mainly central within the main pulmonary arteries in 69 patients (46%) and at the segmental pulmonary arterial level in 60 (40%), while 21 patients (14%) had PE isolated to the subsegmental level.

The patients were divided into (1) high risk and (2) non- high-risk groups according to systolic BP. Table 1 shows a summary of descriptive statistics for studied groups:

The 1^st group included non-high-risk patients who had normal BP. They were 53/150 patients (35%), with mean age 48 ± 18 years, and mean systolic BP of 120 ± 15 mmHg. Mean PAOI for the 2^nd group was 36% ± 13, ranging between 12.5% and 60%.

The 2^nd group included high-risk patients who presented by shock or hypotension. They were 97/150 patients (65%), with mean age 57 ± 12 years, and mean systolic BP of 80 ± 5 mmHg. Mean PAOI for the 1^st group was 66% ± 13%, ranging between 48% and 80%.

There was a significant difference between the 1^st and 2^nd groups regarding PAOI (P < 0.0001) [Figure 1]. Furthermore, significant differences regarding HR, length of hospital stay, RV/LV ratio, sPAP, MPV (P ≤ 0.0001, 0.0003, 0.0006, 0.007, and 0.046, respectively) were detected.

Close

Figure 1:

Box plot showing the difference between pulmonary artery obstruction index for Group I (non-high risk) and Group II (high risk) (P < 0.0001).

Export to PPT

There was a significant correlation found between PAOI and sPESI scoring (P = 0.0001 and r = 0.467). Furthermore, regression analysis revealed significant negative relation between systolic BP and PAOI (P = 0.0002 and r = −0.347), as shown in Figure 2.

Close

Figure 2:

Receiver operating characteristic curve for diagnostic performance of pulmonary artery obstruction index for predicting abnormally increased right to left ventricular ratio >1, thus indicating right ventricular dysfunction (P = 0.0015, area under the curve = 0.739).

Export to PPT

The 1^st group patients who were hemodynamically stable were further subclassified according to RV/LV ratio measured on CT angiography into two groups: Group Ia with RVD (intermediate-risk patients) and Group Ib with no RVD (low-risk patients) [Table 2]. The RVD group included 28/53 patients (53%), 17 (61%) females, 11 (39%) males, with a mean age of 50±18 and mean RVD and LVD 4.9 ± 0.5 cm and 4.1 ± 0.4 cm, respectively. The non-RVD group included 25/53 patients (47%), they were 19 (76%) females, 6 (24%) males, with a mean age of 49 ± 17 and mean RVD and LVD 4.0 ± 0.8 cm and 4.5 ± 0.7 cm, respectively. The mean RVD/LVD ratio for Group Ia and Group Ib was 1.2 ± 0.1 and 0.8 ± 0.1, respectively.

There was a highly significant difference between the two subgroups regarding PAOI (P < 0.0001, r = 0.385). The mean PAOI for Group Ia and Group Ib was 49% ±18% and 33% ± 13%, respectively.

Receiver operating characteristic (ROC) curve was constructed to establish the best cutoff value for PAOI which was associated with RV/LV ratio >1 and showed that PAOI >47% had the best sensitivity and specificity for predicting abnormally increased RV/LV ratio (P = 0.0015, area under the curve = 0.739, 95% confidence interval, 0.593–0.854).

Illustrative cases of studied patients having different degrees of severity of APE are shown in Figures 3-5.

Close

Figure 3:

A 63-year-old male patient presented by central chest pain and fainting. Recent history of lower limb deep vein thrombosis. Blood pressure 80/60 mmHg and heart rate: 130 bpm. High risk according to simplified pulmonary embolism severity index score. (a) Hypodense filling defect of pulmonary embolus seen totally occluding right main pulmonary artery trunk (arrow). (b) The thrombus is seen also totally obstructing the left main pulmonary artery trunk. (c and d) The thrombus is seen extending to main right (arrow) and left segment branches on both sides. (e) Right to left ventricular ratio >1, indicating right ventricular dysfunction. The calculated pulmonary artery obstruction index for this patient was 87.5% (high-risk group).

Export to PPT

Close

Figure 4:

A 45-year-old female patient presented by central chest pain and dyspnea. Recent history of immobilization. Blood pressure 110/60 mmHg and heart rate: 110 bpm. Low risk according to simplified pulmonary embolism severity index score. Non-totally occlusive saddle-shaped hypodense filling defect of pulmonary embolus is seen within the main pulmonary artery trunk. The thrombus is seen extending to the right (a) and left (b) pulmonary artery branches (c) Right to left ventricular ratio >1, indicating right ventricular dysfunction. The calculated pulmonary artery obstruction index for this patient was 50% (intermediate-risk group).

Export to PPT

Close

Figure 5:

A 27-year-old male patient presented by dyspnea and hemoptysis. Blood pressure 110/70 mmHg and heart rate: 105 bpm. Low risk according to simplified pulmonary embolism severity index score. (a) The thrombus is seen within the right middle lobe pulmonary artery (circle) (b) The thrombus is seen extending to a sub-segmental level involving one branch of the right lower lobe pulmonary artery (arrow) (c) Right to left ventricular ratio <1. The calculated pulmonary artery obstruction index for this patient was 22.5% (low-risk group).

Export to PPT

DISCUSSION

Rapid risk stratification and appropriate management are important because of the high mortality rate associated with PE. The ESC 2019 APE guidelines^[1] emphasize that the severity of APE should be based on an individualized assessment of the early death risk associated with APE.^[6] Patients presenting with suspected PE who are hemodynamically unstable (with cardiogenic shock or persistent hypotension) should immediately be stratified as a high risk before any diagnostic workup.^[5] However, patients may lack specific clinical symptoms and may be easily misdiagnosed by clinicians (especially at the beginning of the disease) as acute coronary syndromes or other lung diseases.^[13]

For hemodynamically stable patients, risk stratification, according to the expected PE-related risk of early death, should take place once PE diagnosis is confirmed to ensure appropriate treatment, with the ultimate aim of identifying those patients who may benefit from more aggressive fibrinolytic therapy.^[5]

Conventionally, treatment of submassive PE has been only anticoagulation. However, given the compromise of the RV, other treatments such as systemic low-dose IV thrombolysis and percutaneous interventions such as catheter-directed thrombectomy and thrombolysis may be performed by interventional radiologists.^[5]

The degree of pulmonary vascular obstruction is considered the most important factor determining response to APE. The common methods for PAOI calculation are those used by Mastora embolism index^[14] and Qanadli et al. embolism index.^[12] The evaluation method of Qanadli et al. is widely used for its simple and easy clinical application. In this study, the mean calculated PAOI for patients in our study was 54% ± 18% and ranged between 32.5% and 87.5%. This relatively high PAOI may be due to the referral of most of the patients from emergency, chest, and cardiology departments which led to the inclusion of many high-risk patients.

We detected a high significant correlation between PAOI and sPESI score. Yamashita et al. reported that for patients with sPESI score of 0, the 30-day mortality, recurrent venous thromboembolism, and major bleeding rates were reasonably low.^[15]

A significant negative correlation between PAOI and systolic BP (r = −0.247, P = 0.0002) as well as heart rate (r = 0.185, P = 0.001) in our study were found. The high- risk group patients had a much lower BP, HR, and higher PAOI than non-high-risk patients. This results supported by Jeebun et al.^[16] and Günay et al..^[17] Furthermore, Guo et al.^[13] reported different levels of correlation between PAOI and cardiovascular parameters in APE patients. The right ventricular function was most obvious (r = 0.657, P < 0.05). The clinical hemodynamic parameters, that is, systolic BP and tachycardia are usually strongly related to the size of embolus, degree of vascular occlusion, and anatomical extent on CTPA. This may be the reason why PAOI can predict high-risk patients.

The length of hospital stay and mortality also differed significantly between high-risk and non-high-risk patients (P < 0.0001). Günay et al. found that PAOI was correlated with hospital length of stay which is similar to the results of our study (P < 0.0001 and r = 0.65).^[17] The low PAOI together with other cardiovascular parameters could be useful to identify candidates for early hospital discharge and reducing the risk of hospital-acquired complications.

Researchers found that an increase of PAOI was a risk factor for death in patients with APE. The negative predictive value of death reached 100% when the PAOI was <40%. Acute right-sided heart failure due to increased pulmonary vascular resistance (PVR) is the prime cause of death in massive PE. The rapid rise in afterload causes dilatation of the right ventricle, which, together with systemic hypotension, compromises coronary perfusion, causing ischemia and sometimes even myocardial infarction. The septal shift resulting from the right ventricle dilatation further reduces left ventricular preload, and the patient enters a “vicious cycle” of acute right-sided heart failure.^[18]

On the another hand, Lerche et al.^[19] found that thrombotic obstruction of pulmonary vessels did not correlate well with the severity of PE. This is very unusual, but can be explained by that the mechanisms of heart dysfunction and/or failure in PE are more complex and cannot be explained by only mechanical vessel obstruction. RV pressure overload is predominantly related to the interaction between the mechanical pulmonary vascular obstruction and the underlying cardiopulmonary status. Thus, other comorbidities, age, and medical status of the patient may transform a relatively small embolus, involving limited lung segments, into a clinically serious condition.^[20]

The presence of RVD may indicate a possibly fatal PE.^[14,15] In a meta-analytic study by Vedovati et al.,^[21] they found that increased RV/LV diameter ratio measured on transverse CT images conferred the strongest risk with a 5-fold higher risk for PE-related mortality. In the current work, PAOI was significantly correlated with RV/LV ratio (P < 0.0001, r = 0.482). This finding was similar to Langroudi et al.^[22] who noticed a significant correlation between PAOI and RV/LV ratio in pulmonary thromboembolism. The mechanical obstruction of the pulmonary vascular tree is the main factor in increased PVR, resulting in pulmonary hypertension and subsequent RVD. Additional factors of pulmonary vasoconstriction include neural reflexes, the release of humoral factors from platelets (i.e., serotonin and platelet-activating factor), plasma (i.e., thrombin and vasoactive peptides C3a, C5a), tissue (i.e., histamine), and systemic arterial hypoxia, all of which are associated with increased RV afterload.^[23]

Frémont et al. found that an echocardiographic RV/LV ratio ≥0.9 was shown to be an independent predictor for hospital mortality.^[24] Nevertheless, Lim et al.^[25] found that RVD/LVD ratio of >1 measured on axial sections is indicative of RV strain at pulmonary CTA. Other investigators have proposed a threshold of greater than 1.5.^[26] In this study, for RV/LV ratio >1, ROC curve analysis revealed that the best cutoff value of > 47% for PAOI was associated with the best sensitivity and specificity for predicting RVD. This means that most patients with PAOI above 47% would have RVD.

Attia et al.^[27] reported a cutoff point (>43%) for PAOI to be an independent predictor of worse clinical status and RVD in patients with APE. Rodrigues et al.^[28] found that the best cutoff point of PAOI with great sensitivity and specificity lies between 40% and 49% and values at these levels or greater will identify more than 90% of patients with the right ventricular dilatation. Guo et al.^[13] found that higher PAOI (>52.5%) can predict patients with high risk, for which the sensitivity and specificity were 100% and 81.0%, respectively. Acute RVD would be unlikely when CT obstruction index of <40%.^[21]

There are several studies that have proposed an increased RVD/LVD ratio as a predictor of short-term mortality after PE,^[29,30] although the prognostic value of this sign remains controversial. Kang et al. reported that cardiac chamber measurements are more reproducible than septal bowing and IVC reflux, for establishing RVD in APE as having more limited interobserver agreement.^[31]

In patients without the preexisting cardiopulmonary disease, 25–30% of the pulmonary vasculature must be occluded before the pulmonary artery pressure rises, increasing the RV afterload.^[32] Obstruction of pulmonary vessels more than 30% leads to acute pulmonary hypertension and sudden increase in the right ventricular afterload. This may cause RV dilatation and progression to RV failure. The RV compensates until >50–75% of the pulmonary vasculature is overload is further worsened when hypoxia, induced by the emboli, causes localized vasoconstriction by stimulating the release of vasoactive mediators, such as serotonin, thromboxane, and histamine.^[33] When afterload has reached the critical level, the RV dilates, the LV underfills and decreases supply to the coronary arteries. Perfusion to the right ventricle drops because there is decreased output to the coronary arteries and increased intramuscular pressure impeding the coronary artery flow, leading to the right ventricular ischemia. In severe cases, low right ventricular output results in systemic hypotension and circulation collapse.^[34]

CONCLUSION

Our results further support the use of MDCT pulmonary angiography as a single test for both diagnosis and risk stratification of APE patients. The assessment of pulmonary clot burden through PAOI calculation has a considerable clinical and imaging impact, enabling accurate diagnosis and clot burden estimation. CTPA can be used as an imaging biomarker for the assessment of the right ventricle overload and dysfunction which is associated with increased mortality. This criterion may be of value in selecting cases of submassive/intermediate PE with a poor prognosis who could benefit from thrombolytic treatment.