Dosimetric Uncertainties Research Articles

Bolus Effect Caused by Use of Thermoplastic Masks in Head and Neck Radiotherapy Treatments.

This paper examines the dosimetric uncertainty arising from the use of thermoplastic masks in the treatment of head and neck cancer through radiotherapy. This study was conducted through Monte Carlo simulations using the Monte Carlo N-Particle eXtended (MCNPX code), and the theoretical results are compared with radiochromic films. Using material characterization techniques, the compounds of the thermoplastic mask were identified, confirming that most of the material corresponds to the polymer C10H16O4. The theoretical results show increases ranging from 42% to 57.4% in the surface absorbed dose for 6 and 15 MV photon beams, respectively, compared to the absorbed dose without the mask. The experimental data corroborate these findings, showing dose increases ranging from 18.4% to 52.1% compared to the expected surface absorbed dose without the mask. These results highlight the need to consider the bolus effect induced by thermoplastic masks during the precise and safe planning and application of radiotherapy treatment in order to ensure its therapeutic efficacy and minimize the associated risks to patients.

Impact of delivery variations on 3D dose distributions for volumetric modulated arc therapy plans of various complexity.

Delivery variations during radiotherapy can cause discrepancies between planned and delivered dose distribution. These variations could arise from random and systematic offsets in certain machine parameters or systematic offsets related to the calibration process of the treatment unit. The aim of this study was to present a novel simulation-based methodology to evaluate realistic delivery variations in three dimensions (3D). Additionally, we investigated the dosimetric impact of delivery variations for volumetric modulated arc therapy (VMAT) plans for different treatment sites and complexities. Twelve VMAT plans for different treatment sites (prostate-, head & neck-, lung-, and gynecological cancer) were selected. The clinical plan used for the treatment of each patient was reoptimized to create one plan with reduced complexity (i.e., simple plan) and one of higher complexity (i.e., complex plan). This resulted in a total of 36 plans. Delivery variations were simulated by randomly introducing offsets in multi-leaf collimator position, jaw position, gantry angle and collimator angle simultaneously. Twenty simulations were carried out for each of the 36 plans, yielding 720 simulated deliveries. To explore the impact of individual offsets, additional simulations were conducted for each type of offset separately. A 3D dose calculation was performed for each simulation using the same calculation engine as for the clinical plan. Two standard deviations (2SD) of dose were determined for every voxel for 3D-spatial evaluations. The dose variation in certain DVH metrics, that is, D2% and D98% for the clinical target volume and five different DVH metrics for selected organs at risk, was calculated for the twenty simulated deliveries of each plan. For comparison, the effect of delivery variations was assessed by conducting measurements with the Delta4 phantom. The volume of voxels with 2SD above 1% of the prescribed dose was consistently larger for the complex plans in comparison to their corresponding simple and clinical plans. 2SDs larger than 1% were in many cases, found to accumulate outside the planning target volume. For complex plans, regions with 2SDs larger than 1% were detected also inside the high dose region, exhibiting, on average, a size six times larger volume, than those observed in simple plans. Similar results were found for all treatment sites. Variation in the selected DVH metrics for the simulated deliveries was generally largest for the complex plans with few exceptions. When comparing the 2SD distribution of the measurements with the 2SD distribution from the simulations, the spatial information showed deviations outside the PTV in both simulations and measurements. However, the measured values were, on average, 35% higher for the prostate plans and 10% higher for the head & neck plans compared to the simulated values. The presented methodology effectively quantified and localized dose deviations due to delivery offsets. The 3D analysis provided information that was undetectable using the analysis based on DVH metrics. Dosimetric uncertainties due to delivery variations were prominent at the edge of the high-dose region irrespective of treatment site and plan complexity. Dosimetric uncertainties inside the high-dose region was more profound for plans of higher complexity.

Quantifying the Risk of Technology-Driven Health Disparities in Radiation Oncology

PurposeNew technologies are continuously emerging in radiation oncology. Inherent technological limitations can result in healthcare disparities in vulnerable patient populations. These limitations must be considered for existing and new technologies in the clinic to provide equitable care. Materials and MethodsWe created a health disparity risk assessment metric inspired by failure mode and effects analysis (FMEA). We provide sample patient populations and their potential associated disparities, guidelines for clinics and vendors, and example applications of the methodology. ResultsA disparity risk priority number (dRPN) can be calculated from the product of three quantifiable metrics: the percent of patients impacted (P), the severity of the impact of dosimetric uncertainty or quality of the radiation plan (S), and the clinical dependence on the evaluated technology (C). The dRPN can be used to rank the risk of sub-optimal care due to technical limitations when comparing technologies and to plan interventions when technology is shown to have inequitable performance in the patient population of a clinic. ConclusionThe proposed methodology may simplify the evaluation of how new technology impacts vulnerable populations, help clinics quantify the limitations of their technological resources, and plan appropriate interventions to improve equity in radiation treatments.

Impact of setup and geometric uncertainties on the robustness of free-breathing photon radiotherapy of small lung tumors

Purpose:Respiratory motion and patient setup error both contribute to the dosimetric uncertainty in radiotherapy of lung tumors. Managing these uncertainties for free-breathing treatments is usually done by margin-based approaches or robust optimization. However, breathing motion can be irregular and concerns have been raised for the robustness of the treatment plans. We have previously reported the dosimetric effects of the respiratory motion, without setup uncertainties, in lung tumor photon radiotherapy using free-breathing images. In this study, we include setup uncertainty. Methods:Tumor positions from cine-CT images acquired in free-breathing were combined with per-fraction patient shifts to simulate treatment scenarios. A total of 14 patients with 300 tumor positions were used to evaluate treatment plans based on 4DCT. Four planning methods aiming at delivering 54 Gy as median tumor dose in three fractions were compared. The planning methods were denoted robust 4D (RB4), isodose to the PTV with a central higher dose (ISD), the ISD method normalized to the intended median tumor dose (IRN) and homogeneous fluence to the PTV (FLU). Results:For all planning methods 95% of the intended dose was achieved with at least 90% probability with RB4 and FLU having equal CTV D50% values at this probability. FLU gave the most consistent results in terms of CTV D50% spread and dose homogeneity. Conclusions:Despite the simulated patient shifts and tumor motions being larger than observed in the 4DCTs the dosimetric impact was suggested to be small. RB4 or FLU are recommended for the planning of free-breathing treatments.

Investigation on Individual Variation of Organ Doses for Photon External Exposures: A Monte Carlo Simulation Study

Background: The reference dose coefficients (DCs) of the International Commission on Radiological Protection (ICRP) have been widely used to estimate organ doses of individuals for risk assessments. This approach has been well accepted because individual anatomy data are usually unavailable, although dosimetric uncertainty exists due to the anatomical difference between the reference phantoms and the individuals. We attempted to quantify the individual variation of organ doses for photon external exposures by calculating and comparing organ DCs for 30 individuals against the ICRP reference DCs.Materials and Methods: We acquired computed tomography images from 30 patients in which eight organs (brain, breasts, liver, lungs, skeleton, skin, stomach, and urinary bladder) were segmented using the ImageJ software to create voxel phantoms. The phantoms were implemented into the Monte Carlo N-Particle 6 (MCNP6) code and then irradiated by broad parallel photon beams (10 keV to 10 MeV) at four directions (antero-posterior, postero-anterior, left-lateral, right-lateral) to calculate organ DCsResults and Discussion: There was significant variation in organ doses due to the difference in anatomy among the individuals, especially in the kilovoltage region (e.g., <100 keV). For example, the red bone marrow doses at 0.01 MeV varied from 3 to 7 orders of the magnitude depending on the irradiation geometry. In contrast, in the megavoltage region (1–10 MeV), the individual variation of the organ doses was found to be negligibly small (differences <10%). It was also interesting to observe that the organ doses of the ICRP reference phantoms showed good agreement with the mean values of the organ doses among the patients in many casesConclusion: The results of this study would be informative to improve insights in individualspecific dosimetry. It should be extended to further studies in terms of many different aspects (e.g., other particles such as neutrons, other exposures such as internal exposures, and a larger number of individuals/patients) in the future.

Comparative analysis of dosimetric uncertainty using Gafchromic™ EBT4 and EBT3 films in radiochromic film dosimetry

Advancements in radiotherapy dosimetry aim to enhance precision and accuracy, with Gafchromic™ EBT4 film emerging as a promising tool. However, its performance compared to EBT3 remains underexplored, necessitating a comprehensive analysis of dosimetric uncertainties. This study evaluates the efficacy of EBT4 in radiochromic dosimetry, comparing it to EBT3. Employing a 6 MV beam quality, absorbed doses ranging from 0 to 10 Gy were investigated using different measurement quantities and fitting functions. Gafchromic™ EBT3 and EBT4 films (8′′ × 10″) were irradiated in a water-equivalent phantom, with calibration doses delivered using a Varian Truebeam Stx linear accelerator. Two measurement quantities, pixel values (PV) and net change transmittance (netΔT), were analyzed, along with calibration functions including a rational function for PV and a logarithmic function for netΔT. Uncertainty analyses considered experimental, fitting, and calibration dose uncertainties. Dose-response curves for EBT3 and EBT4 films demonstrated robust fits, emphasizing the reliability of both. EBT4 consistently exhibited lower uncertainties, particularly below 2 Gy, showcasing superior precision in low-dose scenarios. The choice of response quantity and calibration function influenced results, with PV-rational function excelling for doses 2–10 Gy, and netΔT-logarithmic function showing enhanced precision in the 0–2 Gy range. The uncertainty analysis emphasized EBT4's superior performance, especially in low-dose ranges, with lower relative uncertainties and absolute errors, indicating enhanced reliability in dose measurements. The study underscores EBT4's potential to contribute to improved accuracy in various clinical scenarios. However, users should tailor the choice of response quantities and calibration functions to specific dose calibration ranges, suggesting a nuanced approach.

Evaluation of in vitro irradiation setup: Designed for the horizontal beamline at the Danish Centre for Particle Therapy.

Radiobiological experimental setups are challenged by precise sample positioning along depth dose profile, scattering conditions, and practical difficulties that must be addressed in individual designs. The aim of this study was to produce cell survival curves with several irradiation modalities, by using a setup designed at the Danish Centre for Particle Therapy (DCPT) for in vitro proton irradiations using a horizontal beam line and thereby evaluating the setups use for in vitro irradiations experiments. The setup is a water phantom suitable for in vitro research with multiple irradiation modalities, in particular the pencil scanning proton beam available from a horizontal experimental beamline. The phantom included a water tank of 39.0 × 17.0 × 20.5 cm. Cell survival-curves were produced using the cell line V79 Chinese hamster lung fibroblast cells (V79s) in biological triplicates of clonogenic assays. Cell survival curves were produced with both a 18 MeV electron beam, 6 MV photon beam, and a Spread-Out Bragg Peak (SOBP) proton beam formed by pristine energies of 85-111 MeV where three positions were examined. Survival curves with uncertainty areas were made for all modalities. Dosimetric uncertainty amounted to, respectively, 4%, 3% and 3% for proton, electron, and high energy photon irradiations. Cell survival fraction uncertainty was depicted as the standard deviation between replications of the experiment. Cell survival curves could be produced with acceptable uncertainties using this novel water phantom and cellular laboratory workflow. The setup is useful for future in vitro irradiation experiments.

Measurements of the Variation of RFKO Extracted Beam Current for Hitachi Proton Therapy and Carbon Therapy Synchrotrons and Implications for Particle Therapy

Measurements of the extracted beam current (BC) for a Hitachi Carbon Therapy Synchrotron and a Hitachi Compact Proton Therapy Synchrotron are reported for the nominal extracted beam current (BC0) of 10mMUmsec , which was chosen since it is the magnitude of the clinical BC0, and a sample rate of 5 μsec (Carbon) and 8 μsec (Proton). A Noise Power Spectrum (NPS) analysis identifies the source of variation to be beam or power supply related. The rise time in the BC has been modeled and its effect on beam delivery time simulations have been estimated. The impact of the variation in BC from BC0 is shown to cause potentially significant dosimetric uncertainties in treatment delivery for modern particle therapy accelerators using fast Scanning Magnet (SCM) if plans are not simply beam current moderated or robustly optimized. The variation in beam current is shown to be inconsequential for medical physics quality assurance and commissioning measurements using properly biased ion chambers.

Investigation of Effect of Air Gap between Surface and Bolus on Dose Distribution for 6 MV Photon Beam

In radiotherapy, tissue equivalent boluses are frequently used in the treatment of superficially located tumors. The air gap between the patient's skin and the bolus may cause dosimetric uncertainties. This study aims to dosimetrically investigate the effect of the air gap between the surface and the bolus on dose distribution. Computed tomography (CT) images of the phantom were obtained and transferred to the treatment planning system (TPS). In the TPS, a bolus was placed on the phantom surface and then air gaps were created between the bolus and the surface. The effect of the air gaps between the surface and the 5 mm thick bolus on the dose distribution was analyzed with the point doses obtained from the TPS. For the 6 MV X-ray, it was observed that the air gap negatively affected the surface doses calculated by TPS. Accordingly, an inverse correlation was found between air gap and surface dose. It is recommended that bolus use, especially in curved anatomical regions, should be applied before CT scanning as much as possible. When using bolus material in radiotherapy, it is recommended to be careful not to leave an air gap between the surface and the bolus.

Mitigation of Dosimetric Uncertainty in MRI-Based Proton Planning Using Spot-Scanning Proton Arc (SPArc) Technique

MRI-based synthetic CT (SCT) images created using generative adversarial network (GAN) have been demonstrated to be feasible for intensity-modulated proton therapy (IMPT) planning. However, dose calculation accuracy can be uncertain in some regions within/near the target of head and neck patients due to the local CT number estimation error or sharp dose fall-off. This study investigated the feasibility of using the SPArc technique to mitigate such dosimetric uncertainty. A GAN using a 3D U-net as the generator and a 6-layer 3D convolutional neural network as the discriminator was trained with T1-weighted MR-CT image pairs from 162 nasopharyngeal carcinoma patients (14 for validation). The generator was used to generate SCT images from MR images for 7 test patients. For each test patient, the CT image was used to create a SPArc plan and an IMPT plan with the same clinical objectives. The SPArc plans (control point frequency sampling, arc trajectory, etc.) were optimized using a previously developed iterative approach. The dose distributions of both SPArc plans and IMPT plans were re-calculated on the SCT images and compared to the one calculated on the CT images. The dosimetric uncertainty was quantified using the gamma index. The 2%/2mm and 3%/3mm passing rates for SPArc plans were (96.9¡À2.7) % and (98.6¡À1.5) %, while the passing rates for IMPT plans were (94.0¡À3.9) % and (96.4+2.9) %. A significant reduction in dosimetric uncertainty was identified for SPArc plans (p ¡Ü0.021). Table 1 shows the passing rates for the 7 test individuals. SPArc can mitigate the uncertainty of dose calculation in MRI-based proton planning. Further research needs to validate these findings on a larger patient cohort. The study paves the road map for using MRI for SPArc planning.

Towards Automated Dosimetric Analysis of the Bladder Trigone: Deep-Learning-Based Joint Segmentation and Landmark Localization

The bladder trigone dosimetry is hypothesized to have a stronger correlation with post-SBRT urinary toxicity than that of the entire bladder. However, the trigone tends to move significantly between simulation and daily treatment. Its small size, large daily motion, and proximity to the target lead to potentially consequential but unaccounted-for dosimetric uncertainties. Manual segmentation of the structure can be inconsistent and time-consuming, even with MR-guided RT. Here, we propose and demonstrate a deep-learning-based framework for joint segmentation and landmark localization to support deformable registration and comprehensive dosimetric analysis. A total of 30 patients were randomly selected for training, and 20 were held out for testing. Each patient had 1 simulation and 5 daily pre-treatment images obtained from a clinical 0.35T MR Linac. The trigone is defined as the triangular bladder section among three landmarks (2 ureteral orifices and the internal urethral orifice). In the manual contouring process, the 3 landmarks were identified first, followed by trigone segmentation. The proposed joint method uses a modified 3D nnU-Net with 2 decoders, one for segmentation and the other for landmark localization. The shared encoder is expected to extract features useful for both tasks. The joint framework was compared with a baseline method using two separate 3D nnU-Nets for landmark localization and trigone segmentation, respectively. Since the trigone is small (∼2% of the bladder volume), we further experimented with a second-stage prediction mimicking the human contouring process. The predicted landmarks from the first stage were used to crop the trigone region, and a second network was trained on cropped images. Evaluation metrics included the Dice score, 95% Hausdorff distance (HD95), and average surface distance (ASD) for segmentation, and Euclidean distance (ED) between the predicted and ground truth landmarks for localization. The quantification metrics are summarized in the table below. The joint approach shows similar Dice performance to the baseline method but markedly better HD95 by 13%. For landmark localization, the proposed method is similar to the baseline, but the integration of the segmentation task stabilizes the training process. The two-stage approach further improves HD95, ASD, and ED by 27%, 24%, and 19%. Combining segmentation and landmark localization exhibits a synergistic effect. The proposed two-stage approach provided additional improvement. Future studies will explore the deformable registration of the trigone based on the segmentation and landmark detection, as well as analyze cumulated dose distribution.

Feasibility of artificial intelligence-driven interfractional monitoring of organ changes by mega-voltage computed tomography in intensity-modulated radiotherapy of prostate cancer.

High-dose radiotherapy (RT) for localized prostate cancer requires careful consideration of target position changes and adjacent organs-at-risk (OARs), such as the rectum and bladder. Therefore, daily monitoring of target position and OAR changes is crucial in minimizing interfractional dosimetric uncertainties. For efficient monitoring of the internal condition of patients, we assessed the feasibility of an auto-segmentation of OARs on the daily acquired images, such as megavoltage computed tomography (MVCT), via a commercial artificial intelligence (AI)-based solution in this study. We collected MVCT images weekly during the entire course of RT for 100 prostate cancer patients treated with the helical TomoTherapy system. Based on the manually contoured body outline, the bladder including prostate area, and rectal balloon regions for the 100 MVCT images, we trained the commercially available fully convolutional (FC)-DenseNet model and tested its auto-contouring performance. Based on the optimally determined hyperparameters, the FC-DenseNet model successfully auto-contoured all regions of interest showing high dice similarity coefficient (DSC) over 0.8 and a small mean surface distance (MSD) within 1.43 mm in reference to the manually contoured data. With this well-trained AI model, we have efficiently monitored the patient's internal condition through six MVCT scans, analyzing DSC, MSD, centroid, and volume differences. We have verified the feasibility of utilizing a commercial AI-based model for auto-segmentation with low-quality daily MVCT images. In the future, we will establish a fast and accurate auto-segmentation and internal organ monitoring system for efficiently determining the time for adaptive replanning.

Towards simulation-free MR-linac treatment: utilizing male pelvis PSMA-PET/CT and population-based electron density assignments

Objective. This study aimed to investigate the dosimetric impact of using population-based relative electron density (RED) overrides in lieu of simulation computerized tomography (CT) in a magnetic resonance linear accelerator (MRL) workflow for male pelvis patients. Additionally, the feasibility of using prostate specific membrane antigen positron emission tomography/CT (PSMA-PET/CT) scans to assess patients’ eligibility for this proposed workflow was examined. Approach. In this study, 74 male pelvis patients treated on an Elekta Unity 1.5 T MRL were retrospectively selected. The patients’ individual RED values for 8 organs of interest were extracted from their simulation-CT images to establish population-based RED values. These values were used to generate individual (IndD) and population-based (PopD) RED dose plans, representing current and proposed MRL workflows, respectively. Lastly, this study compared RED values obtained from CT and PET-CT scanners in a phantom and a subset of patients. Results. Population-based RED values were mostly within two standard deviations of ICRU Report 46 values. PopD plans were comparable to IndD plans, with the average %difference magnitudes of 0.5%, 0.6%, and 0.6% for mean dose (all organs), D0.1cm 3 (non-target organs) and D95%/D98% (target organs), respectively. Both phantom and patient PET-CT derived RED values had high agreement with corresponding CT-derived values, with correlation coefficients ≥ 0.9. Significance. Population-based RED values were considered suitable in a simulation-free MRL treatment workflow. Utilizing these RED values resulted in similar dosimetric uncertainties as per the current workflow. Initial findings also suggested that PET-CT scans may be used to assess prospective patients’ eligibility for the proposed workflow. Future investigations will evaluate the clinical feasibility of implementing this workflow for prospective patients in the clinical setting. This is aimed to reduce patient burden during radiotherapy and increase department efficiencies.

IMRT in the treatment of locally advanced or inoperable NSCLC in the pre-durvalumab era: clinical outcomes and pattern of relapses, experience from the Oscar Lambret Center.

Intensity-modulated conformal radiotherapy (IMRT) has become the technique of choice for the treatment of locally advanced or inoperable non-small cell lung cancer (NSCLC). Nevertheless, this technique presents dosimetric uncertainties, particularly in treating moving targets such as pulmonary neoplasms. Moreover, it theoretically increases the risk of isolated nodal failure (INF) due to reduced incidental irradiation. The objective of this study was to evaluate the efficacy and safety of IMRT in patients with inoperable NSCLC and to describe the pattern of relapses. Patients with locally advanced NSCLC treated with radiotherapy and chemotherapy between 2015 and 2018 at the Oscar Lambret Center were retrospectively included in the study. Overall and progression-free survival were estimated using the Kaplan-Meier method. The cumulative incidence of the different components of relapse was estimated using the Kalbfleisch and Prentice method. Prognostic factors for relapse/death were investigated using the Cox model. A comparison with literature data was performed using a one-sample log-rank test. Seventy patients were included, and 65 patients (93%) had stage III disease. All the patients received chemotherapy, most frequently with cisplatin and navelbine. The dose received was 66 Gy administered in 33 fractions. The median follow-up and survival were 49.1 and 39.1 months, respectively. A total of 35 deaths and 43 relapses, including 29 with metastatic components, were reported. The overall survival rates at 1 and 2 years were 80.2% (95% confidence interval 68.3%-88.0%) and 67.2% (95% confidence interval 54.2%-77.3%), respectively. Locoregional relapse was observed in 14 patients, including two INF, one of which was located in the lymph node area adjacent to the clinical target volume. Median relapse-free survival was 15.2 months. No variable was statistically associated with the risk of relapse/death in multivariate analysis. Seven patients (10%) experienced grade 3 or higher toxicity. The use of IMRT for locally advanced or inoperable NSCLC led to favorable long-term clinical outcomes. The rate of locoregional relapse, particularly isolated lymph node failure, was low and comparable with that of the three-dimensional radiotherapy series, as was the rate of early and late toxicities.

Viability of the virtual cone technique using a fixed small multi-leaf collimator field for stereotactic radiosurgery of trigeminal neuralgia.

Dosimetric uncertainties in very small (≤1.5×1.5 cm2 ) photon fields are remarkably higher, which undermines the validity of the virtual cone (VC) technique with a diminutive and variable MLC fields. We evaluate the accuracy and reproducibility of the VC method with a very small, fixed MLC field setting, called a fixed virtual cone (fVC), for small target radiosurgery such as trigeminal neuralgia (TGN). The fVC is characterized by 0.5 cm x 0.5cm high-definition (HD) MLC field of 10MV FFF beam defined at 100cm SAD, while backup jaws are positioned at 1.5 cm x 1.5cm. A spherical dose distribution equivalent to 5mm (diameter) physical cone was generated using 10-14 non-coplanar, partial arcs. Dosimetric accuracy was validated using SRS diode (PTW 60018), SRS MapCHECK (SNC) measurements. As a quality assurance measure, 10 treatment plans (SRS) for TGN, consisting of various arc ranges at different collimator angles were analyzed using 6 MV FFF and 10 MV FFF beams, including a field-by-field study (n=130 fields). Dose outputs were compared between the Eclipse TPS and measurements (SRS MapCHECK). Moreover, dosimetric changes in the field defining fVC, prompted by a minute (±0.5-1.0mm) leaf shift, was examined among TPS, diode measurements, and Monte Carlo (MC) simulations. The beam model for fVC was validated (≤3% difference) using SRS MapCHECK based absolute dose measurements. The equivalent diameters of the 50% isodose distribution were found comparable to that of a 5mm cone. Additionally, the comparison of field output factors, dose per MU between the TPS and SRS diode measurements using the fVC field, including±1mm leaf shift, yielded average discrepancies within 5.5% and 3.5% for 6 MV FFF and 10 MV FFF beams, respectively. Overall, the fVC method is a credible alternative to the physical cone (5mm) that can be applied in routine radiosurgical treatment of TGN.

Intentional deep overfit learning for patient-specific dose predictions in adaptive radiotherapy.

The framework of adaptive radiation therapy (ART) was crafted to address the underlying sources of intra-patient variation that were observed throughout numerous patients' radiation sessions. ART seeks to minimize the consequential dosimetric uncertainty resulting from this daily variation, commonly through treatment planning re-optimization. Re-optimization typically consists of manual evaluation and modification of previously utilized optimization criteria. Ideally, frequent treatment plan adaptation through re-optimization on each day's computed tomography (CT) scan may improve dosimetric accuracy and minimize dose delivered to organs at risk (OARs) as the planning target volume (PTV) changes throughout the course of treatment. Re-optimization in its current form is time-consuming and inefficient. In response to this ART bottleneck, we propose a deep learning based adaptive dose prediction model that utilizes a head and neck (H&N) patient's initial planning data to fine-tune a previously trained population model towards a patient-specific model. Our fine-tuned, patient-specific (FT-PS) model, which is trained using the intentional deep overfit learning (IDOL) method, may enable clinicians and treatment planners to rapidly evaluate relevant dosimetric changes daily and re-optimize accordingly. An adaptive population (AP) model was trained using adaptive data from 33 patients. Separately, 10 patients were selected for training FT-PS models. The previously trained AP model was utilized as the base model weights prior to re-initializing model training for each FT-PS model. Ten FT-PS models were separately trained by fine-tuning the previous model weights based on each respective patient's initial treatment plan. From these 10 patients, 26 ART treatment plans were withheld from training as the test dataset for retrospective evaluation of dose prediction performance between the AP and FT-PS models. Each AP and FT-PS dose prediction was compared against the ground truth dose distribution as originally generated during the patient's course of treatment. Mean absolute percent error (MAPE) evaluated the dose differences between a model's prediction and the ground truth. MAPE was calculated within the 10% isodose volume region of interest for each of the AP and FT-PS models dose predictions and averaged across all test adaptive sessions, yielding 5.759% and 3.747% respectively. MAPE differences were compared between AP and FT-PS models across each test session in a test of statistical significance. The differences were statistically significant in a paired t-test with two-tailed p-value equal to and 95% confidence interval (CI) equal to [-2.483, -1.542]. Furthermore, MAPE was calculated using each individually segmented structure as an ROI. Nineteen of 24 structures demonstrated statistically significant differences between the AP and FT-PS models. We utilized the IDOL method to fine-tune a population-based dose prediction model into an adaptive, patient-specific model. The averaged MAPE across the test dataset was 5.759% for the population-based model versus 3.747% for the fine-tuned, patient-specific model, and the difference in MAPE between models was found to be statistically significant. Our work demonstrates the feasibility of patient-specific models in adaptive radiotherapy, and offers unique clinical benefit by utilizing initial planning data that contains the physician's treatment intent.

PO-2142 Dosimetric uncertainties due to inter-fraction organ at risk movement in HDR cervical brachytherapy.

PO-2142 Dosimetric uncertainties due to inter-fraction organ at risk movement in HDR cervical brachytherapy.

Physics Considerations for Evaluation of Dose for Dose-Response Models of Pediatric Late Effects From Radiation Therapy: A PENTEC Introductory Review

We describe the methods used to estimate the accuracy of dosimetric data found in literature sources used to construct the Pediatric Normal Tissue Effects in the Clinic (PENTEC) dose-response models, summarize these findings of each organ-specific task force, describe some of the dosimetric challenges and the extent to which these efforts affected the final modeling results, and provide guidance on the interpretation of the dose-response results given the various dosimetric uncertainties. Each of the PENTEC task force medical physicists reviewed all the journal articles used for dose-response modeling to identify, categorize, and quantify dosimetric uncertainties. These uncertainties fell into 6 broad categories. A uniform nomenclature was developed for describing the "dosimetric quality" of the articles used in the PENTEC reviews. Among the multidisciplinary experts in the PENTEC effort, the medical physicists were charged with the dosimetric evaluation, as they are most expert in this subject. The percentage dosimetric uncertainty was estimated for each late effect endpoint for all PENTEC organ reports. Twelve specific sources of dose uncertainty were identified related to the 6 broad categories. The most common reason for organ dose uncertainty was that prescribed dose rather than organ dose was reported. Percentage dose uncertainties ranged from 5% to 200%. Systematic uncertainties were used to correct the dose component of the models. Random uncertainties were also described in each report and in some cases used to modify dose axis error bars. In addition, the potential effects of dose binning were described. PENTEC reports are designed to provide guidance to radiation oncologists and treatment planners for organ dose constraints. It is critical that these dose constraint recommendations are as accurate as possible, acknowledging the large error bars for many. Achieving this accuracy is important as it enables clinicians to better balance target dose coverage with risk of late effects. Evidence-based dose constraints for pediatric patients have been lacking and, in this regard, PENTEC fills an important unmet need. One must be aware of the limitations of our recommendations, and that for some organ systems, large uncertainties exist in the dose-response model because of clinical endpoint uncertainty, dosimetric uncertainty, or both.

Idiopathic Pneumonitis Syndrome After Total Body Irradiation in Pediatric Patients Undergoing Myeloablative Hematopoietic Stem Cell Transplantation: A PENTEC Comprehensive Review

Pulmonary complications, especially idiopathic pneumonitis syndrome (IPS), are potentially life altering or fatal sequelae of hematopoietic cell transplantation (HCT). Total body irradiation (TBI) as part of the conditioning regimen has been implicated in IPS. A comprehensive PENTEC (Pediatric Normal Tissues in the Clinic) review was performed to increase our understanding of the role of TBI in the development of acute, noninfectious IPS. A systematic literature search was conducted using the MEDLINE, PubMed, and Cochrane library databases for articles describing pulmonary toxicity in children treated with HCT. Data pertaining to TBI and pulmonary endpoints were extracted. Risk of IPS was analyzed in relation to patient age, TBI dose, fractionation, dose rate, lung shielding, timing, and type of transplant, with the goal to better understand factors associated with this complication in children undergoing HCT. A logistic regression model was developed using a subset of studies with comparable transplant regimens and sufficient TBI data. Six studies met criteria for modeling of the correlation of TBI parameters with IPS; all consisted of pediatric patients undergoing allogeneic HCT with a cyclophosphamide-based chemotherapy regimen. IPS was variably defined, but all studies that reported IPS were included in this analysis. The mean incidence of post-HCT IPS was 16% (range, 4%-41%). Mortality from IPS, when it occurred, was high (median, 50%; range, 45%-100%). Fractionated TBI prescription doses encompassed a narrow range of 9 to 14 Gy. Many differing TBI methods were reported, and there was an absence of 3-dimensional dose analysis of lung blocking techniques. Thus, a univariate correlation between IPS and total TBI dose, dose fractionation, dose rate, or TBI technique could not be made. However, a model, built from these studies based on prescribed dose using a normalized dose parameter of equivalent dose in 2-Gy fractions (EQD2), adjusted for dose rate, suggested correlation with the development of IPS (P = .0004). The model-predicted odds ratio for IPS was 24.3 Gy-1 (95% confidence interval, 7.0-84.3). Use of TBI lung dose metrics (eg, midlung point dose) could not be successfully modeled, potentially because of dosimetric uncertainties in the actual delivered volumetric lung dose and imperfections in our modeling process. This PENTEC report is a comprehensive review of IPS in pediatric patients receiving fractionated TBI regimens for allogenic HCT. IPS was not clearly associated with 1 single TBI factor. Modeling using dose-rate adjusted EQD2 showed a response with IPS for allogeneic HCT using a cyclophosphamide-based chemotherapy regimen. Therefore, this model suggests IPS mitigation strategies can focus on not just the dose and dose per fraction but also the dose rate used in TBI. More data are needed to confirm this model and to determine the influence of chemotherapy regimens and contribution from graft-versus-host disease. The presence of confounding variables (eg, systemic chemotherapies) that affect risk, the narrow range of fractionated TBI doses found in the literature, and limitations of other reported data (eg, lung point dose) may have prevented a more straightforward link between IPS and total dose from being observed.

Limitations of phase-sorting based pencil beam scanned 4D proton dose calculations under irregular motion

Objective. 4D dose calculation (4DDC) for pencil beam scanned (PBS) proton therapy is typically based on phase-sorting of individual pencil beams onto phases of a single breathing cycle 4DCT. Understanding the dosimetric limitations and uncertainties of this approach is essential, especially for the realistic treatment scenario with irregular free breathing motion. Approach. For three liver and three lung cancer patient CTs, the deformable multi-cycle motion from 4DMRIs was used to generate six synthetic 4DCT(MRI)s, providing irregular motion (11/15 cycles for liver/lung; tumor amplitudes ∼4–18 mm). 4DDCs for two-field plans were performed, with the temporal resolution of the pencil beam delivery (4–200 ms) or with 8 phases per breathing cycle (500–1000 ms). For the phase-sorting approach, the tumor center motion was used to determine the phase assignment of each spot. The dose was calculated either using the full free breathing motion or individually repeating each single cycle. Additionally, the use of an irregular surrogate signal prior to 4DDC on a repeated cycle was simulated. The CTV volume with absolute dose differences >5% (V dosediff>5%) and differences in CTV V 95% and D 5%–D 95% compared to the free breathing scenario were evaluated. Main results. Compared to 4DDC considering the full free breathing motion with finer spot-wise temporal resolution, 4DDC based on a repeated single 4DCT resulted in V dosediff>5% of on average 34%, which resulted in an overestimation of V 95% up to 24%. However, surrogate based phase-sorting prior to 4DDC on a single cycle 4DCT, reduced the average V dosediff>5% to 16% (overestimation V 95% up to 19%). The 4DDC results were greatly influenced by the choice of reference cycle (V dosediff>5% up to 55%) and differences due to temporal resolution were much smaller (V dosediff>5% up to 10%). Significance. It is important to properly consider motion irregularity in 4D dosimetric evaluations of PBS proton treatments, as 4DDC based on a single 4DCT can lead to an underestimation of motion effects.