Currently there are no licensed targeted treatments available for Dengue Fever (DF). An effective therapeutic intervention is needed, to prevent vascular leakage and the development of complications including organ dysfunction, shock and death.

It has been shown that mild DF can cause pleural effusion as a symptom of vascular leakage (1, 2). Fluid accumulation in the pleural cavity therefore has the potential to be used as a method to monitor vascular leakage and, ultimately, to evaluate therapy efficacy. Commonly, pleural effusion assessment is performed with planar chest x-rays (CXRs). This is however a relatively insensitive technique in which fluid accumulation smaller than 50 ml is difficult to measure (3). As MRI is a 3D technique and is highly sensitive to fluids, it has the potential to produce more accurate volume measurements of pleural effusions. In this study we evaluate whether MRI can be used for measuring changes in fluid accumulation in the pleural cavity over time in DF. A comparison is also made with standard CXRs.

30 consenting subjects (mean age 33 years, range 21-58 years, 27 men) participating in a randomized double blind clinical trial of a novel dengue therapeutic were recruited for baseline imaging, within 72h of fever onset. Follow-up scans were performed 4-8 days post baseline scan. MRI imaging was performed on a Magnetom Prisma 3T scanner (Siemens Healthcare) using a T2-weighted HASTE sequence (TE 195 ms, TR 600 ms, FOV 400 mm, voxel size 1.3×1.3 mm2, slice thickness 8 mm, 2 concatenations of 12 s each). CXRs were acquired with anterior-posterior and left lateral views.

Fluid accumulation in the pleural cavity on MRI was measured with an in-house plug-in incorporated into Medical Imaging Interaction Toolkit (MITK), using a semi-automatic seed-growing algorithm. Pleural effusions for which the combined fluid accumulation in the left and right lung was smaller than 1 ml were not considered. The percentage difference in fluid accumulation between baseline and follow-up was calculated and a two-tailed paired t-test was performed. Clinical readings of the CXRs were performed by Singapore board-certified medical practitioners.

The 30 subjects were scanned at baseline and at 5.3±1.1 days post baseline. Fluid accumulation on MRI was seen for 27 subjects (90%) at baseline and 22 subjects (73%) at follow-up. The effusion was bilateral in 26 subjects (87%) at baseline and 20 subjects (67%) at follow-up. The combined (left and right lung) mean pleural effusion size at baseline was 19.7 ml (s.d. 16.7 ml, range 0-67.2 ml), while it was 9.8 ml (s.d. 12.9 ml, range 0-49.6 ml) at follow-up. This difference was statistically significant (p=0.002). For the majority of the subjects a decrease in fluid accumulation was observed between baseline and follow-up (n=24, 80%). An increase in fluid accumulation was seen in 4 subjects (13%). Figure 1 shows the percentage difference observed in fluid accumulation between the two time points. For the CXRs, all 30 subjects were reported as having clear pleural spaces for both time points. Figure 2 shows an example of one subject’s MRI and left lateral CXR images.

These results confirm that fluid accumulation in the pleural cavity is common in dengue fever and that changes in fluid accumulation do occur over the first week of contracting the infection. Overall, the fluid accumulation noted on MRI was below the 50 ml detectability level of left lateral CXRs.

MRI provides a more sensitive measurement of fluid accumulation in the pleural cavity in dengue fever compared to CXR, with significant changes that can be documented over the course of the disease. MRI therefore has the potential to be a marker of vascular leakage in DF and to be used as an outcome measure for evaluation of potential therapeutic strategies.