Effect of Contour Accuracy On DVH Parameters
A Zhao1*, A Godley2, (1) Cleveland State University, Cleveland, OH, (2) Cleveland Clinic, Cleveland, OHSU-C-211-1 Sunday 1:30:00 PM - 2:15:00 PM Room: 211
Purpose: Dose Volume Histograms (DVH) are the main assessment of IMRT plans, yet they rely on organ contour accuracy. This study quantifies the change in DVH parameters caused by systematic modification of contours to replicate typical inter-observer and auto-contouring errors.
Methods: Eight prostate patients planned for 78 Gy IMRT with margins of 8 mm, (6 mm anterior and posterior) were selected. Prostate contours were uniformly expanded by ±1 and 2 mm, bladder contours uniformly by 1 and 2 mm and in the posterior and inferior by ±1, 1.5 and 2 mm, and rectum contours uniformly by 1 and 2 mm and anteriorly by ±1 and 1.5 mm. This simulated typical contouring errors due to low contrast at the prostate and bladder/rectum borders. The following DVH parameters were calculated for the rectum and bladder contours: V70 and V60 (% and cc), and prostate contours: D100, D98 and mean dose.
Results: For prostate, the contour error had no substantial effect on the DVH. For bladder, at the1 mm level, the deviations in V70 and V60(%) are below 4.6% and for rectum below 7.1%. The bladder, and particularly rectum, showed dramatic (22-56%) increase in absolute volume (cc) receiving 60 and 70 Gy when contour errors were greater than 1 mm, but this absolute change is <12.4% with errors <1 mm. Because V70 and V60 are initially small in the plan, the <12.4% increase was still clinically acceptable. When a 4 mm (2 mm anterior and posterior) margin plan was assessed, the prostate D100 deviation at 1 mm is still <1%, while bladder and rectum deviation were reduced, implying 1 mm accuracy is a robust target level.
Conclusions: Agreement between contours of 1 mm must be achieved to ensure comparable DVHs. Changes in the DVH occur precipitously above this level.