Inverse Planning Simulated Annealing Planning For 3D Image-Based HDR Brachytherapy in Cervical Cancers: A Dosimetric Comparision

Purpose: The objective of this study was to compare the inverse planning algorithms and manual optimization method with the traditional method of prescription to point A in MR image-based dosimetry for HDR intracavitary in cervical cancer. Materials and Methods: Retrospective analysis of MR data sets of 23 patients of cervical cancer consists tandem – ovoid CT/MR compatible HDR applicator were selected for dosimetric study. Fast spin echo T2 weighted MR scans (GE, Signa, Excite, 1.5T) in axial (true), coronal and saggital orientations were taken. Volumes were delineated based on the recommendations of GYN GEC ESTRO. Direct applicator reconstruction was carried out on MR images using of multi-planner reconstruction. Three plans were generated for each patient using 3D treatment planning system (Sunrise, Nucletron): standard clinical plan (SP), manual optimized plan (MOP) and inverse optimized plan (IPSA). SP was based on standard Fletcher loading with dose prescribed to point A, while MOP was carried out by varing dwell times and dwell position. Inverse plan was based on Inverse Planning Simulated Annealing (IPSA) algorithm. The plan objectve of both MOP and IPSA were to obtain optimum HR-CTV coverage while minimizing dose to OARs. A dose of 7 Gy per fraction was prescribed for all three plans. The dose volume constraints used for MOP and IPSA were as follows: HR-CTV: D90(Dose to 90% of HR-CTV) = 7.0 Gy, D2cc (Minimum dose to the most exposed 2cc of bladder) bladder ≤ 6.2 Gy and D2cc of rectum and sigmoid ≤ 4.4Gy. Dose to point-A, TRAK (total reference air kerma), Conformity index (COIN) and dose homogeneity index (DHI) were calculated and documented for all plans. Results: All 23 patients data set were evaluable for the comparison.Tabled 1DVPSPMOPIPSAHR-CTVD90 (Gy)6.3±1.76.0±0.96.0±0.9HR-CTVV100 (%)82.4±8.583.0±5.783.2±5.7HR-CTVV200 (%)38.7±11.736.4±8.141.0±8.1Bladder GyD2cc7.9±1.76.8±1.37.0±1.4Bladder GyD0.1cc11.1±2.99.5±2.19.6±2.3Rectum GyD2cc3.9±0.74.1±1.24.0±1.4Rectum GyD0.1cc5.0±1.15.5±1.85.6±2.1Sigmoid GyD2cc5.5±1.54.9±1.14.8±1.3Sigmoid GyD0.1cc8.0±2.57.1±2.17.4±2.8Abbreviations: DVP=dose volume parameter, All the values given in table of mean with 1SD. Open table in a new tab Abbreviations: DVP=dose volume parameter, All the values given in table of mean with 1SD. The volume of HR-CTV ranged from 10.4-74.2 cc with a mean of 47 cc (±18.6).In summary, both IPSA and MOP resulted in excellent sparing of bladder (p=.000,p=.000) and sigmoid (p=.002, p=.001) without compromising the HR-CTV coverage as compared to SP. Dose to rectum was increased but within tolerance. IPSA and MOP were as homogeneous as SP, with DHI not significant as compared to SP. Similarly IPSA and MOP plans did not result in increase of hot spots, as V200 was not significant as compared to SP. However, dose to point-A was significantly less in MOP(p = 0.011) as compared to SP, while dose to point A in IPSA was comparable to SP. Both IPSA and MOP were highly conformal, COIN was significantly higher in both IPSA (p=.007) and MOP (p = .000) as compared to SP. TRAK for IPSA plan (p = 0.000) was significantly lower as compared to standard plan. Conclusions: IPSA and manual optimized plan resulted in optimal sparing of OARs without compromising the HR-CTV coverage as compared to standard plan. However, as a word of caution, MOP may be operator dependent and is associated with a learning curve.