Po25

Purpose To investigate the clinical utility of inverse optimization for developing time-constrained treatment plans for patients with cancers in the gynecologic tract undergoing HDR interstitial brachytherapy. Methods and Materials Between Jan 2021-Dec 2022, 14 patients were treated for cervical (n=6), endometrial (n=5), vaginal (n=2), and urethral (n=1) cancers using a Syed-template and an average of 16±3 titanium needles per ABS guidelines prescribed to both high-risk (HR) and intermediate risk (IR) clinical target volumes (CTV). Plans were retrospectively developed using volume optimization (Eclipse V15.6, Varian Medical Systems) to cover both targets with prescribed doses while minimizing organ-at-risk (OAR) dose to 2cc volumes (D2cc) without a priori knowledge of OAR doses. Dwell time smoothing and minimization of 150% hotspots (i.e., basal dose optimization) were utilized. A 0.5-cm ring around the combined CTV was used to limit hotspots, D1cc, and D2cc in normal tissue. Target coverage was prioritized over other metrics and all plans were normalized to deliver the prescription dose to HR-CTV. Differences in dose metrics from treated plans were used to calculate averages ± standard deviations. EQD2 doses accumulated with external beam radiotherapy (EBRT) were calculated and compared to EMBRACE II constraints as per clinical practice. Paired t-tests were used to assess significance at p<0.05 with Bonferroni correction for multiple testing. Without knowledge of which plan was used for treatment, the clinician visually inspected the 3D dose distributions in both plans, favored one, and noted the characteristics for such preferences. Results Volumes for HR- and IR-CTVs were 42±25cc and 88±51cc, respectively. The total target volume was 116±39cc, with 6 cases having no overlap between the HR and IR-CTVs. Prescribed doses were 2753±417cGy to HR-CTV and 1643±302cGy to IR-CTV delivered in 5 fractions. Optimized plans (OPT) were comparable to treated (T) plans in terms of target coverage and OAR sparing, while reducing target percentage hotspots volumes (V150% and V200%). Paired t-testing indicated significant reductions in target hotspots by 5±5% and D2cc to bladder by 146±107cGy. EQD2 dose accumulated with EBRT exceeded D2cc in 5 treated plans. Although OPT plans continued to violate constraints in these 5 patients, one OPT plan reduced bladder EQD2 to below the constraint while being preferred by the clinician due to enhanced coverage at the junction of HR and IR CTV (see Figure 1 top row). Using blinded review, the clinician preferred the treated plan over OPT in 50% of the cases; for 3 of these 7 cases the clinician preferred lower doses to the urethra (see Figure 1 bottom row), whose constraints were not included in the optimization. Conclusions Inverse planning currently available in treatment planning systems can produce clinically acceptable plans for interstitial brachytherapy for cancers in the gynecologic tract while meeting multiple planning goals including: creating a dose differential between HR and IR CTV, reducing hotspots, and reducing OAR dose. In addition, the process required less manual intervention therefore potentially providing faster convergence to a solution. This may be important in the clinical setting, particularly for time-constrained cases which are implanted and treated on the same day. To investigate the clinical utility of inverse optimization for developing time-constrained treatment plans for patients with cancers in the gynecologic tract undergoing HDR interstitial brachytherapy. Between Jan 2021-Dec 2022, 14 patients were treated for cervical (n=6), endometrial (n=5), vaginal (n=2), and urethral (n=1) cancers using a Syed-template and an average of 16±3 titanium needles per ABS guidelines prescribed to both high-risk (HR) and intermediate risk (IR) clinical target volumes (CTV). Plans were retrospectively developed using volume optimization (Eclipse V15.6, Varian Medical Systems) to cover both targets with prescribed doses while minimizing organ-at-risk (OAR) dose to 2cc volumes (D2cc) without a priori knowledge of OAR doses. Dwell time smoothing and minimization of 150% hotspots (i.e., basal dose optimization) were utilized. A 0.5-cm ring around the combined CTV was used to limit hotspots, D1cc, and D2cc in normal tissue. Target coverage was prioritized over other metrics and all plans were normalized to deliver the prescription dose to HR-CTV. Differences in dose metrics from treated plans were used to calculate averages ± standard deviations. EQD2 doses accumulated with external beam radiotherapy (EBRT) were calculated and compared to EMBRACE II constraints as per clinical practice. Paired t-tests were used to assess significance at p<0.05 with Bonferroni correction for multiple testing. Without knowledge of which plan was used for treatment, the clinician visually inspected the 3D dose distributions in both plans, favored one, and noted the characteristics for such preferences. Volumes for HR- and IR-CTVs were 42±25cc and 88±51cc, respectively. The total target volume was 116±39cc, with 6 cases having no overlap between the HR and IR-CTVs. Prescribed doses were 2753±417cGy to HR-CTV and 1643±302cGy to IR-CTV delivered in 5 fractions. Optimized plans (OPT) were comparable to treated (T) plans in terms of target coverage and OAR sparing, while reducing target percentage hotspots volumes (V150% and V200%). Paired t-testing indicated significant reductions in target hotspots by 5±5% and D2cc to bladder by 146±107cGy. EQD2 dose accumulated with EBRT exceeded D2cc in 5 treated plans. Although OPT plans continued to violate constraints in these 5 patients, one OPT plan reduced bladder EQD2 to below the constraint while being preferred by the clinician due to enhanced coverage at the junction of HR and IR CTV (see Figure 1 top row). Using blinded review, the clinician preferred the treated plan over OPT in 50% of the cases; for 3 of these 7 cases the clinician preferred lower doses to the urethra (see Figure 1 bottom row), whose constraints were not included in the optimization. Inverse planning currently available in treatment planning systems can produce clinically acceptable plans for interstitial brachytherapy for cancers in the gynecologic tract while meeting multiple planning goals including: creating a dose differential between HR and IR CTV, reducing hotspots, and reducing OAR dose. In addition, the process required less manual intervention therefore potentially providing faster convergence to a solution. This may be important in the clinical setting, particularly for time-constrained cases which are implanted and treated on the same day.