Using a virtual model of the lungs, we investigated the nature of the
'stripe sign' which is sometimes encountered in pulmonary scintigraphy
. A model of the segmental anatomy of the lungs was developed from a n
umber of sources and counts generated within the phantom by Monte-Carl
o simulation of photon emission. Multiple segmental and subsegmental d
efects were created in both lungs and submitted for blinded reporting
of the 'stripe sign'. Images were resubmitted for reporting with the c
ontralateral lung removed. The stripe sign was reported in 32 of the 1
17 studies performed. Nearly half of these were present in defects inv
olving approximately 25% of a segment and the sign was most commonly s
een in the lateral projection. Removal of activity from the contralate
ral lung abolished the sign in only 2 of 32 cases. We conclude that sh
ine through of activity from the contralateral lung is a mechanism rar
ely responsible for the stripe sign. Most occurrences of the sign are
due to interposition of activity from unaffected areas of the same lun
g between the defect and the periphery of the lung. Orientation of the
segments, particularly in the lung bases, accounts for the lateral pr
ojection being the most common view in which the sign is present.