Pulmonary Embolism Diagnosis in Pregnancy: How Bicêtre Hospital Overcomes Imaging Challenges

Alfredo Cantarinha, Prof. Marie France Bellin|2018-05-17

With a mortality rate at approximately 25 percent, pulmonary embolism (PE) is one of the leading causes of maternal death during pregnancy and postpartum.[1-3] Finding the right imaging modality to assess PE in pregnant women is always a priority with not just one patient to think about, but two. Moreover, foetuses are especially sensitive to dose. Over-diagnosis may also result in treatments that might be dangerous for the unborn child and the mother. Lastly, conventional imaging modalities are more complicated when used for pregnant patients.[4]

When PE is suspected, a computed tomography (CT) scan may be recommended in some countries as a quick response. Diagnosing cases of PE presents particular challenges, however; pregnancy-related physiological changes have the potential to impair image quality to such an extent that a diagnosis may no longer be possible. At Bicêtre University Hospital, France, a special scan protocol has been developed to overcome these challenges and to address the specific needs for PE evaluation for pregnant patients in emergency situations.
 

CT image shows bilateral pulmonary embolism in pregnant woman.
Fig. 1: CTPA of bilateral pulmonary embolism in a 25-year-old patient (28 weeks of amenorrhea).

In diagnosing PE, the examination of choice is CT pulmonary angiography (CTPA). Not only is CTPA usually readily available, it also enables comprehensive assessment of the thorax and visualization of the thrombus to support differential diagnoses (Fig. 1).[5,6]

During CTPA, radiologists and radiographers subjectively evaluate injection quality to check for sufficient contrast enhancement of the pulmonary arteries. Objective evidence would be when a region of interest (ROI) placed at the pulmonary artery indicates a density higher than 250 HU.[5] Although there are articles on pulmonary embolism in pregnant women available, there are few recommendations on specific protocols for CTPA acquisition.[7]

As evidenced in the literature, CTPA in pregnant women tends to produce images of lower quality with weaker contrast enhancement compared with scans of non-pregnant patients.[8,9] Several authors report high rates of non-interpretable examinations: up to 27.5% [7,8] versus 6.4% in the general population.[5]

This reduced quality of CTPA in pregnant women is caused almost entirely by pregnancy related physiological changes that can last for several weeks postpartum.[3] These include increased heart rate and cardiac blood flow, increased blood volume, and high venous return from the inferior vena cava.[2,3,10] The indicated physiological characteristics can cause dilution of the contrast bolus and an increase in pressure in the inferior vena cava (IVC), resulting in an interruption of the contrast bolus through unopacified blood from the IVC.[11] This makes it extremely difficult to perform satisfactory CTPA.

At Bicêtre University Hospital in France, the imaging department addressed this problem by developing a protocol optimized for women during pregnancy and postpartum – and named it the GPS protocol (“grossesse-embolie pulmonaire scanner” from the French for “pregnancy pulmonary embolism scanner”). Researchers conducted a survey to find out whether other imaging facilities had a specific protocol for CT pulmonary angiography in pregnant women and how satisfied they were with the quality of these types of examinations. Results showed that very few institutions had a specific protocol and levels of satisfaction varied widely. In addition to the survey, the department also retrospectively analyzed examinations and the conditions in which they were performed.

The researchers identified four fundamental factors that could improve contrast enhancement of the pulmonary arteries during CT angiography in pregnant women: Prohibition of deep inspiration to prevent the Valsalva Maneuver, low voltage scan, appropriate injection volume and rate, and early acquisition.

As a result, they proposed adjustment of these four parameters to improve pulmonary arterial contrast enhancement and, therefore, image quality. This was termed the GPS protocol, designed specifically for imaging in pregnancy to find “a route to the pulmonary arteries”. In the following, we will take a closer look at these four parameters.

Non-diagnostic images can be reduced by the GPS-protocol from Bicêtre Hospital, France.
Fig. 2: Non-optimized CTPA in an emergency setting of a patient 28 weeks of amenorrhea. With an achieved density of 140 HU measured in the left/ right main pulmonary artery, this CT must be rated as non-diagnostic for evaluation of PE.

1) Avoid deep inspiration to prevent the Valsalva Maneuver
When a patient breathes in deeply and pushes against the abdominal muscles at the point of acquisition, it causes non-iodinated blood to flow through the inferior vena cava (IVC) and into the right atrium. This is known as the Valsalva Maneuver. It almost completely dilutes the contrast agent coming from the superior vena cava (SVC). The IVC flow is stronger than the SVC flow, causing a transient interruption of contrast agent from the SVC.[12-13] This dilution can reduce or even eliminate pulmonary artery enhancement (Fig. 2 and 3).[14] The solution proposed was to ask patients to hold their breath during acquisition without deep inspiration.[12]

2) Manage radiation dose
Acquisition at 100 kV yields higher contrast[15] than an acquisition at 120 kV. Voltage affects image quality, particularly when contrast is involved. When voltage is decreased, the photoelectric effect dominates and contrast increases. Conversely, when voltage is increased, the dominant Compton Effect reduces contrast. Moreover, lowering kV from 120 to 100 also considerably reduces radiation, given that the dose delivered to the patient is proportional to the square of the voltage (at an identical mAs).[16] Consequently, the recommendation is to perform acquisitions at low kV if possible.

3) Manage contrast agent parameters
The researchers from Bicêtre Hospital found out, that appropriate injection parameters for their SOMATOM Definition AS+ scanner need to be adjusted to:
• 80 mL contrast agent or less, if the injection is followed by a second bolus of normal saline at 4 to 6 mL/s; and
• earlier start of the acquisition to compensate for the elevated cardiac flow common in pregnancy [10,11] to achieve an increased heart rate and higher blood volume.

Challenges for PE diagnosis in pregnant women and GPS protocol from Bicêtre Hospital in France.
Overview of potential problems and GPS protocol

4) Early acquisition
Today, a chest CT scan takes just a few seconds. This remarkable speed means that time is now also a key factor in pulmonary imaging. Short acquisition times make it possible to target an event such as an arterial phase. The disadvantage is that if the acquisition window misses the point of maximum enhancement, the arteries will not be fully opacified and it becomes difficult to rule out pulmonary embolism. This is why timing is considered so important:[17] Operators must fully master their machine and adapt each examination to the specific characteristics of each patient. In the case of pregnancy, the increased speed of blood flow requires an earlier acquisition that starts when the superior vena cava is opacified vs the opacification of the pulmonary trunk in the general population.

Comparison of HU in the pulmonary arteries shows better outcome with use of the GPS protocol from Bicêtre Hospital in France.
Fig. 4: Comparison of HU in the pulmonary arteries with (orange) and without (purple) use of the GPS protocol in patient cohort (n=30 in June 2016) and control group (n=37).[18]

Since Bicêtre University Hospital began using the GPS protocol on their SOMATOM Definition AS+, the rate of examinations with non-diagnostic image quality has fallen from 33% to 9.6%.[18] This comes close to the rate of non-diagnostic examinations in the general population (6–7%).[5,7] Contrast enhancement of the pulmonary arteries is significantly higher, making interpretation far easier and giving greater diagnostic confidence (Fig. 4). The dose-length product (DLP) was 108–350 mGy cm, with a mean of 229 mGy cm (vs. a mean of 307 mGy cm before optimization).[18]

Implementation of the GPS protocol when imaging pregnant patients with suspected pulmonary embolism in this case has resulted in better CT image quality and therefore quicker initiation of treatment than before. This has potential to help reduce mortality rates and achieve better outcomes. With the improved diagnostic image quality, PE could also potentially be ruled out faster allowing patients to possibly be discharged sooner and given peace of mind.[18] The protocol can be used with any CT scanner since it is adapted to the physiological characteristics of pregnancy and not to the imaging device. For further information, please refer to the original article: Protocole GPS – Une route vers les artères pulmonaires , by Alfredo Cantarinha, published in Manipulateur d’Electroradiologie Médicale (261: 2017) by the Association française du personnel paramédical d’électroradiologie (AFPPE).


Alfredo Cantarinha, Department of Radiology, CHU Bicêtre, France
Prof. Marie France Bellin, Department of Radiology, CHU Bicêtre, France