Commentary

Abstract

A constant issue when trying to read the MR of the thoracic spine is to determine the correct spinal level.  Various centres solve this issue in several ways.  The most robust is to include an acquisition that extends up to the cranio-cervical junction.  This adds time and therefore cost to the basic thoracic MR exam.   Other approaches depend recognition of anatomical landmarks to determine the appropriate level.  While there have always been apprehensions about the accuracy of this approach Connor et al’s effort established beyond doubt the unreliability of these landmarks.  As the author’s suggest always image the cranio cervical junction and count down.