GSOR02 Presentation Time: 12:05 PM

Purpose Brachytherapy (BT) target dose and coverage are critical to the cure of locally advanced cervical cancer (CC). Advances in imaging, volumetric definitions, and BT planning aims/constraints have collectively been associated with improved tumor control outcomes over traditional practices. But some historic conventions remain in BT, such as applying uniform radiobiological parameters (RbP). Having previously generated a/b and T1/2 from CC cell lines, we explored the potential impact of differing RbP against population-based outcomes in an era of MR-BT boost. Materials and Methods All CC patients seen 1/2013 - 12/2015 in the Canadian province of Alberta (AB), and treated with radical (chemo)radiotherapy (external beam (EBRT) + BT), were identified. From health records, demographic, clinicopathologic, treatment and outcomes data were abstracted. Descriptive and Kaplan-Meier methods were supplemented by Cox-regression multivariable analysis (MVA), factoring in experimentally derived (exp) RbP for BT tumor doses (in addition to treated doses using conventional a/b and T1/2 (10 Gy, 1.5 h)). Results 127 women (median age 50.5 y [23-82]) received curative (chemo)EBRT+BT in AB's 2 tertiary cancer centers (2:1), after exclusions (such as n=12 oligometastatic). Over half (73/125) were smokers. Squamous (SCC) predominated (76.4%) over adenocarcinoma-containing (ADC, 16.5%) or higher-risk (3.1%) histologies, with 3.9% incompletely characterized carcinoma. Primaries were mainly FIGO(2009) clinical stage IIB (37.8%, with the remainder IB 29.9%, IIA 10.2%, IIIA 2.4%, IIIB 15.0%, and IVA 3.9%); 61.4% were node(LN)-positive (pelvic and/or para-aortic (PA)) on imaging. Most (n=102) had weekly cisplatin (median 5 cycles [1-6]) with usual pelvic±PA EBRT (45-50.8 Gy EQD2 in 25-28 fx) and image-based intracavitary BT (67.7% PDR 1-2 fx, vs HDR 1-3 fx; 96.9% MRI; 15% hybrid interstitial). N=12 had initial hemostatic 1-2 fx (5-7.2 Gy); EBRT boost was sequential (3.6-21 Gy in 1.8-4 Gy/fx, to LN/parametria/pelvis) in n=26, simultaneous integrated in n=2. Median overall treatment time (OTT) was 48 days [38-112]. 9 of 127 had no chemotherapy. At BT, median HR-CTV volume was 29.8 cm3 [9.6-212.1], with median treated D90, 88.3 Gy EQD2 [53.4-122.1] (Point A 79.6 [56.3-146.2]). Applying a/b 5.6 Gy and T1/2 3.4 h for SCC (4.3 Gy and 1.0 h for ADC), exp HR-CTV D90 and Point A medians were 102.0 [54.6-177.9] and 86.9 [57.7-225.8] Gy EQD2. Median follow-up was 83 months [4-115], with 5y-OS of 69.3% and pelvic LC, 79.9%. Clinical stage group (p=0.001) and OTT (p=0.008) were associated with OS (LN or chemotherapy status did not reach statistical significance). Exp target dose (Point A, p=0.029; D90, p=0.028) but not treated HR-CTV D90 or volume (trend only) also showed association. For ADC (not SCC), 5y-OS varied by exp, not treated, dose (Point A 79.5% vs 61.2%, p=0.013; D90 54.7% vs 80.3%, p=0.006). On MVA, OTT and exp Point A dose were significant predictors of OS after adjusting for age, chemo, stage, HR-CTV volume and smoking status. Pelvic LC differed, on univariate analysis, by stage (p=0.025), LN status (p=0.021), and exp D90 (p=0.022), with trends observed for OTT and HR-CTV volume but not histology, chemotherapy, BT imaging or center. 5y-LC of ADC also differed by exp Point A (93.2% vs 72.4%, p=0.062) and exp D90 (68.5% vs. 92.3%, p=0.019) doses, but not on MVA. Conclusions Real-world late outcomes for a province-wide CC cohort are reflective of treatment factors, more so than disease burden. Further understanding of RbP is needed for optimum application in CC and outcomes-informed EBRT+BT dosing. Brachytherapy (BT) target dose and coverage are critical to the cure of locally advanced cervical cancer (CC). Advances in imaging, volumetric definitions, and BT planning aims/constraints have collectively been associated with improved tumor control outcomes over traditional practices. But some historic conventions remain in BT, such as applying uniform radiobiological parameters (RbP). Having previously generated a/b and T1/2 from CC cell lines, we explored the potential impact of differing RbP against population-based outcomes in an era of MR-BT boost. All CC patients seen 1/2013 - 12/2015 in the Canadian province of Alberta (AB), and treated with radical (chemo)radiotherapy (external beam (EBRT) + BT), were identified. From health records, demographic, clinicopathologic, treatment and outcomes data were abstracted. Descriptive and Kaplan-Meier methods were supplemented by Cox-regression multivariable analysis (MVA), factoring in experimentally derived (exp) RbP for BT tumor doses (in addition to treated doses using conventional a/b and T1/2 (10 Gy, 1.5 h)). 127 women (median age 50.5 y [23-82]) received curative (chemo)EBRT+BT in AB's 2 tertiary cancer centers (2:1), after exclusions (such as n=12 oligometastatic). Over half (73/125) were smokers. Squamous (SCC) predominated (76.4%) over adenocarcinoma-containing (ADC, 16.5%) or higher-risk (3.1%) histologies, with 3.9% incompletely characterized carcinoma. Primaries were mainly FIGO(2009) clinical stage IIB (37.8%, with the remainder IB 29.9%, IIA 10.2%, IIIA 2.4%, IIIB 15.0%, and IVA 3.9%); 61.4% were node(LN)-positive (pelvic and/or para-aortic (PA)) on imaging. Most (n=102) had weekly cisplatin (median 5 cycles [1-6]) with usual pelvic±PA EBRT (45-50.8 Gy EQD2 in 25-28 fx) and image-based intracavitary BT (67.7% PDR 1-2 fx, vs HDR 1-3 fx; 96.9% MRI; 15% hybrid interstitial). N=12 had initial hemostatic 1-2 fx (5-7.2 Gy); EBRT boost was sequential (3.6-21 Gy in 1.8-4 Gy/fx, to LN/parametria/pelvis) in n=26, simultaneous integrated in n=2. Median overall treatment time (OTT) was 48 days [38-112]. 9 of 127 had no chemotherapy. At BT, median HR-CTV volume was 29.8 cm3 [9.6-212.1], with median treated D90, 88.3 Gy EQD2 [53.4-122.1] (Point A 79.6 [56.3-146.2]). Applying a/b 5.6 Gy and T1/2 3.4 h for SCC (4.3 Gy and 1.0 h for ADC), exp HR-CTV D90 and Point A medians were 102.0 [54.6-177.9] and 86.9 [57.7-225.8] Gy EQD2. Median follow-up was 83 months [4-115], with 5y-OS of 69.3% and pelvic LC, 79.9%. Clinical stage group (p=0.001) and OTT (p=0.008) were associated with OS (LN or chemotherapy status did not reach statistical significance). Exp target dose (Point A, p=0.029; D90, p=0.028) but not treated HR-CTV D90 or volume (trend only) also showed association. For ADC (not SCC), 5y-OS varied by exp, not treated, dose (Point A 79.5% vs 61.2%, p=0.013; D90 54.7% vs 80.3%, p=0.006). On MVA, OTT and exp Point A dose were significant predictors of OS after adjusting for age, chemo, stage, HR-CTV volume and smoking status. Pelvic LC differed, on univariate analysis, by stage (p=0.025), LN status (p=0.021), and exp D90 (p=0.022), with trends observed for OTT and HR-CTV volume but not histology, chemotherapy, BT imaging or center. 5y-LC of ADC also differed by exp Point A (93.2% vs 72.4%, p=0.062) and exp D90 (68.5% vs. 92.3%, p=0.019) doses, but not on MVA. Real-world late outcomes for a province-wide CC cohort are reflective of treatment factors, more so than disease burden. Further understanding of RbP is needed for optimum application in CC and outcomes-informed EBRT+BT dosing.