Utility of planning MRI in percutaneous thoracic duct embolization for chylothorax.

Percutaneous thoracic duct embolization (TDE) is an accepted treatment for leaks of the central lymphatic ducts. In this study, we correlate the imaging findings on pre-procedural MRI lymphangiography with findings on conventional lymphangiography, and with operator ability to perform a technically successful TDE. The aim was to examine whether MRI is a good screening mechanism to support an invasive procedure in strong candidates, and avert one in poor candidates. MRI and conventional lymphangiograms of 96 patients (62 male and 34 female; mean age 63 ± 11 years, range 29–92 years) were retrospectively reviewed. The diameter and level of the best target for access were assessed for each study. Technical success rates were evaluated with respect to presence of a cisterna chyli, target duct size, and target level concordance. Presence of a cisterna chyli on MRI significantly increased the likelihood of a successful TDE (68% vs. 42%, p = 0.03). Presence of a duct 4 mm or larger, by either modality, significantly improved the chance of successful TDE (for MRI, 65% vs. 41%, p = 0.04; for lymphangiography, 70% vs. 44%, p = 0.03). MRI was not helpful for localizing a lymphatic target, as less than half were seen within one and one-half vertebrae of the predicted level. There was a weak correlation (Pearson coefficient = +0.30) between duct size as measured on the two modalities. 95% of those without an identifiable target on MRI had a viable target on lymphangiography, and successful TDE was performed in 47% of those patients. Identification of a cisterna chyli and/or 4 mm or greater target on pre-procedural MRI indicated higher likelihood of technically successful TDE. MRI did not help predict unsuccessful TDE procedures. Better target level concordance was not associated with improved technical outcomes. • An identified cisterna chyli or a duct measuring at least 4 mm increases technically successful thoracic duct embolization. • Nearly all patients without an MRI target had one on lymphangiogram and successful embolization was performed in about half. • MRI did not predict the location of a lymphatic target and better target level concordance did not improve technical success. [ABSTRACT FROM AUTHOR]