Comparison of Spleen Volumes Estimated Using Planimetry and Prolate Ellipsoid Methods on MRI in Myelofibrosis Patients

Background: Splenomegaly is one of the most characteristic features of myelofibrosis (MF) with significant impact on the quality of life of patients. About 10% of MF patients present with severely symptomatic splenomegaly at diagnosis; another 50% develop it within 4 years of diagnosis. Pharmaceutical drug trials aimed at treating myelofibrosis have assessment of spleen size as one of its key trial endpoints. While assessment of spleen size (volume) by palpation is routinely performed in the clinic, unfortunately, it is not a precise and reliable method, a key requirement in drug trials. Medical imaging methods such as magnetic resonance imaging (MRI) and computed tomography (CT) provide a non-invasive approach to assessment of spleen volume. Both modalities allow acquisition of cross sectional abdominal images and enable full anatomical coverage of the spleen. Due to radiation concerns, MRI is preferred over CT modality. There are two approaches for analyzing cross sectional images for estimation of spleen volume - the planimetry method, that involves delineating spleen contours on each spleen slice to estimate complete spleen volume and the prolate ellipsoid method, that approximates the spleen to an ellipsoid structure and estimates spleen volume by using the prolate ellipsoid method (described below). The prolate method is a simple and quick way to assess spleen size compared to the planimetry method. However, the former method provides an approximate estimate of spleen size. The goal of this study was to compare the prolate method with planimetry method for estimating spleen volume for use in pharmaceutical clinical trials.

Methods: A retrospective assessment of spleen volume in N=30 subjects with confirmed diagnosis of myelofibrosis was performed. All measurements were performed on T2 weighted MR images (5mm slice thickness/0.0 gap), acquired on MRI scanners from different manufacturers. All images were assessed for image quality prior to including them in the analysis and images without any images quality issues were included in this study. The analysis methods and calculation used for estimating spleen volume were as below:

• Planimetry: Volume = ∑ [(Region of interest outlining the spleen boundary per slice) x (Slice thickness)].

• Prolate Ellipsoid: Volume = 0.52 x (Longest diameter of the spleen) x (longest perpendicular diameter on the same trans-axial slice) x (vertical distance between the most superior and most inferior margins of the spleen)

Results: Spleen volume estimated by planimetry and prolate method was 2113cc±1612cc and 2186cc±1538cc (mean ± Standard Deviation [S.D.]). The range of planimetry spleen volumes in this study was 210 to 5750cc. A plot of spleen volumes estimated using planimetry and prolate method showed high correlation (R2 = 0.97) between the two methods, however, the volume variation tends to increase for large spleen sizes. The Bland-Altman plot for comparing the two methods showed mean difference in measures (73 cc) and difference variation (1 S.D. =300cc).

Conclusions: Prolate method is a simple and useful alternate method for assessing spleen volume that does not require any specialized software for delineating spleen boundaries. However, the implementation of prolate method is operator dependent and requires precise positioning of digital calipers on images. Future work involves comparing the two methods with respect to change in spleen volume compared to baseline, an assessment performed routinely in pharmaceutical clinical trials.