Dose-volume Histogram Constraints Research Articles

Biological optimization for hybrid proton-photon radiotherapy

Objective. Hybrid proton-photon radiotherapy (RT) is a cancer treatment option to broaden access to proton RT. Additionally, with a refined treatment planning method, hybrid RT has the potential to offer superior plan quality compared to proton-only or photon-only RT, particularly in terms of target coverage and sparing organs-at-risk (OARs), when considering robustness to setup and range uncertainties. However, there is a concern regarding the underestimation of the biological effect of protons on OARs, especially those in close proximity to targets. This study seeks to develop a hybrid treatment planning method with biological dose optimization, suitable for clinical implementation on existing proton and photon machines, with each photon or proton treatment fraction delivering a uniform target dose. Approach. The proposed hybrid biological dose optimization method optimized proton and photon plan variables, along with the number of fractions for each modality, minimizing biological dose to the OARs and surrounding normal tissues. To mitigate underestimation of hot biological dose spots, proton biological dose was minimized within a ring structure surrounding the target. Hybrid plans were designed to be deliverable separately and robustly on existing proton and photon machines, with enforced uniform target dose constraints for the proton and photon fraction doses. A probabilistic formulation was utilized for robust optimization of setup and range uncertainties for protons and photons. The nonconvex optimization problem, arising from minimum monitor unit constraint and dose-volume histogram constraints, was solved using an iterative convex relaxation method. Main results. Hybrid planning with biological dose optimization effectively eliminated hot spots of biological dose, particularly in normal tissues surrounding the target, outperforming proton-only planning. It also provided superior overall plan quality and OAR sparing compared to proton-only or photon-only planning strategies. Significance. This study presents a novel hybrid biological treatment planning method capable of generating plans with reduced biological hot spots, superior plan quality to proton-only or photon-only plans, and clinical deliverability on existing proton and photon machines, separately and robustly.

Improving Dose Volume Histogram (DVH) Based Analysis of Clinical Outcomes Using Modern Statistical Techniques: A Systematic Answer to Multiple Comparisons Concerns

DVH constraints are essential in the clinical practice of radiation therapy. Historically, DVH constraints were found through sparse sampling of all possible DVH indices to find one that appeared to be most predictive for clinical toxicity. This approach can lead to inconsistent results among studies and to multiple comparison concerns. We aim to solve both problems by examining a full array of DVH indices using statistical methods that account for strong correlations among DVH indices and incorporate radiobiological knowledge constraints. We extracted a dense array of V%_D indices from a treatment planning system using ESAPI interface, with V%_D corresponding to the volume fraction irradiated to dose D, or higher. We used Fused Lasso as the base model to compensate for correlations among DVH indices because it applies a penalty on the difference between DVH variables with adjacent dose. The base model was augmented with additional constraints based on radiobiological considerations: the positivity constraint (beta_i > 0) which assumes that any tissue irradiation cannot reduce the risk of toxicity, and monotonicity constraint (beta_i+1 > = beta_i) which assumes that higher dose to a fixed volume fraction cannot be associated with a lower risk of toxicity. We called the hybrid model KC-Lasso (Knowledge Constrained Lasso) and applied it to two clinical examples: grade 2 acute rectal toxicity in conventionally fractionated RT for 79 prostate cancer patients (77.4 Gy + MR based boost to 81-83 Gy) and cardiac toxicity in conventionally fractionated RT for 119 locally advanced Non-small Cell Lung Cancer (NSCLC) patients (Median prescribed dose 62 Gy). We further examined alternative data driven models to determine the importance of knowledge constraints. KC-Lasso detected two distinct dose thresholds for grade 2 rectal toxicity, at 35 Gy and 78 Gy. A threshold of 51 Gy was detected for reduced overall survival due to cardiac irradiation in NSCLC patients. An examination of KC-Lasso models at varying step size suggested that a single mid-range index can be used as a treatment planning constraint while full model can be used for confirmatory, final plan evaluation. Alternative models which lack knowledge constraints show patterns of negative and isolated coefficients which are difficult to interpret and are not likely to be generalizable. A more systematic approach to the analysis of correlations between DVH constraints and clinical toxicity can lead to greater consistency of results among different studies, better understanding of true dose thresholds and results which are more generalizable.

Consensus Quality Measures and Dose Constraints for Lung Cancer From the Veterans Affairs Radiation Oncology Quality Surveillance Program and ASTRO Expert Panel

For patients with lung cancer, it is critical to provide evidence-based radiation therapy to ensure high-quality care. The US Department of Veterans Affairs (VA) National Radiation Oncology Program partnered with the American Society for Radiation Oncology (ASTRO) as part of the VA Radiation Oncology Quality Surveillance to develop lung cancer quality metrics and assess quality of care as a pilot program in 2016. This article presents recently updated consensus quality measures and dose-volume histogram (DVH) constraints. A series of measures and performance standards were reviewed and developed by a Blue-Ribbon Panel of lung cancer experts in conjunction with ASTRO in 2022. As part of this initiative, quality, surveillance, and aspirational metrics were developed for (1) initial consultation and workup; (2) simulation, treatment planning, and treatment delivery; and (3) follow-up. The DVH metrics for target and organ-at-risk treatment planning dose constraints were also reviewed and defined. Altogether, a total of 19 lung cancer quality metrics were developed. There were 121 DVH constraints developed for various fractionation regimens, including ultrahypofractionated (1, 3, 4, or 5 fractions), hypofractionated (10 and 15 fractionations), and conventional fractionation (30-35 fractions). The devised measures will be implemented for quality surveillance for veterans both inside and outside of the VA system and will provide a resource for lung cancer-specific quality metrics. The recommended DVH constraints serve as a unique, comprehensive resource for evidence- and expert consensus-based constraints across multiple fractionation schemas.

Consensus Quality Measures and Dose Constraints for Prostate Cancer From the Veterans Affairs Radiation Oncology Quality Surveillance Program and American Society for Radiation Oncology Expert Panel

PurposeThere are no agreed upon measures to comprehensively determine the quality of radiation oncology (RO) care delivered for prostate cancer. Consequently, it is difficult to assess the implementation of scientific advances and adherence to best practices in routine clinical practice. To address this need, the US Department of Veterans Affairs (VA) National Radiation Oncology Program established the VA Radiation Oncology Quality Surveillance (VA ROQS) Program to develop clinical quality measures to assess the quality of RO care delivered to Veterans with cancer. This article reports the prostate cancer consensus measures. Methods and MaterialsThe VA ROQS Program contracted with the American Society for Radiation Oncology to commission a Blue Ribbon Panel of prostate cancer experts to develop a set of evidence-based measures and performance expectations. From February to June 2021, the panel developed quality, aspirational, and surveillance measures for (1) initial consultation and workup, (2) simulation, treatment planning, and delivery, and (3) follow-up. Dose-volume histogram (DVH) constraints to be used as quality measures for definitive and post-prostatectomy radiation therapy were selected. The panel also identified the optimal Common Terminology Criteria for Adverse Events, version 5.0 (CTCAE V5.0), toxicity terms to assess in follow-up. ResultsEighteen prostate-specific measures were developed (13 quality, 2 aspirational, and 3 surveillance). DVH metrics tailored to conventional, moderately hypofractionated, and ultrahypofractionated regimens were identified. Decision trees to determine performance for each measure were developed. Eighteen CTCAE V5.0 terms were selected in the sexual, urinary, and gastrointestinal domains as highest priority for assessment during follow-up. ConclusionsThis set of measures and DVH constraints serves as a tool for assessing the comprehensive quality of RO care for prostate cancer. These measures will be used for ongoing quality surveillance and improvement among veterans receiving care across VA and community sites. These measures can also be applied to clinical settings outside of those serving veterans.

Consensus Quality Measures and Dose Constraints for Breast Cancer From the Veterans Affairs Radiation Oncology Quality Surveillance Program and American Society for Radiation Oncology Expert Panel

Using evidence-based radiation therapy to direct care for patients with breast cancer is critical to standardize practice, improve safety, and optimize outcomes. To address this need, the Veterans Affairs (VA) National Radiation Oncology Program (NROP) established the VA Radiation Oncology Quality Surveillance Program to develop clinical quality measures (QMs). The VA NROP contracted with the American Society for Radiation Oncology to commission 5 Blue Ribbon Panels for breast, lung, prostate, rectal, and head and neck cancers. The Breast Cancer Blue Ribbon Panel experts worked collaboratively with the NROP to develop consensus QMs for use throughout the VA system, establishing a set of QMs for patients in several areas, including consultation and work-up; simulation, treatment planning, and treatment; and follow-up care. As part of this initiative, consensus dose-volume histogram (DVH) constraints were outlined. In total, 36 QMs were established. Herein, we review the process used to develop QMs and final consensus QMs pertaining to all aspects of radiation patient care, as well as DVH constraints. The QMs and expert consensus DVH constraints are intended for ongoing quality surveillance within the VA system and centers providing community care for Veterans. They are also available for use by greater non-VA community measures of quality care for patients with breast cancer receiving radiation.

Consensus Quality Measures and Dose Constraints for Rectal Cancer From the Veterans Affairs Radiation Oncology Quality Surveillance Program and American Society for Radiation Oncology (ASTRO) Expert Panel

PurposeEnsuring high quality, evidence-based radiation therapy for patients with cancer is of the upmost importance. To address this need, the Veterans Affairs (VA) Radiation Oncology Program partnered with the American Society for Radiation Oncology and established the VA Radiation Oncology Quality Surveillance program. As part of this ongoing effort to provide the highest quality of care for patients with rectal cancer, a blue-ribbon panel comprised of rectal cancer experts was formed to develop clinical quality measures. Methods and MaterialsThe Rectal Cancer Blue Ribbon panel developed quality, surveillance, and aspirational measures for (a) initial consultation and workup, (b) simulation, treatment planning, and treatment, and (c) follow-up. Twenty-two rectal cancer specific measures were developed (19 quality, 1 aspirational, and 2 surveillance). In addition, dose-volume histogram constraints for conventional and hypofractionated radiation therapy were created. ConclusionsThe quality measures and dose-volume histogram for rectal cancer serves as a guideline to assess the quality of care for patients with rectal cancer receiving radiation therapy. These quality measures will be used for quality surveillance for veterans receiving care both inside and outside the VA system to improve the quality of care for these patients.

A mathematical and dosimetric approach to validate auto‐contouring by Varian Smart segmentation for prostate cancer patients

AbstractPurposeThe aim of this study was to quantify the discrepancies in geometrical and dosimetric impacts (in volumetric modulated arc therapy) between manually segmented (MS) contours and smart segmentation (SS) auto‐contours (by Varian Eclipse Treatment Planning System SS v13.5) for prostate cancer patients.MethodsThe automated segmentation was carried out by Eclipse Treatment Planning System (Varian, version 13.5) Smart Segmentation (SS) workspace of 10 prostate cancer patients for four regions of interest; such as, bladder, rectum, femoral head left, and femoral head right. The geometric and dosimetric deviation between SS and MS contours have been quantified in the form of different parameters. The organ‐wise correlation between different validation parameters was addressed.ResultsThe organ‐wise correlation analysis showed the good and consistent correlation between different geometric validation parameters for the bladder. The hypothesis test for checking compliance of different parameters with AAPM 132 tolerance was addressed and validated between MS and SS bladder with p‐value = 0.01 and 0.05. There was no significant dosimetric difference between the dose–volume histogram (DVH) estimated for the SS bladder and standard DVH constraints protocol (as per the TMH PRIME trial) with p‐value = 0.01 and 0.05. The difference between DVH estimated for MS and SS bladder was also not significant, with p‐value = 0.05.ConclusionThis study shows that “well correlated validation parameters infer correctly about the matching or coincidence between auto and manually segmented contours,” and the bladder contouring by Smart Segmentation and plan optimization can achieve acceptable DVH constraints.

Dose volume histogram constraints in patients with soft tissue sarcomas of the extremities and the superficial trunk treated with surgery and perioperative HDR brachytherapy

BackgroundWound healing complications (WHC), osteoradionecrosis (ORN), and nerve damage (ND) are common adverse effects in adult patients with soft tissue sarcomas of the extremities and the superficial trunk treated with surgery and perioperative high dose rate brachytherapy (PHDRB) alone or combined with external beam radiotherapy (EBRT). RationaleAnalysis of the treatment factors contributing to these complications can potentially minimize their occurrence and severity. PatientsA total of 169 patients enrolled in two parallel prospective studies were included in this analysis. Previously Unirradiated cases (Group 1; n = 139) were treated with surgical resection, 16–24 Gy of PHDRB and 45 Gy of EBRT. Adjuvant chemotherapy was given to selected patients with high-grade tumors. Previously irradiated cases (Group 2; n = 30) were treated with surgical resection and 32–40 Gy of PHDRB without further EBRT. MethodsPatient factors, tumor factors, surgical factors, PHDRB factors and EBRT factors were analyzed using Cox univariate and multivariate analysis. ResultsIn Previously Unirradiated cases, WHC, ORN and ND occurred in 38.8%, 5.0% and 19.4%. Multivariate analysis indicated that WHC increased with CTV size (p = 0.02) and CTV2cm3 Physical dose (p = 0.02). ORN increased with Bone2cm3 EQD2 ≥ 67 Gy(p = 0.01) and ND was more frequent in patients with TV100DVH-based dose (tissue volume encompassed by the 100% isodose) ≥ 84 Gy (p < 0.01). In Previously Irradiated cases, WHC, ORN and ND occurred in 63.3%, 3.3% and 23.3%. Multivariate analysis showed that WHC was more frequent in patients with Skin2cm3Lifetime EQD2 ≥ 84 Gy (p = 0.01) and ND was more frequent after CTVD90 Physical Doses ≥ 40 Gy (p < 0.01). ConclusionsWHC in Previously Unirradiated patients can be minimized by using a more conservative CTV definition together with a meticulous implant technique and planning aimed to minimize hyperdose CTV2cm3 areas. In Previously Irradiated patients WHC may be mimimized considering Lifetime EQD2 Skin2cm3 doses. ORN can be reduced by using the Bone2cm3 EQD2 constraint. ND occurs more frequently in patients with large tumors receiving high treated volume doses, but no specific constraints can be recommended due to the lack of peripheral nerve definition during brachytherapy planning.

Dosimetric Effects and Estimated Osteoradionecrosis Development in Oropharynx Cancer Patients Treated With Intensity Modulated Radiotherapy With the Adoption of Dose Volume Histogram Based Mandibular Constraints

<h3>Purpose/Objective(s)</h3> Dose constraint recommendations to limit mandible (MN) osteoradionecrosis (ORN) are largely unchanged since the advent of head and neck intensity modulated radiotherapy (IMRT). Updated MN dose volume histogram (DVH) constraints have been proposed but not validated in a prospective dataset. We hypothesize that after adopting more stringent MN constraints at our institution, we lowered the estimated ORN rate without compromising target coverage / organ at risk (OAR) avoidance. <h3>Materials/Methods</h3> This is a single institution retrospective study of a prospective database of non-metastatic oropharynx cancer (OPC) patients ≥ 18 years treated with standard fractionation adjuvant or definitive IMRT from 01 / 2014 – 08 / 2020. In 09 / 2017, MN dose constraint was changed from 0.1 cm³ < 70 Gy – historical constraints (HC) – to V44 < 42%, V58 < 25%, 0.5cc < 70 Gy – modified constraints (MC). No other target / OAR criteria were altered. ORN was defined as having exposed MN bone for ≥ 3 months without recurrence. Impact on ORN rate and resulting dosimetric changes in MN and other OARs (parotid and submandibular glands, oral cavity, pharyngeal constrictors) between HC and MC groups were evaluated. Continuous variables were compared via the Wilcoxon test, discrete variables via Fisher's exact test. Due to shorter follow-up (FU), regression modeling was used to estimate ORN cases in MC group. <h3>Results</h3> There were 174 patients total, 71 in MC group. The mean prescription dose was 67.7 Gy and 68.5 Gy in HC and MC groups, respectively. More of the HC group (37%) had clinical T3 - T4 tumors compared to MC group (25%). More patients in the MC group met both the V44 < 42% (62 v 87%, <i>P</i> < 0.01) and the V58 < 25% constraints (73 v 92%, <i>P</i> < 0.01). Percent of patients meeting neither, one, or both MN constraints in the HC group was 23%, 13%, and 63%, versus 8%, 4%, and 87% in the MC group, respectively (<i>P</i> < 0.01). Mandible V44 and V58 were significantly associated with ORN (<i>P</i> < 0.01 and <i>P</i> = 0.03, respectively) across entire patient cohort. In HC group, mean MN V44 was significantly associated with ORN (62% in ORN cases vs 38%, <i>P</i> = 0.01) while mean MN V58 was approaching significance (<i>P</i> < 0.06). There were too few ORN cases in the MC group for similar analysis. Mean dose to OARs was not increased in MC group. There were 7 cases (76%) of ORN in the HC group at median FU 39 months (range: 2 - 78 months) and 2 cases (33%) of ORN in the MC group at median FU 11 months (range: 2 - 39 months). To account for shorter FU in MC group, logistic regression of ORN based on V44 in HC patients resulted in a model to predict ORN cases in MC group. This model estimates a total of 3.2 ORN cases (55%) in MC group (95% CI: 0.00 - 6.4). <h3>Conclusion</h3> We confirm MN V44 and V58 are predictive of ORN development. The goals of MN V44 < 42% and V58 < 25% were achievable in 87% of cases without sacrificing target coverage or other OAR constraints. Regression modeling estimates fewer ORN cases in MC group.

Biologically Effective Dose and Rectal Bleeding in Definitive Proton Therapy for Prostate Cancer.

Purpose and ObjectivesWith increasing use of hypofractionation and extreme hypofractionation for prostate cancer, rectal dose-volume histogram (DVH) parameters that apply across dose fractionations may be helpful for treatment planning in clinical practice. We present an exploratory analysis of biologically effective rectal dose (BED) and equivalent rectal dose in 2 Gy fractions (EQD2) for rectal bleeding in patients treated with proton therapy across dose fractionations.Materials and MethodsFrom 2016 to 2018, 243 patients with prostate cancer were treated with definitive proton therapy. Rectal DVH parameters were obtained from treatment plans, and rectal bleeding events were recorded. The BED and EQD2 transformations were applied to each rectal DVH parameter. Univariate analysis using logistic regression was used to determine DVH parameters that were significant predictors of grade ≥ 2 rectal bleeding. Youden index was used to determine optimum cutoffs for clinically meaningful DVH constraints. Stepwise model-selection criteria were then applied to fit a “best” multivariate logistic model for predicting Common Terminology Criteria for Adverse Events grade ≥ 2 rectal bleeding.ResultsConventional fractionation, hypofractionation, and extreme hypofractionation were prescribed to 117 (48%), 84 (34%), and 42 (17.3%) patients, respectively. With a median follow-up of 20 (2.5-40) months, 10 (4.1%) patients experienced rectal bleeding. On univariate analysis, multiple rectal DVH parameters were significantly associated with rectal bleeding across BED, EQD2, and nominal doses. The BED volume receiving 55 Gy > 13.91% was found to be statistically and clinically significant. The BED volume receiving 55 Gy remained statistically significant for an association with rectal bleeding in the multivariate model (odds ratio, 9.81; 95% confidence interval, 2.4-40.5; P = .002).ConclusionIn patients undergoing definitive proton therapy for prostate cancer, dose to the rectum and volume of the rectum receiving the dose were significantly associated with rectal bleeding across conventional fractionation, hypofractionation, and extreme hypofractionation when using BED and EQD2 transformations.

Dose volume histogram constraints in patients with head and neck cancer treated with surgery and adjuvant HDR brachytherapy: A proposal of the head and neck and skin GEC ESTRO Working group

BackgroundThe Head and Neck and Skin (HNS) Working group of the GEC-ESTRO acknowledges the lack of widely accepted Dose Volume Histogram (DVH) constraints in adjuvant head and neck brachytherapy and issues recommendations to minimize mandibular Osteoradionecrosis (ORN) and Soft Tissue Necrosis (STN). MethodsA total of 227 patients with the diagnosis of head and neck cancer treated with surgery and adjuvant HDR brachytherapy alone or combined with other treatment modalities during the period 2000–2018 were analyzed. ResultsSTN was observed in 28 out of 227 cases (12.3%) with an average time to appearance of 4.0 months. In previously unirradiated cases, there was a positive correlation between CTV size and STN (p = 0.017) and a trend towards significance between Total EQD2-DVH TV100 dose and STN (p = 0.06). The risk of STN in the absence of both factors (i.e, CTV < 15 cm3 and Total EQD2-DVH TV100 dose < 87 Gy) was 2%, with one factor present 15.7% and with both factors 66.7% (p = 0.001). ORN was observed in 13 out of 227 cases (5.7%) with an average time to appearance of 26.2 months. In unirradiated cases, ORN correlated with Total Physical Dose to Mandible2cm3 (p = 0.027). Patients receiving Total Physical Doses greater than 61 Gy had a 20-fold increased risk of ORN. ConclusionsIn Unirradiated patients the panel recommends to avoid implantation of postoperative CTVs exceeding 15 cm3 at Total EQD2-DVH TV100 doses in excess of 87 Gy as well as to limit the irradiation of the Mandible2cm3 to 61 Gy. In previously irradiated patients the panel cannot make a recommendation based on the available results.

Robust radiobiological optimization of ion beam therapy utilizing Monte Carlo and microdosimetric kinetic model

To develop a Monte Carlo (MC)-based and robust ion beam therapy optimization system that separates the optimization algorithm from the relative biological effectiveness (RBE) modeling. Robustly optimized dose distributions were calculated and compared across three ion therapy beams (proton, helium, carbon). The effect of different averaging techniques in calculating RBE in mixed beams was also investigated. Ion particles were transported in TOPAS MC. The microdosimetric-kinetic model (MKM) parameter, saturation corrected specific energy (), was calculated with a customized MKM implementation in TOPAS MC. Intensity modulated ion therapy robust optimization was performed by a quasi-Newton iterative method based on dose-volume objective function. The robust optimization took setup and range uncertainties into account. In the present work, the biological dose for each individual spot was calculated, and then summed together to calculate total biological dose. In other published works, radiosensitive parameters, such as , were first averaged over all beam spots within a mixed-beam field, after which biological dose was calculated using the averaged radiosensitive parameters. The difference between the two mixed-beam biological dose calculations was quantified. Robust plans were achieved with the three particle types. The effect of averaging depended on particle type. The difference between biological doses calculated with individual and averaged may be greater than 3% for a carbon beam. MC based radiobiological and robust optimization was made flexible to incorporate dose-volume histogram constraints and to be independent of RBE models. Iterative optimization with RBE models was feasible. Evaluation of the RBE calculation for mixed beam could be necessary if better accuracy was demanded.

Evaluation of the introduction of a 'traffic light protocol' for dose-volume histogram constraints of organs at risk in definitive radiotherapy at Liverpool and Macarthur Cancer Therapy Centres.

Protocols have been developed in our department with recommended dose constraints for organs at risk (OAR) for each tumour site receiving definitive radiotherapy. We have developed a colour coding system to indicate when constraints are meeting objectives (green), have minor variation from planning objectives (yellow) and have major variation from planning objectives (red). We performed a quality audit to assess adherence to the protocol and to determine the rate of acute and subacute toxicities. All definitive radiotherapy dose-volume histogram (DVH) reports generated in the first 6months of 2017 at Liverpool and Macarthur cancer therapy centres were collected. For each radiotherapy group, the overridden dose constraints were evaluated and categorized to red and yellow. For all patients in our data set, follow-up documents/assessments were searched for grade 3 or higher acute or subacute radiotherapy toxicity and compared with those who had overridden dose constraints. There were 210 (34%) plans accepted with at least one major variation and 161 (26%) plans with minor variation. Head and neck group had the most rate of major variations (77%). The best groups in adherence to protocol were lymphoma and breast groups. In general, grade 3 toxicity was observed in 1%, 4% and 9% of patients who were in green, yellow and red categories. Overall, we noted a correlation with grade 3 toxicities between acceptable plans (green) and ones with a minor or major variation (yellow or red) (1% vs. 7% P=0.0001). In conclusion this study showed an increased risk of higher grade toxicities when DVHs were beyond our departmental constraints using a 'Traffic Light System'. With this new colour coding system, we can facilitate auditing of the dose constraints in order to improve the quality of radiotherapy plans and potentially provide benchmarking for reducing toxicities in radiotherapy treatments.

Operating a treatment planning system using a deep-reinforcement learning-based virtual treatment planner for prostate cancer intensity-modulated radiation therapy treatment planning.

In the treatment planning process of intensity-modulated radiation therapy (IMRT), a human planner operates the treatment planning system (TPS) to adjust treatment planning parameters, for example, dose volume histogram (DVH) constraints' locations and weights, to achieve a satisfactory plan for each patient. This process is usually time-consuming, and the plan quality depends on planer's experience and available planning time. In this study, we proposed to model the behaviors of human planners in treatment planning by a deep reinforcement learning (DRL)-based virtual treatment planner network (VTPN), such that it can operate the TPS in a human-like manner for treatment planning. Using prostate cancer IMRT as an example, we established the VTPN using a deep neural network developed. We considered an in-house optimization engine with a weighted quadratic objective function. Virtual treatment planner network was designed to observe an intermediate plan DVHs and decide the action to improve the plan by changing weights and threshold dose in the objective function. We trained the VTPN in an end-to-end DRL process in 10 patient cases. A plan score was used to measure plan quality. We demonstrated the feasibility and effectiveness of the trained VTPN in another 64 patient cases. Virtual treatment planner network was trained to spontaneously learn how to adjust treatment planning parameters to generate high-quality treatment plans. In the 64 testing cases, with initialized parameters, quality score was 4.97 (±2.02), with 9.0 being the highest possible score. Using VTPN to perform treatment planning improved quality score to 8.44 (±0.48). To our knowledge, this was the first time that intelligent treatment planning behaviors of human planner in external beam IMRT are autonomously encoded in an artificial intelligence system. The trained VTPN is capable of behaving in a human-like way to produce high-quality plans.

Hybrid proton-photon inverse optimization with uniformity-regularized proton and photon target dose

The goal of radiation therapy is to deliver tumoricidal dose to clinical target volume (CTV) while sparing organs-at-risk (OAR). We hypothesize that the joint use of proton and photon radiation therapy via appropriate hybrid proton-photon inverse planning method will be more favorable than proton or photon therapy alone, in terms of optimized combination of CTV coverage and OAR sparing. This work develops hybrid proton-photon inverse optimization method that simultaneously optimizes proton and photon variables. To account for delivery uncertainty, proton dose is targeted at CTV using robust optimization, and photon dose is targeted at either CTV using robust optimization or planning target volume (PTV) using the same setup shifts. The optimization objectives enforce OAR sparing and uniform CTV coverage for the total dose, while imposing uniform-dose regularization at targets for both the proton and photon component in order for both components to be individually deliverable. The hybrid problem with dose-volume-histogram (DVH) constraints is nonconvex and solved by iterative convex relaxations of DVH constraints and alternating direction method of multipliers (ADMM). Preliminary results suggest the hybrid proton-photon planning potentially improves proton or photon planning in terms of optimized combination of CTV coverage and OAR sparing.

Using benchmarked lung radiation dose constraints to predict pneumonitis risk: Developing a nomogram for patients with mediastinal lymphoma

PurposeWe identified lung dosimetric constraints to assist in predicting the radiation pneumonitis (RP) risk in patients with mediastinal lymphoma and then identified the clinical prognostic factors that were associated with the achievement of key dosimetric constraints. Methods and MaterialsIn 190 patients who received mediastinal intensity modulated radiation therapy, we used univariate χ2 and multivariate logistic models to identify the predictors of RP and achievement of lung dose-volume histogram (DVH) constraints and build a predictive nomogram for RP. ResultsAn increased risk of RP was strongly associated with mean lung dose (MLD) > 13.5 Gy (odds ratio [OR]: 8.13; 95% confidence interval [CI], 3.01-21.93; P < .001) and the percent of lung volume receiving ≥5 Gy (V5) > 55% (OR: 7.01; 95% CI, 2.94-16.72; P < .001). Therefore, patients had low RP risk (8%) if both MLD ≤13.5 and V5 ≤55 constraints were achieved, moderate risk (24%) if only MLD was achieved, and the highest risk (48%) if MLD was not achieved. Deep-inspiration breath-hold (DIBH) technique during treatment strongly prognosticated achieving MLD and V5 DVH constraints (OR,3.88; 95% CI, 1.84-8.19; P < .001). Specifically, 86% of patients who were treated with DIBH versus 63% without DIBH achieved DVH constraints (P < .001). This translated into a “number needed to treat” with DIBH of 4 patients to enable 1 additional patient to achieve both constraints. In comparison, the clinical characteristics were marginal prognosticators: DVH constraints were more likely achieved in nonbulky disease (OR: 3.01; 95% CI, 0.89-4.53; P = .09) and patients who had not previously received salvage chemotherapy (OR, 2.44; 95% CI, 0.98-6.11; P = .06). Nomogram-predicted risks of RP ranged from 4% to 60% on the basis of MLD and V5, total radiation dose, and use of salvage chemotherapy. ConclusionsAchieving mean lung and V5 DVH constraints is critical to reduce RP risk in patients with lymphoma who receive mediastinal intensity modulated radiation therapy. The use of the DIBH technique is a promising risk-modifying treatment approach in patients with mediastinal lymphoma and especially in patients with a history of nonmodifiable risk factors for RP such as bulky disease and salvage chemotherapy.

An in silico comparative dosimetric study of postmastectomy locoregional irradiation using intensity-modulated vs 3-dimensional conventional radiotherapy

An in silico dosimetric evaluation of intensity-modulated radiation therapy (IMRT) vs 3-dimensional conventional radiation therapy (3D-CRT) treatment plans in postmastectomy radiation therapy (PMRT) to the chest wall and regional lymphatics was conducted. Twenty-five consecutive patients with breast cancer referred for locoregional PMRT, stages T2-4 with N1-3, were planned to receive 50 Gy in 25 fractions with IMRT. Additionally, a 3D-CRT plan was generated using identical contours for the clinical target volumes (CTV), planning target volumes (PTV), and organs at risk (OAR). Treatment plans were assessed using dose-volume histogram (DVH) parameters of D98, D95, D50, D2, and homogeneity index for individual CTVs and PTVs. OARs evaluated were ipsilateral and contralateral lungs, heart, spinal cord, and opposite breast. Most DVH parameters pertaining to CTVs and PTVs significantly favored IMRT. V20 for ipsilateral and contralateral lungs, D33 of heart and maximum dose to spinal cord favored IMRT (all p < 0.001). The mean dose to the opposite breast was significantly lesser with 3D-CRT (5.8 ± 1.8 Gy vs 2.0 ± 1.0 Gy, p < 0.001). Thus, except for the mean dose to the opposite breast, the compliance to DVH constraints applied to PTV and OARs were significantly better with IMRT. At a median follow-up of 76 months (7-91), none had locoregional failure or pulmonary or cardiac morbidity. For PMRT, requiring comprehensive irradiation to both chest wall and regional lymphatics, IMRT offers superior dosimetric advantages over 3D-CRT. This was also corroborated by long-term outcomes in these patients treated with IMRT.

Vertebral Bone Marrow Irradiation Contributes to Hematologic Toxicity During Chemoradiation Therapy for Esophageal Cancer

Hematological toxicity (HT) commonly occurs during chemoradiation therapy (CRT) for esophageal cancer, with both modalities contributing to the decline in blood counts. Since leukopenia commonly precludes chemotherapy delivery, we sought to determine the dose volume histogram (DVH) constraints that correlate with decline in leukocyte counts due to comprehensive vertebra (CV) irradiation during CRT. 32 esophageal cancer patients treated with weekly neoadjuvant CRT were selected to be in the study. The typical chemotherapy regimen consisted of weekly intravenous carboplatin (area under the curve = 2) and paclitaxel (50 mg/m2). Radiation therapy (RT) was delivered using 3-dimensional conformal radiation therapy (3D-CRT) or intensity modulated radiation therapy (IMRT). Patients who missed CRT for reasons other than neutropenia or received colony-stimulating factors were excluded. HT was scored using the Common Terminology Criteria for Adverse Events version 4.0. CV were contoured from C2 to L2 vertebra on 4DCT simulation scans. DVH data in both percentages and cubic centimeters was collected for CV V10-V60 (CVV) in increments of 10s. CVV5 and mean vertebral dose (MVD) were included as well. A DVH parameter of Vx was defined as either the percentage or volume in cubic centimeters of organ receiving at least x Gy of radiation. Univariate linear regression was performed to identify associations between leukopenia nadirs and DVH parameters. β represents the change in blood counts in k/μL for every 1 unit increase in DVH parameter. Receiver operator curves demonstrate cutoffs to avoid grade ≥ 3 leukopenia. Of the 32 esophageal patients, 6 patients developed grade 1 leukopenia, 12 developed grade 2 leukopenia, 8 developed grade 3 leukopenia, and 3 developed grade 4 leukopenia. CVV30, CVV20, CVV10, and MVD were significantly associated with decreasing WBC when calculated either as a percentage or by volume of the organ on univariate analysis. CVV5 was significantly associated with decreasing WBC counts only when calculated as a percentage of the volume and not as volume by cubic centimers. Associations with leukopenia were not seen with higher DVH values. CVV60, CVV50, and CVV40 did not significantly influence WBC levels during CRT. Cutoffs to avoid grade ≥ 3 leukopenia were CVV20 < 44.3%, CVV10 < 54.2%, CVV20 < 225 cc, CVV10 < 260 cc, and MVD < 18.8 Gy.< table class=”abstracttable>< tbody>< tr> CVV30 CVV20 CVV10 CVV5 MVD WBC (DVH by %) β p-value . -0.00528 0.0154 . -0.00767 0.0016 . -0.00826 0.0005 . -0.00828 0.0005 . -0.00014 0.0125 WBC (DVH by cc) . . . . . β p-value -0.00111 0.022 -0.00109 0.0269 -0.00107 0.0268 -0.000804 0.0577 -0.00014 0.0125 Leukopenia is associated with higher RT doses to the CVV30, CVV20, CVV10, CVV5, and MVD during CRT for esophageal cancer patients. Improved low dose radiation sparing of the CV may decrease HT.

Influence of daily imaging on plan quality and normal tissue toxicity for prostate cancer radiotherapy

BackgroundModern radiotherapy offers various possibilities for image guided verification of patient positioning. Different clinically relevant IGRT (image guided radiotherapy) scenarios were considered with regard to their influence on dosimetric plan quality and normal tissue complication probability (NTCP).MethodsThis study is based on treatment plans of 50 prostate patients. We evaluate the clinically performed IGRT and simulate the influence of different daily IGRT scenarios on plan quality. Imaging doses of planar and cone-beam-CT (CBCT) images for three different energies (6 MV, 1 MV and 121 kV) were added to the treatment plans. The plan quality of the different scenarios was assessed by a visual inspection of the dose distribution and dose-volume-histogram (DVH) and a statistical analysis of DVH criteria. In addition, an assessment of the normal tissue complication probability was performed.ResultsDaily 1MV-CBCTs result in undesirable high dose regions in the target volume. The DVH shows that the scenarios with actual imaging performed, daily kV-CBCT and daily 6MV imaging (1x CBCT, 4x planar images per week) do not differ exceedingly from the original plan; especially imaging with daily kV-CBCT has little influence to the sparing of organs at risk. In contrast, daily 1MV- CBCT entails an additional dose of up to two fraction doses. Due to the additional dose amount some DVH constraints for plan acceptability could no longer be satisfied, especially for the daily 1MV-CBCT scenario. This scenario also shows increased NTCP for the rectum.ConclusionDaily kV-CBCT has negligible influence on plan quality and is commendable for the clinical routine. If no kV-modality is available, a daily IGRT scenario with one CBCT per week and planar axial images on the other days should be preferred over daily MV-CBCT.

Cardiac Dose and Survival After Stereotactic Body Radiotherapy for Early-stage Non–Small-cell Lung Cancer

IntroductionRecent analyses have identified cardiac dose as an important predictor of overall survival (OS) after chemoradiation for locally advanced non–small-cell lung cancer (NSCLC). However, the survival influence of the cardiac dose after stereotactic body radiotherapy (SBRT) is unknown. We performed a dose–volume histogram (DVH) analysis of patients treated with SBRT for early stage NSCLC to examine survival and cardiac toxicity. Materials and MethodsWe reviewed the medical records of patients who had undergone SBRT for early-stage NSCLC from June 2007 to June 2015 and documented the cardiac DVH parameters, including the maximum and mean dose and percentage of volume receiving >5, >10, >20, and >30 Gy (V5, V10, V20, and V30, respectively). The biologically effective doses and 2-Gy equivalent doses were also calculated. The DVH parameters were assessed as predictors of OS using Cox regression analysis. ResultsWe identified 102 patients with 118 treated tumors. At a median follow-up period of 27.2 months (range, 9.8-72.5 months), the 2-year OS estimate was 70.4%. The cardiac DVH parameters were as follows: maximum dose, median, 14.2 Gy (range, 0.3-77.8 Gy); mean dose, median, 1.6 Gy (range, 0-12.6 Gy); and V5, median, 8.7% (range, 0%-96.4%). We identified no correlation between OS and any cardiac dose parameter. No patient developed acute (within 3 months) cardiac toxicity. Four patients died of cardiac causes; all had had preexisting heart disease. ConclusionIn our cohort, cardiac dose was not a predictor of OS after lung SBRT, despite a subset of patients receiving high maximum cardiac doses. The findings from our limited cohort showed that high doses to small volumes of the heart appear safe. Analyses of larger patient cohorts with longer follow-up durations are needed to better delineate the safe cardiac DVH constraints for SBRT.