Background: Optically stimulated luminescent dosimeters (OSLDs) are used in clinical radiation dosimetry, but their sensitivity changes with the accumulated dose, limiting their reusability. Different studies have reported inconsistent results regarding the changes in sensitivity at different accumulated doses, and the reasons for this inconsistency remain unclear. In addition, the extent to which the OSLD linearity and element sensitivity correction factor (k(s,i)) change with the accumulated dose has not been well established. Purpose: We sought to characterize how the individual ks, i and dose non-linearity correction factor (k(L)) change with accumulated dose and how such changes affect the measurement of the dose to water (D-w). Determining the extent to which these parameters change with the accumulated dose can help clinics, institutions, and the Imaging and Radiation Oncology Core (IROC) at MD Anderson extend the OSLD accumulated dose limit above 10 Gy and thus increase the efficiency of use of these detectors, including for the remote audit output audits run by IROC for the National Clinical Trials Network. Methods: 95 nanoDot OSLDs were individually irradiated in 4-Gy cycles with 90 cGy used to determine sensitivity, irradiated between 0.25- to 3-Gy to determine linearity, and then irradiated to return all dosimeters to the same total dose history. Repeated measurements between 10 and 23 Gy were performed with the same batch (N = 129) and with a different batch (batch 2, N = 130). Dosimeters were read with a microSTARii system 5-8 h following each irradiation, and each OSLD was bleached for 24 h in the IROC bleaching box prior to every irradiation. A combination of single-factor ANOVA, Levene tests, and linear regressions was used to quantify changes in OSLD characteristics as a function of accumulated dose. The Dw was calculated using Equation (1)in the American Association of Physicists in Medicine Task Group 191 report. Results: The signal response of the OSLDs showed stability within 1% up to an accumulated dose of 23 Gy before decreasing with dose. The k(s,i) remained the same, while linearity changed with accumulated dose but did so in a linear and predictable manner. Conclusion: Our findings improve our understanding of accumulated dose and could help improve overall process efficiency by extending the reusability of OSLDs.

BackgroundOptically stimulated luminescent dosimeters (OSLDs) can be bleached and reused, but questions remain about the effects of repeated bleaching and fractionation schedules on OSLD performance.PurposeThe aim of this study was to investigate how light sources with different wavelengths and different fractionation schemes affect the performance of reused OSLDs.MethodsOSLDs (N = 240) were irradiated on a cobalt-60 beam in different step sizes until they reached an accumulated dose of 50 Gy. Between irradiations they were bleached using light sources of different wavelengths: the Imaging and Radiation Oncology Core (IROC) bleaching system (our control); monochromatic red, green, yellow, and blue lights; and a polychromatic white light. Sensitivity and linearity-based correction factors were determined as a function of dose step-size. The rate of signal removal from different light sources was characterized by sampling these OSLDs at various time points during their bleaching process. Relative doses were calculated according to the American Association of Physicists in Medicine Task Group-191. Signal repopulation was investigated by irradiating OSLDs (N = 300) to various delivered doses of 2, 10, 20, 30, 40, and 50 Gy in a single fraction, bleached with one of the colors, and read over time. Fractionation effects were evaluated by irradiating OSLDs up to 30 Gy in different size steps. After reading, the OSLDs were bleached following IROC protocol. OSLDs (N = 40) received irradiations in 5, 10, 15, 30 Gy fractions until they had an accumulated dose of 30 Gy; The sensitivity response of these OSLDs was compared with reference OSLDs that had no accumulated dose.ResultsLight sources with polychromatic spectrums (IROC and white) bleached OSLDs faster than did sources with monochromatic spectra. Polychromatic light sources (white light and IROC system) provided the greatest dose stability for OSLDs that had larger amounts of accumulated dose. Signal repopulation was related to the choice of bleaching light source, timing of bleaching, and amount of accumulated dose. Changes to relative dosimetry were more pronounced in OSLDs that received larger fractions. At 5-Gy fractions and above, all OSLDs had heightened sensitivity, with OSLDs exposed to 30-Gy fractions being 6.4% more sensitive than reference dosimeters.ConclusionsThe choice of bleaching light plays a role in how fast an OSLD is bleached and how much accumulated dose an OSLD can be exposed to while maintaining stable signal sensitivity. We have expanded upon investigations into signal repopulation to show that bleaching light plays a role in the migration of deep traps to dosimetric traps after bleaching. Our research concludes that the bleaching light source and fractionation need to be considered when reusing OSLD.

Purpose/Objective(s) Definitive chemoradiation with pelvic external beam radiotherapy (EBRT) followed by a brachytherapy (BT) boost remains the standard of care for locally advanced cervical cancer (LACC). Stereotactic body radiotherapy (SBRT) has been explored as a substitute for BT boost in patients who are not candidates for BT, however due to interfraction motion, SBRT often results in higher dose to organs at risk (OAR). We hypothesize that daily CT-guided online adaptive radiotherapy (OART) using the Ethos adaptive platform allows for more effective delivery of SBRT that better emulates the dose distribution of BT. Materials/Methods This is a single-institution retrospective dosimetric analysis of patients with pathologically confirmed LACC, stages IB-IVA, treated with EBRT followed by BT boost from September 2016 through October 2023. We identified 12 patients who received a BT boost dose of 28Gy in 4 fractions following 45 Gy EBRT, and also had cone-beam CT (CBCT) imaging from their EBRT course capturing their entire cervix. We used each of the patients’ last 4 CBCT images to simulate the creation of 4 unique OART plans. OART plans were generated by contouring CTV-HR, bowel, sigmoid, rectum and bladder on each of the CBCTs. A 3 mm PTV margin was added to CTV-HR. These volumes were imported into the Ethos platform where 4 unique daily OART plans were generated. BT dosimetry data was extracted from our BT planning software. For both BT and OART plans, each OAR was considered to have received 45 Gy in 25 fractions during EBRT. The total EQD2 (a/b = 3) to D 2.0 cc was calculated for each treatment course. The dosimetric goal for bowel and sigmoid were EQD2 <70 Gy, for rectum EQD2 <65 Gy and bladder EQD2 < 80 Gy. Statistical analysis was performed using a paired two-sided T test and α of 0.05. Results Twelve patients with LACC who received pelvic EBRT followed by BT boost to 28 Gy in 4 fractions were analyzed for a total of 48 fractions. OART plans achieved 100% prescription dose coverage to at least 90% of the CTV-HR. The mean EQD2 for each OAR and their P values can be found in Table 1. OART plans had a statistically significant lower dose to sigmoid and higher dose to rectum likely due to the absence of vaginal packing that is routinely placed at the time of BT. Despite the higher dose to rectum, the mean EQD2 of 61.3 Gy was still below our goal of 65 Gy and only 2 patients exceeded this constraint (17%). There were no significant differences in cumulative bowel or bladder dose. Conclusion Daily online adaptive SBRT can achieve comparable dosimetric outcomes compared to BT by minimizing the impact of interfraction variability on both target coverage and OAR dose compared to non-adapted SBRT. The use of daily OART to deliver SBRT in patients who are not BT candidates appears promising and further investigation is warranted.

Purpose/Objective(s) Acute hospitalization during or after cancer treatment negatively impacts quality of care and causes significant patient morbidity. In patients receiving radiation therapy (RT) for any malignancy, we hypothesized that a machine learning approach would enable prediction of hospitalizations during or shortly after RT. Materials/Methods We analyzed 35,810 courses of RT provided to treat a cancer diagnosis at a large multisite academic department in a major metropolitan environment (all cancer patients treated with RT between 8/1999—1/2022 regardless of disease site or treatment intent). Approximately 150 clinical/treatment variables were extracted and processed into analytic format by a proprietary oncology analytics platform connected to the Electronic Medical Record System (inpatient and outpatient), Oncology Information System, and PACS. Such variables included past medical history, recent laboratory data, prior systemic cancer therapies, prior RT, recent hospitalization, and RT details. A patient was labeled as having suffered an acute hospitalization if they had an encounter classified as inpatient or emergency during RT or in the 30 days following completion of RT. A machine learning model was trained on a subset (75%) of the data reserved for training (26857 cases). Five-fold cross-validation was used to select model type and hyperparameters, using area under the ROC curve (AUC) to measure performance. The final model was evaluated for accuracy, AUC, precision, recall, and F1 score on an independent test set (25%) excluded from the training process (8953 cases). Model calibration was assessed by visual inspection of calibration plots. An AUC>0.70 was considered clinically valid. Results Among the 35,810 courses of RT, the incidence of acute hospitalization was 9.1% (9.2% training set; 9.0% test set). Model performance on the test set is shown in Table 1. Variables deemed to be significant predictors for hospitalization included recent lab values (PT INR, sodium, potassium) recent hospitalization (within 30 days prior to RT start), and patient age Conclusion In cancer patients undergoing RT, a machine learning model identified patients at risk of 30-day hospitalization. Predictive analytics may be a key tool to help providers identify high-risk patients and optimize interventions, while improving quality and value of care.

AbstractBackgroundLong‐term breast cancer survivors are at risk for cardiotoxicity after treatment, but there is insufficient evidence to provide long‐term (~10 years) cardiovascular disease (CVD) screening recommendations. We sought to evaluate a tri‐modality CVD screening approach.MethodsThis single‐arm, feasibility study enrolled 201 breast cancer patients treated ≥6 years prior without CVD at diagnosis. Patients were sub‐grouped: cardiotoxic (left‐sided) radiation (RT), cardiotoxic (anthracycline‐based) chemotherapy, both cardiotoxic chemotherapy and RT, and neither cardiotoxic treatment. Patients underwent electrocardiogram (EKG), transthoracic echocardiogram with strain (TTE with GLS), and coronary artery calcium computed tomography (CAC CT). The primary endpoint was preclinical or clinical CVD.ResultsMedian age was 50 (29–65) at diagnosis and 63 (37–77) at imaging; median interval was 11.5 years (6.7–14.5). Among sub‐groups, 44% had no cardiotoxic treatment, 31.5% had cardiotoxic RT, 16% had cardiotoxic chemotherapy, and 8.5% had both. Overall, 77.6% showed preclinical and/or clinical CVD and 51.5% showed clinical CVD. Per modality, rates of any CVD and clinical CVD were, respectively: 27.1%/10.0% on EKG, 50.0%/25.3% on TTE with GLS, and 50.8%/45.8% on CAC CT. No statistical difference was seen among the treatment subgroups (NS, χ2 test, p = 0.58/p = 0.15).ConclusionThis study identified a high incidence of CVD in heterogenous long‐term breast cancer survivors, most >10 years post‐treatment. Over half had clinical CVD findings warranting follow‐up and/or intervention. Each imaging test independently contributed to the detection rate. This provides early evidence that long‐term cardiac screening may be of value to a wider group of breast cancer survivors than previously recognized.

Purpose Management of head and neck cancers (HNC) in older adults is a common but challenging clinical scenario. We assess the impact of Stereotactic Body Radiation Therapy (SBRT) on survival utilizing the Geriatric-8 (G8) questionnaire. Materials and methods 171 HNC patients, deemed medically unfit for definitive treatment, were treated with SBRT ± systemic therapy. G8 questionnaires were collected at baseline, at 4–6 weeks, and at 2–3 months post-treatment. Patients were stratified according to their baseline G8 score: <11 as 'vulnerable', 11–14 as 'intermediate', and >14 as 'fit'. Overall survival (OS) was assessed through univariate Kaplan Meier analysis. Repeated measures ANOVA was used to determine if baseline characteristics affected G8 score changes. Results Median follow-up was seventeen months. 60% of patients presented with recurrent HNC, 30% with untreated HNC primaries, and 10% with metastatic non-HNC primaries. Median age was 75 years. Median Charlson Comorbidity Index score was 2. 51% of patients were 'vulnerable', 37% were 'intermediate', and 12% were 'fit' at baseline, with median survival of 13.2, 24.3, and 41.0 months, respectively (p = .004). Patients who saw a decrease in their follow-up G8 score (n = 69) had significantly lower survival than patients who had stable or increased follow-up G8 scores (n = 102), with median survival of 8.6 vs 36.0 months (p < .001). Conclusion The G8 questionnaire may be a useful tool in upfront treatment decision-making to predict prognosis and prevent older patients from receiving inappropriate anti-cancer treatment. Decline in follow-up G8 scores may also predict worse survival and aid in goals of care following treatment.

To evaluate rectal wall doses sparing in prostate cancer patients undergoing combination therapy of external beam radiotherapy (EBRT) and low-dose-rate (LDR) brachytherapy with hydrogel spacer insertion. Fifty-five patients were enrolled in this retrospective study, in which 25 patients were applied with injection of polyethylene-glycol hydrogel and others without the insertion as control. Patient was prescribed with EBRT in dose of 45Gy to primary tumor site and nodal regions over 25 fractions, and LDR in dose of 100Gy on prostate. Two weeks after spacer placement, patient underwent CT scan for VMAT planning. Intraoperative prostate seed (Pd-103) implant (PSI) was performed two weeks after EBRT completed. Doses to the target and rectal wall were examined in each plan. BED calculation was applied in evaluation of sum dose of the rectal wall in each patient according to AAPM TG-137. On average, 1.04cm interspace between prostate and rectum was achieved by spacer insertion. Rectal wall was defined as 0.4cm thickness inside of rectum contour, and the volume was in average of 13.9cm3 in patients with spacer and 12.6cm3 in those without spacer. Target dose coverage was not seen in significant difference in VMAT and PSI plans between two groups of the patients. In VMAT plan, rectal wall doses were in average of 44.8Gy, 43.6Gy, 42.3Gy, 40.0Gy and 28.8Gy to the volume of 0.1, 0.5, 1, 2 and 5cm3, respectively, in patients with no spacer. In patients with spacer, the doses reduced significantly in average of 42.7Gy, 39.9Gy, 37.7Gy, 34.2Gy and 24.5Gy to rectal volume of 0.1, 0.5, 1, 2 and 5cm3, respectively (p<0.01 or <0.05 in t-test). In postimplant dosimetry, in patient without spacer, rectal wall doses were in average of 129.1Gy, 84.2Gy, 64.8Gy, 43.8Gy and 15.3Gy to the volume of 0.1, 0.5, 1, 2 and 5cm3, respectively. In patient with spacer, the doses decreased to 46.3Gy, 35.3Gy, 28.8Gy, 21.1Gy and 9.3Gy to rectal wall volume of 0.1, 0.5, 1, 2 and 5cm3, respectively. Dose decrease in the volume 0.1 - 5cm3 was significant in patients with spacer, compared with those without insertion (p<0.01 or <0.05 in t-test). Calculation of BED was performed in order to evaluate sum dose of the rectal wall in combination of both VMAT and PSI plans. In patient without spacer, BEDsum was in average of 126.9Gy, 106.7Gy, 84.3Gy and 45.5Gy to rectal volume of 0.5, 1, 2 and 5cm3, respectively. In contrast, BEDsum was seen in decrease of 20% to 39% in patient with spacer, in which was 76.8Gy, 68.2Gy, 57.2Gy and 36.2Gy to the rectal volume of 0.5, 1, 2 and 5cm3, respectively. Insertion of hydrogel spacer significantly improved rectal dose sparing, about 20% to 39%, in prostate cancer patient in EBRT and LDR combination therapy.

Breast cancer survivors are at increased risk of late cardiotoxicity years after left-sided radiation (RT) which can manifest as coronary artery disease (CAD) or other morbidities; the risk increases with both dosage and time. Classically, left-sided 3-D tangent RT fields deliver dose to the anterior heart and coronary vessels. As part of a multi-modality screening study, we hypothesized an increased incidence of CAD in patients (pts) who received left-sided RT compared to those who received other treatment(s). The CAROLE (CArdiac Related Oncologic Late Effects) Study, a prospective, single-center screening study enrolled 201 pts treated for breast cancer (any stage/treatment), ≥6 years prior (2004-2011), age 18-65 at diagnosis, and with no known cardiac disease at the time of cancer diagnosis. All pts received an electrocardiogram, echocardiogram, and a coronary artery calcium CT scan (CAC CT). Presence of CAD was defined as an Agatston score of >0 on CAC-CT. The individual vessel (left main [LM], left anterior descending [LAD], left circumflex [LCX], and right coronary arteries [RCA]) scores and overall Agatston sum scores were assessed by a single experienced cardiologist. CAD incidence was evaluated among +/-left-sided RT pts using a Chi-squared test. Following 6/2017-7/2018 enrollment, 199/201 pts had sufficient data for CAC CT analysis, 79 pts with left-sided RT and 120 with other treatment(s). Among all, median age at imaging was 63 (37-77). Median interval from diagnosis to imaging was 11.5 years (6.7-14.5). Overall, 38.6% had CAD on CAC CT; 43.6% in left-sided RT pts vs. 36.1% without (p=.70). For individual vessels (left-sided RT vs. without), incidence of CAD was: 17.9% vs. 8.4% LM (p=.04), 39.2% vs. 32.7% LAD (p=.35), 18.9% vs. 12.5% LCX (p=.21), and 18.9% vs. 20% RCA (p=.85). Among breast cancer survivors a median of 11.5 years since diagnosis, 43.6% of left sided RT pts and 36.1% of other pts (NS) had evidence of CAD on CAC CT screening. In this pilot study, there was a trend towards higher CAD in the LM of left-sided RT pts (p=.04, NS after Bonferroni adjustment); other vessels and overall scores were not statistically different among groups. This data provides evidence of high rates of occult CAD in a heterogenous group of breast cancer survivors and warrants consideration of CAC CT for cardiac screening for patients ∼10 years after cancer treatment. Further prospective evaluation with a larger cohort is warranted.

11569 Background: Long-term breast cancer survivors are at risk of late effects from cardiotoxic (tox) radiation (RT) and chemotherapy (chemo). However, there is a paucity of data to recommend appropriate cardiac screening for those ~10 years (yrs) after diagnosis (dx). This phase II, single arm study assessed multimodality screening in a heterogeneous cohort of long-term survivors. Methods: Our prospective, single center pilot study enrolled 201 (of 200 planned) eligible patients (pts): age 18-65 at dx, with treatment (tx) for breast cancer (any stage/tx) in 2004-2011 (≥6 yrs prior), with no cardiac disease (dz) at dx. The primary endpoint was presence of any cardiac dz (preclinical or clinical) on electrocardiogram (EKG), echocardiogram (echo), or coronary artery calcium CT (CAC CT); secondary endpoint included clinical (clin) dz alone. Subgroups assessed were: no tox chemo or RT (no tox), tox RT (left-sided) alone, tox chemo (anthracycline) alone, and both tox chemo and tox RT (tox chemoRT). Results: After enrollment (6/2017-7/2018), 200 pts had sufficient data for analysis. Median age was 50 (29-65) at dx, 63 (37-77) at imaging, and interval (dx to imaging) was 11.5 (6.7-14.5) yrs. There were 44% no tox, 31.5% tox RT, 16% tox chemo, and 8.5% tox chemoRT pts. Among all pts, 77.6% displayed evidence of any cardiac dz, 51.5% had clin dz. Per modality, rates of any and clin dz were: 27.1%/10% EKG, 50.0%/25.3% echo, and 50.8%/45.8% CAC CT. Among subgroups, these rates were: 73.9%/53.4% no tox, 82.5%/58% tox RT, 75%/38.7% tox chemo, and 82.4%/35.3% tox chemoRT (NS on χ2 test, p = .58/p = .15). Multivariable logistic regression with tox RT, tox chemo, and age as covariates, showed no significant tox RT (NS) or tox chemo effect (NS), however, increasing age was significantly associated with higher incidence of cardiac dz (p < 0.01; OR = 1.12 per yr, 95% CI: 1.07-1.18). Conclusions: Among pts with a median of > 10 yrs post breast cancer tx, multimodality CAROLE screening was feasible and showed high rates of cardiac dz (any/clin = 77.6%/51.5%) among all subgroups, including those without tox tx. This study provides evidence for long-term cardiac screening in a diverse group of breast cancer survivors and provides a novel pathway for evaluation. Clinical trial information: NCT03235427.

Stereotactic Body Radiation Therapy (SBRT) has been established as an option for treatment of recurrent cancers of the head and neck (H&N) in previously irradiated patients. There has been limited data published on effective means of salvaging patients who fail SBRT. Here we report on the safety and efficacy of repeat SBRT (R-SBRT). From October 2012 to May 2018, 9 patients with local relapse of primary H&N cancer who failed the first course of SBRT (F-SBRT) were retreated with R-SBRT. Median age at time of R-SBRT was 67 years (range, 51-90). 7 patients were men and 2 were women. Median F-SBRT dose was 40 Gy delivered in five biweekly fractions. 5 patients received concurrent systemic therapy with F-SBRT. Local recurrences were diagnosed by CT and/or PET scans showing increasing size and/or avidity and confirmed pathologically via biopsy. Median interval between F-SBRT and R-SBRT was 9 months (range, 5-27). Median R-SBRT dose was 40 Gy in 5 biweekly fractions. Median PTV volume of R-SBRT was 34.7 cubic centimeters (range, 8.6-90.2). The median percentage volume of overlap between the two SBRT courses was 64% (range, 26-100). All patients received concurrent systemic therapy with R-SBRT. Control and survival outcomes were captured for all patients. Toxicity was scored using CTCAE version 5.0. After a median follow-up of 12 months (range, 6-36), rates of local control and overall survival were both 56%. Median time to recurrence was 4 months (range, 2-13). 2 patients experienced grade 3 toxicities, which included fatigue in one patient and hearing loss in another, an expected side effect given that the cochlea fell within the PTV for both courses of SBRT. There were no grade 4 or 5 toxicities. Repeat SBRT was well-tolerated, with rates of grade 3+ toxicity similar to those previously reported on SBRT for head and neck cancer. Repeat SBRT is a viable option for those who fail initial SBRT.

To evaluate hydrogel spacer for improved rectal dose sparing in prostate cancer patients in both LDR brachytherapy and EBRT radiotherapy. Eleven patients with prostate cancer treated using both Pd-103 seed implantation and VMAT plan were selected in this retrospective study. Patient was prescribed with LDR brachytherapy in dose of 100Gy on primary cancer site, and then treated using VMAT plan in dose of 45Gy to primary tumor and nodal regions over 25 fractions. Briefly, after completion of intraoperatively planned Pd-103 seed implantation, patient had hydrogel spacer placed between prostate and rectum. Postimplant dosimetry was analyzed using Variseed software on CT scan after identifying seeds and contouring target and OAR. VMAT plan was generated using the same CT scan, in optimization with gEUD objective for target and OARs. In plan evaluation, PTV dose coverage and OAR dose constraint were adequate to satisfy departmental planning directive following RTOG0526 and RTOG0815 guidelines. Dose to rectal wall was examined in both postimplant and VMAT plan. Rectum was outlined as a solid organ, and rectal wall was defined as 0.4cm thickness inside of rectum contour. Volume of each contour, i.e. cm3, was measured using a treatment planning system. On average, 1.06cm (SD 0.32) separation between prostate and rectum was achieved by insertion of the spacer. Rectal wall volumes were in average of 21.8cm3 and 24.1cm3 in patients with spacer and without spacer, respectively. In postimplant assessment, we found, in patient without spacer, that rectal wall doses were in average of 66.6Gy (SD 23.1), 43.4Gy (SD 12.8), 15.4Gy (SD 4.5) and 4.4Gy (SD 1.8) to the volume of 1, 2, 5 and 10cm3, respectively. In those with spacer, on average, doses were 25.4Gy (SD 7.1), 19.0Gy (SD 5.3), 9.1Gy (SD 3.4) and 3.4Gy (SD 1.1) to rectal wall volume of 1, 2, 5 and 10cm3, respectively. On the volume 1 - 5cm3, rectal wall doses decreased about 40-60% in patients with spacer, compared with those having no insertion (p<0.01 in t-test). In VMAT plan, doses were in average of 43.5Gy (SD 2.0), 41.7Gy (SD 2.6), 35.5Gy (SD 4.3) and 24.6Gy (SD 5.3) to rectal wall volume of 1, 2, 5 and 10cm3, respectively, in patients with no spacer. In patients having spacer, rectal wall doses were in average of 39.6Gy (SD 4.5), 37.2Gy (SD 4.8), 30.7Gy (SD 4.6) and 21.2Gy (SD 4.0) to the volume of 1, 2, 5and 10cm3, respectively. Decrease of rectal wall doses was about 10-15% in the patients having spacer insertion (p<0.01 in t-test). Injection of hydrogel spacer improved rectal dose sparing in prostate cancer patients in both LDR brachytherapy and EBRT radiotherapy. Analysis of biological effective doses (BED) combining EBRT and brachytherapy will be presented.

Peer review is an essential component of quality assurance in radiation therapy. In 2012, we instituted multidisciplinary prospective daily contour rounds where the care team discusses contours and plans before initiating treatment. The purpose of this work was to examine peer review recommendations to assess if peer review generally results in recommending more conservative target volumes or increased organ sparing. The peer group evaluated appropriateness of planning target volumes (PTVs), organs-at-risk (OARs), and care plans for each prescription according to evidence-based departmental directives. Peer recommendations were compiled using in-house software. We reviewed 7645 prescriptions presented at daily rounds between September 2013 and March 2017. We isolated recommendations for PTV or OAR modification and classified each as an expansion, reduction, both, or could not be determined. Recommendations were analyzed by treatment technique, anatomical site, and physician experience. Seven (7) junior and 8 senior radiation oncologists were included in the review. PTV or OAR modifications were recommended for 750 of 7645 prescriptions (9.7%). The peer group recommended PTV modifications for 534 prescriptions (7.0%). Of these, 306 were expansions (57.3%), 119 were reductions (22.3%), 14 were both (2.6%), and 95 could not be determined (17.8%). Reasons for PTV expansions included increased nodal coverage, incorporation of supplemental imaging, or inadequate PTV margins due to motion or setup uncertainty. The peer group recommended OAR modifications for 216 prescriptions (2.8%). Of these, 102 were expansions (47.2%), 23 were reductions (10.6%), 2 were both (0.9%), and 89 could not be determined (41.2%). Reasons for OAR expansions included missing critical structures, inaccurate contours, or inadequate extent as per departmental standardization. Head and neck represented the largest percentage of PTV and OAR recommendations (28.8% and 18.5% respectively). IMRT plans received the most PTV and OAR recommendations (66.8% and 74.5% respectively). Of these, 39.2% were PTV expansions, 14.8% were PTV reductions, 15.2% were OAR expansions, and 3.7% were OAR reductions. The majority of PTV adjustment recommendations were directed towards senior faculty (65.5%). Expansions and reductions constituted 60.5% and 18.5% of the recommendations respectively, comparecd to 51.1% and 28.8% respectively for the junior faculty. Peer review resulted in PTV or OAR change recommendations for approximately 10% of cases. Expansions of PTV were recommended over 2.5 times more frequently than reductions and over 3 times more frequently than OAR expansions. This general trend was observed for all the variables examined (technique, site, and experience). Over 65% of PTV recommendations were for senior faculty. Overall, peer review tended to recommend more conservative target volumes.

The Radiation Oncology-Incident Learning System (RO-ILS) is a web based incident reporting system that provides a global mechanism for collaborative learning within radiation oncology. The RO-ILS mission encourages all members to report events as they happen in the department in a non-punitive environment. In September 2017, our multi-site radiation oncology department converted completely to RO-ILS following a decade of usage of an in-house system. The purpose of this work is to share our preliminary experience with RO-ILS, specifically focusing on comparative trends in event reporting, classification and learning. Between October 2017 and January 2018, events submitted were classified into the following categories: Therapeutic radiation incident, other safety incident, near-miss, unsafe condition, operational/process improvement. A single RO-ILS coordinator reviewed and further characterized these events into the following sub-categories for more detailed analysis and follow up. A multi-disciplinary quality management (QM) team met biweekly to investigate and develop follow-up actions which were communicated to the entire staff at monthly meetings at each geographical site to “close the loop” and demonstrate the effectiveness of event reporting. In a 4-month time period, staff reported 325 events. The distribution of events was as follows: Therapeutic radiation incident (53), Other Safety incident (10), Near-miss (19), Unsafe condition (5), and Operational/Process Improvement (238). Due to the default categories being too shallow, events were reclassified into: Communication, Prescription and Directive mismatch, Film Review, Re-simulation, New Plan, On Treatment Visit, Dosimeter, Physics, Equipment Malfunction, Information Technology, Patient on Break, Consent, No physicist on site, Patient Safety, Treatment terminated, Physics weekly chart checks not completed, Pathology, SOARIAN error, Scheduling, and Accident. Event reporting systems serve as an important function in raising awareness and generating a culture of safety. To date, our department has successfully and effectively rolled out RO-ILS and its key components. However, meaningful expansion of default classifications are recommended. RO-ILS allows our team to “track and trend” events, identify root causes, and implement targeted changes to improve patient care. We have provided feedback to the front lines monthly, closing the loop and encouraging further reporting. The frontline staff feels as though their reporting makes a difference, effecting positive changes in the department. As one of the few large multi-site institutions to fully implement RO-ILS, the amount of entries and expanded classifications submitted will provide valuable feedback to the radiation oncology community.

To develop a model to predict the possibility of achieving planning objectives for both PTV coverage and OAR dose-sparing based on the overlap volumes between PTVs and OARs, prior to the planning process. Forty VMAT plans on head-and-neck cancer patients were analyzed in this study. Patient was treated using 2-arc VMAT plan in prescriptions of 70Gy, 63Gy and 58.1Gy to primary tumor, high-risk nodal regions and low-risk nodal regions in bilateral areas of the neck, respectively, over 35 fractions. A VMAT plan was generated using a commercially available treatment planning system, with the use of the biological optimization objective (gEUD) for OARs and targets during optimization. Plans were evaluated by target dose coverage (D95, Dmax, conformity index) and mean or maximum doses on parotid, submandibular gland, larynx, spinal cord and brain stem. With a range of volume overlaps between OARs and PTVs and dose constraints applied, results showed that each PTV dose (D95) met dose coverage requirement. D95 for PTV70 ranged from 69Gy to 70.1Gy. The conformity index (CI) for PTV70 ranged from 81% to 96.6%. D95 for PTV63 was from 60.7Gy to 62.7Gy. CI was from 79% to 94.4%. Similarly, for PTV58, D95 was from 56.4Gy to 58.2Gy, and CI was from 76.5% to 95%. Dose conformity index at above 70% was acceptable and consistent with other reports. Parotid mean dose (Dmean) and its volumetric overlap with PTV58 was the main focus of the study. The relationship of the mean dose and volume of overlap can be fit in a linear regress formula: Dmean = A ∗ Voverlap + B (cGy) Where Dmean is the mean dose of the OAR, and Voverlap is the percentage volume overlapped with PTV58. A and B are regression constants. On average, the Dmean was seen at 15.6Gy, 21.1Gy and 24.6Gy for the parotid overlap volume of <5%, <10% and <20%, respectively. As expected, parotids dose-sparing became more difficult as the increase of overlap volume. When the overlap exceeded 20%, mean dose of 26Gy or less was no longer achievable. For parotids, the constant A in above is 54.8 and B is 1584.4 with regression R2 = 0.95. For submandibular glands (SMG), an average Dmean 27.6Gy was found in those having <10% overlap with PTV58, and 36.1Gy in the glands with <20% overlap. For SMG, the constant A is 41 and B is 2775.1 with regression R2 = 0.97. The mean doses to larynx (average Dmean =26.6Gy), and maximum doses to spinal cord (average Dmax=34.9Gy) and brain stem (average Dmax=31.5Gy) were also examined in those plans. The results demonstrate a strong relationship between the mean dose to parotid and SMG and its volume overlapped with the PTV. The linear model of the regression may serve as the predicting factors for early determinations of achievable dosimetric outcomes for these plans.

Cardiac toxicity is a well-established late effect of radiation to the chest. There is no established method for accurately quantifying coronary artery disease which may occur years after radiation treatment. Our objective is to utilize deformable registration to fuse radiation treatment CT scans with recent coronary artery calcium (CAC) CT scans (performed >5 years after treatment) to be able to accurately correlate dose with Agatston scores from the CAC CT. We identified a cohort of patients with thoracic radiation and archived treatment plan information who had also received CAC CT scans. Treatment plans were recreated in a commercially available treatment planning system; initial treatment dose-volume histograms (DVHs) were reviewed and compared to reconstructed plan DVHs (when available). CAC CTs were fused to cardiac volumes on the planning CT scans via deformable registration to an area including the heart. The contouring of heart and cardiac structures (including atria and ventricles) were performed by a radiation oncologist using the Michigan Heart Atlas on both simulation CT and the CAC CT (both contrast and noncontrast series). Heart and cardiac structure conformality index (CI), a measure of volumetric similarity, was evaluated. The individual vessel Agatston scores for patients determined to have calcification present were reviewed by an experienced cardiologist and then compared to treatment dose. A group of 8 patients who received treatment from 2003-2015 were reconstructed from initial planning information. Treatment areas included breast, lung, stomach, liver, and esophagus. The mean and maximum heart doses from reconstructed DVHs were within a ±10% margin of error of initial DVHs in all available cases (n=3), all with a mean dose of <100 cGy for all patients. Heart volumes contoured on CT sim and CAC CTs were found to have a CI of >0.90, except for 1 patient whose heart was initially autocontoured (CI= 0.84). Left atrium, left ventricle, right atrium, and right ventricle mean CIs were 0.60, 0.78, 0.82, and 0.78, respectively. Among patients with a >5-year period between treatment and CAC CT scans, 66% had coronary calcium present, all within the high dose receiving region of the heart (all left anterior descending (LAD) artery). Our deformable fusion method for evaluating coronary calcification could be performed accurately in patients with a small-to-moderate margin of error, except in a single patient for whom autocontouring was used. This method was successful in a range of dose patterns and thoracic and abdominal treatment sites. Cardiac CT deformable fusion may be employed in a larger patient population in the future to establish a quantitative correlation of coronary calcium and radiation dose for appropriate patient populations who receive significant radiation to the heart.

Purpose: Optically stimulated luminescent dosimeters (OSLDs) are utilized for in vivo dosimetry (IVD) of modern radiation therapy techniques such as intensity modulated radiation therapy (IMRT) and volumetric modulated arc therapy (VMAT). Dosimetric precision achieved with conventional techniques may not be attainable. In this work, we measured accuracy and precision for a large sample of clinical OSLD-based IVD measurements.Methods and materials: Weekly IVD measurements were collected from 4 linear accelerators for 2 years and were expressed as percent differences from planned doses. After outlier analysis, 10,224 measurements were grouped in the following way: overall, modality (photons, electrons), treatment technique (3-dimensional [3D] conformal, field-in-field intensity modulation, inverse-planned IMRT, and VMAT), placement location (gantry angle, cardinality, and central axis positioning), and anatomical site (prostate, breast, head and neck, pelvis, lung, rectum and anus, brain, abdomen, esophagus, and bladder). Distributions were modeled via a Gaussian function. Fitting was performed with least squares, and goodness-of-fit was assessed with the coefficient of determination. Model means (mu) and standard deviations (sigma) were calculated. Sample means and variances were compared for statistical significance by analysis of variance and the Levene tests (alpha = 0.05).Results: Overall, mu +/- sigma was 0.3 +/- 10.3%. Precision for electron measurements (6.9%) was significantly better than for photons (10.5%). Precision varied significantly among treatment techniques (P < .0001) with field-in-field lowest (sigma = 7.2%) and IMRT and VMAT highest (sigma = 11.9% and 13.4%, respectively). Treatment site models with goodness-of-fit greater than 0.90 (6 of 10) yielded accuracy within +/- 3%, except for head and neck (mu = -3.7%). Precision varied with treatment site (range, 7.3%-13.0%), with breast and head and neck yielding the best and worst precision, respectively. Placement on the central axis of cardinal gantry angles yielded more precise results (sigma = 8.5%) compared with other locations (range, 10.5%-11.4%).Conclusions: Accuracy of +/- 3% was achievable. Precision ranged from 6.9% to 13.4% depending on modality, technique, and treatment site. Simple, standardized locations may improve IVD precision. These findings may aid development of patient-specific tolerances for OSLD-based IVD. (C) 2016 American Society for Radiation Oncology. Published by Elsevier Inc. All rights reserved.