Clinical Dose-volume Histogram Research Articles

Enhancing Precision in Cardiac Segmentation for Magnetic Resonance-Guided Radiation Therapy Through Deep Learning

IntroductionCardiac substructure dose metrics are more strongly linked to late cardiac morbidities than whole-heart metrics. MR-guided radiation therapy (MRgRT) enables substructure visualization during daily localization, allowing potential for enhanced cardiac sparing. We extend a publicly available state-of-the-art deep learning (DL) framework, nnU-Net, to incorporate self-distillation (nnU-Net.wSD) for substructure segmentation for MRgRT. MethodsEighteen (Institute A) patients who underwent thoracic or abdominal radiation therapy on a 0.35 T MR-guided linac were retrospectively evaluated. On each image, one of two radiation oncologists delineated reference contours of 12 cardiac substructures (chambers, great vessels, and coronary arteries) used to train (n=10), validate (n=3), and test (n=5) nnU-Net.wSD leveraging a teacher-student network and comparing to standard 3D U-Net. The impact of using simulation data or including 3-4 daily images for augmentation during training was evaluated for nnU-Net.wSD. Geometric metrics (Dice similarity coefficient (DSC), mean distance to agreement (MDA), and 95% Hausdorff distance (HD95)), visual inspection, and clinical dose volume histograms (DVHs) were evaluated. To determine generalizability, Institute A's model was tested on an unlabeled dataset from Institute B (n=22) and evaluated via consensus scoring and volume comparisons. ResultsnnU-Net.wSD yielded a DSC (reported mean ± standard deviation) of 0.65±0.25 across the 12 substructures (Chambers: 0.85±0.05, Great Vessels: 0.67±0.19, and Coronary Arteries 0.33±0.16, mean MDA <3 mm, and mean HD95 <9 mm) while outperforming the 3D U-Net (0.583±0.28, p<0.01). Leveraging fractionated data for augmentation improved over a single MR-SIM timepoint (0.579±0.29, p<0.01). Predicted contours yielded DVHs that closely matched the clinical treatment plans where mean and D0.03cc doses deviated by 0.32±0.5 Gy and 1.42±2.6 Gy respectively. No statistically significant differences between Institute A and B volumes (p>0.05) for 11 of 12 substructures with larger volumes requiring minor changes and coronary arteries exhibiting more variability. ConclusionsThis work is a critical step to rapid and reliable cardiac substructure segmentation to improve cardiac sparing in low-field MRgRT.

Domain knowledge driven 3D dose prediction using moment-based loss function

Objective. To propose a novel moment-based loss function for predicting 3D dose distribution for the challenging conventional lung intensity modulated radiation therapy plans. The moment-based loss function is convex and differentiable and can easily incorporate clinical dose volume histogram (DVH) domain knowledge in any deep learning (DL) framework without computational overhead. Approach. We used a large dataset of 360 (240 for training, 50 for validation and 70 for testing) conventional lung patients with 2 Gy × 30 fractions to train the DL model using clinically treated plans at our institution. We trained a UNet like convolutional neural network architecture using computed tomography, planning target volume and organ-at-risk contours as input to infer corresponding voxel-wise 3D dose distribution. We evaluated three different loss functions: (1) the popular mean absolute error (MAE) loss, (2) the recently developed MAE + DVH loss, and (3) the proposed MAE + moments loss. The quality of the predictions was compared using different DVH metrics as well as dose-score and DVH-score, recently introduced by the AAPM knowledge-based planning grand challenge. Main results. Model with (MAE + moment) loss function outperformed the model with MAE loss by significantly improving the DVH-score (11%, p < 0.01) while having similar computational cost. It also outperformed the model trained with (MAE + DVH) by significantly improving the computational cost (48%) and the DVH-score (8%, p < 0.01). Significance. DVH metrics are widely accepted evaluation criteria in the clinic. However, incorporating them into the 3D dose prediction model is challenging due to their non-convexity and non-differentiability. Moments provide a mathematically rigorous and computationally efficient way to incorporate DVH information in any DL architecture. The code, pretrained models, docker container, and Google Colab project along with a sample dataset are available on our DoseRTX GitHub (https://github.com/nadeemlab/DoseRTX)

Data-Driven Dose-Volume Histogram Prediction

PurposeTo evaluate dose-volume histogram (DVH) prediction from prior radiation therapy data. Methods and MaterialsAn Oncospace radiation therapy database was constructed including images, structures, and dose distributions for patients with advanced lung cancer. DVH data was queried for total lungs, esophagus, heart, and external body contours. Each query returned DVH data for the N-most similar organs at risk (OARs) based on OAR-to-planning-target-volume (PTV) geometry via the overlap volume histogram (OVH). The DVHs for 5, 20, and 50 of the most similar OVHs were returned for each OAR for each patient. The OVH(0cm) is the relative volume of the OAR overlapping with the PTV, and the OVH(2cm) is the relative volume of the OAR 2 cm away from the PTV. The OVH(cm) and DVH(%) queried from the database were separated into interquartile ranges (IQRs), nonoutlier ranges (NORs) (equal to 3 × IQR), and the average database DVH (DVH-DB) computed from the NOR data. The ability to predict the clinically delivered DVH was evaluated based on percentiles and differences between the DVH-DB and the clinical DVH (DVH-CL) for a varying number of returned patient DVHs for a subset of patients. ResultsThe ability to predict the clinically delivered DVH was excellent in the lungs and body; the IQR and NOR were <4% and <16%, respectively, in the lungs and <1% and <5%, respectively, in the body at all distances less than 2 cm from the PTV. For 21/23 patients considered, the differences in lung DVH-DB and DVH-CL were <4.6% and in 14/23 cases, <3%. In esophagus and heart, the ability to predict DVH-CL was weaker, with mean DVH differences >10% for 12/23 esophagi and 10/23 hearts. In esophagus and heart queries, the NOR was often 10% to 100% volume in dose ranges between 0% and 50% of prescription, independent of the number of patients queried. ConclusionsUsing prior data to predict clinical dosimetry is increasingly of interest, but model- and data-driven methods have limitations if based on limited data sets. This study's results showed that prediction may be reasonable in organs containing tumors with known overlap, but for nonoverlapped OARs, planning preference and plan design may dominate the clinical dose.

Reduction of Intra-Dosimetrist Variation, and Improvement of Overall Plan Quality as Assessed by Dose-Volume Histograms, Through Training and Implementation of Advanced Optimization Tools

Our Cancer Center network comprises a research Cancer Center as well as 10 independent community practices, with 18 total Dosimetrists. Many steps have been taken to increase treatment planning standardization, while dosimetry techniques and practices across the system have remained independent, with limited use of advanced optimization tools, knowledge-based planning (KBP) and multi-criteria optimization (MCO). With the objective to improve standardization of techniques, identify best practices, and improve overall quality, we implemented an in-house dosimetry plan challenge and solicited participation from 18 dosimetrists within the network.We anonymized an existing intermediate risk prostate adenocarcinoma clinical T1c N0 M0 patient dataset with physician drawn target contours (prostate plus proximal 1cm seminal vesicles) and normal structures, to create an identical test case for each dosimetrist. Clinically relevant planning objectives were provided in addition to instructions to produce a plan of clinical quality, committing typical clinical time/effort, using only standard optimization tools. Following submission of all plans, the results were discussed with the dosimetry team, including evaluations of dose-volume histograms (DVH). Instruction was provided on the advanced optimization tools KBP and MCO. Dosimetrists were instructed to repeat the challenge with the explicit instruction to use KBP and MCO. KBP/MCO plan results were compared to the plan results without KBP/MCO use to show the overall impact of advanced tools on achieved plan quality and intra-dosimetrist variation. All plans were explicitly normalized to the PTV D100% = 95% to allow direct comparison of DVHs.All initial plans submitted by 16 dosimetrists met the prescribed objectives outlined for clinical practice. However, plans demonstrated a wide range of acceptable dose-volume histograms. In addition to visual evaluation of the DVHs, multiple dose points for the Target (PTV), Rectum and Bladder, were exported for comparison between groups of plans. The group average achieved bladder doses were reduced from 52.8%Rx to 41.9%Rx, 23.9%Rx to 15.5%Rx, and 6.3%Rx to 5.5%Rx at D15%, D30%, and D60% respectively. The group average achieved rectal doses were reduced from 62.3%Rx to 45.8%Rx, 49.3%Rx to 32.7%Rx, and 33.9%Rx to 19.9%Rx at D10%, D20%, and D40% respectively. Similarly, the standard deviations also showed marked decrease with those for the bladder reduced from 0.084 to 0.044, 0.067 to 0.02, and 0.014 to 0.010 for D15%, D30%, and D60% respectively, and those for rectum reduced from 0.106 to 0.0690, 0.109 to 0.066, 0.086 to 0.038 at D10%, D20%, and D40% respectively.Standardization of clinical treatment planning objectives and consistent use of advanced optimization tools can improve overall planning quality and reduce intra-dosimetrist variation of treatment planning results based on clinical DVH metrics.

A practical method to quantify knowledge-based DVH prediction accuracy and uncertainty with reference cohorts.

The adoption of knowledge‐based dose‐volume histogram (DVH) prediction models for assessing organ‐at‐risk (OAR) sparing in radiotherapy necessitates quantification of prediction accuracy and uncertainty. Moreover, DVH prediction error bands should be readily interpretable as confidence intervals in which to find a percentage of clinically acceptable DVHs. In the event such DVH error bands are not available, we present an independent error quantification methodology using a local reference cohort of high‐quality treatment plans, and apply it to two DVH prediction models, ORBIT‐RT and RapidPlan, trained on the same set of 90 volumetric modulated arc therapy (VMAT) plans. Organ‐at‐risk DVH predictions from each model were then generated for a separate set of 45 prostate VMAT plans. Dose‐volume histogram predictions were then compared to their analogous clinical DVHs to define prediction errors Vclin,i‐Vpred,i (ith plan), from which prediction bias μ, prediction error variation σ, and root‐mean‐square error RMSEpred≡1N∑iVclin,i‐Vpred,i2≅σ2+μ2 could be calculated for the cohort. The empirical RMSEpred was then contrasted to the model‐provided DVH error estimates. For all prostate OARs, above 50% Rx dose, ORBIT‐RT μ and σ were comparable to or less than those of RapidPlan. Above 80% Rx dose, μ < 1% and σ < 3‐4% for both models. As a result, above 50% Rx dose, ORBIT‐RT RMSEpred was below that of RapidPlan, indicating slightly improved accuracy in this cohort. Because μ ≈ 0, RMSEpred is readily interpretable as a canonical standard deviation σ, whose error band is expected to correctly predict 68% of normally distributed clinical DVHs. By contrast, RapidPlan's provided error band, although described in literature as a standard deviation range, was slightly less predictive than RMSEpred (55–70% success), while the provided ORBIT‐RT error band was confirmed to resemble an interquartile range (40–65% success) as described. Clinicians can apply this methodology using their own institutions’ reference cohorts to (a) independently assess a knowledge‐based model's predictive accuracy of local treatment plans, and (b) interpret from any error band whether further OAR dose sparing is likely attainable.

Revisiting the formalism of equivalent uniform dose based on the linear-quadratic and universal survival curve models in high-dose stereotactic body radiotherapy

PurposeTo examine the equivalent uniform dose (EUD) formalism using the universal survival curve (USC) applicable to high-dose stereotactic body radiotherapy (SBRT).Materials and methodsFor nine non-small-cell carcinoma cell (NSCLC) lines, the linear-quadratic (LQ) and USC models were used to calculate the EUD of a set of hypothetical two-compartment tumor dose–volume histogram (DVH) models. The dose was varied by ±5%, ±10%, and ±20% about the prescription dose (60 Gy/3 fractions) to the first compartment, with fraction volume varying from 1% and 5% to 30%. Clinical DVHs of 21 SBRT treatments of NSCLC prescribed to the 70–83% isodose lines were also considered. The EUD of non-standard SBRT dose fractionation (EUDSBRT) was further converted to standard fractionation of 2 Gy (EUDCFRT) using the LQ and USC models to facilitate comparisons between different SBRT dose fractionations. Tumor control probability (TCP) was then estimated from the LQ- and USC-EUDCFRT.ResultsFor non-standard SBRT fractionation, the deviation of the USC- from the LQ-EUDSBRT is largely limited to 5% in the presence of dose variation up to ±20% to fractional tumor volume up to 30% in all NSCLC cell lines. Linear regression with zero constant yielded USC-EUDSBRT = 0.96 × LQ-EUDSBRT (r2 = 0.99) for the clinical DVHs. Converting EUDSBRT into standard 2‑Gy fractions by the LQ formalism produced significantly larger EUDCFRT than the USC formalism, particularly for low alpha /beta ratios and large fraction dose. Simplified two-compartment DVH models illustrated that both the LQ- and USC-EUDCFRT values were sensitive to cold spot below the prescription dose with little volume dependence. Their deviations were almost constant for up to 30% dose increase above the prescription. Linear regression with zero constant yielded USC-EUDCFRT = 1.56 × LQ-EUDCFRT (r2 = 0.99) for the clinical DVHs. The clinical LQ-EUDCFRT resulted in median TCP of almost 100% vs. 93.8% with USC-EUDCFRT.ConclusionA uniform formalism of EUD should be defined among the SBRT community in order to apply it as a single metric for dose reporting and dose–response modeling in high-dose-gradient SBRT because its value depends on the underlying cell survival model and the model parameters. Further investigations of the optimal formalism to derive the EUD through clinical correlations are warranted.

Deep Learning Based Dosimetry Evaluation at Organs-at-Risk in Esophageal Radiation Treatment Planning.

Rapid esophageal radiation treatment planning is often obstructed by manually adjusting optimization parameters. The adjustment process is commonly guided by the dose-volume histogram (DVH), which evaluates dosimetry at planning target volume (PTV) and organs at risk (OARs). DVH is highly correlated with the geometrical relationship between PTV and OARs, which motivates us to explore deep learning techniques to model such correlation and predict DVHs of different OARs. Distance to target histogram (DTH) is chosen to measure the geometrical relationship between PTV and OARs. DTH and DVH features are then undergone dimension reduction by autoencoder. The reduced feature vectors are finally imported into deep belief network to model the correlation between DTH and DVH. This correlation can be used to predict DVH of the corresponding OAR for new patients. Validation results revealed that the relative dose difference of the predicted and clinical DVHs on four different OARs were less than 3%. These promising results suggested that the predicted DVH could provide near-optimal parameters to significantly reduce the planning time.

EP-1721 Sensitivity study between gamma index passing rate and clinical dose volume histogram

EP-1721 Sensitivity study between gamma index passing rate and clinical dose volume histogram

Linking dose delivery accuracy and planning target margin in radiosurgery based on dose-volume histograms derived from measurement-guided dose reconstruction

In radiosurgery (SRS), the geometric uncertainties of machine-related delivery including image-guidance and hence the planning target volume (PTV) are often evaluated by the end-to-end gamma (γ) comparison that carries no information about the clinical relevance of deviations of individual SRS plans during delivery quality assurance (DQA). A proof-of-concept method was proposed to derive the PTV against both the plan- and the machine-specific delivery errors directly from the clinically relevant dose-volume histograms (DVHs) using measured-guided dose reconstruction (MGDR) during DQA. A liquid-filled detector array and a rotating phantom were used to measure sixteen arc-based radiosurgery treatments with 1 and 2 mm gross tumor volume (GTV)-to-PTV margins, producing MGDR-3D dose distribution on both the phantom and the patient CT for γ index and clinical DVH evaluations, respectively. The PTV was considered optimal when the MGDR showed the desired prescription dose coverage (V pres ) of the GTV (100% in this study). Associations of the binary V pres outcomes (<or =100%) of the GTV with the acceptance level of percent γ pass rate (γPR%) at 90 versus 95% were assessed. Further receiver operator characteristic (ROC) analysis was performed to assess the distance-to-agreement (DTA) and local dose difference (ΔD) criteria that may be suitable for treatment acceptance. From the MGDR, 100% GTV V pres was achieved in 68.8% and 100% of plans with 1 and 2 mm PTV, respectively. V pres outcomes were neither associated with γPR% at 1–2 mm DTA and 1%–3% ΔD nor the acceptance level for MGDR in the patient CT. ROC analysis shows statistically significant AUC values from 0.78–0.84 and 0.79–0.80 for MGDR phantom and patient doses, respectively. DQA by MGDR-DVH objectives offers the unique opportunity of direct assessment of the dose delivery accuracy and hence the optimal PTV without subject to the statistical correlation between γPR% and clinical metrics. Based on multi-criteria DVH objectives, clinical decision can be instantly made to adjust the treatment plan prescription.

Knowledge-based automated planning for oropharyngeal cancer.

The purpose of this study was to automatically generate radiation therapy plans for oropharynx patients by combining knowledge-based planning (KBP) predictions with an inverse optimization (IO) pipeline. We developed two KBP approaches, the bagging query (BQ) method and the generalized principal component analysis-based (gPCA) method, to predict achievable dose-volume histograms (DVHs). These approaches generalize existing methods by predicting physically feasible organ-at-risk (OAR) and target DVHs in sites with multiple targets. Using leave-one-out cross validation, we applied both models to a large dataset of 217 oropharynx patients. The predicted DVHs were input into an IO pipeline that generated treatment plans (BQ and gPCA plans) via an intermediate step that estimated objective function weights for an inverse planning model. The KBP predictions were compared to the clinical DVHs for benchmarking. To assess the complete pipeline, we compared the BQ and gPCA plans to both the predictions and clinical plans. To isolate the effect of the KBP predictions, we put clinical DVHs through the IO pipeline to produce clinical inverse optimized (CIO) plans. This approach also allowed us to estimate the complexity of the clinical plans. The BQ and gPCA plans were benchmarked against the CIO plans using DVH differences and clinical planning criteria. Iso-complexity plans (relative to CIO) were also generated and evaluated. The BQ method tended to predict that less dose is delivered than what was observed in the clinical plans while the gPCA predictions were more similar to clinical DVHs. Both populations of KBP predictions were reproduced with inverse plans to within a median DVH difference of 3Gy. Clinical planning criteria for OARs were satisfied most frequently by the BQ plans (74.4%), by 6.3% points more than the clinical plans. Meanwhile, target criteria were satisfied most frequently by the gPCA plans (90.2%), and by 21.2% points more than clinical plans. However, once the complexity of the plans was constrained to that of the CIO plans, the performance of the BQ plans degraded significantly. In contrast, the gPCA plans still satisfied more clinical criteria than both the clinical and CIO plans, with the most notable improvement being in target criteria. Our automated pipeline can successfully use DVH predictions to generate high-quality plans without human intervention. Between the two KBP methods, gPCA plans tend to achieve comparable performance as clinical plans, even when controlling for plan complexity, whereas BQ plans tended to underperform.

Inverse optimization of objective function weights for treatment planning using clinical dose-volume histograms

We developed and evaluated a novel inverse optimization (IO) model to estimate objective function weights from clinical dose-volume histograms (DVHs). These weights were used to solve a treatment planning problem to generate ‘inverse plans’ that had similar DVHs to the original clinical DVHs. Our methodology was applied to 217 clinical head and neck cancer treatment plans that were previously delivered at Princess Margaret Cancer Centre in Canada. Inverse plan DVHs were compared to the clinical DVHs using objective function values, dose-volume differences, and frequency of clinical planning criteria satisfaction. Median differences between the clinical and inverse DVHs were within 1.1 Gy. For most structures, the difference in clinical planning criteria satisfaction between the clinical and inverse plans was at most 1.4%. For structures where the two plans differed by more than 1.4% in planning criteria satisfaction, the difference in average criterion violation was less than 0.5 Gy. Overall, the inverse plans were very similar to the clinical plans. Compared with a previous inverse optimization method from the literature, our new inverse plans typically satisfied the same or more clinical criteria, and had consistently lower fluence heterogeneity. Overall, this paper demonstrates that DVHs, which are essentially summary statistics, provide sufficient information to estimate objective function weights that result in high quality treatment plans. However, as with any summary statistic that compresses three-dimensional dose information, care must be taken to avoid generating plans with undesirable features such as hotspots; our computational results suggest that such undesirable spatial features were uncommon. Our IO-based approach can be integrated into the current clinical planning paradigm to better initialize the planning process and improve planning efficiency. It could also be embedded in a knowledge-based planning or adaptive radiation therapy framework to automatically generate a new plan given a predicted or updated target DVH, respectively.

EP-1758: A correlation study between gamma index passing rate and clinical dose volume histogram

EP-1758: A correlation study between gamma index passing rate and clinical dose volume histogram

Clinical Dose-Volume Histogram Analysis for Radiation-Induced Proximal Bronchial Tree Toxicity in Patients With Non–small Cell Lung Cancer

To study risk factors for radiation-induced proximal bronchial tree (PBT) toxicity in patients treated with thoracic radiotherapy (RT) for non-small cell lung cancer (NSCLC). A total of 100 patients with medically inoperable/unresectable NSCLC treated with definitive thoracic RT in prospective studies were included in the study. PBT and PBT wall (PBTW) were contoured consistently per RTOG 1106 Atlas. The primary endpoint was any grade of PBT toxicity, including bronchial stricture/obstruction/fistula and atelectasis, which was diagnosed and graded based on clinical and radiographic presentation according to the National Cancer Institute's Common Terminology Criteria for Adverse Events, version 4.0. Dosimetric parameters including mean dose, maximum doses in 0.03 cc, 0.5 cc and 1.0 cc and volumes receiving doses greater than 5-90 Gy (V5-V90) were extracted for both PBT and PBTW. Equivalent uniform doses (EUD) were generated using a power-law based model (a value = 19). Cox proportional hazards model was used to estimate hazard ratios (HR) with 95% confidence interval (95% CI). Receiver operator characteristic curves were constructed to assess the predictive accuracy of positive factors for PBT toxicity and the areas under the ROC curve (AUC) were calculated. Of the 100 patients, 73 (73%) received 70 Gy or above, 17 (17%) developed PBT toxicity (Grade 1, 9%; Grade 2, 5% and Grade 4, 3%) after a minimal follow-up of 6 months. The median time interval between treatment initiation and onset of PBT toxicity was 8.4 months. Under univariate analysis, none of clinical factors, such as age, gender, tumor volume, tumor stage, was significantly associated with PBT toxicity. Dosimetric factors including EUD, mean dose, maximum doses, V20-V50, V75 and V80 of PBT/PBTW were all significantly associated with PBT toxicity (all P-values < 0.05). Under multivariate analysis, however, only V75_PBT (P = 0.013, HR = 1.03, 95% CI, 1.00-1.06), EUD_PBT (P = 0.022, HR = 1.06, 95% CI, 1.01-1.12), and EUD_PBTW (P = 0.022, HR = 1.06, 95% CI, 1.06-1.12) remained significant. AUCs of V75_PBT, EUD_PBT and EUD_PBTW were 0.66, 0.70 and 0.71, respectively. The dosimetric thresholds to limit the incidence of PBT toxicity to < 30% were 33%, 77 Gy, 78 Gy for V75_PBT, EUD_PBT and EUD_PBTW, respectively. V75 of PBT and EUD of PBT or PBTW are important significant dosimetric parameters in predicting PBT toxicity in thoracic RT planning. PBT toxicity may be limited by dosimetric constraints. Key words: Proximal bronchial tree; Toxicity; Dose-volume histogram; Equivalent uniform dose; Lung cancer.

On the feasibility of extracting dose–response curves from clinical DVH data using correlation and regression analysis

Purpose: To quantify the ability of correlation and regression analysis to extract the normal lung dose-response function from dose volume histogram (DVH) and radiation pneumonitis (RP) data. Methods: A local injury model is adopted, in which radiation-induced damage (functional loss) G is the integral of the DVH with function R(D). RP risk is H(G) where H() is the sigmoid cumulative distribution of functional reserve. RP incidence is a Bernoulli function of risk. A homogeneous patient cohort is assumed, allowing non-dose-related factors to be ignored. Clinically realistic DVHs are combined with the injury model to simulate RP data. Results: Correlation analysis is often used to identify a subset of predictor variables that are significantly correlated with outcome, for inclusion in a predictive model. In the local injury model, all DVH metrics VD contribute to damage through the integral with R(D). Correlation analysis therefore has limited value. The subset of VD that are most significantly correlated with incidence varies randomly from trial to trial as a result of random variations in the DVH set, and does not necessarily reveal anything useful about the patient cohort or the underlying biological dose-response relationship. Regression or matrix analysis has the potential to extract R(D) from damage or risk data, provided smoothness regularization is employed. Extraction of R(D) from incidence data was not successful, due to its higher level of statistical variability. Conclusions: To the authors’ knowledge, smoothness regularization has not been applied to this problem, so represents a novel approach. Dose–response functions can be successfully extracted from measurements of integral (as opposed to regional) lung damage G, suggesting value in re-visiting available measurements of ventilation, perfusion and radiographic damage. The techniques developed here can potentially be used to extract the dose–response functions of different tissues from multiple types of quantitative volumetric imaging data.

Extracting the normal lung dose–response curve from clinical DVH data: a possible role for low dose hyper-radiosensitivity, increased radioresistance

In conventionally fractionated radiation therapy for lung cancer, radiation pneumonitis’ (RP) dependence on the normal lung dose-volume histogram (DVH) is not well understood. Complication models alternatively make RP a function of a summary statistic, such as mean lung dose (MLD). This work searches over damage profiles, which quantify sub-volume damage as a function of dose. Profiles that achieve best RP predictive accuracy on a clinical dataset are hypothesized to approximate DVH dependence.Step function damage rate profiles R(D) are generated, having discrete steps at several dose points. A range of profiles is sampled by varying the step heights and dose point locations. Normal lung damage is the integral of R(D) with the cumulative DVH. Each profile is used in conjunction with a damage cutoff to predict grade 2 plus (G2+) RP for DVHs from a University of Michigan clinical trial dataset consisting of 89 CFRT patients, of which 17 were diagnosed with G2+ RP.Optimal profiles achieve a modest increase in predictive accuracy—erroneous RP predictions are reduced from 11 (using MLD) to 8. A novel result is that optimal profiles have a similar distinctive shape: enhanced damage contribution from low doses (<20 Gy), a flat contribution from doses in the range ~20–40 Gy, then a further enhanced contribution from doses above 40 Gy. These features resemble the hyper-radiosensitivity / increased radioresistance (HRS/IRR) observed in some cell survival curves, which can be modeled using Joiner’s induced repair model.A novel search strategy is employed, which has the potential to estimate RP dependence on the normal lung DVH. When applied to a clinical dataset, identified profiles share a characteristic shape, which resembles HRS/IRR. This suggests that normal lung may have enhanced sensitivity to low doses, and that this sensitivity can affect RP risk.

Monte Carlo simulations of patient dose perturbations in rotational-type radiotherapy due to a transverse magnetic field: a tomotherapy investigation.

Several groups are exploring the integration of magnetic resonance (MR) image guidance with radiotherapy to reduce tumor position uncertainty during photon radiotherapy. The therapeutic gain from reducing tumor position uncertainty using intrafraction MR imaging during radiotherapy could be partially offset if the negative effects of magnetic field-induced dose perturbations are not appreciated or accounted for. The authors hypothesize that a more rotationally symmetric modality such as helical tomotherapy will permit a systematic mediation of these dose perturbations. This investigation offers a unique look at the dose perturbations due to homogeneous transverse magnetic field during the delivery of Tomotherapy(®) Treatment System plans under varying degrees of rotational beamlet symmetry. The authors accurately reproduced treatment plan beamlet and patient configurations using the Monte Carlo code geant4. This code has a thoroughly benchmarked electromagnetic particle transport physics package well-suited for the radiotherapy energy regime. The three approved clinical treatment plans for this study were for a prostate, head and neck, and lung treatment. The dose heterogeneity index metric was used to quantify the effect of the dose perturbations to the target volumes. The authors demonstrate the ability to reproduce the clinical dose-volume histograms (DVH) to within 4% dose agreement at each DVH point for the target volumes and most planning structures, and therefore, are able to confidently examine the effects of transverse magnetic fields on the plans. The authors investigated field strengths of 0.35, 0.7, 1, 1.5, and 3 T. Changes to the dose heterogeneity index of 0.1% were seen in the prostate and head and neck case, reflecting negligible dose perturbations to the target volumes, a change from 5.5% to 20.1% was observed with the lung case. This study demonstrated that the effect of external magnetic fields can be mitigated by exploiting a more rotationally symmetric treatment modality.

SU-E-T-92: Creation of a Comprehensive Head and Model Using Knowledge Based Planning

Purpose: The purpose of this study was to evaluate if a commercial implementation of Knowledge Based Planning (KBP) software can be used to estimate Dose Volume Histogram (DVHs) on a comprehensive data set of Head and Neck (HN) patients. Methods: KBP is a tool capable of estimating DVHs for Organs At Risk (OARs) based on the DVHs of plans of similar patients treated in the past. This study used a newly developed commercial implementation of KBP to create a HN model. The model was trained using a database of retrospectively treated HN patients. This database covered the spectrum of cases expected to be found in the clinic, including multiple targets and 18 different dose prescription combinations. A set of independent validation patients was used to quantify the accuracy of DVHs estimated using the model and covered the same spectrum of HN cases. Results: The accuracy of the model was calculated by comparing the volumes of the estimated and clinical DVHs at doses equal to 50%, 85% and 99% of the maximum OAR dose. This allowed us to quantify the accuracy of the estimated DVHs even in cases when the OAR was receiving a low dose. The highest accuracy was obtained in the estimation of the DVHs for the Brain (<1% on average). The accuracy for the Brainstem, Cord, Mandible, Oral Cavity, Parotids and the Pharyngeal Constrictor ranged from -2% to 9% on average. Conclusion: This study shows the feasibility to estimate DVHs in a wide range of HN cases using a novel KBP algorithm. The use of a comprehensive set of patients results in a robust HN model which can be used in a wide range of clinical cases. Further planning is required to confirm if the current accuracy is sufficient to guide the planning process. The authors are Clinical Evaluators/Consultants for Varian Medical Systems. This study was partially funded by Varian Medical Systems.

Sensitivity of volumetric modulated arc therapy patient specific QA results to multileaf collimator errors and correlation to dose volume histogram based metrics

This study investigates the impact of systematic multileaf collimator (MLC) positional errors on gamma analysis results used for quality assurance (QA) of Rapidarc treatments. In addition, this study evaluates the relationship of these gamma analysis results and clinical dose volume histogram metrics (DVH) for Rapidarc treatment plans. Five prostate plans were modified by the introduction of systematic MLC errors. The MLC shifts to each individual active leaf introduced were 0.25, 0.5, 0.75, and 1 mm. All QA verification plans were delivered and estimated 3D patient dose or high density phantom dose were obtained based on the ArcCHECK measurement files. QA gamma analysis of 3%/3 mm and 2%/2 mm were implemented and relationships to dose differences in DVH metrics encountered due to MLC errors were determined. Tolerances of 3% and 5% for DVH metric were implemented to determine the sensitivity of gamma analysis to MLC errors. A calculation of sensitivity was determined from the number of incidences of false negative and false positive cases in gamma analysis results. The sensitivity of global gamma analysis for criteria of 3%/3 mm was 0.78 and for 2%/2 mm was 0.82. A number of instances occurred for an acceptable VMAT QA gamma index which did not indicate a DVH metric dose error greater than 5%. The correlation between global gamma analysis using criteria 3%/3 mm and DVH metric dose error were all <0.8 indicating less than a strong correlation. There is a greater sensitivity for detection of dosimetric errors occurring in a Rapidarc plan using gamma criteria of 2%/2 mm than 3%/3 mm. However, there is lack of consistently strong correlation between global gamma indexes and clinical DVH metrics for PTV and bladder and rectum for Rapidarc plans. It is recommended that the sole use of gamma index for Rapidarc QA plan evaluation could be insufficient and a methodology for evaluation of delivered dose to patient is required.

Pretreatment patient‐specific IMRT quality assurance: A correlation study between gamma index and patient clinical dose volume histogram

The aim of this work is to investigate the predictive power of a common conventional intensity modulated radiation therapy (IMRT) quality assurance (QA) performance metric, the gamma passing rate (%GP), through the analysis of the sensitivity and of the correlation between %GP and different dose discrepancies between planned dose-volume histogram (DVH) and perturbed DVH. The perturbed DVH is calculated by using a dedicated software, 3DVH (Sun Nuclear Corporation, Melbourne, FL), which is able to modify the dose distribution calculated by the treatment planning system (TPS) according to the dose discrepancies detected with planar measurements in order to predict the delivered 3D dose distribution in the patient. Twenty-seven high-risk prostate cancer (PP) patients and 15 head and neck (HN) cancer patients, treated with IMRT technique, were analyzed. Pretreatment verifications were performed for all patients' plans by acquiring planar dose distributions of each treatment field with 2D-diode array. Measured dose distributions were compared to the calculated ones using the gamma index (GI) method applying both global (Van Dyk) and local normalization, and %GP were generated for each pair of planar doses using the following acceptance criteria: 1%∕1, 2%∕2, and 3%∕3 mm. Planar dose distributions acquired during pretreatment verifications, together with patient's DICOM RT plan, RT structure set, and RT dose files from TPS were loaded into the 3DVH software. Percentage dose differences (%DE) between DVHs, obtained by TPS and by 3DVH, were calculated; statistical correlation between %DE and %GP was studied by using Pearson's correlation coefficient (r). This analysis was performed, for each patient, on planning target volumes and on some typical organs at risk of the prostatic and head and neck anatomical district. The sensitivity was calculated to correctly identify the pretreatment plans with high dose errors and to quantify the incidence of false negatives, on varying the gamma index method. Analysis of %DE vs %GP showed that there were only weak correlations (Pearson's r-values < 0.8). The results also showed numerous instances of false negatives (cases where high IMRT QA passing rates did not imply good agreement in anatomy dose metrics) and the reverse, mainly for the 3%∕3 mm global gamma passing rate. The lack of correlation between conventional IMRT QA performance metrics gamma passing rates and dose errors in DVHs values and the low sensitivity of 3%∕3 mm global gamma method show that the most common published acceptance criteria have disputable predictive power for per-patient IMRT QA.

Predicting dose‐volume histograms for organs‐at‐risk in IMRT planning

The objective of this work was to develop a quality control (QC) tool to reduce intensity modulated radiotherapy (IMRT) planning variability and improve treatment plan quality using mathematical models that predict achievable organ-at-risk (OAR) dose-volume histograms (DVHs) based on individual patient anatomy. A mathematical framework to predict achievable OAR DVHs was derived based on the correlation of expected dose to the minimum distance from a voxel to the PTV surface. OAR voxels sharing a range of minimum distances were computed as subvolumes. A three-parameter, skew-normal probability distribution was used to fit subvolume dose distributions, and DVH prediction models were developed by fitting the evolution of the skew-normal parameters as a function of distance with polynomials. Cohorts of 20 prostate and 24 head-and-neck IMRT plans with identical clinical objectives were used to train organ-specific average models for rectum, bladder, and parotids. A sum of residuals analysis quantifying the integrated difference between the clinically approved DVH and predicted DVH evaluated similarity between DVHs. The ability of the average models to prospectively predict DVHs was evaluated on an independent validation cohort of 20 prostate plans. Statistical comparison of the sums of residuals between training and validation cohorts quantified the accuracy of the average model. Restricted sums of residuals (RSR) were used to identify potential outliers, where large values of RSR indicate a clinical DVH that exceeds the predicted DVH by a considerable amount. A refined model was obtained for each organ by excluding outliers with large RSR values from the training cohort. The refined model was applied to the original training cohort and restricted sums of residuals were utilized to estimate potential DVH improvements. All cases were replanned and evaluated by the physician that approved the original plan. The ability of the refined models to correctly identify outliers was assessed using the residual sum between the original and replanned DVHs to quantify dosimetric gains realized under replanning. Statistical analysis of average sum of residuals for rectum (SR(rectum)=0.003±0.037), bladder (SR(bladder)=-0.008±0.037), and parotid (SR(parotid)=-0.003±0.060) training cohorts yielded mean values near zero and small with respect to the standard deviations, indicating that the average models are capturing the essential behavior of the training cohorts. The predictive abilities of the average rectum and bladder models were statistically indistinguishable between the training and validation sets, with SR(rectum)=0.002±0.044 and SR(bladder)=-0.018±0.058 for the validation set. The refined models' ability to detect outliers and predict achievable OAR DVHs was demonstrated by a strong correlation between predicted gains (RSR) and realized gains after replanning with sample correlation coefficients of r = 0.92 for the rectum, r = 0.88 for the bladder, and r = 0.84 for the parotid glands. The results demonstrate that our mathematical framework and modest training cohorts successfully predict achievable OAR DVHs based on individual patient anatomy. The models correctly identified suboptimal plans that demonstrated further OAR sparing after replanning. This modeling technique requires no manual intervention except for appropriate selection of a training set with identical evaluation criteria. Clinical implementation is in progress to evaluate impact on real-time IMRT QC.