Purpose
Intensity‐modulated radiation therapy (IMRT) quality assurance (QA) measurements are routinely performed prior to treatment delivery to verify dose calculation and delivery accuracy. In this ...work, we applied a machine learning‐based approach to predict portal dosimetry based IMRT QA gamma passing rates.
Methods
182 IMRT plans for various treatment sites were planned and delivered with portal dosimetry on two TrueBeam and two Trilogy LINACs. A total of 1497 beams were collected and analyzed using gamma criteria of 2%/2 mm with a 5% threshold. The datasets for building the machine learning models consisted of 1269 beams. Ten‐fold cross‐validation was utilized to tune the model and prevent “overfitting.” A separate test set with the remaining 228 beams was used to evaluate model performance. Each beam was characterized by a set of 31 features including both plan complexity metrics and machine characteristics. Three tree‐based machine learning algorithms (AdaBoost, Random Forest, and XGBoost) were used to train the models and predict gamma passing rates.
Results
Both AdaBoost and Random Forest had 98% of predictions within 3% of the measured 2%/2 mm gamma passing rates with a maximum error less than 4% and a mean absolute error < 1%. XGBoost showed a slightly worse prediction accuracy with 95% of the predictions within 3% of the measured gamma passing rates and a maximum error of 4.5%. The three models identified the same nine features in the top 10 most important ones that are related to plan complexity and maximum aperture displacement from the central axis or the maximum jaw size in a beam.
Conclusion
We have demonstrated that portal dosimetry IMRT QA gamma passing rates can be accurately predicted using tree‐based ensemble learning models. The machine learning based approach allows physicists to better identify the failures of IMRT QA measurements and to develop proactive QA approaches.
We measured tissue temperature changes during ex vivo microwave ablation (MWA) procedures for bovine liver tissue. Tissue temperature increased rapidly at the beginning of the MW power application. ...It came to a plateau at 100 degC to 104 degC before it increased again. We split the changes of tissue temperature versus time into four phases. This suggests that tissue temperature changes may be directly related to tissue water related phenomena during MWA, including evaporation, diffusion, condensation and tissue water composition. An additional analysis indicated the lesion boundary at ~50 degC to 60 degC temperature. We also measured the water content of ablated tissue lesions and examined the relationship of tissue water content and tissue temperature by mapping temperature to remaining tissue water after ablation. The results demonstrate significant tissue water content changes and lead to a better understanding of tissue water movement
To characterize potential advantages of online-adaptive magnetic resonance (MR)-guided stereotactic body radiation therapy (SBRT) to treat oligometastatic disease of the non-liver abdomen and central ...thorax.
Ten patients treated with RT for unresectable primary or oligometastatic disease of the non-liver abdomen (n=5) or central thorax (n=5) underwent imaging throughout treatment on a clinical MR image guided RT system. The SBRT plans were created on the basis of tumor/organ at risk (OAR) anatomy at initial computed tomography simulation (P
), and simulated adaptive plans were created on the basis of observed MR image set tumor/OAR "anatomy of the day" (P
). Each P
was planned under workflow constraints to simulate online-adaptive RT. Prescribed dose was 50 Gy/5 fractions, with goal coverage of 95% planning target volume (PTV) by 95% of the prescription, subject to hard OAR constraints. The P
was applied to each MR dataset and compared with P
to evaluate changes in dose delivered to tumor/OARs, with dose escalation when possible.
Hard OAR constraints were met for all P
based on anatomy from initial computed tomography simulation, and all P
based on anatomy from each daily MR image set. Application of the P
to anatomy of the day caused OAR constraint violation in 19 of 30 cases. Adaptive planning increased PTV coverage in 21 of 30 cases, including 14 cases in which hard OAR constraints were violated by the nonadaptive plan. For 9 P
cases, decreased PTV coverage was required to meet hard OAR constraints that would have been violated in a nonadaptive setting.
Online-adaptive MRI-guided SBRT may allow PTV dose escalation and/or simultaneous OAR sparing compared with nonadaptive SBRT. A prospective clinical trial is underway at our institution to evaluate clinical outcomes of this technique.
To study the seismic performance of wood frame structures filled with light wood shear walls, three full-size single-layer single span wooden frame structures with infill walls were designed and ...manufactured. The beams and columns were connected by mortise-tenon joints, and quasi-static tests were conducted on the specimens under reversed cyclic load. The failure modes and load-displacement hysteresis performance of structures with door opening infill wall, window opening infill wall, and solid infill wall were investigated. The seismic performance was analyzed using indicators such as strength, ductility, and equivalent viscous damping ratio. The failure modes of light wood frame filling walls were the tearing of sheathing panel and the failure of nail connections. The filled wall with the opening initially exhibited inclined cracks at the corner of the opening, and then they extended to the periphery. Compared with the solid filled wall, the positive and negative bearing capacity of the structure with door opening decreased, and that of the structure with window opening decreased also. Because the specimens with opening in the filled wall were more conducive to the deformation of the structure when the bearing capacity was not significantly reduced, the ductility of the specimen with door opening was the highest.
In collinear factorization, light-cone distribution amplitudes (LCDAs) are key ingredients to calculate the production rate of a hadron in high energy exclusive processes. For a doubly-heavy meson ...system (such as Bc, J/ψ, Y etc), the LCDAs contain perturbative scales that can be integrated out and then are re-factorized into products of perturbatively calculable distribution parts and non-relativistic QCD matrix elements. In this re-factorization scheme, the LCDAs are known at next-to-leading order in the strong coupling constant αs and at leading order in the velocity expansion. In this work, we calculate the Ov2 corrections to twist-2 LCDAs of S-wave Bc mesons. These results are applicable to heavy quarkonia like ηc,b, J/ψ and Y by setting mb = mc. We apply these relativistically corrected LCDAs to study their inverse moments and a few Gegenbauer moments which are important for phenomenological study. We point out that the relativistic corrections are sizable, and comparable with the next-to-leading order radiative corrections. These results for LCDAs are useful in future theoretical analyses of the productions of heavy quarkonia and Bc mesons.
To demonstrate the feasibility of online adaptive magnetic resonance (MR) image guided radiation therapy (MR-IGRT) through reporting of our initial clinical experience and workflow considerations.
...The first clinically deployed online adaptive MR-IGRT system consisted of a split 0.35T MR scanner straddling a ring gantry with 3 multileaf collimator-equipped (60)Co heads. The unit is supported by a Monte Carlo-based treatment planning system that allows real-time adaptive planning with the patient on the table. All patients undergo computed tomography and MR imaging (MRI) simulation for initial treatment planning. A volumetric MRI scan is acquired for each patient at the daily treatment setup. Deformable registration is performed using the planning computed tomography data set, which allows for the transfer of the initial contours and the electron density map to the daily MRI scan. The deformed electron density map is then used to recalculate the original plan on the daily MRI scan for physician evaluation. Recontouring and plan reoptimization are performed when required, and patient-specific quality assurance (QA) is performed using an independent in-house software system.
The first online adaptive MR-IGRT treatments consisted of 5 patients with abdominopelvic malignancies. The clinical setting included neoadjuvant colorectal (n=3), unresectable gastric (n=1), and unresectable pheochromocytoma (n=1). Recontouring and reoptimization were deemed necessary for 3 of 5 patients, and the initial plan was deemed sufficient for 2 of the 5 patients. The reasons for plan adaptation included tumor progression or regression and a change in small bowel anatomy. In a subsequently expanded cohort of 170 fractions (20 patients), 52 fractions (30.6%) were reoptimized online, and 92 fractions (54.1%) were treated with an online-adapted or previously adapted plan. The median time for recontouring, reoptimization, and QA was 26 minutes.
Online adaptive MR-IGRT has been successfully implemented with planning and QA workflow suitable for routine clinical application. Clinical trials are in development to formally evaluate adaptive treatments for a variety of disease sites.
Purpose:
In our clinic, physicists spend from 15 to 60 min to verify the physical and dosimetric integrity of radiotherapy plans before presentation to radiation oncology physicians for approval. The ...purpose of this study was to design and implement a framework to automate as many elements of this quality control (QC) step as possible.
Methods
: A comprehensive computer application was developed to carry out a majority of these verification tasks in the PhilipsPINNACLE treatment planning system (TPS). This QC tool functions based on both PINNACLE scripting elements and PERL sub-routines. The core of this technique is the method of dynamic scripting, which involves a PERL programming module that is flexible and powerful for treatment plan data handling. Run-time plan data are collected, saved into temporary files, and analyzed against standard values and predefined logical rules. The results were summarized in a hypertext markup language (HTML) report that is displayed to the user.
Results
: This tool has been in clinical use for over a year. The occurrence frequency of technical problems, which would cause delays and suboptimal plans, has been reduced since clinical implementation.
Conclusions:
In addition to drastically reducing the set of human-driven logical comparisons, this QC tool also accomplished some tasks that are otherwise either quite laborious or impractical for humans to verify, e.g., identifying conflicts amongst IMRT optimization objectives.
Purpose
Clinical treatment planning systems for proton therapy currently do not calculate monitor units (MUs) in passive scatter proton therapy due to the complexity of the beam delivery systems. ...Physical phantom measurements are commonly employed to determine the field‐specific output factors (OFs) but are often subject to limited machine time, measurement uncertainties and intensive labor. In this study, a machine learning‐based approach was developed to predict output (cGy/MU) and derive MUs, incorporating the dependencies on gantry angle and field size for a single‐room proton therapy system. The goal of this study was to develop a secondary check tool for OF measurements and eventually eliminate patient‐specific OF measurements.
Method
The OFs of 1754 fields previously measured in a water phantom with calibrated ionization chambers and electrometers for patient‐specific fields with various range and modulation width combinations for 23 options were included in this study. The training data sets for machine learning models in three different methods (Random Forest, XGBoost and Cubist) included 1431 (~81%) OFs. Ten‐fold cross‐validation was used to prevent “overfitting” and to validate each model. The remaining 323 (~19%) OFs were used to test the trained models. The difference between the measured and predicted values from machine learning models was analyzed. Model prediction accuracy was also compared with that of the semi‐empirical model developed by Kooy (Phys. Med. Biol. 50, 2005). Additionally, gantry angle dependence of OFs was measured for three groups of options categorized on the selection of the second scatters. Field size dependence of OFs was investigated for the measurements with and without patient‐specific apertures.
Results
All three machine learning methods showed higher accuracy than the semi‐empirical model which shows considerably large discrepancy of up to 7.7% for the treatment fields with full range and full modulation width. The Cubist‐based solution outperformed all other models (P < 0.001) with the mean absolute discrepancy of 0.62% and maximum discrepancy of 3.17% between the measured and predicted OFs. The OFs showed a small dependence on gantry angle for small and deep options while they were constant for large options. The OF decreased by 3%–4% as the field radius was reduced to 2.5 cm.
Conclusion
Machine learning methods can be used to predict OF for double‐scatter proton machines with greater prediction accuracy than the most popular semi‐empirical prediction model. By incorporating the gantry angle dependence and field size dependence, the machine learning‐based methods can be used for a sanity check of OF measurements and bears the potential to eliminate the time‐consuming patient‐specific OF measurements.
Purpose:
The development of iterative image reconstruction algorithms for cone-beam computed tomography (CBCT) remains an active and important research area. Even with hardware acceleration, the ...overwhelming majority of the available 3D iterative algorithms that implement nonsmooth regularizers remain computationally burdensome and have not been translated for routine use in time-sensitive applications such as image-guided radiation therapy (IGRT). In this work, two variants of the fast iterative shrinkage thresholding algorithm (FISTA) are proposed and investigated for accelerated iterative image reconstruction in CBCT.
Methods:
Algorithm acceleration was achieved by replacing the original gradient-descent step in the FISTAs by a subproblem that is solved by use of the ordered subset simultaneous algebraic reconstruction technique (OS-SART). Due to the preconditioning matrix adopted in the OS-SART method, two new weighted proximal problems were introduced and corresponding fast gradient projection-type algorithms were developed for solving them. We also provided efficient numerical implementations of the proposed algorithms that exploit the massive data parallelism of multiple graphics processing units.
Results:
The improved rates of convergence of the proposed algorithms were quantified in computer-simulation studies and by use of clinical projection data corresponding to an IGRT study. The accelerated FISTAs were shown to possess dramatically improved convergence properties as compared to the standard FISTAs. For example, the number of iterations to achieve a specified reconstruction error could be reduced by an order of magnitude. Volumetric images reconstructed from clinical data were produced in under 4 min.
Conclusions:
The FISTA achieves a quadratic convergence rate and can therefore potentially reduce the number of iterations required to produce an image of a specified image quality as compared to first-order methods. We have proposed and investigated accelerated FISTAs for use with two nonsmooth penalty functions that will lead to further reductions in image reconstruction times while preserving image quality. Moreover, with the help of a mixed sparsity-regularization, better preservation of soft-tissue structures can be potentially obtained. The algorithms were systematically evaluated by use of computer-simulated and clinical data sets.
We apply the discrete S3 flavor symmetry to both lepton and quark sectors of the Standard Model extended by introducing one Higgs triplet and realizing the type-II seesaw mechanism for finite ...neutrino masses. The resultant mass matrices of charged leptons (Ml), neutrinos (Mν), up-type quarks (Mu) and down-type quarks (Md) have a universal form consisting of two terms: one is proportional to the identity matrix I and the other is proportional to the democracy matrix D. We argue that the textures of Ml, Mu and Md are dominated by the D term, while that of Mν is dominated by the I term. This hypothesis implies a near mass degeneracy of three neutrinos and can naturally explain why the mass matrices of charged fermions are strongly hierarchical, why the quark mixing matrix is close to I and why the lepton mixing matrix contains two large angles. We discuss a rather simple perturbation ansatz to break the S3 symmetry and obtain more realistic mass spectra of leptons and quarks as well as their flavor mixing patterns. We stress that the I term, which used to be ignored from Ml, Mu and Md, is actually important because it can significantly modify the smallest lepton flavor mixing angle θ13 or three quark flavor mixing angles.