Dose-volume Histograms Research Articles

Quantitative Evaluation of a Fully Automated Planning Solution for Prostate-Only and Whole-Pelvic Radiotherapy

Background/Objectives: To evaluate an end-to-end pipeline for normo-fractionated prostate-only and whole-pelvic cancer treatments that requires minimal human input and generates a machine-deliverable plan as an output. Methods: In collaboration with TheraPanacea, a treatment planning pipeline was developed that takes as its input a planning CT with organs-at-risk (OARs) and planning target volume (PTV) contours, the targeted linac machine, and the prescription dose. The primary components are (i) dose prediction by a single deep learning model for both localizations and (ii) a direct aperture VMAT plan optimization that seeks to mimic the predicted dose. The deep learning model was trained on 238 cases, and a held-out set of 86 cases was used for model validation. An end-to-end clinical evaluation study was performed on another 40 cases (20 prostate-only, 20 whole-pelvic). First, a quantitative evaluation was performed based on dose–volume histogram (DVH) points and plan parameter metrics. Then, the plan deliverability was assessed via portal dosimetry using the global gamma index. Additionally, the reference clinical manual plans were compared with the automated plans in terms of monitor unit (MU) numbers and modulation complexity scores (MCSv). Results: The automated plans provided adequate treatment plans (or minor deviations) with respect to the dose constraints, and the quality of the plans was similar to the manual plans for both localizations. Moreover, the automated plans showed successful deliverability and passed the portal dose verification. Despite higher median total MUs, no statistically significant correlation was observed between any of the gamma criteria tested and the number of MUs or MCSv. Conclusions: This study shows the feasibility of a deep learning-based fully automated treatment planning pipeline that generates high-quality plans that are competitive with manually made plans and are clinically approved in terms of dosimetry and machine deliverability.

FLIP: a novel method for patient-specific dose quantification in circulating blood in large vessels during proton or photon external beam radiotherapy treatments.

To provide a novel and personalized method (FLIP, FLowandIrradiation Personalized) using patient-specific circulating blood flows and individualized time-dependent irradiation distributions, to quantify the dose delivered to blood in large vessels during proton or photon external beam radiotherapy. Patient-specific data were obtained from ten cancer patients undergoing radiotherapy, including the blood velocity field in large vessels and the temporal irradiation scheme using photons or protons. The large vessels and the corresponding blood flow velocities are obtained from phase-contrast MRI sequences. The blood dose is obtained discretizing the fluid into individual blood particles (BPs). A Lagrangian approach was applied to simulate the BPs trajectories along the vascular velocity field flowlines. Beam delivery dynamics was obtained from beam delivery machine measurements. The whole irradiation sequence is split into a sequence of successive irradiation elements, each one with its constant dose rate, as well as its corresponding initial and final time. Calculating the dose rate and knowing the spatiotemporal distribution of BPs, the dose is computed by accumulating the energy received by each BP as the time-dependent irradiation beams take place during the treatment. Blood Dose Volume Histograms (DVHs) from proton therapy and photon radiotherapy (RT) patients were assessed. The irradiation times distribution is obtained for BPs in both modalities. Two dosimetric parameters are presented: (i) D3%, representing the minimum dose received by the 3% of BPs receiving the highest doses, and (ii) V0.5Gy, denoting the blood volume percentage that has received at least 0.5 Gy. A novel methodology is proposed for quantifying the circulating blood dose along large vessels. This methodology involves the use of patient-specific vasculature, blood flow velocity field, and dose delivery dynamics recovered from the irradiation machine. Relevant parameters that affect the dose received, as the distance between large vessels and CTV, are identified.

Treatment Planning Comparison of Gantry-based and Fixed Beams for the Treatment of Liver Tumors With Carbon Ion Therapy.

This study aimed to compare the use of a rotating gantry in liver tumor carbon-ion radiotherapy using of a fixed-port for treatment planning. Thirty patients with liver tumors were analyzed. Three treatment plans were developed for each case: one with a rotating gantry with a 360° angle, one with fixed ports of 0° and 90° with a ±20° couch rolling setting, and one with fixed ports of 45° and 90° with a ±20° couch rolling setting. The dose-volume histogram parameters of the clinical target volume (CTV) and organs at risk (OARs) for each treatment plan were compared. Significant differences in the volume of the liver-gross tumor volume (GTV) of normal liver irradiated with 5 Gy to 15 Gy were found between the gantry treatment plans and fixed-port treatment plans. There were no significant differences in the OARs, except for the CTV and liver GTV, between the gantry and fixed-port treatment plans. The study results support the potential of using a rotating gantry to reduce liver doses, especially in the low-to-medium dose range, while maintaining target and OAR doses except for the liver. A rotating gantry could be especially useful in cases in which the relationship between the tumor and OAR is complicated by location.

Dosimetric optimization for dynamic mixed beam arc therapy (DYMBARC).

Non-coplanarity and mixed beam modality could be combined to further enhance dosimetric treatment plan quality. We introduce dynamic mixed beam arc therapy (DYMBARC) as an innovative technique that combines non-coplanar photon and electron arcs, dynamic gantry and collimator rotations, and intensity modulation with photon multileaf collimator (MLC). However, finding favorable beam directions for DYMBARC is challenging due to the large solution space, machine component constraints, and optimization parameters, posing a highly non-convex optimization problem. To establish DYMBARC and solve the pathfinding challenge by employing direct aperture optimization (DAO) to determine the table angles and gantry angle ranges of photon and electron arcs for different clinically motivated cases. The method starts by generating a grid of beam directions based on user-defined resolutions along the gantry and table angle axes for each beam quality considered. Beam directions causing collisions or entering through the end of CT are excluded. For electrons, a fixed source-to-surface distance of 80cm is used to reduce in-air scatter. Electron beam energies with insufficient range to reach the target or beam directions impinging on the table before reaching the patient are excluded. The remaining beam directions form the pathfinding solution space. Promising photon and electron MLC-defined apertures, with associated monitor unit (MU) weights, are iteratively added using a hybrid-DAO algorithm. This algorithm combines column generation to add apertures and simulated annealing to further refine aperture shapes and weights. Apertures are added until the requested number of paths are formed and the user-defined maximum total gantry angle range is reached. Paths are resampled to a finer gantry angle resolution and subject to DAO for simultaneous optimization of beam intensities along the photon/electron arcs. Subsequent final dose calculation and MU weight reoptimization result in a deliverable DYMBARC plan. DYMBARC plans are created for three clinically motivated cases (brain, breast, and pelvis) and compared to DYMBARC variants: colli-DTRT (dynamic collimator trajectory radiotherapy) using non-coplanar photon arcs; and Arc-MBRT (mixed beam radiotherapy) using photons and electrons but restricted to coplanar setup. Additionally, a manually defined volumetric modulated arc therapy (VMAT) setup serves as a reference clinical technique. Dose distributions, dose-volume histograms, and dosimetric endpoints are evaluated. Dosimetric validation with radiochromic film measurements (gamma evaluation, 3% / 2mm (global), 10% dose threshold) is performed on a TrueBeam system in developer mode for one case. While maintaining similar target coverage and homogeneity, DYMBARC reduced mean doses to organs-at-risk compared to VMAT by an average of 3.2, 0.5, and 2.9Gy for the brain, breast, and pelvis cases, respectively. Similar or smaller mean dose reductions were observed for Arc-MBRT or colli-DTRT, compared to VMAT. Electron contributions to the mean planning target volume dose ranged from 2% to 34% for DYMBARC and from 11% to 40% for Arc-MBRT. Measurement validation showed>99.7% gamma passing rate. DYMBARC was successfully established using a dosimetrically optimized pathfinding approach, combining non-coplanarity with mixed beam modality. DYMBARC facilitated the determination of photon and electron contributions on a case-by-case basis, enhancing more personalized treatment modalities.

Monte Carlo dose calculation for photon and electron radiotherapy on dynamically deforming anatomy.

Dose calculation in radiotherapy aims to accurately estimate and assess the dose distribution of a treatment plan. Monte Carlo (MC) dose calculation is considered the gold standard owing to its ability to accurately simulate particle transport in inhomogeneous media. However, uncertainties such as the patient's dynamically deforming anatomy can still lead to differences between the delivered and planned dose distribution. Development and validation of a deformable voxel geometry for MC dose calculations (DefVoxMC) to account for dynamic deformation in the dose calculation process of photon- and electron-based radiotherapy treatment plans for clinically motivated cases. DefVoxMC relies on the subdivision of a regular voxel geometry into dodecahedrons. It allows shifting the dodecahedrons' corner points according to the deformation in the patient's anatomy using deformation vector fields (DVF). DefVoxMC is integrated into the Swiss Monte Carlo Plan (SMCP) to allow the MC dose calculation of photon- and electron-based treatment plans on the deformable voxel geometry. DefVoxMC is validated in two steps. A compression test and a Fano test are performed in silico. Delta4 (for photon beams) and EBT4 film measurements in a cubic PMMA phantom (for electron beams) are performed on a TrueBeam in Developer Mode for clinically motivated treatment plans. During these measurements, table motion is used to mimic rigid dynamic patient motion. The measured and calculated dose distributions are compared using gamma passing rate (GPR) (3% / 2mm (global), 10% threshold). DefVoxMC is used to study the impact of patient-recorded breathing motion on the dose distribution for clinically motivated lung and breast cases, each prescribed 50Gy to 50% of the target volume. A volumetric modulated arc therapy (VMAT) and an arc mixed-beam radiotherapy (Arc-MBRT) plan are created for the lung and breast case, respectively. For the dose calculation, the dynamic deformation of the patient's anatomy is described by DVFs obtained from deformable image registration of the different phases of 4DCTs. The resulting dose distributions are compared to the ones of the static situation using dose-volume histograms and dose differences. DefVoxMC is successfully integrated into the SMCP to enable the MC dose calculation of photon- and electron-based treatments on a dynamically deforming patient anatomy. The compression and the Fano test agree within 1.0% and 0.1% with the expected result, respectively. Delta4 and EBT4 film measurements agree with the calculated dose by a GPR>95%. For the clinically motivated cases, breathing motion resulted in areas with a dose increase of up to 26.9Gy (lung) and up to 7.6Gy (breast) compared to the static situation. The largest dose differences are observed in high-dose-gradient regions perpendicular to the beam plane, consequently decreasing the planning target volume coverage (V95%) by 4.2% for the lung case and 2.0% for the breast case. A novel method for MC dose calculation for photon- and electron-based treatments on dynamically deforming anatomy is successfully developed and validated. Applying DefVoxMC to clinically motivated cases, we found that breathing motion has non-negligible impact on the dosimetric plan quality.

Replacing manual planning with automatic iterative planning for locally advanced rectal cancer VMAT treatment.

To develop and implement a fully automatic iterative planning (AIP) system in the clinical practice, generating volumetric-modulated arc therapy plans combined with simultaneous integrated boost technique VMAT (SIB-VMAT) for locally advanced rectal cancer (LARC) patients. The designed AIP system aimed to automate the entire planning process through a web-based service, includingauxiliary structuregeneration,plan creation,field configuration,plan optimization,dose calculation, andplan assessment. The system was implemented based on theEclipse scriptingapplication programming interface and an efficientiterative optimization algorithm was proposed to reduce the required iterations in the optimization process. To verify the performance of the implemented AIP system, we retrospectively selected a total of 106 patients and performed dosimetric comparisons between the automatic plans (APs) and the manual plans (MPs), in terms of dose-volume histogram (DVH) metrics, homogeneity index (HI), and conformity index (CI) for different volumes of interest. The AIP system has successfully created 106 APs within clinically acceptable timeframes. The average planning time per case was 36.8±6.5 min, with an average iteration number of 6.8(±1.1) in plan optimization. Compared to MPs, APs exhibited better performance in the planning target volume conformity and hotspot control ( ). The organs at risk (OARs) sparing was significantly improved in APs, with mean dose reductions in the femoral heads, the bone marrow, and the SmallBowel-Avoid of 0.53Gy, 1.18Gy, and 1.00Gy, respectively ( ). Slight improvement was also observed in the urinary bladder and the small bowel . Additionally, quality variation between plans from different planners was observed in DVH metrics while the APs represented better plan quality consistency. An AIP system has been implemented and integrated into the clinical treatment planning workflow. The AIP-generated SIB-VMAT plans for LARC have demonstrated superior plan quality and consistency compared with the manual counterparts. In the meantime, the planning time has been reduced by the AIP approach. Based on the reported results, the implemented AIP framework has been proven to improve plan quality and planning efficiency, liberating planners from the laborious parameter-tuning in the optimization phase.

Impact of contrast-enhanced CT in the dosimetry of SBRT for liver metastases treated with MR-Linac

BackgroundTo investigate the impact of using contrast-enhanced computed tomography (CHCT) in the dosimetry of stereotactic body radiation therapy (SBRT) for liver metastases treated with MR-Linac.MethodsA retrospective study was conducted on 21 liver cancer patients treated with SBRT (50 Gy in 5 fractions) using a 1.5 Tesla Unity MR-Linac. The clinical treatment plans optimised on plain computed tomography (pCT) were used as reference. The electronic density (ED) of regions of interest (ROIs) including the liver, duodenum, esophagus, spinal cord, heart, ribs, and lungs, from pCT and CHCT, was analysed. The average ED of each ROI from CHCT was used to generate synthetic CT (sCT) images by assigning the average ED value from the CHCT to the pCT. Clinical plans were recalculated on sCT images. Dosimetric comparisons between the original treatment plan (TPpCT) and the sCT plan (TPsCT) were performed using dose-volume histogram (DVH) parameters, and gamma analysis.ResultsSignificant ED differences (p < 0.05) were observed in the liver, great vessels, heart, lungs, and spinal cord between CHCT and pCT, with the lungs showing the largest differences (average deviation of 11.73% and 12.15% for the left and right lung, respectively). The target volume covered by the prescribed dose (VDpre), and the dose received by 2% and 98% of the volume (D2%, and D98%, respectively) showed statistical differences (p < 0.05), while the gradient index (GI) and the conformity index (CI) did not. Average deviations in target volume dosimetric parameters were below 1.02%, with a maximum deviation of 5.57% for. For the organs at risk (OARs), significant differences (p < 0.05) were observed for D0.35cc and D1.2cc of the spinal cord, D10cc for the stomach, D0.5cc for the heart, and D30% for the liver-GTV, with mean deviations lower than 1.83% for all the above OARs. Gamma analysis using 2%-2 mm criteria yielded a median value of 95.64% (range 82.22–99.65%) for the target volume and 99.40% (range 58–100%) for the OARs.ConclusionThe findings suggest that the use of CHCT in the SBRT workflow for liver metastases may result in minor target volume overdosage, indicating its potential for adoption in clinical settings. However, its use should be further explored in a broader context and tied to personalized treatment approaches.

Investigation of interfractional range variation owing to anatomical changes with beam directions based on water equivalent thickness in proton therapy for pancreatic cancer.

To assess the interfractional anatomical range variations (ARVs) with beam directions and their impact on dose distribution in intensity modulated proton therapy, we analyzed water equivalent thickness (WET) from 10 patients with pancreatic cancer. The distributions of the interfractional WET difference ($\Delta{\mathrm{WET}}^{\theta }$) across 360° were visualized using polar histograms. Interfractional ARVs were evaluated using the mean absolute error and ΔWET pass rate, indicating the percentage of $\Delta \mathrm{WE}{\mathrm{T}}^{\theta }$ < thresholds. The impact on dose distribution in proton therapy was evaluated based on two treatment plans for 40Gy(RBE)/5 fractions: 'Plan A', using two beam angles, in which the target was closest to the body surface among four perpendicular directions; and 'Plan B', using two beam angles with small ARVs. Analysis revealed individual variations in angular trends of interfractional ARVs. Three distinct trends were identified: Group 1 exhibited small ARVs around posterior directions; Group 2 exhibited small ARVs except ~60°; Group 3 demonstrated minimal ARVs only ~90°. In dose evaluation, while 150° and 210° were selected in Plan B for 9 out of 10 patients, for the remaining patient, 60° and 90° were chosen. Comparing dose volume histogram parameters for all patients, Plan B significantly reduced target coverage loss while maintaining organ-at-risk sparing comparable to Plan A. These results demonstrated that selecting beam angles with small interfractional ARVs for each patient enhances the robustness of dose distribution, reducing target coverage loss.

Enhancing stereotactic ablative boost radiotherapy dose prediction for bulky lung cancer: A multi-scale dilated network approach with scale-balanced structure loss.

Partial stereotactic ablative boost radiotherapy (P-SABR) effectively treats bulky lung cancer; however, the planning process for P-SABR requires repeated dose calculations. To improve planning efficiency, we proposed a novel deep learning method that utilizes limited data to accurately predict the three-dimensional (3D) dose distribution of the P-SABR plan for bulky lung cancer. We utilized data on 74 patients diagnosed with bulky lung cancer who received P-SABR treatment. The patient dataset was randomly divided into a training set (51 plans) with augmentation, validation set (7 plans), and testing set (16 plans). We devised a 3D multi-scale dilated network (MD-Net) and integrated a scale-balanced structure loss into the loss function. A comparative analysis with a classical network and other advanced networks with multi-scale analysis capabilities and other loss functions was conducted based on the dose distributions in terms of the axial view, average dose scores (ADSs), and average absolute differences of dosimetric indices (AADDIs). Finally, we analyzed the predicted dosimetric indices against the ground-truth values and compared the predicted dose-volume histogram (DVH) with the ground-truth DVH. Our proposed dose prediction method for P-SABR plans for bulky lung cancer demonstrated strong performance, exhibiting a significant improvement in predicting multiple indicators of regions of interest (ROIs), particularly the gross target volume (GTV). Our network demonstrated increased accuracy in most dosimetric indices and dose scores in different ROIs. The proposed loss function significantly enhanced the predictive performance of the dosimetric indices. The predicted dosimetric indices and DVHs were equivalent to the ground-truth values. Our study presents an effective model based on limited datasets, and it exhibits high accuracy in the dose prediction of P-SABR plans for bulky lung cancer. This method has potential as an automated tool for P-SABR planning and can help optimize treatments and improve planning efficiency.

Evaluation of cumulative dose distributions from external beam radiation therapy using CT-to-CBCT deformable image registration (DIR) for cervical cancer patients.

To investigate dose differences between the planning CT (pCT) and dose calculated on pre-treatment verification CBCTs using DIR and dose summation for cervical cancer patients. Cervical cancer patients treated at our institution with 45Gy EBRT undergo a pCT and 5 CBCTs, once every five fractions of treatment. A free-form intensity-based DIR in MIM was performed between the pCT and each CBCT using the "Merged CBCT" feature to generate an extended FOV-CBCT (mCBCT). DIR-generated bladder and rectum contours were adjusted by a physician, and dice similarity coefficients (DSC) were calculated. After deformation, the investigated doses were (1) recalculated in Eclipse using original plan parameters (ecD), and (2) deformed from planning dose (pD) using the deformation matrix in MIM (mdD). Dose summation was performed to the first week's mCBCT. Dose distributions were compared for the bladder, rectum, and PTV in terms of percent dose difference, dose volume histograms (DVHs), and gamma analysis between the calculated doses. For the 20 patients, the mean DSC was 0.68±0.17 for bladder and 0.79±0.09 for rectum. Most patients were within 5% of pD for D2cc (19/20), Dmax (17/20), and Dmean (16/20). All patients demonstrated a percent difference>5% for bladder V45 due to variations in bladder volume from the pCT. D90 showed fewer differences with 19/20 patients within 2% of pD. Gamma rates between pD and ecD averaged 94% for bladder and 94% for rectum, while pD and mdD exhibited slightly better performance for bladder (93%) and lower for rectum (85%). Using DIR with weekly CBCT images, the MIM deformed dose (mdD) was found to be in close agreement with the Eclipse calculated dose (ecD). The proposed workflow should be used on a case-by-case basis when the weekly CBCT shows marked difference in organs-at-risk from the planning CT.

Evaluation of Radiation Pneumonitis in a Phase II Study of Consolidation Immunotherapy with Nivolumab and Ipilimumab or Nivolumab Alone Following Concurrent Chemoradiotherapy for Unresectable Stage IIIA/IIIB Non-Small Cell Lung Cancer (NSCLC): Big Ten Cancer Research Consortium BTCRC-LUN16-081

Purpose/Objective(s)The addition of immunotherapy (IO) after concurrent chemoradiation (CCRT) for unresectable non-small cell lung cancer (NSCLC) has become common practice in eligible patients. Approaches to further improve outcomes and reduce treatment-related toxicity for these patients are needed. This study evaluates the risk of radiation pneumonitis after CCRT and its correlation with the radiation dose distribution, immunotherapy regimen (nivolumab vs nivolumab plus ipilimumab), and patient demographics across BTCRC-LUN16-081. Materials/MethodsPatients with unresectable stage III NSCLC after completion of CCRT were enrolled on BTCRC-LUN16-081, a randomized phase II trial to assess efficacy and tolerability of consolidative nivolumab versus nivolumab plus ipilimumab for six months. Radiation dose parameters, patient demographics, and toxicity events were evaluated among treatment arms for risk and severity of pneumonitis. ResultsOne-hundred and five patients were enrolled into two treatment arms, 54 patients received nivolumab alone and 51 patients received nivolumab plus ipilimumab. Of these, 104 patients had dose volume histogram (DVH) information available. Within this cohort, 65 patients (62.5%) had stage IIIA and 39 patients (37.5%) had stage IIIB NSCLC disease per AJCC 7th edition. During the study, 29 patients (27.9%) were diagnosed with grade 2 or greater pneumonitis. Utilizing logistic regression and evaluating different cut-offs for lung V20, patients with V20 >23% demonstrated significantly higher grade 2 or greater pneumonitis rates (37.1% vs. 16.2%, P = .031). No significant difference in rates of pneumonitis between arms were identified. Traditional lung DVH cut-offs (V5 of >65%, V20 of >35%, mean >20 Gy) were not associated with pneumonitis. ConclusionIn patients receiving nivolumab or nivolumab plus ipilimumab after definitive CCRT, lung V20 of >23% was associated with an increased risk of grade 2 or greater pneumonitis. Radiation dose constraints for lungs in patients receiving consolidative IO after CCRT should continue to be evaluated and optimized when feasible.

Deep Learning-Based Synthetic Computed Tomography for Low-Field Brain Magnetic Resonance-Guided Radiation Therapy

PurposeMR-guided Radiation Therapy (MRgRT) enables online adaptation to address intra- and inter-fractional changes. To address the need of high-fidelity synthetic CT (synCT) required for dose calculation, we developed a conditional generative adversarial network (cGAN) for synCT generation from low-field MRI in the brain. Methods and MaterialsSimulation MR-CT pairs from twelve glioma patients imaged with a head and neck surface coil and treated on a 0.35T MR-linac were prospectively included to train the model consisting of a 9-block residual network generator and a PatchGAN discriminator. Four-fold cross validation was implemented. SynCT was quantitatively evaluated against real CT using mean absolute error (MAE), Peak Signal-to-Noise Ratio (PSNR) and Structural Similarity (SSIM). Dose was calculated on synCT applying original treatment plan. Dosimetric performance was evaluated by dose-volume histogram (DVH) metric comparison and local three-dimensional gamma analysis. To demonstrate utilization in treatment adaptation, longitudinal synCTs were generated for qualitative evaluation, and one offline adaptation case underwent two comparative plan evaluations. Secondary validation was conducted with 9 patients on a different MR-linac using a high-resolution brain coil. ResultsOur model generated high-quality synCTs with MAE, PSNR and SSIM of 70.9±10.4 HU, 28.4±1.5 d.B. and 0.87±0.02 within the field-of-view, respectively. Underrepresented post-surgical anomalies challenged model performance. Nevertheless, excellent dosimetric agreement was observed with the mean difference between real and synCT DVH metrics of -0.07±0.29 Gy for target D95 and within [-0.14, 0.02] Gy for organs at risk. Significant differences were only observed in the right lens D0.01cc with negligible overall difference (<0.13 Gy). Mean gamma analysis pass rates were 92.2%±3.0%, 99.2%±0.7% and 99.9%±0.1% at 1%/1mm, 2%/2mm and 3%/3mm, respectively. Secondary validation yielded no significant differences in synCT performance for whole brain MAE, PSNR, and SSIM with comparable dosimetric results. ConclusionsOur cGAN model generated high-fidelity brain synCTs from low-field MRI with excellent dosimetric performance. Secondary validation suggests great promise of implementing synCTs to facilitate robust dose calculation for online adaptive brain MRgRT.

Comparing robust proton versus online adaptive photon radiotherapy for short-course treatment of rectal cancer

Background and purposeImage-guided proton beam therapy (IG-PBT) and cone-beam CT (CBCT)-based online adaptive photon radiotherapy (oART) have potentials to restrict radiation toxicity. They are both hypothesised to reduce therapy limiting bowel toxicity in the multimodality treatment of locally advanced rectal cancer (LARC). This study aimed to quantify the difference in relevant dose-volume metrics for these modalities. Material and MethodsSix-degrees-of-freedom IG-PBT and oART short-course radiotherapy (SCRT) were simulated for 18 LARC patients. Relative biological effectiveness (RBE) was 1.1 for IG-PBT. Delivered dose was evaluated using post-CBCTs. Target coverage was considered robust if average dose to 99 % of the clinical target volume was ≥ 95 % of the prescription. Organ at risk (OAR) doses were compared using dose-volume histograms and severe bowel toxicity estimated using dose–response modelling. ResultsTarget coverage was robust in all patients for oART and all but one patient for IG-PBT. For the main OARs, IG-PBT increased the volume exposed to ≥ 15 Gy (RBE), but reduced volumes exposed to lower doses. Both low- and high-dose exposure to bowel loops were significantly different between the modalities (median (interquartile range) IG-PBT-V8.9Gy(RBE) = 92 cm3 (51–156), oART-V8.9Gy(RBE) = 166 cm3 (107–234), p < 0.001; IG-PBT-V23Gy(RBE) = 62 cm3 (25–106), oART-V23Gy(RBE) = 38 cm3 (18–75), p < 0.001), translating into similar total grade ≥ 3 bowel toxicity risk. ConclusionIG-PBT and oART delivered comparable and satisfying target coverage in SCRT for LARC with similar estimated risk of severe bowel toxicity. Volumes of OAR exposed to 15 Gy (RBE) or more were reduced by oART, while IG-PBT reduced the volumes receiving doses below this level.

Dosimetric analysis of LAD dose in left-sided breast cancer radiotherapy with deep inspiratory breath hold

Objectives To analyze the dose to the left anterior descending artery (LAD) in patients who have received radiotherapy for left breast cancer with Deep Inspiratory Breath Hold (DIBH) technique and compare it with other cardiac dosimetric parameters, as well as the accepted dose constraints. Materials and Methods 20 patients (10 prospective and 10 retrospective) were selected for this study. All patients underwent 2 non-contrast radiation planning CT scans of 2.5 mm thickness - one with DIBH and one with free breathing. Contouring was done using the Radiation Therapy Oncology Group (RTOG) guidelines. LAD was delineated and given a PRV of 3 mm and 5 mm. Dose-volume histograms (DVH) were used to obtain the data from the approved plans. Results The lung volume receiving 17 Gy in percentage, Dmean of the heart, LAD Dmean and Dmax, and the Dmean and Dmax received by 3 mm and 5 mm PRVs were both very well achieved when compared to the dose constraints given by the DBCG HYPO trial. The study found a higher correlation between the mean heart dose and the 5 mm PRV dose (R2 = 0.81 and 0.71 respectively for the mean and max dose) than the 3 mm PRV, and a positive correlation between the heart dose and LAD making it a useful structure for predicting acute cardiac events. Conclusion The study of 20 patients found that DIBH is effective to minimize cardiac dose and potentially cardiac toxicity, with heart and LAD doses being comparable or lower compared to other studies. The LAD doses recorded were significantly less than those in non-DIBH studies, demonstrating the feasibility of routine contouring and recording LAD dose in left-sided breast radiation patients. Further research is needed to determine the dosimetry and clinical consequences of the Dmean and Dmax of the 5mm PRV to the LAD.

Dosimetric Evaluation of Hippocampus Sparing Intensity Modulated Radiation Therapy in Patients With Stage T1-T2 and Stage T3-T4 Nasopharyngeal Carcinoma

PurposeTo compare the hippocampus (HPC) dose reduced by HPC-sparing IMRT plans between nasopharyngeal carcinoma (NPC) patients of stages T1-T2 and T3-T4, and to investigate the correlation between the dose of the hippocampus and the volume of PTVnx70 (the planning target volume of the primary tumor in the nasopharynx received 70 Gy). Methods and MaterialsFifty-eight NPC patients were retrospectively evaluated. HPC non-sparing IMRT or sparing IMRT for each patient was designed according to the protocol for NPC. Dose-volume histogram (DVH) was used to evaluate the IMRT plans for each patient. The difference values of HPC parameters (e.g., Dmin[NS] - Dmin[S]) between HPC sparing and non-sparing plans in stage T1-T2 Group and stage T3-T4 Group were compared. The correlations between the dose of the hippocampus and the volume of PTVnx70 were analyzed. ResultsThere was no significance between HPC-sparing and non-sparing IMRT plans. Compared with the HPC non-sparing plans, the HPC sparing plans significantly decreased both dosimetric and volumetric parameters for the HPC (P < 0.05), except for Dmin, D98%, and V5. The medians of Dmedian[NS] - Dmedian[S], Dmean[NS] - Dmean[S], D40%[NS] - D40%[S], V30[NS] - V30[S], V40[NS] - V40[S] and V50[NS] - V50[S] in the T1-T2 Group were significantly lower than in the T3-T4 Group (P < 0.05), respectively. Both dosimetric and volumetric parameters for the HPC were positively correlated with the volume of PTVnx70 in HPC-sparing and non-sparing plans (P < 0.05). The volume of PTVnx70 was positively correlated with Dmedian[NS] - Dmedian[S], Dmean[NS] - Dmean[S], D40%[NS] - D40%[S], V40[NS] - V40[S] and V50[NS] - V50[S] (P < 0.05). ConclusionsHPC-sparing IMRT plans may play a more significant role in decreasing Dmedian, Dmean, D40%, and V30 ∼ V50 of HPC in NPC patients with stages T3-T4 than those in stages T1-T2. PTVnx70 volume of NPC patients are positively correlated with all dosimetric and volumetric parameters of HPC and the reduction of specific dosage parameters by HPC-sparing IMRT plans.

Large-language-model empowered 3D dose prediction for intensity-modulated radiotherapy.

Treatment planning is currently a patient specific, time-consuming, and resource demanding task in radiotherapy. Dose-volume histogram (DVH) prediction plays a critical role in automating this process. The geometric relationship between DVHs in radiotherapy plans and organs-at-risk (OAR) and planning target volume (PTV) has been well established. This study explores the potential of deep learning models for predicting DVHs using images and subsequent human intervention facilitated by a large-language model (LLM) to enhance the planning quality. We propose a pipeline to convert unstructured images to a structured graph consisting of image-patch nodes and dose nodes. A novel Dose Graph Neural Network (DoseGNN) model is developed for predicting DVHs from the structured graph. The proposed DoseGNN is enhanced with the LLM to encode massive knowledge from prescriptions and interactive instructions from clinicians. In this study, we introduced an online human-AI collaboration (OHAC) system as a practical implementation of the concept proposed for the automation of intensity-modulated radiotherapy (IMRT) planning. The proposed DoseGNN model was compared to widely employed DL models used in radiotherapy, including Swin Transformer, 3D U-Net CNN, and vanilla MLP. For PTV, DoseGNN achieved the mean absolute error (MAE) of , , , and between true plans and predicted plans that were 64%, 53%, 64%, 61% of the best baseline model. For the worst case among OARs (left lung, right lung, chest wall, heart, spinal cord), DoseGNN achieved the mean absolute error of , , that were 85%, 91%, 80% of the best baseline model. Moreover, the LLM-empowered DoseGNN model facilitates seamless adjustment to treatment plans through interaction with clinicians using natural language. We developed DoseGNN, a novel deep learning model for predicting delivered radiation doses from medical images, enhanced by LLM to allow adjustment through seamless interaction with clinicians. The preliminary results confirm DoseGNN's superior accuracy in DVH prediction relative to typical DL methods, highlighting its potential to facilitate an online clinician-AI collaboration system for streamlined treatment planning automation.

Deep learning–based statistical robustness evaluation of intensity-modulated proton therapy for head and neck cancer

A dense, dilated 3D U-net was trained to predict the 5th and 95th percentile distributions of planning dose plan using the nominal dose and planning CT images. The data set comprised proton therapy plans for 582 H&N cancer patients. Ground truth percentile values were estimated for each patient through 600 dose recalculations, representing randomly sampled uncertainty scenarios. The comprehensive comparisons of different models were conducted for H&N cancer patients, considering those with and without a beam mask (BM) and diverse beam configurations, including varying beam angles, couch angles, and beam numbers. The performance of our model trained based on a mixture of patients with H&N and prostate cancer was also assessed in contrast with models trained based on data specific for patients with cancer at either site. The DL-based model's predictions of percentile dose distributions exhibited excellent agreement with the ground truth dose distributions. The average gamma index with 2 mm/2%, consistently exceeded 97% for both 5th and 95th percentile dose volumes. Mean dose-volume histogram error analysis revealed that predictions from the combined training set yielded mean errors and standard deviations that were generally similar to those in the specific patient training data sets. Our proposed DL-based method for evaluation of the robustness of proton therapy plans provides precise, rapid predictions of percentile dose for a given confidence level regardless of the beam arrangement and cancer site. This versatility positions our model as a valuable tool for evaluating the robustness of proton therapy across various cancer sites.

Deep learning-based prediction of the dose-volume histograms for volumetric modulated arc therapy of left-sided breast cancer.

The advancements in artificial intelligence and computational power have made deep learning an attractive tool for radiotherapy treatment planning. Deep learning has the potential to significantly simplify the trial-and-error process involved in inverse planning required by modern treatment techniques such as volumetric modulated arc therapy (VMAT). In this study, we explore the ability of deep learning to predict organ-at-risk (OAR) dose-volume histograms (DVHs) of left-sided breast cancer patients undergoing VMAT treatment based solely on their anatomical characteristics. The predicted DVHs could be used to derive patient-specific dose constraints and dose objectives, streamlining the treatment planning process, standardizing the quality of the plans, and personalizing the treatment planning. This study aimed to develop a deep learning-based framework for the prediction of organ-specific dose-volume histograms (DVH) based on structures delineated for left-sided breast cancer treatment. We used a dataset of 249 left-sided breast cancer patients treated with tangential VMAT fields. We extracted delineated structures and dose distributions for each patient and derived slice-by-slice DVHs for planning target volume (PTV) and organs-at-risk. The patients were divided into training (70%, n=174), validation (10%, n=24), and test (20%, n=51) sets. Collected data were used to train a deep learning model for the prediction of the DVHs based on the delineated structures. The developed deep learning model comprised a modified DenseNet architecture followed by a recurrent neural network. In the independent test set (n=51), the point-wise differences in the slice-by-slice DVHs between the clinical and predicted DVHs were small; the mean squared errors were 3.53, 1.58, 2.28, 3.37, and 1.44 [×10-4] for PTV, heart, ipsilateral lung, contralateral lung, and contralateral breast, respectively. With the derived cumulative DVHs, the mean absolute difference±standard deviation of mean doses between the clinical and the predicted DVH were 0.08± 0.04Gy, 0.24± 0.22Gy, 0.73± 0.46Gy, 0.07± 0.06Gy, and 0.14± 0.14Gy for PTV, heart, ipsilateral lung, contralateral lung, and contralateral breast, respectively. The deep learning-based approach enabled automatic and reliable prediction of the DVH based on delineated structures. The predicted DVHs could potentially serve as patient-specific clinical goals used to aid treatment planning and avoid suboptimal plans or to derive optimization objectives and constraints for automated treatment planning.

Reducing treatment time for robotic radiosurgery by considering path efficiency during node selection.

Robotic radiosurgery treatments allow for precise non-coplanar beam delivery by utilizing a robot equipped with a linac that traverses through a set of predetermined nodes. High quality treatment plans can be produced but treatment times can grow large, with one substantial component being the robot traversal time. The aim of this study is to reduce the treatment time for robotic radiosurgery treatments by introducing algorithms for reducing the robot traversal time. The algorithms are integrated into a commercial treatment planning system. First, an optimization framework for robotic radiosurgery planning is detailed, including a heuristic optimization method for node selection. Second, two methods aimed at reducing the traversal time are introduced. One utilizes a centrality measure focusing on the structure of the node network, while the other is based on the direct computation of traversal times during optimization. A comparison between plans with and without the time-reducing algorithms is made for three brain cases and one liver case with basis in treatment time, plan quality, monitor units, and network structure of the selectednodes. Large decreases in traversal times are obtained by the traversal time reducing algorithms, with reductions of up to 49 % in the brain cases and 31 % in the liver case. The resulting reductions in treatment times are up to 30 % and 13 %, respectively. Small differences in plan quality are observed, with similar dose-volume histograms, dose distributions, and conformity/gradientindices. The total treatment time of the robotic radiosurgery treatments can be reduced by selecting nodes with more efficient robot traversal paths, while maintaining planquality.

Nested CNN architecture for three-dimensional dose distribution prediction in tomotherapy for prostate cancer.

The hypothesis of changing network layers to increase the accuracy of dose distribution prediction, instead of expanding their dimensions, which requires complex calculations, has been considered in our study. Atotal of 137 prostate cancer patients treated with the tomotherapy technique were categorized as 80% training and validating as well as 20% testing for the nested UNet and UNet architectures. Mean absolute error (MAE) was used to measure the dosimetry indices of dose-volume histograms (DVHs), and geometry indices, including the structural similarity index measure (SSIM), dice similarity coefficient (DSC), and Jaccard similarity coefficient (JSC), were used to evaluate the isodose volume (IV) similarity prediction. To verify a statistically significant difference, the two-way statistical Wilcoxon test was used at alevel of 0.05 (p < 0.05). Use of anested UNet architecture reduced the predicted dose MAE in DVH indices. The MAE for planning target volume (PTV), bladder, rectum, and right and left femur were D98% = 1.11 ± 0.90; D98% = 2.27 ± 2.85, Dmean = 0.84 ± 0.62; D98% = 1.47 ± 12.02, Dmean = 0.77 ± 1.59; D2% = 0.65 ± 0.70, Dmean = 0.96 ± 2.82; and D2% = 1.18 ± 6.65, Dmean = 0.44 ± 1.13, respectively. Additionally, the greatest geometric similarity was observed in the mean SSIM for UNet and nested UNet (0.91 vs.0.94, respectively). The nested UNet network can be considered asuitable network due to its ability to improve the accuracy of dose distribution prediction compared to the UNet network in an acceptable time.