Cardiopulmonary Imaging
Original Research
Resolution of Pulmonary Embolism on CT Pulmonary Angiography
Audio Available | 
Share
OBJECTIVE. The objective of our study was to retrospectively determine the rate of resolution of pulmonary emboli (PEs) in individual vessels and the rate of complete resolution of PEs on CT angiography.
MATERIALS AND METHODS. Follow-up CT pulmonary angiograms, obtained during the period from January 2006 through May 2009, of 69 patients with acute PE from two hospitals were assessed. Initial and follow-up CT angiograms were reread together by one radiologist at both of the hospitals. Images were obtained using a 10-, 16-, 40-, or 64-MDCT angiography unit with a 0.5-mm collimation, 1.25- to 2.0-mm reconstruction, 0.3- to 0.5-second rotation time, and 7.5-mm/rotation table speed. All CT angiograms were obtained using a PE protocol.
RESULTS. Follow-up CT angiograms were obtained in 35 men and 34 women who ranged in age from 17 to 92 years (mean age, 58 ± 17 [SD] years). Complete CT angiographic resolution of PE was seen in six of 15 patients (40%) 2–7 days after diagnostic imaging. After day 28, complete resolution occurred in 17 of 21 patients (81%). The main pulmonary arteries showed complete PE resolution during days 2–7 in seven of nine patients (78%) and after day 28 in 34 of 36 (94%). The lobar pulmonary arteries showed complete resolution of PE during days 2–7 in 23 of 33 patients (70%) and after 28 days in 44 of 48 (92%). The segmental pulmonary arteries showed complete resolution during days 2–7 in eight of 21 patients (38%) and after day 28 in 38 of 38 (100%).
CONCLUSION. Most patients (81%) showed complete resolution of PE on CT angiography after 28 days. PEs resolved faster in the main and lobar pulmonary arteries than in the segmental branches.
Keywords: CT angiography, pulmonary embolism
| Introduction |   |
|---|
Pulmonary emboli (PEs) resolve because of natural thrombolytic processes [1, 2]. Follow-up with CT pulmonary angiography of patients with acute PE gives an opportunity for detailed evaluation of the rate of resolution of PE with only anticoagulant therapy. The focus of this investigation was to retrospectively assess the rate of clearance of PE from individual vessels and the rate of complete resolution of PE on CT angiograms in a subset of patients who underwent follow-up CT pulmonary angiography. This information would be useful in determining the extent to which PE can be expected to be seen in various pulmonary vessels as time passes after an acute episode. It would also be useful in determining the rate at which a PE completely resolves on CT angiograms in patients in whom imaging is delayed or in whom additional CT angiograms are obtained for follow-up.
| Materials and Methods | |
|---|
This is a retrospective collaborative study from two hospitals: St. Joseph Mercy Oakland Hospital in Pontiac, MI, and William Beaumont Hospital in Troy, MI. Imaging of acute PE at William Beaumont Hospital was performed using a 10-, 16-, 40-, or 64-MDCT unit. At St. Joseph Mercy Oakland Hospital, imaging was performed using a 16- or 64-detector unit. Patients with acute PE shown on CT angiograms who underwent follow-up CT pulmonary angiography were included if they did not receive thrombolytic therapy and if the images were of interpretable quality. Only follow-up CT pulmonary angiograms obtained with a PE protocol were evaluated. All patients except one received anticoagulant therapy. The remaining patient underwent surgery for placement of an inferior vena cava filter and did not receive anticoagulant therapy. The reason for additional CT angiograms for follow-up, when stated, was to evaluate continuing or recurrent symptoms or to exclude a recurrence of PE.
At St. Joseph Mercy Oakland Hospital, from October 2006 through May 2009, CT angiograms showed PE in 332 patients, among whom 38 (11%) had additional CT angiograms for follow-up. At William Beaumont Hospital, from January 2006 through September 2008, CT angiograms showed PE in 482 patients, among whom 31 (6%) had additional CT angiograms for follow-up. All patients who, during the period of review, had additional CT angiograms for follow-up using a PE protocol were assessed. In six patients, more than one additional CT angiogram was obtained for follow-up. Only the first follow-up CT angiogram was included in this investigation. The attending physicians selected patients for additional CT angiography follow-up. Approval by the institutional review boards of both hospitals was obtained.
Patients with acute PE were identified by a computer analysis of the hospitals' records. The initial CT angiogram of each patient was reread with the follow-up CT angiogram by a local radiologist in both of the hospitals. Both radiologists had more than 20 years' experience with CT at the time of the study. They had access to the original requisition for an additional CT angiogram for follow-up, but no other clinical or radiologic information. The reviewers read the CT angiogram obtained at the time of acute PE diagnosis first without knowing the results of the follow-up CT angiogram. Then, the follow-up CT angiogram was brought up on a screen for a simultaneous comparison. At St. Joseph Mercy Oakland Hospital, a Centricity RA1000 workstation (GE Health Integrated IT Solutions) was used. At William Beaumont Hospital, an IMPAX workstation (AGFA) was used.
Each CT pulmonary angiogram was interpreted for the presence of complete or partial occlusion of the following vessels: main pulmonary artery, right pulmonary artery, left pulmonary artery, right upper lobe pulmonary artery, right middle lobe pulmonary artery, right lower lobe pulmonary artery, left upper lobe pulmonary artery, lingular pulmonary artery, left lower lobe pulmonary artery, and segmental branches. The resolution of PEs in subsegmental branches was not assessed. We chose to exclude these cases because the positive predictive value of MDCT angiographic readings in patients in whom the largest branch showing PE was a subsegmental branch was only 25% in the Prospective Investigation of Pulmonary Embolism Diagnosis II (PIOPED II) [3].
 View larger version (16K) | Fig. 1 —Proportion of CT pulmonary angiograms of 69 patients showing complete angiographic resolution according to time of follow-up CT angiogram. Number of patients evaluated at each time interval is shown in each bar. |
Complete angiographic resolution was defined as a CT pulmonary angiogram on follow-up that showed no evidence of PE in any vessel. The patient, therefore, no longer showed PE. Complete vessel resolution was defined as a CT pulmonary angiogram on follow-up that showed no PE in the vessel of interest. Other sites, however, may have continued to show PE. If the PE in the vessel of interest did not show complete resolution, there may have been no resolution or incomplete resolution. No attempt was made to characterize the degree of residual thrombi. No attempt was made to assess for the presence of webs or strictures.
If the follow-up CT angiogram showed a new PE, which was defined as a PE in a larger vessel than the PE shown on the initial CT angiogram, the patient was excluded from investigation of complete angiographic resolution and the vessels in question were excluded from investigation of vessel resolution. Contralateral vessels, however, were evaluated for vessel resolution provided that no new PE was seen in a larger vessel.
Contrast-enhanced CT, both for the initial and follow-up CT angiographic examinations, was performed after the injection of 90–135 mL of low-osmolar nonionic contrast material at a rate of 3.5–4.0 mL/s. After an injection-to-scan delay of 4–28 seconds, as determined by bolus-tracking software, or after a 15-mL test injection, CT angiography was performed at 120 kVp and 300–400 mA. The collimation was 0.5 mm with a 1.25- to 2.0-mm reconstruction. The rotation time was 0.3–0.5 second and table speed, 7.5 mm/rotation. The diagnostic criteria for acute PE by CT angiography were those of PIOPED II [3]: first, failure to opacify the entire lumen due to a central filling defect (the artery may enlarge compared with its peers); second, partial filling defect surrounded by contrast material on a cross-sectional image; third, “railway tracking” defined as contrast material between the central filling defect and the artery wall on an in-plane longitudinal image; and, fourth, a peripheral intraluminal filling defect (i.e., the eccentric filling defect makes an acute angle with the artery wall).
The chi-square test was used to assess differences of proportions (InStat version 3.0, GraphPad Software).
| Results | |
|---|
Follow-up CT angiograms were obtained in 69 patients (35 men, 34 women; age range, 17–92 years; mean age, 58 ± 17 years [SD]). The CT angiograms were obtained on day 1 of follow-up in one patient, on days 2–7 in 15 patients, on days 8–14 in 10 patients, on days 15–21 in 11 patients, on days 22–28 in 11, and after 28 days in 21.
Only one patient had a follow-up CT angiogram on day 1 after the initial diagnostic image and no changes were observed. Among the patients who had follow-up CT angiograms on days 2–7, complete CT angiographic resolution was seen in six of 15 patients (40%) (Fig. 1). Among the patients who underwent CT angiograms on days 8–14, complete angiographic resolution occurred in five of 10 (50%). Between days 15 and 21, complete resolution occurred in nine of 11 (82%). Between days 22 and 28, complete resolution was shown in seven of 11 (64%). After 28 days (mean, 83 days; median, 46 days; range, 29–290 days), complete angiographic resolution occurred in 17 of 21 patients (81%).
Complete resolution of PE in the main, left, and right pulmonary arteries of one patient did not occur within 1 day (0 of 3 [0%]) (Fig. 2A, 2B, 2C). During days 2–7, seven of nine (78%) main pulmonary arteries showed complete resolution of PE. From days 8 through 28, complete resolution occurred in 10 of 12 (83%) main pulmonary arteries and after day 28, complete resolution occurred in 34 of 36 (94%). There was no instance in which a PE in the main pulmonary artery resolved but resulted in new emboli in either the lobar or segmental branches. An example of resolution of a PE in the main and left pulmonary arteries is shown in Figure 3A, 3B.
 View larger version (15K) | Fig. 2A —Resolution of pulmonary emboli (PEs) according to vessel and number of days after initial CT angiography. Number of vessels evaluated at each time interval is shown in each bar. Bar graphs show time of follow-up imaging revealing resolution of PEs in main, right, and left pulmonary arteries (A); lobar pulmonary arteries (B); and segmental pulmonary arteries (C). |
 View larger version (14K) | Fig. 2B —Resolution of pulmonary emboli (PEs) according to vessel and number of days after initial CT angiography. Number of vessels evaluated at each time interval is shown in each bar. Bar graphs show time of follow-up imaging revealing resolution of PEs in main, right, and left pulmonary arteries (A); lobar pulmonary arteries (B); and segmental pulmonary arteries (C). |
 View larger version (15K) | Fig. 2C —Resolution of pulmonary emboli (PEs) according to vessel and number of days after initial CT angiography. Number of vessels evaluated at each time interval is shown in each bar. Bar graphs show time of follow-up imaging revealing resolution of PEs in main, right, and left pulmonary arteries (A); lobar pulmonary arteries (B); and segmental pulmonary arteries (C). |
Complete resolution of PE in four lobar pulmonary arteries of one patient did not occur within 1 day (0 of 4 (0%]) (Fig. 2A, 2B, 2C). During days 2–7, 23 of 33 (70%) lobar pulmonary arteries showed complete resolution of PE. From days 8 through 28, complete resolution occurred in 44 of 56 (79%) lobar pulmonary arteries and after day 28, in 44 of 48 (92%) lobar pulmonary arteries. An example of resolution of a PE in the right descending pulmonary artery just beyond the takeoff of the pulmonary artery branch to the right upper lobe is shown in Figure 4A, 4B. There was no instance in which a PE in a lobar pulmonary artery resolved but resulted in new emboli in the segmental branches.
In patients who had follow-up studies on day 1, PEs in the segmental pulmonary arteries were not observed on the first CT angiogram (Fig. 2A, 2B, 2C). During days 2–7, eight of 21 (38%) segmental pulmonary arteries showed complete resolution of PEs. From days 8 through 28, complete resolution occurred in 67 of 96 (70%) segmental pulmonary arteries and after day 28, in 38 of 38 (100%). An example of resolution of a PE in a segmental branch of the left pulmonary artery is shown in Figure 4A, 4B and of a PE in a segmental branch of the right lower lobe in Figure 5A, 5B.
During days 2–7, the PEs resolved at a faster rate in the main or lobar pulmonary arteries than in the segmental branches: 30 of 42 (71%) versus eight of 21 (38%), respectively (p < 0.01) (Fig. 2A, 2B, 2C). However, complete resolution after day 28 occurred in 92–100% of these vessels regardless of the order (size) of the vessels with PE.
 View larger version (123K) | Fig. 3A —70-year-old man with stroke, chronic obstructive pulmonary disease, deep venous thrombosis, and acute pulmonary embolism. Initial CT angiogram shows thrombus (arrow) in main and left pulmonary arteries. |
 View larger version (120K) | Fig. 3B —70-year-old man with stroke, chronic obstructive pulmonary disease, deep venous thrombosis, and acute pulmonary embolism. Follow-up CT angiogram obtained 30 days after A shows complete resolution. |
| Discussion | |
|---|
The proportion of patients showing complete resolution of PEs on the follow-up CT angiograms increased with the time after the initial diagnostic CT angiogram. CT pulmonary angiograms showed complete resolution in 40% of patients on days 2–7, 50% of patients on days 8–14, and 82% of patients on days 15–21. After 28 days, there was no meaningful increase in the proportion of patients in whom complete angiographic resolution occurred, 81%. During days 2–7, PEs resolved at a faster rate in the main and lobar pulmonary arteries than those in the segmental branches, 71% versus 38%, respectively. However, complete resolution after day 28 occurred in 92–100% of the vessels regardless of the order of the vessels.
 View larger version (130K) | Fig. 4A —67-year-old woman with several episodes of dyspnea on exertion. Initial CT angiogram shows acute pulmonary embolism (arrow) with thrombus in right descending pulmonary artery just beyond takeoff of pulmonary artery branch to right upper lobe. In addition, there is thrombus in segmental branch of left pulmonary artery. |
 View larger version (125K) | Fig. 4B —67-year-old woman with several episodes of dyspnea on exertion. Follow-up CT angiogram obtained 3 months after A shows complete resolution of all emboli. |
 View larger version (137K) |
 View larger version (129K) |
We did not assess the rate of PE resolution in the subsegmental branches. Images showing the subsegmental vessels were often unclear, so interpretation was uncertain. The difficulties of diagnosing PEs in the subsegmental vessels on CT angiography and even on conventional pulmonary angiography have been well documented [3, 4]. Kappa statistics for the diagnosis of PE in the subsegmental vessels have been shown to be low on conventional pulmonary angiograms [4]. With this background in mind, we decided not to evaluate the subsegmental vessels for PE.
The results of previous studies in the literature with regard to the complete resolution of PEs on images (CT angiograms, conventional angiograms, and perfusion lung scans) are outlined in Table 1. The data in Table 1 are arranged according to the time interval between the diagnostic image and the follow-up image. We are not aware of any prior studies showing the rate of resolution according to the order of vessels.
Our results related to complete resolution are similar to those of López-Baret et al. [5] (i.e., complete resolution in 91% of patients within 30 days). However, our results showed a higher proportion of patients with complete resolution after 28 days than the results reported by Van Rossum et al. [6] (32%) and Remy-Jardin et al. (48%) [7].
Regarding complete resolution of PE on conventional pulmonary angiograms or CT angiograms, previous investigators, based on pooled data, showed complete resolution after 1–7 days in two of 44 (5%) patients [2, 8–11] (Table 1). In our experience, a larger proportion of patients showed complete resolution in 1–7 days: six of 16 patients (38%). In 10–21 days, previous investigators showed complete resolution in six of 15 patients (40%) [2, 8], whereas our results showed complete resolution in 14 of 21 patients (67%) in 8–21 days. After 1 month, the proportion of patients showing complete resolution varied widely, but on average, PEs had resolved in 85 of 136 patients (63%) [2, 5–7]. PEs showed complete resolution after 28 days in 17 of 21 patients (81%).
The rate of complete resolution on perfusion lung scans is also shown in Table 1. The proportion of patients with complete resolution of perfusion defects within 7–10 days was six of 89 (7%) [12], and based on pooled data, the proportion with complete resolution at 3–6 months was 78 of 240 patients (33%) [12, 13]. After 1–4 years, complete resolution of perfusion scan defects, based on pooled data, was shown in 261 of 382 patients (68%) [1, 14–16] (Table 1). Wartski and Collignon [13] showed that 35 of 77 patients (46%) showed complete PE resolution on perfusion scans at 3 months if the initial pulmonary vascular obstruction was < 50% [13]. If the initial pulmonary vascular obstruction was ≥ 50%, only 17 of 80 patients (21%) showed complete resolution on the perfusion scan. Menéndez et al. [12] showed that the size of residual defects at 6 months depended mainly on the size of defects at 7–10 days. Although the size of the perfusion defect affected the proportion of patients with residual defects [13], our data showed that the size of individual vessels with PE did not affect the proportion of vessels that showed complete resolution after 28 days. This finding suggests that the total number of occluded end-arteries—and not the size of the vessels or order of the vessels—affect the proportion of patients with residual defects.
A strength of our investigation is the number of patients investigated, which is high in comparison with previous investigations. Another strength is that we evaluated the rate or resolution according to the order of the vessel involved, which has not been done previously to our knowledge. There were several limitations of this investigation. The investigation was retrospective. Only a small proportion of the patients with PE had follow-up CT angiograms. Therefore, there may have been bias in the selection of patients for follow-up imaging. A patient who is referred for follow-up CT pulmonary angiography may be at higher risk for residual and recurrent PEs than one who is not, and this bias may have impacted the results in favor of a higher prevalence of residual thrombi. Because the study was retrospective, we were not always able to obtain some clinical information, such as the reason for the follow-up CT angiogram. Only one radiologist interpreted each CT angiogram. We did not look for sequelae of chronic PE (webs, strictures, mosaic perfusion, enlarged central pulmonary arteries, or pruning of the peripheral pulmonary arteries).
A need for additional follow-up studies to more accurately define the rate of resolution after documented PE has been identified [17]. Published rates of resolution differ, and the proportion of patients with residual thrombi also differs. Routine imaging with perfusion scans after cessation of anticoagulant therapy in patients with acute PE to establish a new baseline has been recommended [13]. In patients with persisting thromboembolic obstruction or with persisting cardiopulmonary complaints, one should be alert for the development of chronic thromboembolic hypertension [17]. It had been thought that chronic thromboembolic pulmonary hypertension occurs in only 0.1–0.5% of patients with acute PE [18]; however, more recent data indicate an incidence of 3.8% [19].
Recognition that PEs resolve at a faster rate in the main and lobar pulmonary arteries than in the segmental branches assists in recognizing the difficulties in diagnosing PE if the CT angiogram is delayed several days after the onset of symptoms. This information is also important for interpreting the significance of continuing thromboembolic obstruction when obtaining follow-up studies.
In conclusion, after 28 days, 81% of our patients who underwent follow-up CT angiography showed complete resolution of PE on CT pulmonary angiograms. Emboli resolved at a faster rate in the main and lobar pulmonary arteries than in the segmental branches. However, complete resolution occurred in more than 90% of the PEs in the main, lobar, and segmental pulmonary arteries after 28 days regardless of the order of the vessels.
Address correspondence to P. D. Stein ([email protected]).
| References | |
|---|
