CT diagnostics of pulmonary embolism: Does iodine delivery rate still affect image quality in iterative reconstruction?

1 Introduction

Pulmonary embolism (PE), an obstruction of the pulmonary arteries due to the washout of an embolus from the venous circulation, is a potentially life-threatening event. The gold standard method for diagnosing PE is computed tomographic pulmonary angiography (CTPA) [1–5], although recent advances have shown that ultrasound may serve as an possible alternative [6]. Using iodine-based contrast agent (CM), emboli can be visualized indirectly as filling defects in the pulmonary arteries thus allowing the diagnosis of PE [1, 7] (Fig. 1). To minimize the rate of false-positive and false-negative results, studies reported a cut-off value of 250 Hounsfield units (HU) for a diagnostically conclusive CTPA examination [8–10]. Less contrast-attenuated lung arteries may not allow the diagnosis of PE (Fig. 2).

OPEN IN VIEWER

Fig. 1

Axial CT image of hemodynamically relevant PE in A. pulmonalis dextra and A. pulmonalis sinistra in an axial CTPA image. PE marked by black arrows.

OPEN IN VIEWER

Fig. 2

Axial CTPA images of the pulmonary trunk (marked by star) (A and C) and the lower lobe arteries (B and D) after the injection of 50 ml contrast. The measured mean density values are 959 vs. 180 HU for the pulmonary trunk and 923 HU vs. 156 HU in the segmental arteries of the left lower lobe. The improper image contrast (C and D) result in reduced sensitivity for PE diagnosis.

Various contrast agents with different iodine concentrations and effects are commercially available [4, 11–15]. Previous studies have found that increased injection rate and thus increased iodine delivery rate (IDR) leads to higher mean density values in the pulmonary arteries [3, 16]. The IDR is determined by the contrast agent’s iodine concentration and the injection rate. The injection rate is limited by the patient’s cubital vein catheter and the injector [4, 15–17]. In parallel, various methods such as the use of alternative breathing maneuvers, administration of the CM via the inferior vena cava inflow area, and different contrast injection protocols have already been investigated to optimize mean density values in the pulmonary arteries at the time of image acquisition [3, 18–20]. In recent years a new reconstruction method called “iterative reconstruction” (IR) has been introduced, which can improve computed tomography-based diagnosis [16, 22]. Iterative reconstruction can be used to improve image quality and reduce radiation dose [21–24]. The use of iterative reconstruction is reported to allow dose savings of up to 40%compared to filtered back projection (FBP) without the loss of adequate image quality [22, 24–26].

Based on these recent developments, the aim of this study was to evaluate the subjective and objective effect of higher IDR achieved by high concentrated contrast agent in PE diagnostics using IR.

3 Results

3.2 Objective image quality

The mean density value in the pulmonary trunk was 483.0±166.4 HU in group A and 393.4±132.5 HU in group B (p < 0.001). In all other more peripherally located pulmonary arteries (APS, APD, segmental arteries) as well as the ascending aorta, the use of a high concentrated contrast agent also resulted in significantly increased contrast enhancement (Fig. 3). The p-value was p < 0.001 for AA, APD, APS, ASP, respectively. 6.3%(31/493) of the images in group A and 14.0%(80/572) of the images in group B had mean density values < 250 HU in the pulmonary trunk (p < 0.001).

OPEN IN VIEWER

Fig. 3

Contrast enhancement in the lung arteries and the aorta for both groups. Circle indicates a mild outlier. Star indicates an extreme outlier. Triangle indicates statistical significance (p < 0.05) between the two groups.

4 Discussion

In this study, we found that an IDR increased through the use of a high concentrated contrast agent in CTPA examinations with iterative image reconstruction resulted in a better contrast enhancement of central and peripheral pulmonary arteries. Increasing the iodine dose by 12.5%from 35 g per 100 ml to 40 g per 100 ml resulted in 22.8%greater contrast. This was evident not only in the large, central pulmonary arteries (TP, APD, APS) but also in the small, more peripherally located segmental arteries and in the ascending aorta. Despite the use of IR, an increased IDR significantly decreased the proportion of diagnostically inconclusive images with contrast < 250 HU. In group A 6.3%of the images were diagnostically inconclusive while in group B it 14.0%of the images were diagnostically inconclusive.

The influence of IDR was already demonstrated by Mühlenbruch et al, Keil et al, and Behrendt et al. [11, 28] without IR. They found that increasing the iodine concentration of the CM without increasing the IDR did not lead to any change in the contrast enhancement of the pulmonary arteries. Accordingly, the goal of high contrast in pulmonary arteries must be achieved by optimal IDR. Increasing IDR is possible in two ways: One way is to increase the contrast injection rate. This could increase the risk of CM extravasation and requires a minimum cubital vein catheter size, which is not always feasible in clinical practice. A second possibility is to increase the iodine concentration of the used CM.

Several studies have already investigated the effect of higher IDR on on the contrast in pulmonary arteries. They found that higher IDR resulted in significantly greater contrast of TP, APD, and APS [3, 29]. Commonly cited limitations of these studies are the small case numbers (80–120 consecutive patients). The authors recommended a larger study population to determine the true clinical benefit. Our study confirms the described results with a significantly higher case number. The use of IR instead of FPB in our study did not affect these results.

Most PE are localized at the lobar (upper, middle, and lower lobe arteries) and segmental (segmental arteries and their branches) levels [30]. We also examined the contrast enhancement of the segmental arteries. The reason is that smaller emboli with transient or no symptoms are reported to be overlooked, although they might have a high risk of recurrence and can be seen as a warning sign of larger emboli [2, 31]. Detection of smaller, peripherally located emboli could protect patients from future, life-threatening PE by initiating appropriate treatment early. We demonstrated that IR-based reconstruction had no effect on higher IDR which resulted in significantly higher mean density values of the peripherally located segmental arteries.

To our knowledge the study performed by Kosmala et al. in 2020 is the first study on this topic in which iterative image reconstruction was used [16]. In contrast to our study, CT-scanning was performed with a dual-energy CT scanner. In their retrospective analysis with 198 patients, they demonstrated that high IDR resulted in the strongest contrast enhancement in the pulmonary arteries. In their discussion, the authors pointed out that all patients with diagnosed PEs as well as not further specified artifacts were excluded from the statistical analysis. This meant that 198 of the 340 CTPA examinations performed were analyzed. Transfereability of the study to clinical practise might not be given. We examined 1065 patients on a monoenergetic CT scanner (widely used) and included all patients with PE. Because CTPA is performed primarily to diagnose PE, we consider the inclusion of this patient group as essential to demonstrate the true clinical benefit.

Despite the large number of cases, there was no statistically significant difference between the two groups for the subjectively assessed contrast enhancement of the apical and basal pulmonary arteries. However, we demonstrated by objective measurements that an increase of the IDR resulted in a significantly higher contrast of the peripheral pulmonary arterial pathway.

There are limitations of this study. Because of the retrospective nature of this study, the use of identical peripheral vein catheters cannot be guaranteed. Nevertheless, both our hospital’s standard protocols include a fixed contrast inflow rate of 4 ml/s. The automated injectors adjust the injection pressure to the targeted injection flow rate so that an achievement of 4 ml/s can be assumed. If this could not be guaranteed (technically incorrect bolus tracking), patients were excluded from this study.

In this study, we did not record any possible cardiovascular (arterial hypertension, valvular heart disease, persistent foramen ovale) or pulmonary diseases (pulmonary hypertension) of the patients, which could have had an influence on the cardiac output and thus on the distribution of the contrast agent in the entire pulmonary arterial pathway.

In addition, it was not possible to determine retrospectively whether patients understood the breathing command “Please inhale and hold your breath” and if they were able to sustain it for the full length of the examination. We suspect that some patients in both groups were unable to execute the breathing command correctly because of pain or dyspnea. This may have resulted in nondiagnostic images with mean density values < 250 HU. Due to the comparable patient population this effect should have occurred in both groups; an influence on the results seems therefore unlikely.

5 Conclusion

The use of high-concentration contrast agent at the same injection rate of CM resulted in significantly higher contrast enhancement in all centrally and peripherally located pulmonary arteries in iteratively reconstructed CTPA examination images. In both cohorts, mean density values above 300 HU were achieved. Nevertheless, the number of diagnostically inconclusive images was significantly reduced with increased iodine concentration.