Dosimetric Data Research Articles

Post-radiation primary hypothyroidism in patients with head and neck cancer: External validation of thyroid gland dose–volume constraints with long-term endocrine outcomes

BackgroundPost-radiation primary hypothyroidism is a common late complication in head and neck cancer (HNC) survivors. No radiation dose-volume constraint of the thyroid gland has been externally validated for predicting long-term thyroid function outcomes. Materials and MethodsThis external validation study evaluated the diagnostic properties of 22 radiation dose–volume constraints of the thyroid gland proposed in the literature. Radiation dosimetric data from 488 HNC patients who underwent neck irradiation from January 2013 to December 2015 at two tertiary oncology centers were reviewed. The diagnostic metrics of candidate constraints were computed by inverse probability of censoring weighting and compared using time-dependent receiver operating characteristic (ROC) curves with death designated as a competing event. Multivariable regression analyses were performed using the Fine–Gray sub-distribution hazard model. ResultsOver a median follow-up period of 6.8 years, 205 (42.0 %) patients developed post-radiation primary hypothyroidism. The thyroid volume spared from 60 Gy (VS60) had the largest area under ROC curve of 0.698 at 5 years after radiotherapy. Of all evaluated constraints, VS60 at a cutoff value of 10 cc had the highest F-score of 0.53. The 5-year hypothyroidism risks of patients with thyroid VS60 ≥ 10 cc and < 10 cc were 14.7 % and 38.2 %, respectively (p < 0.001). The adjusted sub-hazard ratio for post-radiation primary hypothyroidism for VS60 < 10 cc was 1.87 (95 % confidence interval, 1.22–2.87; p < 0.001). ConclusionThyroid VS60 is the best radiation dose–volume parameter to predict the long-term risk of primary hypothyroidism in patients with HNC who underwent neck irradiation. VS60 ≥ 10 cc is a robust constraint that limits the 5-year primary hypothyroidism risk to less than 15 % and should be routinely employed during radiotherapy optimization.

Dosimetric analysis of patients receiving radiotherapy with VMAT technique in localized prostate cancer and its correlation with side effects.

The aim was to study the relationship between dosimetric data of localized prostate cancer patients who have been treated with curative radiotherapy (RT) and gastrointestinal (GIS), genitourinary (GUS), anal and sexual side effects, and whether there was a difference between dosimetric data and clinical findings between risk groups. Eighty-seven patients who received curative radiotherapy for localized prostate cancer between 2014 and 2019 were included in the study. Dosimetrically; whether there was a relationship between V30, V40, V50, V60, V65, V70, V75 for rectum and bladder; D90 for the penile bulb, V72, V74, V76 for the bulbomembranous urethra, V30, V45, V53, Dmax for the anus, and V45 (cc) for the intestine data and the side effects were analyzed. It was evaluated whether there was a relationship between testosterone values and sexual side effects. The Kolmogorov-Smirnov test, one-way analysis of variance (ANOVA) (F-test), and paired-sample t-test were used as statistical methods. For statistical significance, P < 0.05 was accepted. : The mean age of the patients was 69 (50-86), the mean Prostat specific antigen (PSA) (ng/dL) before RT was 25.1 (0.9-339), the median RT dose was 76 Gy (74-78 Gy), and the mean follow-up period was 38.2 months. PTVmax, PTVmean, PTVmin, bladder V40, bladder V50, rectum V30, rectum V40, rectum V50, and intestinal V45 (cc) were determined as dosimetric data showing differences between risk groups. A statistically significant relationship was found between rectum V30 (P = 0.017), V60 (P = 0.019), V65 (P = 0.008), V70 (P = 0.007), and V75 (P = 0.034) and chronic GIS side effects. G2 GIS side effects were observed in four patients (4.6%) in the entire patient group during the acute period. A statistically significant relationship was found between the patients receiving hormonotherapy (P = 0.021) and testosterone values at the last control (P ≤ 0.001) and chronic sexual side effects. Attention should be paid to the rectum V30, V60, V65, V70, and V75 values to minimize the long-term GIS side effects in patients who have undergone RT. Testosterone level and ADT status affect chronic sexual toxicity.

GSOR3 Presentation Time: 9:10 AM: Adding Free Needles and Optimization to Tandem and Ovoids in High Dose Rate Brachytherapy Treatment for Cervical Cancer: How Much Do They Matter?

<h3>Purpose</h3> With increased MRI utilization, targets for cervical cancer are more customized. This in turn makes hybrid applicators more essential, creating treatment plans which deviate significantly from classical pear-shaped dose distributions. This study analyzes the dosimetric impact of integrating free needles into tandem and ovoid (T&O) HDR-brachytherapy treatment for locally advanced cervical cancer. <h3>Materials and Methods</h3> 29 patients with cervical cancers (FIGO stage IB-IIIC2) were treated with 125 fractions. HR_CTV and OARs (bladder, rectum, sigmoid) were contoured by one physician. All delivered plans were optimized using a previously described ‘two-step' optimization procedure, with the goal of maximizing D90(HR-CTV), minimizing dose to OARs, while maintaining the pear aspect of the Prescription Dose iso-surface. These plans were then compared to two additional plans, created using only T&O, re-optimized to either match the original plans for coverage (D90 to HR-CTV) or for OAR sparing (D2cc). Dosimetric information was extracted from TPS and distributions evaluated using histograms and statistical t-test. <h3>Results</h3> The CTV_HR volume for this series was 53.58±23.46 cm3. The dose coverage of the target for the clinical plans was excellent, at 99.14±6.62 %PD. There was a negative correlation between D90 and HR-CTV, in that larger volume of HR-CTV in HDR brachytherapy cervical cancer cases tends to have an outcome of lower dose coverage (shown in figure 1). When clinical plans, using T&O and free needles, were compared with T&O only plans optimized to match D90, the mean D2cc calculated from plans without needles was higher than those from original plans by an average of 29%, 22%, 27% for bladder, rectum and sigmoid, respectively. A t-test between the two populations, with a p-values < 0.00006 shows that they are significantly different. The highest difference of D2cc between original plans and re-optimized plans were 261%, 294%, and 193% for bladder, rectum, and sigmoid, respectively. When clinical plans were compared to re-optimized T&O only, matching OAR toxicity, the D90(HR-CTV) coverage of the plans without needles, was as expected, lower by an average of 21% for the entire series. The T-test between the two populations had a p-value < 2.2e-16. The maximum percentage difference of the dose coverage was 128% compared between re-optimized T&O only and original plans with matched toxicity. <h3>Conclusions</h3> Free handed needles are an easy to implement, inexpensive add-on to a classical T&O, in order to create a ‘hybrid' applicator. Based on this series of patients and the dosimetric data from their clinical plans, we believe that integrating free-handed needles and optimization methods into traditional T&O cervical cancer brachytherapy treatment is highly beneficial for controlling organs-at-risk toxicity while achieving optimal coverage for HR-CTV and preserving an overall ‘pear'-shape-like aspect of the overall dose distribution. Our findings suggest that the use of additional needles should be considered for cases with a larger or asymmetric volume of HR-CTV. This study, is to our knowledge, the first to quantify the dosimetric benefits of free-needles augmented T&O when compared to classical T&O. We recommend in addition to adding free needles the two step optimization method to best take advantage of the degrees of freedom one has by using hybrid applicators.

Dosimetry and Toxicity Outcomes in Patients Treated with Hypofractionated Regional Nodal Irradiation for Breast Cancer: What is the Best Dose-Volume Limit to Minimize Risks of Radiation Pneumonitis?

Although published data have supported the use of hypofractionated regional nodal irradiation (HF-RNI) for breast cancer, limited dosimetric data exist to evaluate predictors of lung toxicity. The ongoing RT CHARM trial limits the percentage of ipsilateral lung volume that receives ≥18 Gy to 35 to 40%. We assessed dosimetry, toxicity, and disease outcomes in patients with breast cancer treated with HF-RNI with a particular focus on pneumonitis. We retrospectively reviewed all patients with breast cancer treated with HF-RNI (40-43 Gy in 15-16 fractions) after either lumpectomy or mastectomy at The University of Pittsburgh Medical Center from September 2018 to December 2021 to collect dosimetric and outcomes data. All post-radiation therapy chest computed tomography (CT) scans were manually reviewed for evidence of acute (≤6 months postradiation) or chronic (>6 months postradiation) pneumonitis. One-hundred-ninety-one patients qualified with a median follow-up of 20.3 months (range, 5.1-42.2). Acute grade 1 (G1) pneumonitis was observed in 6.8% of the overall cohort (13 of 191 patients) and 39.4% of the patients (13 of 33) who received a chest CT ≤6 months postradiation therapy. Only 1 patient developed acute G2 pneumonitis. Chronic G1 pneumonitis was observed in 29.8% of the overall cohort (57 of 191 patients) and 77% of patients (57 of 74 patients) who received a chest CT >6 months postradiation therapy. No patients developed acute G3+ or chronic G2+ pneumonitis. Rates of symptomatic pneumonitis were low in this cohort of patients treated with HF-RNI, even with integration of HER2/neu-directed therapy, chemotherapy, hormone therapy, and internal mammary nodal irradiation. Lung V20Gy <26% appeared safe in this cohort to limit symptomatic pneumonitis, though this is not meant to represent the safe upper limit. Given the low event rate of symptomatic pneumonitis, data from larger cohorts will be needed to assess dosimetric predictors and the safe upper limit of lung dose.

Tumor Control Probability Modeling and Systematic Review of Oligometastatic Prostate Cancer Treated with Stereotactic Ablative Radiotherapy

<h3>Purpose/Objective(s)</h3> Oligometastatic prostate adenocarcinoma is increasingly treated with stereotactic ablative body radiotherapy (SABR), and treatment efficacy requires achieving adequate local control (LC). Reported dose and fractionation schemes vary and evidence supporting a consensus treatment schema is lacking. We sought to define a dose-response curve based on meta-analysis of reported SABR regimens for prostate cancer oligometastases. <h3>Materials/Methods</h3> A systematic literature review was performed through PubMed using MESH terms "neoplasm metastasis," "radiotherapy," and "prostatic neoplasms." Inclusion criteria included radiologically- or biopsy- identified prostate cancer metastases, histologically confirmed prostate cancer diagnosis, metastasis-directed treatment with SABR, follow-up with radiologic assessment of treated metastases, and reporting of SABR dose and fractionation. Review articles were excluded. Doses were converted to equivalent doses of 2 Gy per fraction (EQD₂) with alpha/beta ratio of 1.5. 90% and 95% tumor control probability (TCP) at 6 months, 1 year, and 2 years was calculated using the Poisson model, with TCP = (1/2)<i>^e</i>^{[2γ50 (1-D/D50)/ln2]} and fit determined by least squares analysis in Proc NLIN of a data management and decision management software. The basic parameters of each model are D₅₀, which is the equivalent dose for a LC rate of 50%; and γ, which is the maximum normalized slope (gradient) of the dose response curve. <h3>Results</h3> We screened 733 studies and 16 studies met inclusion criteria. 1,214 metastatic prostate lesions were treated with SABR. The median EQD₂ was 230 Gy (IQR 202.9-286.7) with a median prescription dose of 30 Gy (IQR 24-33.8) in 3 fractions (range 1-10). There was poor model fit at 6 months and 1 year (p>0.05). At 2 years, the median LC was 95.7% (IQR 93.3%-100%). Chi-square test for our model fit was <0.0001, with estimated D₅₀ of 56.35 (95% CI 12.18-100.50) and γ of 0.30 (95% CI -0.05-0.65). TCP of 90% may be achieved with EQD₂ of 179.0 Gy, and TCP of 95% with EQD₂ of 225.8 Gy. This corresponds to a prescription of 41.1 Gy in 3 fractions or 52.3 Gy in 5 fractions for 90% TCP and 46.5 Gy in 3 fractions or 59.2 Gy in 5 fractions for 95% TCP at 2 years. <h3>Conclusion</h3> Although significant dose and fractionation variation is present, excellent TCP for oligometastatic prostate cancer treated with SABR was reported. Our study was limited by availability of prescription doses, rather than dosimetric data, and an individual patient-level meta-analysis to expand on this work is planned. There are many areas of need for continued research as SABR is increasingly utilized for oligometastatic prostate cancer, including durability of LC with longer follow-up, integration of novel imaging techniques into disease assessment, and the impact of systemic therapy on LC assessment.

Cardiac and Pulmonary Dosimetric Parameters in Lung Cancer Patients Undergoing Post-Operative Radiation Therapy in the Real-World Setting

<h3>Purpose/Objective(s)</h3> The recently published Lung ART trial reported increased rates of cardiac and pulmonary toxicity in the post-operative radiation therapy arm. It remains unknown whether the dosimetric parameters reported in Lung ART are representative of real-world practice. The purpose of this study is to examine heart and lung dose exposure in patients receiving post-operative radiation therapy for non-small cell lung cancer (NSCLC) across a statewide consortium. <h3>Materials/Methods</h3> From 2012 to 2020, 377 patients at 27 academic and community centers within the Michigan Radiation Oncology Quality Consortium (MROQC) underwent surgical resection followed by post-operative radiation therapy for non-metastatic NSCLC. Demographic and dosimetric data were prospectively collected for these patients. Rates of 3D-CRT and IMRT use were analyzed. Mean heart dose (MHD), heart V5, heart V35, mean lung dose (MLD), lung V20, target volume and minimum dose to 95% PTV were calculated for these patients and the reported dosimetric parameters were stratified by treatment modality. <h3>Results</h3> 51% of patients in this cohort had N2 disease at the time of surgery, 18% had a positive margin. 65.8% of patients were treated with IMRT compared to 32.1% treated with 3D-CRT. Average MHD for all patients was 10.3 Gy, mean Heart V5 was 40.3% and mean heart V35 was 12.6%. Average MLD was 11.2 Gy and mean lung V20 was 18.9%. These dosimetric parameters did not significantly differ based on treatment modality, with MHD and MLD 9.9 Gy and 10.1 Gy, respectively, for patients treated with 3D-CRT compared to 10.6 Gy and 11.8 Gy for patients treated with IMRT. <h3>Conclusion</h3> Cardiac and lung dosimetric parameters for patients receiving post-operative radiation therapy for NSCLC are similar to the dosimetric characteristics reported in Lung ART. The mean heart and mean lung doses observed are slightly lower (MHD 10.3 Gy, MLD 11.2 Gy) compared to Lung ART (MHD 13 Gy, MLD 13 Gy), possibly owing to increased use of IMRT. These data support application of Lung ART's findings outside of the clinical trial setting.

Development of a Support Vector Machine-Based Tool for Survival Prediction after Stereotactic Body Radiotherapy for Liver Metastases Using Multi-Institutional Data

<h3>Purpose/Objective(s)</h3> Stereotactic body radiotherapy (SBRT) has been increasingly used as well-tolerated, effective treatment for metastatic liver lesions. Machine learning can predict outcomes and aid in developing treatment regimens; however, few studies have attempted to predict liver SBRT outcomes and most have been limited to a single institution. Support vector machine (SVM) algorithms can address complex classification by mapping data into multi-dimensional space. This study aims to develop an SVM-based model to predict survival after SBRT for liver metastases using multi-institutional data. <h3>Materials/Methods</h3> The Comprehensive Liver Pooled SBRT Outcomes (CaLiPSO) is an international consortium of four high-volume SBRT academic centers. Patients who received SBRT for liver metastases were identified from the CaLiPSO database; baseline demographic, clinical, and dosimetric data were collected. Outcomes of interest were overall survival at 12, 18, and 24 months (12mo-OS, 18mo-OS, and 24mo-OS). We investigated a SVM algorithm using a Gaussian kernel and 10-fold cross-validation schema. Performance was measured by accuracy defined as the ratio of correct classifications to samples and by computing the area under the receiving operator characteristic curve (AUC of ROC). Training was performed using a programming environment. <h3>Results</h3> A total of 132 patients with metastatic liver lesions treated with SBRT were included in this study, 77 females (58.3%) and 55 males (41.7%) with median age of 65 years (range 27-88). The median 2-Gy Equivalent Dose (EQD2) was 60 Gy. At the time of SBRT, 62 patients (47.0%) had extrahepatic disease and 70 (53.0%) did not. The number of patients alive at 12, 18, and 24 months were 64 (48.5%), 52 (39.4%), and 41 (31.1%), respectively. Patients were randomly divided into training (n=92) and testing (n=40) sets. Predictive factors included in the SVM were age, EQD2, planning target volume (PTV), and percentage of normal liver volume. Training set AUCs were above 0.95; results for the testing set are listed in Table 1. Performance at 18 and 24 months had AUC of 0.74 and 0.68, respectively, and was improved compared to at 12 months. <h3>Conclusion</h3> Using clinical and dosimetric variables, promising results were obtained for accurate prediction of liver SBRT outcomes. To our knowledge, this is the first SVM-based tool for SBRT outcome prediction in liver metastases validated using international, multi-institutional data for over 100 patients. This tool can help improve care and facilitate appropriate treatment escalation; it can be implemented online for the radiation oncology community as it uses commonly collected data and is easy to compute. The tool can be further refined and validated with larger cohorts in prospective trials.

Contrast-Enhanced 4D MRI for Internal Target Volume Generation in Treatment Planning for Liver Tumors

<h3>Purpose/Objective(s)</h3> Treatment planning for liver tumors is challenging, as they have variable respiratory motion and are often invisible on four-dimensional commuted tomography (4DCT), even with use of intravenous contrast. Thus, imperfect imaging surrogates are used to estimate the motion of tumors. 4D magnetic resonance (MR) with hepatobiliary contrast has the potential to allow for direct visualization of tumors through the respiratory cycle. Thus, we assessed multiple reconstruction algorithms for directly visualizing liver tumor motion for radiotherapy planning and compared treatment plans with those generated using 4DCT. <h3>Materials/Methods</h3> Patients were simulated using a 3T MR and CT scanner. Three prototype binning algorithms (phase, amplitude, and two-directional) were applied to the 4DMRs. Phase and amplitude binning averaged signal from both the inhale and exhale trajectories for theoretically improved signal to noise ratio, while two-directional binning separated trajectories for theoretically improved spatial resolution. Image quality for tumor delineation was assessed using a qualitative clarity score and quantitative sharpness score. Radiation plans were generated for internal target volumes (ITVs) derived using 4DMR and 4DCT, and the dosimetry compared. Paired t-tests were used to compare sharpness scores and dosimetric data. <h3>Results</h3> Twelve patients with 17 liver tumors were scanned between May and November 2021. Compared to phase binning, two-directional binning demonstrated equal or better clarity and sharpness scores (end-expiration: 0.33 vs. 0.38, p=0.018, end-inspiration: 0.28 vs. 0.31, p=0.010). Compared to amplitude binning, two-directional binning captured hysteresis (anterior displacement of the tumor during respiration) of ≥3 mm in 35% of patients. Dosimetric evaluation of 4DCT-derived plans revealed resulting underdosage of the PTV where the 4DCT did not capture the full extent of 4DMRI-defined PTV, (PTV receiving 90% of prescription: 75.56% vs. 89.38%, p=0.002). <h3>Conclusion</h3> Using contrast-enhanced 4D MRI is feasible for directly delineating liver tumors throughout the respiratory cycle. Two-directional binning capture hysteresis in 1/3 of patients. The current standard of creating radiation plans optimized for 4DCT-derived targets achieved lower coverage of MRI-derived targets, suggesting that adopting MRI for motion management may improve radiation treatment of liver lesions and reduce the risk of marginal misses.

Bone Marrow Toxicity and Tolerance of Sandwich Therapy for Advanced-Stage Endometrial Cancer: A Single-Institution Experience

<h3>Purpose/Objective(s)</h3> Adjuvant therapy for advanced-stage endometrial cancer (EC) involve chemotherapy (CT) and/or radiation therapy (RT). Institutional practices vary regarding the sequencing of these therapies. While recent trials favored concurrent chemoradiation followed by consolidating CT, there is no randomized comparison of efficacy between ‘sandwich' (ST) versus concurrent CT. Previous studies show BM dose is predictive of hematological toxicities (HT) in concurrent chemoradiation. In this study, we sought to determine whether bone marrow dose contributes to hematological toxicity in ST for treatment of advanced-stage EC. <h3>Materials/Methods</h3> Data for patients who received ST (CT with carboplatin/paclitaxel then 45 Gy external beam radiation therapy [EBRT] with/without paraaortic [PA] fields then additional CT) at a single academic institution for EC cases was abstracted. Full ST was defined as 3 CT cycles, then RT, followed by a final 3 CT cycles. Grade 3 (G3) higher hematologic toxicities (ANC <1000, Hgb <8.0, platelet count <50,000) were recorded for each patient based at the time of infusion or radiation therapy. T-test and Chi-square test were used as appropriate to determine association. <h3>Results</h3> 98 EC patients that underwent ST were included; 66.0% were stage IIIC. 84% of patients completed full ST, and 26.3% of cases experienced ≥1 treatment delay. Among cycle delays cases, 57.7% had ≥1 G3 hematologic toxicity: 10 neutropenia, 4 anemia, and 3 thrombocytopenia. HT rate was 49.0%: 34% neutropenia, 13% anemia, 12% thrombocytopenia. Frequency of cycle delay (31.3% vs. 21.6%) or ST non-completion (13.7% vs. 18.4%) were not significantly associated with HT presence (p>0.05, all). Significantly fewer patients who had HT both pre- and post-EBRT completed post-sandwich CT compared to toxicities occurring at other time intervals (Table). While patients with ≥2 HT during their treatment received a significant larger mean bone marrow radiation dose (69.0 vs 72.5%, p=0.037), overall dosimetric data were similar regardless of HT timing during treatment. PA boost did display more frequent HT following RT (21.1% vs. 51.7%, p = 0.033) but completed ST at similar rates (63.2% vs 36.8%, p=0.441). There was a 10% mortality rate; overall survival was similar despite presence of HT occurrence, or with HT requiring a cycle delay or cycle non-completion (p>0.05, all cases). <h3>Conclusion</h3> In our study, BM dose does matter in determining HT even in ST, but does not affect overall treatment completion. The treatment completion rate is comparable to those reported historically for CT only arm. Low rates of cycle interruption and similar overall survival point to a promising safety and efficacy profile with ST.

Radiation-Induced Acute Cardiotoxicity in Women with Left Sided Breast Cancer: Results from a Phase II Prospective Clinical Trial

<h3>Purpose/Objective(s)</h3> Women with left sided breast cancer (BC) who receive adjuvant radiotherapy (RT) have increased risk of subsequent cardiac events, which may be mitigated by maintaining mean heart dose (MHD) below a threshold of 3 Gy. We hypothesized that early MRI evaluation of cardiac functional changes can identify acute cardiac insult and predict for the risk of long-term cardiac complications. Here we report the findings of a phase II prospective clinical trial evaluating RT-related LV structure and function using cardiac MRI in women with left sided BC treated with adjuvant RT. <h3>Materials/Methods</h3> Patients with left sided BC were enrolled into a single institutional phase II study. After informed consent, patients were allocated for breast RT with free breathing (FB) or with deep inspiratory breath hold (DIBH) delivery, if the predicted MHD was ≤3 Gy (≤ 5 Gy, if also regional nodal RT) or > 3 Gy (> 5 Gy, if regional nodal RT), respectively, on initial FB planning. Cardiac MRI was performed at baseline (pre-RT), and at 16 and 52 weeks. Cardiac toxicity was defined as change in the indexed LV end diastolic volume (LVEDVi), LV ejection fraction (LVEF), indexed LV end systolic volume (LVESVi) and indexed LV mass (LV MASSi). Estimation of multiple serum cardiac biomarkers were also done at the three time points. Continuous variables were expressed as mean ± standard deviation, and ANOVA analysis was used to identify correlation between changes in cardiac parameters from baseline to 16 and 52 weeks, when compared to cardiac radiation dosimetric data. <h3>Results</h3> Of the 69 participants, 25 received RT with FB technique. 44 received RT with DIBH, of which 23 also had received adjuvant chemotherapy (CT) prior to RT. For the FB, DIBH (no CT), and DIBH (with CT) cohorts, MHD was 1.70 ± 0.93 Gy, 1.93 ± 1.30 Gy, and 2.26 ± 1.15 Gy and mean LV dose was 2.47 ± 1.48 Gy, 3.18 ± 1.93 Gy, and 3.23 ± 1.82 Gy, respectively. Evaluation of LVEDVi, LVEF, LVESVi, and LV MASSi showed no statistical difference between the 3 treatment arms. No statistical correlation was noted between LVEF changes (≥ 5% and ≥ 10% from baseline) and the cardiac dosimetric data; all p values > 0.1. Prior CT did not measurably impact cardiac function. <h3>Conclusion</h3> In this prospective study, no significant post-RT cardiac events were demonstrable by functional imaging. This supports the significance of maintaining MHD ≤ 3 Gy to minimize or prevent RT related cardiac toxicity in women with left sided BC. Our observation warrants further investigation in a large prospective setting to evaluate late cardiac events. Analysis of serum cardiac biomarkers is currently ongoing.

Do HDR Tandem and Ring Interfractional Displacements Predict the Need to Re-Plan?

<h3>Purpose/Objective(s)</h3> To evaluate the effect of changes in HDR tandem and ring applicator position on doses to nearby organs at risk. <h3>Materials/Methods</h3> CT imaging studies (29) and dosimetry of 6 women ages 44 to 86 undergoing tandem/ring based HDR for cervical cancer were used retrospectively. Measures of applicator positional changes (translation and rotation) relative to the pelvis were obtained using image fusion in a commercial image analysis software. Dosimetric data were calculated in the commercial image analysis software using the original (first fraction) dose distribution applied to organs at risk (rectum and bladder) via deformable registration of the original contours. The new contours were examined visually for appropriateness. Differences in translation and rotation between the planning CT and subsequent individual fractions were recorded and plotted against dose changes between each fraction of treatment and the original (first) fraction. Percent differences in positional shifts vs dosimetric changes were calculated and a correlation coefficient between dose parameters and measured positional shifts of the applicator was determined. <h3>Results</h3> Dosimetric parameters for rectum and bladder (D_2cc and V₅₀) calculated using a commercial treatment planning system for HDR brachytherapy and commercial image analysis software gave similar results within 1% of the original plan (Fraction 1 of 5). A summary of positional changes and average and maximum dose changes for each dosimetric endpoint is found in Table 1. Average percent difference between planned dose and fractional doses obtained through deformable registration were 12%, 16%, 4% and 49% for bladder D_2cc, Bladder V₅₀, Rectum D_2cc and Rectum D₅₀, respectively. Correlation between dose parameters and measured positional changes for bladder D₅₀ and displacement and bladder D₅₀ and sagittal plane rotation gave an r² of 0.3. Applicator displacement of approximately 1.6 times greater than average applicator displacement across all plans studied have shown a significant increase in bladder dose. Bladder dose increased by a maximum of 30 Gy and 54.5 Gy for bladder D_2cc and bladder V₅₀, respectively for the patient with the largest catheter shifts. <h3>Conclusion</h3> We have quantified the effects of applicator positional changes on dose differences for bladder and rectum. Even large changes did not result in significant changes in dose to organs at risk, meaning adaptive re-planning is not required in most cases below a certain displacement threshold.

Initial Clinical Experience of CBCT-Based Adaptive Online Radiotherapy for SAbR of Thoracic Malignancies

<h3>Purpose/Objective(s)</h3> There has been an increasing trend in the availability and utilization of cone beam computed tomography (CBCT)-based adaptive online radiotherapy (AORT) in radiation oncology clinics. In this study, we report our initial experience of utilizing AORT to deliver stereotactic ablative radiation therapy (SAbR) for either early-stage non-small cell lung cancers (NSCLCs) or metastases to the thorax. <h3>Materials/Methods</h3> We recruited 18 patients treated with daily AORT whose tumors were treated with a total prescribed dose ranging from 33-60Gy treated over 3-5 fractions. During the AORT fraction, the physician contoured an ITV and critical OARs on the CBCT and two plans were created for the physician to choose from, a scheduled plan and an adaptive plan. The scheduled plan was a re-calculation of the reference plan based on the anatomy of the patient on that day's fraction while the adaptive plan was a re-optimization of the reference plan. For each scheduled and adaptive plan, we recorded PTV and ITV coverage (V100%), dose conformality (R50 and R100), and the maximum dose (D_max) received by the esophagus, bronchus, trachea, heart, and spinal cord. Timing data for each fraction from CBCT acquisition to approved plan and patient ready for treatment were also recorded. Dosimetric data between the scheduled and adaptive plans were compared relative to the scheduled plan. Paired t-tests were used to evaluate differences between the scheduled and adaptive plans. <h3>Results</h3> Among the 68 fractions analyzed, there was a reduction in ITV volume over the course of the adaptive treatments in 54.4% of cases (range: 0.7%-82.4%, mean: 24.3%), with physicians selecting the adaptive over scheduled plan 88.2% of the time. Relative to scheduled plans, mean difference in PTV and ITV coverage was 2.99% (p=0.009) and 1.07% (p=0.005) higher, respectively. Dose conformality, R50 and R100, were improved in 62.5% (p=0.05) and 66.7% (p=0.098) of adaptive compared to the scheduled plans. Mean difference in D_max relative to the scheduled plan for critical OARs ranged from -5.62% to 4.03%, only heart and spinal cord showing significance (p=0.004, p=0.023). Timing data demonstrated an average of 9 min spent on contouring, 5 min on plan optimization, and 5 min for online quality assurance. <h3>Conclusion</h3> A significant proportion of the patients in our SAbR-treated CBCT-adaptive therapy cohort demonstrated ITV volume changes during overall course of treatment. AORT-based lung SAbR treatments significantly improved PTV/ITV prescription coverage (compared to scheduled plan deliveries), dose conformity, and OAR sparing, with <15 min added to the overall treatment time for online contour and reoptimization. Overall, CBCT based-AORT is a robust approach to offer high quality online reoptimized plans for SAbR treatment of thoracic malignancies.

Can Superficial Parotid Sparing Alone Reduce Xerostomia? A Retrospective Study Correlating Type of Tobacco Use, Chronic Xerostomia and Parotid Sparing

<h3>Purpose/Objective(s)</h3> Head and neck cancer is one of the most common cancers in India. Tobacco chewing is the most widespread form of tobacco consumption. Majority of the patients present in locally advanced stage and achieving parotid sparing with IMRT in these patients can be challenging. We aimed to correlate the incidence of chronic xerostomia with the type of tobacco use and parotid volumes. <h3>Materials/Methods</h3> Patients with squamous carcinomas of head and neck region treated with conventionally fractionated curative-intent radiotherapy (bilateral neck) to a dose of 60-70 Gy between June 2017 and March 2020 were included in the study. The superficial and deep lobes of parotid were contoured retrospectively and the dosimetric data along with physician-assessed Radiation Therapy Oncology Group (RTOG) xerostomia toxicity grades at one year were retrieved. The individual, superficial and deep lobe parotids were considered spared if the mean dose was less than 26Gy. Patients were divided into five groups: Group 1 constituted bilateral parotid-spared, Group 2 of contralateral parotid-spared, Group 3 of contralateral superficial parotid-spared, Group 4 of bilateral superficial parotid-spared, and Group 5 where no sparing could be achieved. The groups were compared using chi-square test. <h3>Results</h3> 174 patients were included in the study. The median age was 59 years (IRQ 20 - 89). Chewing (40%), followed by smoking (35%) were the most frequent forms of tobacco consumption. The most common subsite was oropharynx (30.5%) followed by hypopharynx (26.4%), and about 83% had locally advanced disease. The mean bilateral parotid volume was 55.9 cc (18-130 cc). Tobacco chewers had significantly smaller mean parotid (53cc vs 60cc, p=0.02) and submandibular gland volumes (6cc vs 14cc, p<0.001) as compared to smokers. Bilateral or contralateral parotid sparing was achieved in 62.7%, bilateral or contralateral superficial lobe in 27.6%, and no sparing in 9.8% of patients. At one year, grade 0/I xerostomia was reported by 89.5%, and ≥grade II xerostomia in 11.5%. The grade of xerostomia was similar in smokers and chewers (p=0.95). The ipsilateral (26.5 Gy vs 37.8 Gy, p <0.001) and contralateral (23.3 Gy vs 34.1 Gy, p <0.001) superficial parotid mean dose in patients with grade 0/I and ≥grade II xerostomia was significantly different. Patients with bilateral or contralateral superficial lobe sparing had lower rates of grade II/III xerostomia compared to no sparing group (p<0.001). <h3>Conclusion</h3> Tobacco chewers have smaller parotid and submandibular glands compared to smokers. This did not impact the sparing of glands or the severity of xerostomia. Contralateral or bilateral superficial parotid sparing translated into superior salivary gland function at 1 year.

Multi-Institutional Dosimetric Analysis of Toxicity from Definitive SBRT in Cervical Cancer

<h3>Purpose/Objective(s)</h3> Brachytherapy is standard of care in locally advanced cervical cancer. In patients unable to undergo brachytherapy due to complex medical comorbidities, technical reasons, or refusal, treatment with stereotactic body radiotherapy (SBRT) is an option. We report a retrospective analysis of dosimetric parameters and toxicities for definitive cervix patients who received SBRT boost across three institutions. <h3>Materials/Methods</h3> Eligible patients had FIGO IB2-IVB cervical cancer and did not receive brachytherapy. Patient were treated with concurrent chemotherapy and pelvic external beam (EBRT) followed by SBRT with patient immobilization and fiducial image guidance. GTV and HR-CTV were defined according to EMBRACE II guidelines for brachytherapy. Clinicopathological data, dosimetric data, toxicity (CTCAE V 4.0), and outcome data was collected. Spearman correlations completed in R v4.2.1. <h3>Results</h3> 95 patients were included. All patients received EBRT to the pelvis with doses ranging from 45- 51.75Gy in 25-29 fx. SBRT doses to HR-CTV ranged from 16.8-40 Gy given in 4-6 fractions. Median (IQR) HR-CTV volume was 72.6 (58.7). The median total EQD2 were 104.25 (20.3) and SBRT EQD2 60 (20.3). Total EQD2 to 2cc of rectum and bladder were 65.67 (11.9) and 75.8 (21.5). Rates of acute and late grade 1-2 and grade 3 toxicities are reported in Table 1; there were no grade 4+ toxicities. Low grade urinary frequency was the most common reported toxicity and decreased over time. GU stricture (ureteral) was the most common acute grade 3 toxicity and was correlated with EQD2 of EBRT, SBRT, and total dose (P < 0.001), in addition to bladder 2cc SBRT (p = 0.04) and total EQD2 (p=0.02). Fistulas were rare, though some grade 3 fistula occurred. Acute grade 1-2 recto-vaginal fistulas were correlated with EQD2 of EBRT (P < 0.01), SBRT (p= 0.02), and total EQD2 (p=0.02). Late grade 3 fistulas were significantly correlated with HR-CTV volume, EQD2 of EBRT and SBRT (p=0.01) in addition to rectum 2cc SBRT dose (p=0.012), SBRT EQD2 (p=0.01), and total EQD2. <h3>Conclusion</h3> These data suggest that higher bowel/bladder 2cc SBRT and higher total EQD2 is associated with greater rates of acute and late grade 3 bowel and bladder toxicity in a cohort of women receiving SBRT for cervical cancer. Overall rates of ≥ grade 3 toxicities were small and SBRT can be a safe option for patients who cannot receive standard of care brachytherapy, especially when careful attention is given to D2cc volumes.

Near-Maximum Rib Dose is the Most Relevant Risk Factor for Ipsilateral Spontaneous Rib Fracture: A Dosimetric Analysis of Breast Cancer Patients after Radiotherapy

<h3>Purpose/Objective(s)</h3> Spontaneous rib fracture (SRF) is a common late complication in treated breast cancer patients. This study evaluated the incidence and risk factors of ipsilateral SRF after radiotherapy (RT) in breast cancer patients. In addition, we identified dosimetric parameters that were significantly associated with ipsilateral SRF. <h3>Materials/Methods</h3> We retrospectively reviewed 2,204 patients with breast cancer who underwent RT between 2014 and 2016, and were followed up with bone scans. We evaluated clinical risk factors for ipsilateral SRF. Dose-volume histogram analysis was also performed for patients (n = 538) whose dosimetric data were available. All ipsilateral ribs were manually delineated, and dosimetric parameters of the ribs were converted into the equivalent dose in 2 Gy fractions (EQD2). <h3>Results</h3> Ipsilateral SRF occurred in 14.5% of patients 3 years after RT. The median time to develop ipsilateral SRF was 15 months. In multivariate analysis of all patients, a hypofractionated dose scheme and abnormally low bone density were significant clinical risk factors for ipsilateral SRF. In dosimetric analysis, near-maximum rib dose (D2cc) best predicted ipsilateral SRF. The cutoff value of D2cc was EQD2 52 Gy, as determined by receiver operating characteristic analysis. In multivariate analysis including dosimetric variables, D2cc EQD2 ≥ 52 Gy was the only significant risk factor for ipsilateral SRF. <h3>Conclusion</h3> Our data demonstrated that near-maximum rib dose was the best dosimetric parameter to predict ipsilateral SRF in RT-treated breast cancer patients. Additionally, we suggest that the dose for the rib be constrained to D2cc EQD2 < 52 Gy to minimize the risk of ipsilateral SRF.

Dosimetric Analysis of Radiation Treatment Plans Based on a Deep Learning Auto Contouring Model for Patients with Localized Prostate Cancer

<h3>Purpose/Objective(s)</h3> To evaluate the feasibility and safety of creating radiation treatment plans based on a convolutional neural network auto-contouring (AC) model for patients receiving radiation for localized prostate cancer with a radio-opaque rectal spacer hydrogel. <h3>Materials/Methods</h3> The deep learning model produces ACs for target volumes (prostate, proximal seminal vesicle), OARs (rectum, bladder, penile bulb, femoral heads), and rectal hydrogel spacers. ACs were compared to manual contours (MC) by MDs for clinical use. Individual volumes and overall performance (composite score) were previously evaluated and reported, qualitatively using a 1 (minor discrepancy), 2 (moderate discrepancy), 3 (significant discrepancy), and 4 (rejected, gross error) scale and quantitatively using Dice Similarity Coefficient (DSC) and mean distance to agreement. VMAT treatment plans were created on unedited AC volumes from five composite 1 (C1) cases and five composite 2 (C2) cases and dose volume metrics were computed for the final MCs. Dosimetry comparison was performed to assess the safety of delivering a treatment plan created from unedited ACs. Cases were planned to 50.4 Gy in 28 fractions to the SVs, with a SIB of 70 Gy the prostate. PTV expansions were CTV+5 mm (CTV+4 mm posteriorly). <h3>Results</h3> Of 62 initial cases evaluated, 11 were scored as a C1, 28 were scored as a C2, and 23 received a C3. 10 cases were randomly selected for treatment planning. For prostate (CTV P) in selected C1 and C2 cases, mean score was 1 and 1.8 and mean DSC was 0.919 and 0.876, respectively. Clinical goals for planning and dosimetric data from treatment plans are shown in Table 1. The mean V70 Gy for prostate CTV/PTV was 99.83%/90.05% for C1 cases and 99.46%/89.15% for C2 cases. All clinical goals for OARs were met in both cohorts. <h3>Conclusion</h3> 63% of all cases evaluated received a C1 or C2 score. Our results indicate the treatment plans generated from unedited ACs on these cases respected all clinical goals for OARs. Although C1 plans produced better PTV coverage than C2, almost all plans had sub-optimal PTV coverage, indicating the need for MDs to edit target volumes before treatment planning.

Dose-Escalated Ablative Radiotherapy as Definitive Treatment for Unresectable and Recurrent Intrahepatic Cholangiocarcinoma

<h3>Purpose/Objective(s)</h3> Treatment of intrahepatic cholangiocarcinoma remains a clinical challenge with noted short survival. Ablative doses of radiation may provide durable response in this setting. We hypothesized that dose escalated radiotherapy in the setting intrahepatic cholangiocarcinoma would provide disease control and minimal toxicity. Here, we report the efficacy and safety of hypofractionated ablative radiotherapy (RT) in the setting of intrahepatic cholangiocarcinoma from a single institution. <h3>Materials/Methods</h3> Consecutive patients treated with definitive-intent RT for a diagnosis of intrahepatic cholangiocarcinoma between 2016-2022 were identified. Clinical and dosimetric data were retrospectively collected. Patients treated with a biological equivalent dose (BED₁₀) of less than 80 Gy were excluded. Toxicities were evaluated based on Common Terminology Criteria for Adverse Events (CTCAE) v5.0 criteria. Kaplan Meier analysis was utilized to calculate progression free survival (PFS) and overall survival (OS). <h3>Results</h3> Twenty-five patients were identified. Two patients (7%) received prior surgery with local recurrence and 12 patients (48%) received chemotherapy prior to RT. RT was delivered by MRI-guidance in 16 patients (64%) and CT-guidance in the remaining 9 patients (36%). RT doses ranged between 50 to 75 Gy (median 67.5 Gy) in 5 to 25 fractions (median 15) for a median BED₁₀ of 90.2 (range: 87.3 - 113.7 Gy). Median PFS and OS for all patients were 24.8 months (range: 12.5 – 37.1), and 50.6 months (range: 39.7 – 62.5 months) respectively. Overall treatment was well tolerated. Most common side effects reported included grade 1 fatigue and grade 1 nausea seen in 8 (32%) and 5 (20%) patients respectively. One patient (4%) exhibited grade 3 abdominal pain and ascites requiring hospitalization. No grade 4 or 5 toxicities were observed. <h3>Conclusion</h3> Dose escalation in the setting of intrahepatic cholangiocarcinoma appears to be a safe and effective treatment. Prolonged survival with definitive-intent ablative radiotherapy is possible in appropriately selected patients. Further studies are warranted to investigate this approach in a prospective setting.

Mature Outcomes of Stereotactic Body Radiotherapy for Spinal Metastases with 30 Gy in 4 Fractions

<h3>Purpose/Objective(s)</h3> At our institution, 30 Gy in 4 spine stereotactic body radiotherapy (SBRT) fractions is typically delivered for larger volume when multiple segments are included in the treatment volume and/or for the retreatment of spinal metastases. We report MRI-based local failure (LF) and vertebral compression fracture (VCF) rates for patients (pts) treated with 30 Gy/4Fx of spine SBRT. <h3>Materials/Methods</h3> A retrospective analysis of all pts with spine metastases treated with 30 Gy/4Fx from 2010 to 2021 from an institutional registry was performed. Demographic, clinical, dosimetric, and radiological data were summarized, and the primary endpoint was the MRI-based LF rate. Secondary endpoints included the incidence of VCF and overall survival (OS). Kaplan Meier was used to calculate LF and VCF per segment and OS per patient. <h3>Results</h3> Were included 116 pts with 245 treated segments in this analysis. The median number of consecutive segments in the treatment volume was 3 (1-7), and the median clinical target volume (CTV) was 126 cc (range, 10-863). Kidney (25%), lung (20%), breast (19%), prostate (19%), and colon (10%) cancer were the most common primary histologic types. 38% of pts had oligometastatic disease and 24% spine metastases only. 15% (17/116) of the patients were treated with postoperative SBRT. 31% of segments were re-irradiated for conventional palliative radiation, and 26% prior SBRT, failures. 25% of segments had a baseline VCF, 46% epidural disease, and 53% paraspinal tumor extension. The median follow-up per patient was 18.5 (range, 0.1-61) and per segment 10.7 mos (range, 0.1-59). The LF rates at 12 and 24 mos were 10.7% (95% CI 7.1-15.2) and 16% (95% CI 11.5-21.2), and for VCF were 7.3% (95% CI 4.4-11.2) and 11.2% (95% CI 7.5-15.8), respectively. 60% of failures had an epidural, and 23% a paraspinal, component. Age <68y (HR=0.43, 95% CI 0.19-0.95, p=0.038), volume of CTV <72cc (HR=0.09, 95% CI 0.01-0.70, p=0.021), and prior surgical stabilization (HR=0.25, 95% CI 0.07-0.81, p=0.021) were protective predictors for VCF on multivariable analyses (MVA). In particular, the VCF rate at 24 mos was 1.8% for pts with CTV volume <72 vs 14.6% for ≥72cc. Median OS was 20.3 mos (95% CI 14.8-27.1). Low grade or no epidural disease (HR=0.42, 95% CI 0.21-0.84, p=0.014), paraspinal disease (HR=3.07, 95% CI 1.68-5.62, p<0.001) and kidney primary cancer (HR=2.43, 95% CI 1.08-5.44, p=0.031) were predictors for OS on MVA, with a trend for oligometastatic disease (HR=0.56, 95% CI 0.31-1.02, p=0.059). No radiation-induced myelopathy was observed. <h3>Conclusion</h3> This first report for 30 Gy/4Fx as a novel fractionation for spine SBRT suggests high rates of local control and a low rate of VCF, particularly for those target volumes with a CTV<72 cc. However, our rate of VCF in very large treatment volumes (≥72cc) is comparable to the SBRT-induced VCF literature, where optimal management remains an active area of investigation.

Predictive Factors for Survival and Radiation Necrosis in Patients with Recurrent High-Grade Glioma Treated with Re-Irradiation

<h3>Purpose/Objective(s)</h3> To analyze the predictors of survival and radiation necrosis in adult patients (pts) with recurrent high-grade gliomas (rHGG) who have undergone re-irradiation (ReRT). <h3>Materials/Methods</h3> All adult pts with rHGG who had ReRT from 2009 to 2020 at one institution were retrospectively reviewed. Demographic, clinical, dosimetric, and radiological data were obtained from the electronic medical records. The primary outcome was to identify predictors of overall survival (OS) and radiation necrosis (RN). The secondary outcome was to identify patterns of failure after ReRT, which was defined as in-field, marginal or distant if >95% of the recurrence volume was in the 80% isodose line (IDL), between 80%-20% IDL, or outside 20% IDL, respectively. OS, progression-free survival (PFS), and RN were estimated by the Kaplan-Meier method. Toxicity was recorded according to CTCAE V5.0. <h3>Results</h3> Were included 79 pts with a median age of 52 yrs (range, 19-79), 62% were male, 85% had grade 4 glioma at presentation and 98% at ReRT. IDH was wildtype/mutated/unknown in 73%/11%/15% of pts. 92% had concurrent/adjuvant temozolamide at the primary treatment. 34% had re-resection prior to ReRT. 16% had concurrent bevacizumab at ReRT. The most common fractionation schedules at the primary treatment were 40 Gy/15Fx (9%) and 50-60 Gy/28-33Fx (91%), and at ReRT were 17-24 Gy/1Fx (9%), 20-35 Gy/5Fx (38%), 25-35 Gy/10Fx (48%) and 36-54 Gy/18-30Fx (5%). The median cumulative equivalent dose in 2 Gy fractions (EQD2, a/b=2) was 103 Gy (range, 81-216). The median OS and PFS were 9.9 (95% CI 8.3-11.6) and 4.1 mos (95% CI 3.6-5.4), respectively. The OS and PFS rate at 6/12 months were 69.6%/34.2% and 29.1%/7.5%, respectively. The prognostic factors for OS on multivariate analyses (MVA) were interval from initial treatment to first progression ≥16.3mos (HR=0.35, 95% CI 0.20-0.59, p<0.001), re-resection prior ReRT (HR=0.43, 95% CI 0.25-0.73, p=0.002), ECOG ≥1 at ReRT (HR=1.90, 95% CI 1.10-3.30, p=0.022) and PTV volume at ReRT ≥112cc (HR=2.63, 95% CI 1.55-4.44, p≤0.001). Toxicities G2 and G3 were 22% (8.8% RN) and 5% (2.5% RN), respectively. Concurrent use of bevacizumab (p<0.001) and EQD2 ≤98 Gy (p<0.001) were predictors for lower incidence of RN on MVA. Exploratory analysis suggested three risk groups for RNs based on cumulative EQD2: ≤100 Gy (RN = 4%), 100-111 Gy (RN = 13%), and ≥111 Gy (RN = 20%). The failures after ReRT were in-field, marginal or distant in 67%, 6%, and 27% of pts who had follow-up MRI, respectively. <h3>Conclusion</h3> Re-irradiation is a safe and effective treatment for GBM. We describe predictive factors for OS and RN to guide patient section. Focus on pts with the most favorable OS may aid in identifying pts most likely to benefit from ReRT.

Multi-Factor Analysis and Prediction of Late Rectal Toxicity after Radiotherapy for Cervical Cancer

<h3>Purpose/Objective(s)</h3> This study is to explore the role of clinical and dosimetric parameters in predicting grade 2-4 rectal toxicity after intensity modulated radiotherapy and three-dimensional image-guided brachytherapy for patients with locally advanced cervix cancer. <h3>Materials/Methods</h3> A total of 104 patients with cervical cancer who were treated with radiotherapy and experienced late rectal toxicity from January 2016 to October 2021 were collected. late toxicity was defined as toxicity occurring or lasting more than 3 months after initiation of treatment and graded according to Common Toxicity Criteria for Adverse Events (CTCAE version 5.0). Collect the general clinical characteristics (age, pathological type, clinical stage, history of Cardiovascular diseases, radiotherapy type(radical/adjuvant radiotherapy), the dose of External therapy and brachytherapy, cycles of concurrent or adjuvant chemotherapy, etc.),and dosimetric factors data of the rectum(V5-V50, Dmin, Dmean, Dmax, D0.1 cc, D1 cc, D2 cc, D5 cc, etc.) of patients after radiotherapy, and statistically analyze the relationship between clinical factors, dosimetric factors, and grade 2-4 rectal toxicity. the statistical methods we used include independent sample t test, chi-square test, non-parametric rank sum test, binary logistic regression analysis, etc. <h3>Results</h3> The ratio of grade 2-4 rectal toxicity in the whole group was 42.31%(44/104).Univariate analysis showed that the occurrence of grade 2-4 rectal toxicity was significantly correlated with radiotherapy type (radical/adjuvant radiotherapy),brachytherapy dose, rectum V45,D2cc and D5cc.Factors with statistical significance (P <0.05) from univariate analysis were included in the multivariate analysis to show that rectum V45,D2cc and D5cc were independent risk factors for late rectal toxicity of grade 2-4(P=0.04,0.032 and 0.014).The ROC curve analysis is used to evaluate the combined detection of rectum V45,D2cc and D5cc,the area under the ROC curve for the dose, rectum occurrence of grade 2-4 rectal toxicity was 0.727 (95 %CI: 0.630-0.823,P<0.001),the best cut-off value was 0.355, the sensitivity was 0.707, and the specificity was 0.650. <h3>Conclusion</h3> In this study, V45, D2cc and D5cc were independent risk factors for late grade 2-4 rectal toxicity. A prediction mode was created on multivariable analysis, with three dose-volume parameters predictive of late grade 2-4 rectal toxicity.