Friday, June 17, 20223:30 PM - 3:45 PMPO12: Time Investment and Planning Benefit of Serial Treatment Planning MRI Scans for HDR Intracavitary Brachytherapy for Cervix Cancer

Abstract

<h3>Purpose</h3> To estimate the time investment and planning benefit of acquiring treatment planning MRI scans for each insertion for tandem and ring (T&R) cases. <h3>Materials and Methods</h3> Fourteen (14) consecutive cervical cancer patients who underwent intracavitary HDR brachytherapy with T&R applicator were retrospectively analyzed. All patients received 28Gy in 4 fractions delivered in 2 insertions (Ins1 and Ins2), with 2 fractions per insertion, after completion of EBRT. All insertions were performed under general anesthesia with intraoperative CT and MRI guidance. Target (Cx) and normal tissues were contoured on each MRI scan at the time of the procedure. Four patients who did not have both insertions using MRI planning were excluded. Image registration of Ins2 to Ins1 was retrospectively performed in Eclipse (Varian Medical Systems, Palo Alto, CA) based on applicator position. Target contour from Ins2 (Cx2) was copied to the Ins1 structure set (Cx2reg) using this registration. The following metrics were recorded: volume of Cx1, Cx2, and dice similarity coefficient of the two [DSC = 2*(volCx1+volCx2)/(overlap between Cx1 and Cx2reg)] (0 indicates no overlap between the contours; 1 indicates perfect overlap). Target D90 using MRI for each insertion (D90 Cx1 on plan of Ins1 + D90 Cx2 on plan of Ins2) and using only Ins1 MRI (D90 Cx1 on plan of Ins1 + D90 Cx2reg on plan of Ins1). One patient had a replan for the second fraction of the first insertion was therefore excluded from the D90 analysis. Additional time required for MRI-based planning was calculated using the timestamps of the last in-room CT and of the T2 MRI. <h3>Results</h3> Cx1 and Cx2 volumes [median, interquartile range (IQR)] were 26.0 cm³ (IQR 11.3 cm³) and 22.9 cm³(IQR 16.7 cm³). DSC was 0.73 (IQR 0.05). Target D90 (as % of prescription dose) using MRI for each insertion vs. using only Ins1 MRI information for Ins2 was 109.6% (IQR 4.8%) vs. 110.1% (IQR 7.5%). Additional time for MRI-based planning was 38.8 minutes (IQR 6.3 minutes). <h3>Conclusions</h3> The addition of an MRI for Ins2 added only modest alterations in the target volume and D90, however, did add approximately 40 minutes of added time under anesthesia. The projected target coverage if Cx1 was used in insertion 2 instead of Cx2 would be sufficient and could be more efficient in terms of scan and, potentially, contouring time. This study supports the hypothesis that the time commitment required for MRI acquisition for Ins2 may not be justified by a planning advantage. Limiting factors of this study: contouring and fusion uncertainties; our methods did not consider changes in organs-at-risk between insertions; limited number of patients.