Clinical applications of MRI in the diagnosis of pulmonary embolism

Abstract

Pulmonary embolism (PE) is a potentially serious medical condition and is the third most common cause of death among cardiovascular diseases. The diagnosis of PE is made with imaging due to its nonspecific clinical signs and symptoms. Historically, pulmonary angiography has been the gold standard until the emergence of computed tomography pulmonary angiography (CTPA), which has now become the go-to modality due to its fast imaging, availability, and high diagnostic accuracy. However, as with pulmonary angiography, CTPA has its drawbacks, such as ionizing radiation and iodinated contrast agents. It is estimated that approximately 20% of patients with suspected PE have a contraindication to this method.

The overall aim of this doctoral thesis can be divided into three parts: 1. To evaluate the diagnostic accuracy of our in-house developed MRI protocol for diagnosing acute PE. 2. To explore other areas where this method could be of benefit since it allows for multiple examinations without any risk to the patients. 3. To understand the emergency doctor's decisions regarding imaging requests for suspected PE.

Studies 1–3 are based on a native standard SSFP sequence under free breathing and without respiratory or cardiac gating, repeated five times at each anatomical position. This is done to catch/image the vessels at different breathing and cardiac cycles. Study 4 is a retrospective analysis of the protocol used by physicians to order CTPA in the clinical setting.

In Study 1, we looked at the one-year outcome of patients who underwent our MRI protocol as the only diagnostic method for suspected PE due to contraindication to CTPA. This study used clinical outcome instead of an imaging modality as the reference. Our results showed that out of 45 patients with a negative MRI result for PE, only one was diagnosed with DVT within three months. In Study 2, we looked at the feasibility of the MRI method for looking at the natural history of acute PE. We examined 18 patients within 36 hours of PE diagnosis with CTPA and then at one week, one, three, and six months. Our results showed that most of the resolution happens within the first few weeks of the treatment. In Study 3, we evaluated the diagnostic accuracy of our in-house developed MRI protocol in reference to CTPA. A total of 243 cases were included, and two radiologists read the MRI exams. Our results showed a sensitivity of 87% and 89% for Readers 1 and 2, specificity of 100% for both, and a kappa value of 0.88.

In Study 4, we looked at PE from a clinician's perspective. Therefore, we retrospectively calculated the clinical decision support system (CDSS) points by extracting data from the Electronic Medical Records (EMR) to examine whether the radiology requests were based on a clinical hunch or the available and recommended clinical decision support systems. Our results showed that clinician bypass these CDSS, which unfortunately leads to lower yield.

To conclude, this doctoral thesis has shown that our MRI method is a viable option for diagnosing PE in patients with contraindication to CTPA and can also be used in treatment follow-up of these patients. However, insights from study number 4 inthis thesis showed that the diagnosis of PE is complicated, and more research is needed to improve the diagnostic accuracy for these patients.