Clinical variability of target volume description in conformal radiotherapy planning.

Abstract

PURPOSE: The pivotal step in radiation planning is delineation of the target volume and production of a treatment plan to encompass this. This study assesses the variation of physicians in creation of these volumes. METHODS AND MATERIALS: Three radiologists and eight radiation oncologists outlined the gross tumour volume (GTV) on the planning CT scans of four cases with T3 bladder cancer. In addition, the radiation oncologists (RO) created a planning target volume according to a set protocol for all cases. Volumes were produced and comparison of these volumes and the position of the isocenters were analysed. In addition, the margins allowed were measured and compared. RESULTS: There was a maximum variation ratio (largest to smallest volume outlined) of the GTV in the four cases of 1.74 among radiologists and 3.74 among oncologists. There was a significant difference (p = 0.01) in mean GTV between RO and the radiologists. The mean GTV of the RO exceeded the radiologists by a factor of 1.29 with a mean difference of 13.4 cm3. The variation ratio in PTV among oncologists ranged from 1.25 to 3.33. There was no significant difference in mean PTV values between the two groups of ROs divided by specialization in uro-oncology. The mean variation in location of the isocenter from the centroid of the radiologists' volume in the four cases was from 2.6 to 5.7 mm. There was, however, a wide range of values from 1.4 mm to 24.1 mm. Median margin per case ranged from 14.7 to 18.7 mm. Minimum margins allowed in each case varied from minus 7 mm to 9 mm. CONCLUSION: This study demonstrates significant interphysician variability in producing target volumes and radiation plans for conformal radiotherapy. The scale of this difference is clearly of significance, with up to 3-fold variation in volumes delineated by clinicians. The factors leading to these differences will be further addressed. The existence of such variability, however, clearly needs to be accepted as a factor in the overall uncertainty analysis in conformal radiotherapy planning.