• Benjamin Harris, Dale L Bailey, Peter Chicco, Elizabeth A Bailey, Paul J Roach, and Gregory G King.
• Department of Respiratory Medicine, Royal North Shore Hospital, St Leonards, Sydney, Australia. bharris@med.usyd.edu.au
• Clin Physiol Funct Imaging. 2008 Jan 1; 28 (1): 14-26.

PurposeLung scintigraphy using single photon emission computed tomography (SPECT) allows accurate regional measurement of the ventilation/perfusion (V/Q) relationship. Objective V/Q analysis has been shown to be useful in the diagnosis of pulmonary embolism (PE). By using anatomical information provided by co-registered computed tomography, we describe methodology for determining the extent of V/Q heterogeneity at a lobar level. We investigate this methodology using simulated data, and demonstrate its potential application in the clinical setting of PE.MethodsData representing an incremental perfusion defect involving the right lung, together with an unaffected ventilation dataset, were modelled using Monte Carlo simulation. For each increase in the size of the perfusion defect, the whole lung V/Q relationship was objectively determined. In addition, using an image mask of the pulmonary lobes, lobar V/Q relationships were also determined. V/Q heterogeneity was characterized using the log(10) standard deviation of the V/Q ratio (log SDVQR), ventilation (log SDV) and perfusion (log SDQ) distributions. Finally, this methodology was explored in clinical cases.ResultsAs an increasing number of segments were involved by perfusion defects, there was a progressive increase in all objective parameters of V/Q heterogeneity. The relative change was greatest for log SDV. Analysis of both the simulated and clinical studies demonstrated sensitive changes in the lobar V/Q profiles to the presence of PE.ConclusionsSegmentation and analysis of SPECT ventilation-perfusion scintigraphy at a lobar level can be used to quantify regional V/Q relationships. This objective methodology is sensitive to the presence of PE, and may be useful in a clinical setting.

24 keywords...

hide…