Purpose: Multichannel cylinder provides channel optimization to treat an asymmetric vaginal tumor. However, the effect of the accuracy of the position, especially rotation of the MCC on the plan robustness has not been fully investigated.
Methods: We explored three patient cases. In each case, we generated two series of plans: one including the central channel (CC) and 3-4 peripheral channels; the other using only 3-4 outside channels (NC). All the plans met clinical requirements with prescription of 21Gy in 3 fractions. We rotated the MCC around its central axis by ±1°, ±3°, ±5°, ±10°, and ±15° (positive being clockwise while negative meaning counterclockwise) and we generated 20 corresponding plans for each patient (10 for CC and 10 for NC.) We compared the CTV dose coverage and the max point doses for Bladder, Rectum, and Bowels. Results from CC and NC plans were compared and tested by paired T-test.
Results: CTV coverage changes are significant (P=0.01) with rotation direction (clockwise vs counterclockwise) due to asymmetric tumor geometry. The max dose of OARs is less sensitive with small rotation angle (1° to 3°) versus large rotation angle (5° to 15°). Max point dose to Bladder and Rectum have inverse correlation with rotations. Comparing CC and NC plans, the difference between CTV coverage on each rotation angle was insignificant for case 1, 2 and 3 (P=0.794, P=0.942, P= 0.975). The difference in max point doses for Bladder, Rectum and Bowels were also insignificant.
Conclusion: Both CC and NC plans show lack of robustness with rotations for all clinical cases studied. Due to the asymmetric tumors and their locations, the direction of rotation has a significant impact on dose coverage and max dose for OARs. Further investigation including more cases with different tumor locations is needed.