Purpose/Objective(s) Radiation therapy (RT) for localized pancreatic ductal adenocarcinoma (PDAC) remains controversial, with conflicting data likely resulting in variable practice patterns. The Canopy Cancer Collective (CCC) is a learning health network of 14 institutions with high-volume multidisciplinary PDAC teams. To characterize the current state of RT for PDAC at high-volume, academic centers, radiation oncology faculty at CCC member institutions were surveyed. Materials/Methods A survey focusing on RT practice patterns for localized PDAC was distributed to 18 faculty with expertise in gastrointestinal RT at CCC sites. Survey results are descriptively reported. Results The survey was completed by 17/18 (94%) faculty. Among respondents, the number of localized PDAC patients treated per year with RT included >50 (n=5, 29%), 26-50 (n=8, 47%), 10-25 (n=3, 18%), and <10 (n=1, 6%). Two (12%) respondents reported that they routinely offer pre-operative RT for resectable PDAC. Majority of respondents (n=15, 88%) indicated that pre-operative RT is offered to certain patients with borderline resectable pancreatic cancer (BPRC), but the specific proportion of BRPC patients receiving pre-operative RT was variable across respondents (median 70%, range: 10-100%). Similarly, majority of respondents (n=16, 94%) indicated that RT is offered as definitive treatment for certain patients with locally advanced pancreatic cancer (LAPC), but the specific proportion of LAPC patients receiving RT was also variable (median 80%, range: 20-100%), as was the estimated proportion of LAPC patients who eventually undergo resection (median 25%, range 2-50%). Table 1 summarizes the distribution of preferred planning techniques and target volumes used by respondents in both the pre-operative and definitive settings. Variation was also seen in use of fiducials (n=12, 71%), preferred motion management strategy (breath-hold: n=12, 71%; gating: n=5, 29%), systematic use of adjunctive imaging beyond CT for target delineation (MRI: n=6, 35%; PET: n=3, 18%), and use of prophylactic anti-emetics (n=5, 29%) and proton pump inhibitors (n=11, 65%). Conclusion RT practice patterns for the treatment of localized PDAC remain disparate, with significant variation in institutional standards across stage of disease as well as in basic treatment parameters such as prescription dose and target volume design. A better understanding of the basis for such variation may help guide future avenues of study.

For select LAPC patients, dose escalation to the target dose of 33Gy in 3 fractions resulted in no DLTs and disease outcomes comparable to conventional RT. These results warrant further exploration of hypofractionated SBRT schemes to maximize tumor control while enabling efficient integration of RT with systemic therapy for more expedient treatment options for these high-risk LAPC patients.

BACKGROUND:Use of neoadjuvant therapy for elderly patients with pancreatic cancer has been debatable. With FOLFIRINOX (folinic acid, 5-fluorouracil, irinotecan, oxaliplatin) or gemcitabine plus nab-paclitaxel (GnP) showing tremendous effects in improving the overall survival of patients with borderline resectable and locally advanced pancreatic cancer, there is no definitive consensus regarding the use of this regimen in the elderly.METHODS:This study evaluated the eligibility of elderly patients with borderline resectable or locally advanced pancreatic cancer for neoadjuvant therapy. Patients registered in the database of pancreatic cancer at the University of Colorado Cancer Center, who underwent neoadjuvant treatment between January 2011 and March 2019, were separated into three age groups (less than 70, 70-74, 75 or more years) and respective treatment outcomes were compared.RESULTS:The study included 246 patients with pancreatic cancer who underwent neoadjuvant treatment, of whom 154 and 71 received chemotherapy with FOLFIRINOX and GnP respectively. Among these 225 patients, 155 were younger than 70 years, 36 were aged 70-74 years, and 34 were aged 75 years or older. Patients under 70 years old received FOLFIRINOX most frequently (124 of 155 versus 18 of 36 aged 70-74 years, and 12 of 34 aged 75 years or more; P < 0.001). Resectability was similar among the three groups (60.0, 58.3, and 55.9 per cent respectively; P = 0.919). Trends towards shorter survival were observed in the elderly (median overall survival time 23.6, 18.0, and 17.6 months for patients aged less than 70, 70-74, and 75 or more years respectively; P = 0.090). After adjusting for co-variables, age was not a significant predictive factor.CONCLUSION:The safety and efficacy of multiagent chemotherapy in patients aged 75 years or over were similar to those in younger patients. Modern multiagent regimens could be a safe and viable treatment option for clinically fit patients aged at least 75 years.

SC-TNT demonstrated a greater rate of cCR compared to LC-TNT, without DFS difference on MVA. Although future studies are warranted to compare SC-TNT with a CRT and consolidative chemotherapy TNT approach, our data support SC-TNT as a suitable regimen for LARC NOM.

Historically, radiation to the thorax has been associated with cardiac toxicity and mortality in breast cancer, lymphoma, and more recently, non-small cell lung cancer. The purpose of the present study was to evaluate whether FDG-PET-based metabolic changes in the heart exhibited a dose response in a cohort of esophageal cancer patients treated with chemoradiation (CRT), and if these changes could be used as an imaging biomarker to predict for overall survival (OS). Furthermore, we sought to validate our previous findings shown in a cohort of lung cancer patients. 28 patients with esophageal cancer treated with radiation who underwent pre- and post-treatment FDG-PET imaging were retrospectively evaluated. Patients were treated with definitive or preoperative CRT to a median radiation dose of 50 Gy (range 30 to 56) in 25 fractions (range 10-28). Post-CRT PET scans were acquired at a median time of 41 days (range 3 to 692 days) after treatment. Pre- and post-treatment PET-CT scans were rigidly registered to the planning CT, dose, and structure set for each patient. Pre-treatment to post-treatment mean Standardized Uptake Values (SUVmean) in the heart were compared to assess dose-response. A dose-response curve was generated by binning each voxel in the heart contour into 10 Gy dose-bins and analyzing the SUVmean changes in each dose-bin. Univariate and multivariate cox proportional hazard models were used to assess whether pre-to-post treatment SUVmean changes predicted for OS. The cardiac SUV dose-response curve demonstrated increasing changes in SUV as a function of dose with relative SUV changes of -4.5%, -0.3%, 5.8%, 12.0%, 12.1%, and 9.4% in the 10, 20, 30, 40, 50, and 60 Gy dose-bins, respectively. The SUV dose-response curve demonstrated an average increase of 3.2% for every 10 Gy (p = 0.02, linear regression). Median follow-up was 19 months (range 4-71). Sixteen of 16/28 patients (57.1%) were alive at last follow-up; these patients had an average increase of +14.3% in cardiac SUVmean while patients that did not survive had an average decrease in SUVmean of -6.2%. The mean pre to post-treatment SUV changes approached significance as predictors of OS on univariate (p = 0.06) and multivariate (p = 0.07) analysis. We found that a cardiac SUV dose-response is demonstrated in esophageal cancer patients treated with CRT, with a trend toward cardiac SUV changes significantly predicting for OS. These data are in-line with previous findings in a lung cancer cohort. Pre- and post-CRT PET imaging is frequently obtained for esophageal cancer patients. If validated, our data show the potential for PET cardiac changes to be an early biomarker predictive of clinical outcomes for esophageal cancer patients undergoing CRT.

Nonoperative management (NOM) is increasingly utilized for patients with locally advanced rectal cancer (LARC) who achieve a complete clinical response (cCR) to neoadjuvant therapy. Close surveillance for tumor regrowth is a mandatory component of NOM but requires a high burden of studies, and limited data exist on patient and provider compliance. This study reports compliance and early outcomes for a single institution cohort. Patients with LARC who had a cCR to neoadjuvant therapy and elected NOM from 4/2015 to 2/2020 were included. An institutional surveillance protocol was developed as follows: 1. DRE, flexible sigmoidoscopy, and CEA every 3 months for 2 years, then every 6-12 months for 5 years. Scar biopsy not required unless suspicion for regrowth 2. MRI pelvis every 6 months for 2 years, then annually for 5 years 3. CT chest/abdomen/pelvis every 12 months for 5 years Compliance was evaluated in terms of follow up retention (rate of patients meeting all indicated follow up visits), patient adherence (number of studies completed/number of studies ordered or recommended), provider adherence (number of studies ordered or recommended/number of studies indicated per protocol), overall compliance (number of studies completed/number of studies indicated per protocol), and safety (rate of salvageable local recurrences). Two-year progression-free survival (PFS) and overall survival (OS) were calculated using the Kaplan-Meier method. Twenty patients were included. All patients underwent neoadjuvant chemoradiation and 16 (80%) received neoadjuvant chemotherapy. Median follow up and time on NOM were 19 months (range, 10-57) and 11 months (1-54), respectively. Seventeen patients (85%) were retained to follow up (1 moved, 2 lost for unknown reasons). Median patient adherence was 100% (25-100), provider adherence was 81% (22-100), and overall compliance was 67% (6-100). Three patients received surveillance at outside institutions and had overall compliance rates of 6%, 47%, and 67%. When excluding these patients, median overall compliance was 80% (44-100). Two patients (10%) had local progression detected by flexible sigmoidoscopy after 3 and 20 months on NOM, respectively. Both underwent salvage surgery with no evidence of disease after an additional 5 and 4 months of follow up. One patient had distant progression detected by elevated CEA and CT c/a/p. PFS was 73% and OS was 100%. Six patients were deemed to require APR at the time of diagnosis, and 5 (of 6, 83%) were spared. In a small cohort with short follow up, this institutional NOM surveillance protocol appears to be safe with acceptable compliance rates, particularly for patients completing all surveillance at this experienced multidisciplinary cancer center. Future work involves education to improve provider adherence and development of a prospective registry to obtain data on long term outcomes, optimization of cCR rates, improved patient selection, QOL outcomes, and cost-effectiveness.

Abdominal motion can increase the risk of toxicity and hinder dose-escalation in Stereotactic Body Radiotherapy (SBRT) of pancreatic tumors. Real-time imaging and tracking is an emerging technique to increase the accuracy of delivery. We report on a large, retrospective cohort of pancreatic patients treated with real-time, fiducial-based kV image guidance. The purpose of our study was to quantify the impact of real-time target tracking in pancreatic SBRT on clinical workflow, treatment accuracy, and tumor dose. 68 patients were treated with pancreatic SBRT under real-time kV image guidance to visualize the location of implanted fiducial markers. Corrections were made to target localization if the markers were observed >3 mm from the expected reference position. To understand impact on treatment accuracy and clinical workflow, we retrospectively analyzed all treatment interruptions and corrections made based on this imaging. To assess the dosimetric impact of the real-time imaging, an artificial neural network dosimetric model was trained with prior clinical plans. The dosimetric impact was evaluated by assessing point dose differences to the tumor resulting from target re-localization triggered by real-time imaging. Real-time imaging resulted in 0.81 pauses per fraction of treatment. 60% of the treatment pauses were due to having to adjust the gating thresholds and 40% were due to having to re-localize the target. The average time per pause was 1.9 ± 1.8 minutes. Treatment pauses that required patient re-alignment due to real-time tumor tracking occurred during 32% of all fractions. The median shifts for patient re-alignment were 0.8 mm (AP), 4.0 mm (SI), and 1.2 mm (LR). The median radial (3D) shift was 5.2 mm. 41% of all patients had at least one shift throughout the course of treatment with magnitude >5 mm, and 16% of all fractions had at least one treatment pause that required an alignment >5 mm. 45% of shifts resulted in dosimetric differences to the tumor; of these, the median point dose difference was 23% ± 22% of prescription dose (max 94%). The number of pauses per fraction was significantly higher in patients treated with respiratory gating (vs. abdominal compression) and in patients with greater treatment time. The current work demonstrated the feasibility of fiducial-based real-time target tracking for pancreatic SBRT treatment, and quantified the benefits of this imaging to increase the accuracy of pancreatic SBRT. This dataset represents, to our knowledge, the largest experience treating these tumors with fiducial marker-guided in-treatment kV imaging. Our data indicate that real-time tumor tracking leads to patient re-alignment in 32% of cases and results in significant dosimetric benefit to target coverage. The increased accuracy of real-time target tracking may potentially enable safe dose escalation in pancreatic SBRT.

The importance of the heart has been emphasized in recent thoracic chemo-radiation (CRT) literature. A lung cancer dose escalation study (RTOG 0617) noted that heart doses were significantly associated with overall survival (OS). The purpose of this study was to characterize pre-to-post treatment cardiac metabolic changes using FDG-PET scans and to evaluate whether PET-based cardiac imaging changes predicted for OS. Lung cancer patients enrolled on a multi-institutional, prospective clinical trial for functional avoidance thoracic radiotherapy were analyzed. 39 patients who had undergone pre and post chemo-radiation FDG-PET imaging were studied. Post-CRT PET scans were acquired at a median time of 97 days (range 11 to 477 days) after treatment. Patients on the study were treated with definitive CRT with radiation doses ranging from 45-60 Gy in 15-30 fractions. For each patient, the pre-treatment PET-CT and post-treatment PET-CT were rigidly registered to the planning CT, dose, and structure set. PET-based dose-response was assessed by comparing pre-treatment to post-treatment Standardized Uptake Values (SUV) in the heart contour. A dose-response curve was generated by binning each voxel in the heart contour into 10 Gy dose-bins and analyzing the average SUV changes in each dose-bin. Patients on the study were followed for OS for up to 14 months after completing chemo-radiation. We evaluated whether the pre-to-post treatment changes in the average or maximum SUV were predictive of OS. The OS analysis was performed by comparing mean SUV changes for patients that were alive or had died at last follow-up and by using a univariate cox proportional hazard model to assess whether pre-to-post treatment SUV changes were a significant predictor of OS. The cardiac SUV dose-response curve revealed increasing changes in SUV as a function of dose with relative SUV increases of 11.4%, 9.8%, 9.4%, 12.2%, 14.2%, and 20.1% in the 10, 20, 30, 40, 50, and 60 Gy dose-bins, respectively. The SUV dose-response curve demonstrated an average increase of 1.7% for every 10 Gy. Median follow-up was 410 days (range 181 to 541) with 30/39 patients alive at last follow-up. Patients that were alive at follow up had an average increase of +17.2% in cardiac SUV while patients that did not survive had a decrease in SUV of -13.5% (p=0.048). Both the mean and maximum pre to post-treatment SUV changes were significant predictors (p<0.03) of OS. Functional imaging has been used for target delineation and evaluating normal lung function in thoracic CRT but has yet to be used for cardiac treatment response assessment. In our dataset, post-treatment changes in cardiac PET were significant indicators of dose-response and potential predictors of OS. Pre and post-treatment PET imaging is frequently obtained for lung cancer patients and if validated by multivariate analysis, our data show the potential for PET cardiac changes to be an early predictor for cardiac death, allowing for intervention.

Stereotactic body radiation therapy (SBRT) is emerging as an attractive treatment option for patients with advanced pancreatic cancer. However, retrospective analysis of the RTOG-9704 phase III trial recently revealed that failure to adhere to radiation therapy protocols was significantly associated with reduced median survival. This suggests that ensuring high quality radiation therapy is crucial, and that future pancreatic radiation therapy trials should include more robust methods for providing patient-specific plan quality validation. This study trains artificial neural network dose models (ANN-DMs) for pancreatic SBRT, and validates their ability to predict patient-specific dose distributions similar to physician-approved treatment plans. Arc-based SBRT treatment plans for 49 pancreatic cancer patients were prepared, delivering 30-33 Gy in five fractions. Treatments were overseen by one of two physicians, each with their own treatment protocol including dosage prescribed, volumes treated, and primary organs-at-risk (OARs). Physician-approved treatment plans were used to train artificial neural network dose models (ANN-DMs) that could predict physician-approved dose distributions based on a set of geometric and plan parameters. For each individual voxel, these parameters were used as inputs to a neural network, and a single output of dose was compared against that voxel’s dose, as calculated by the treatment planning system. Patient datasets were randomly allocated, with roughly 2/3rds used for training and 1/3rd used for validation. Differences between clinical and ANN-DM dose distributions were used to evaluate model performance. Mean dose errors were less than 5% at all distances from the PTV, and mean absolute dose errors were on the order of 5%, but no more than 10%. Dose-volume histogram errors demonstrated good model performance above 25 Gy, but larger errors were seen at lower doses. Remarkable improvements in ANN-DM accuracy (from >30% to <5% mean absolute dose error) were achieved by training separate dose models for each set of OAR constraints and prescription doses. Following one treatment protocol, 6 plans prepared at another institution were used for model validation without being included in model training and no significant increases in model errors were observed. Dose distributions predicted by trained pancreatic SBRT ANN-DMs showed excellent overall agreement with physician-approved dose distributions, and predictive accuracy was substantially improved by developing separate neural network models for differing treatment protocols. Model accuracy for a given treatment protocol was also maintained for patients planned and treated at another institution. For future large-scale trials of pancreatic SBRT, such a model could be trained using consensus guidelines for high-quality plans, allowing for patient-specific plan quality validation in a multi-institutional setting.

Jeffrey Olsen, MD1, Jennifer Moughan, MS2, Robert Myerson, MD, PhD3, Andre Abitbol, MD4, Desiree E Doncals, MD5, Douglas Johnson, MD6, Tracey E. Schefter, MD1, Yuhchyau Chen, MD, PhD7, Barbara Fisher, MD8, Jeff Michalski, MD3, Samir Narayan, MD9, Albert Chang, MD, PhD10, Christopher H. Crane, MD11, and Lisa Kachnic, MD12 1University of Colorado Denver, Denver, CO 2NRG Oncology Statistics and Data Management Center, Philadelphia, PA 3Washington University, St. Louis, MO 4Baptist Hospital of Miami, Miami, FL 5Summa Akron City Hospital accruals for Akron City Hospital, Akron, OH 6Florida Radiation Oncology Group – Baptist Regional, Jacksonville, FL 7University of Rochester Medical Center, Rochester, NY 8London Regional Cancer Program — University of Western Ontario, London, ON 9Michigan Cancer Research Consortium CCOP, Ann Arbor, MI 10University of California San Francisco, San Francisco, CA 11Memorial Sloan Kettering Cancer Center, New York, NY 12Vanderbilt University Medical Center, Nashville, TN

Objectives: The treatment of choice for locally advanced cervical cancer is definitive chemoradiation (CRT). Hysterectomy is not indicated due to higher-rates of cut-through resections leaving gross disease behind, requiring additional therapy with increasing morbidity and no benefit in overall survival (OS). The objectives of this study were to determine factors associated with cut-through hysterectomies and evaluate OS outcomes. Materials and Methods: The National Cancer Database (NCDB) was queried for patients 18 years and older with clinical Federation of Gynecology and Obstetrics stage IB2 to IVA. All patients underwent upfront hysterectomy and had known margin status. Cut-through hysterectomy was classified as presence of microscopic or macroscopic disease at the margin. Results: A total of 11,638 patients were included; 993 (8.5%) had positive margins. In patients with positive margins, 560 (56.4%) received postoperative CRT and 148 (14.9%) underwent postoperative radiation. Five-year OS was worse for those with cut-through resections when compared with those with negative margins, 66.0% versus 86.7%, respectively (hazard ratios, 3.08; P<0.001). Under multiple logistic regression, African American race (odds ratio [OR], 1.45; P=0.001), older age (OR per year increase, 1.03; P<0.001), patients with government insurance (OR, 1.21; P=0.019), and those treated at community practices (OR, 1.31; P=0.001) were more likely to undergo cut-through hysterectomies. Conclusions: A review of national patterns of care over the past decade confirms women with positive margins after hysterectomy for cervical cancer have significantly worse OS. Disparities in surgical results for women with cervical cancer exist. In response, further causality evaluation and corrective action are warranted to address these inequalities.

This phase II trial demonstrated the utility of DP-IMRT in reducing the acute toxicity of 5FU/MMC chemoradiation for anal cancer. This is the report of the secondary endpoints estimating long-term efficacy and evaluating late effects. Patients with T2-4N0-3M0 anal cancer received 5FU/MMC days 1 and 29 of DP-IMRT, prescribed per stage - T2N0: 42 Gy elective nodal and 50.4 Gy anal tumor planning target volumes (PTVs) in 28 fractions; T3-4N0-3: 45 Gy elective nodal, 50.4 Gy ≤3 cm or 54 Gy >3 cm metastatic nodal and 54 Gy anal tumor PTVs in 30 fractions. Disease-free survival (DFS), colostomy-free survival (CFS) and overall survival (OS) were estimated with Kaplan-Meier, and local-regional failure (LRF), colostomy failure (CF) and distant failure (DF) with cumulative incidence methods. Univariate Cox proportional hazards models were used to identify the impact of gender, tumor diameter [<4 vs. ≥4 cm], and T, N and overall stage on DFS/OS, with Fine-Gray regression models used for LRF/DF. Late adverse events (>90 days from the start of treatment) were graded with CTCAE v3.0. Sixty-three patients accrued, with 52 evaluable. AJCC sixth edition stage included: 54% II, 25% IIIA, and 21% IIIB. Median f/u is 7.9 years (0.02-9.2), compared to 6.1 years (0.05-11.8) for the MMC arm of RTOG 9811 (65% II, 15% IIIA, 16% IIIB, 4% unknown). Five and eight year efficacy outcomes are shown in Table 1. On 0529 univariate analysis, female gender significantly predicted for decreased LRF (HR = 0.13, 95% CI = 0.03-0.54; P = 0.0048) and improved OS (HR = 0.33, 95% CI = 0.12-0.91; P = 0.032); tumor size >4 cm was associated with increased LRF (HR = 3.78, 95% CI = 0.78-18.40; P = 0.099) and poorer OS (HR = 2.41, 95% CI = 0.87-6.65; P = 0.089). Worst overall 0529 late effects included: 18% grade 1, 55% grade 2, 16% grade 3, 0% grade 4, 2% grade 5 (sinus bradycardia), and two possibly associated second primaries (myelodysplasia, prostate). Nine thousand eight hundred eleven displayed similar grade 3+ profiles, however, was scored using the RTOG/EORTC Late Radiation Morbidity Schema. Further analysis of 0529 recurrence patterns and late effects will be provided at the annual meeting. Chemoradiation using DP-IMRT for anal canal cancer provided reduced acute morbidity with comparable long-term efficacy and late effects as compared to non-conformal radiation delivery.Tabled 1Abstract 138; Table 10529(n = 52)9811 MMC Arm(n = 325)EndpointTotalEvents5y-%(95% C.I.)8y-%(95% C.I.)TotalEvents5y-%(95% C.I.)8y-%(95% C.I.)LRF816 (7, 27)16 (7, 27)6720 (16, 25)22 (17, 27)CF610 (4, 20)12 (5, 23)3812 (9, 16)12 (9, 16)DF1116 (7, 27)22 (12, 34)4613 (10, 17)16 (12, 21)DFS2068 (53, 79)60 (45, 72)12268 (62, 73)57 (50, 63)CFS1774 (59, 84)66 (51, 77)10672 (67, 77)63 (57, 69)OS1676 (61, 86)68 (53, 79)8778 (73, 83)69 (62, 74) Open table in a new tab

For locally advanced early stage cervical cancer, chemoradiation alone has been shown to have equivalent outcomes with reduced toxicity when compared to upfront surgery with or without postoperative radiation (PORT). In this study, we conduct an analysis evaluating predictors for upfront surgery plus PORT in place of definitive chemoradiation (CRT). The National Cancer Database (NCDB) was queried for patients ≥ 18 years with FIGO IB2-IIB cervical cancer treated from 2004-2012. All patients underwent either upfront hysterectomy followed by PORT or definitive CRT. Logistic regression was used to assess variables associated with modality of treatment (surgery + PORT vs. CRT). Of the 8,574 patients included, 1,287 (15.0%) underwent combined modality therapy with upfront surgery followed by PORT and 7,287 (85.0%) received definitive CRT. Median follow-up was 36.3 months (1-130). The majority were squamous cell (n = 6,683; 78.0%), followed by adenocarcinoma (n = 1,290; 15.0%), adenosquamous (n = 311; 3.6%), and carcinoma NOS (n = 290; 3.4%). Of those undergoing surgery, 292 (23%) had positive margins; 352 (27%) had positive nodes; 5,260 (61.3%) had clinical involvement of the parametrium. Under multiple logistic regression accounting for patient and disease characteristics, rates of surgery + PORT significantly increased between 2004 (7.8%) to 2012 (21.2%) (OR = 1.15; 1.12-1.18; P < 0.001). Upfront surgery followed by PORT was more commonly performed in rural counties (OR, 1.80; 95% CI = 1.17-2.74; P = 0.007), for adenocarcinoma (OR, 2.14; 1.83-2.49; P < 0.001) and adenosquamous (OR = 2.99; 2.28-3.93; P < 0.001) histologies, and at community hospitals (OR = 1.30; 1.12-1.51; P = 0.001). Black women (OR = 0.76; 0.63-0.92; P = 0.005) and those with government (OR = 0.64; 0.56-0.74; P < 0.001) or no insurance (OR = 0.48; 0.40-0.63; P < 0.001) were less likely to undergo surgery. Rates of combined modality therapy with upfront surgery + PORT relative to definitive CRT significantly increased over the past decade (7.8% to 21.2%). Women from rural counties and those receiving treatment at community hospitals were more likely to undergo surgery + PORT. In the setting of level one, evidence supporting the use of definitive CRT alone for these women, the rising rates of upfront surgery raises concern for both higher rates of treatment related morbidity and greater healthcare costs. These findings further support the need for a multidisciplinary approach for women with locally advanced cervical cancer.

Many centers are exploring dose-escalated pancreatic stereotactic body radiotherapy (SBRT). The increased local control of SBRT relies on very precise delivery of high dose, and one of the main challenges is related to respiratory-induced motion. With respiration, the tumor can potentially move out of the focal SBRT field and radiosensitive structures may move into the high-dose region, increasing toxicity. The purpose of this study was to determine the optimal strategy for motion mitigation, and to quantify effect of motion on dose-escalated treatments. Specifically, we compared abdominal compression, respiratory gating, and real-time image guidance in terms of tumor motion mitigation and dosimetric effects. We analyzed imaging data from 38 patients treated with 30-33 Gy to pancreatic tumors in 5 fractions. Imaging data included CBCT scans and real-time kV images acquired during treatment. Motion was mitigated using either abdominal compression or end-exhale respiratory gating. Compression was achieved using an inflatable, indexed compression belt. The daily motion trajectory of pancreatic tumors during CBCT acquisition was reconstructed using an in-house technique that calculated the instantaneous 3D tumor position over time from the position of implanted fiducial markers in each 2D CBCT projection image. The instantaneous 3D trajectories were used both to analyze the effect of motion mitigation strategies, and to reconstruct the true dose received by the tumors. The effects of dose-escalated therapy were simulated by linearly increasing the planned dose distribution and comparing the reconstructed dose in cases with high (>5 mm) vs low (<5 mm) tumor motion amplitude. Compared to compression, respiratory gating significantly reduced the range of superior-inferior (SI) tumor motion during treatment (mean 8.5 to 5.5 mm, range 1.6-12 to 1.5-8.4 mm, p<0.01). In a simulated scenario using reduced SI margins (3 mm), gating increased tumor coverage (mean 88% to 98%, range 36-100% to 90-100%, p<0.01). In dosimetric simulations, using real-time kV imaging to gate the beam based on the position of implanted fiducial markers outperformed traditional end-exhale respiratory gating (tumor D95 99±2% vs 97±7%, p<0.01). In a simulation of dose-escalation, the role of motion on toxicity increased as dose became higher. At doses above 100 Gy BED, high motion contributed over 200% to duodenual NTCP vs low motion. Motion management is critical to the success of dose-escalated pancreatic SBRT. In our data, respiratory gating decreased motion compared to abdominal compression. We found a significant advantage to strategies that measure the position of the tumor in real-time, including real-time kV imaging and gating. As dose to the tumor is escalated, the effect of motion only grows, and enhanced motion mitigation could allow for significant widening of the therapeutic window.