Self-channel-and-spatial-attention neural network for automated multi-organ segmentation on head and neck CT images

Accurate segmentation of organs at risk (OARs) is necessary for adaptive head and neck (H&N) cancer treatment planning, but manual delineation is tedious, slow, and inconsistent. A self-channel-and-spatial-attention neural network (SCSA-Net) is developed for H&N OAR segmentation on CT images. To simultaneously ease the training and improve the segmentation performance, the proposed SCSA-Net utilizes the self-attention ability of the network. Spatial and channel-wise attention learning mechanisms are both employed to adaptively force the network to emphasize the meaningful features and weaken the irrelevant features simultaneously. The proposed network was first evaluated on a public dataset, which includes 48 patients, then on a separate serial CT dataset, which contains ten patients who received weekly diagnostic fan-beam CT scans. On the second dataset, the accuracy of using SCSA-Net to track the parotid and submandibular gland volume changes during radiotherapy treatment was quantified. The Dice similarity coefficient (DSC), positive predictive value (PPV), sensitivity (SEN), average surface distance (ASD), and 95% maximum surface distance (95SD) were calculated on the brainstem, optic chiasm, optic nerves, mandible, parotid glands, and submandibular glands to evaluate the proposed SCSA-Net. The proposed SCSA-Net consistently outperforms the state-of-the-art methods on the public dataset. Specifically, compared with Res-Net and SE-Net, which is constructed from squeeze-and-excitation block equipped residual blocks, the DSC of the optic nerves and submandibular glands is improved by 0.06, 0.03 and 0.05, 0.04 by the SCSA-Net. Moreover, the proposed method achieves statistically significant improvements in terms of DSC on all and eight of nine OARs over Res-Net and SE-Net, respectively. The trained network was able to achieve good segmentation results on the serial dataset, but the results were further improved after fine-tuning of the model using the simulation CT images. For the parotids and submandibular glands, the volume changes of individual patients are highly consistent between the automated and manual segmentation (Pearson's correlation 0.97-0.99). The proposed SCSA-Net is computationally efficient to perform segmentation (sim 2 s/CT).

Analysis of Geometric Performance and Dosimetric Impact of Using Automatic Contour Segmentation for Radiotherapy Planning

Purpose: To analyze geometric discrepancy and dosimetric impact in using contours generated by auto-segmentation (AS) against manually segmented (MS) clinical contours. Methods: A 48-subject prostate atlas was created and another 15 patients were used for testing. Contours were generated using a commercial atlas-based segmentation tool and compared to their clinical MS counterparts. The geometric correlation was evaluated using the Dice similarity coefficient (DSC) and Hausdorff distance (HD). Dosimetric relevance was evaluated for a subset of patients by assessing the DVH differences derived by optimizing plan dose using the AS and MS contours, respectively, and evaluating with respect to each. A paired t-test was employed for statistical comparison. The discrepancy in plan quality with respect to clinical dosimetric endpoints was evaluated. The analysis was repeated for head/neck (HN) with a 31-subject atlas and 15 test cases. Results: Dice agreement between AS and MS differed significantly across structures: from (L:0.92/R: 0.91) for the femoral heads to seminal vesical of 0.38 in the prostate cohort, and from 0.98 for the brain, to 0.36 for the chiasm of the HN group. Despite the geometric disagreement, the paired t-tests showed the lack of statistical evidence for systematic differences in dosimetric plan quality yielded by the AS and MS approach for the prostate cohort. In HN cases, statistically significant differences in dosimetric endpoints were observed in structures with small volumes or elongated shapes such as cord (p = 0.01) and esophagus (p = 0.04). The largest absolute dose difference of 11 Gy was seen in the mean pharynx dose. Conclusion: Varying AS performance among structures suggests a differential approach of using AS on a subset of structures and focus MS on the rest. The discrepancy between geometric and dosimetric-end-point driven evaluation also indicates the clinical utility of AS contours in optimization and evaluating plan quality despite of suboptimal geometrical accuracy.

Practical Safety Considerations for Integration of Magnetic Resonance Imaging in Radiation Therapy

Interest in integrating magnetic resonance imaging (MRI) in radiation therapy (RT) practice has increased dramatically in recent years owing to its unique advantages such as excellent soft tissue contrast and capability of measuring biological properties. Continuous real-time imaging for intrafractional motion tracking without ionizing radiation serves as a particularly attractive feature for applications in RT. Despite its many advantages, the integration of MRI in RT workflows is not straightforward, with many unmet needs. MR safety remains one of the key challenges and concerns in the clinical implementation of MR simulators and MR-guided radiation therapy systems in radiation oncology. Most RT staff are not accustomed to working in an environment with a strong magnetic field. There are specific requirements in RT that are different from diagnostic applications. A large variety of implants and devices used in routine RT practice do not have clear MR safety labels. RT-specific imaging pulse sequences focusing on fast acquisition, high spatial integrity, and continuous, real-time acquisition require additional MR safety testing and evaluation. This article provides an overview of MR safety tailored toward RT staff, followed by discussions on specific requirements and challenges associated with MR safety in the RT environment. Strategies and techniques for developing an MR safety program specific to RT are presented and discussed.

Fraction-variant beam orientation optimization for intensity-modulated proton therapy

PURPOSE

To achieve a superior balance between dosimetry and the delivery efficiency of intensity-modulated proton therapy (IMPT) using as few beams as possible in a single fraction, we optimally vary beams in different fractions.

METHODS

In the optimization, 400~800 feasible noncoplanar beams were included in the candidate pool. For each beam, the doses of all scanning spots covering the target volume and a margin were calculated. The fraction-variant beam orientation optimization (FVBOO) problem was formulated to include three terms: two quadratic dose fidelity terms to penalize the deviation of planning target volume fractional dose and organs at risk (OAR) cumulative doses from prescription, respectively; an L2,1/2-norm group sparsity term to control the number of active beams per fraction to between 1 and 4. The Fast Iterative Shrinkage-Thresholding Algorithm (FISTA) was applied to solve this problem. FVBOO was tested on a patient with base-of-skull (BOS) tumor of 5 fractions (5f) and 30 fractions (30f) with an average number of active beams per fraction varying between 4 and 1. In addition, one bilateral head-and-neck (H&N) patient, and one esophageal cancer (ESG) patient of 30f were tested with about three active beams per fraction. The results were compared with IMPT plans that use fixed beams in each fraction. The fixed beams were selected using the group sparsity term with a fraction-invariant BOO (FIBOO) constraint.

RESULTS

Varying beams were chosen in either the 5f or 30f FVBOO plans. While similar number of beams per fraction was selected as the FIBOO plan, the FVBOO plans were able to spare the OARs better, with an average reduction of [Dmean, Dmax] from the FIBOO plans by [0.85, 2.08] Relative Biological Effective Gy (GyRBE) in the 5f plan and [1.87, 4.06] GyRBE in the 30f plans. While reducing the number of beams per fraction in the BOS patient, a three-beam/fraction 5f FVBOO plan performs comparably as the four-beam FIBOO plan and a two-beam/fraction 30f FVBOO plan still provides superior dosimetry.

CONCLUSION

Fraction-variant beam orientation optimization allows the utilization of a larger beam solution space for superior dose distribution in IMPT while maintaining a practical number of beams in each fraction.

Artificial intelligence in radiotherapy: a technological review

Radiation therapy (RT) is widely used to treat cancer. Technological advances in RT have occurred in the past 30 years. These advances, such as three-dimensional image guidance, intensity modulation, and robotics, created challenges and opportunities for the next breakthrough, in which artificial intelligence (AI) will possibly play important roles. AI will replace certain repetitive and labor-intensive tasks and improve the accuracy and consistency of others, particularly those with increased complexity because of technological advances. The improvement in efficiency and consistency is important to manage the increasing cancer patient burden to the society. Furthermore, AI may provide new functionalities that facilitate satisfactory RT. The functionalities include superior images for real-time intervention and adaptive and personalized RT. AI may effectively synthesize and analyze big data for such purposes. This review describes the RT workflow and identifies areas, including imaging, treatment planning, quality assurance, and outcome prediction, that benefit from AI. This review primarily focuses on deep-learning techniques, although conventional machine-learning techniques are also mentioned.

Self-paced DenseNet with boundary constraint for automated multi-organ segmentation on abdominal CT images

Automated multi-organ segmentation on abdominal CT images may replace or complement manual segmentation for clinical applications including image-guided radiation therapy. However, the accuracy of auto-segmentation is challenged by low image contrast, large spatial and inter-patient anatomical variations. In this study, we propose an end-to-end segmentation network, termed self-paced DenseNet, for improved multi-organ segmentation performance, especially for the difficult-to-segment organs. Specifically, a learning-based attention mechanism and dense connection block are seamlessly integrated into the proposed self-paced DenseNet to improve the learning capability and efficiency of the backbone network. To heavily focus on the organs showing low soft-tissue contrast and motion artifacts, a boundary condition is utilized to constrain the network optimization. Additionally, to ease the large learning pace discrepancies of individual organs, a task-wise self-paced-learning strategy is employed to adaptively control the learning paces of individual organs. The proposed self-paced DenseNet was trained and evaluated on a public abdominal CT data set consisting of 90 subjects with manually labeled ground truths of eight organs (including spleen, left kidney, esophagus, gallbladder, stomach, liver, pancreas, and duodenum). For quantitative evaluation, the Dice similarity coefficient (DSC) and average surface distance (ASD) were calculated. An average DSC of 84.46% and ASD of 1.82 mm were achieved on the eight organs, which outperforms the state-of-the-art segmentation methods 2.96% on DSC under the same experimental configuration. Moreover, the proposed segmentation method shows notable improvements on the duodenum and gallbladder, obtaining an average DSC of 69.26% and 80.94% and ASD of 2.14 mm and 2.24 mm, respectively. The results are markedly superior to the average DSC of 63.12% and 76.35% and average ASD of 3.87 mm and 4.33 mm using the vanilla DenseNet, respectively, for the two organs. We demonstrated the effectiveness of the proposed self-paced DenseNet to automatically segment abdominal organs with low boundary conspicuity. The self-paced DenseNet achieved consistently superior segmentation performance on eight abdominal organs with varying segmentation difficulties. The demonstrated computational efficiency (<2 s/CT) makes it well-suited for online applications.

Linear energy transfer weighted beam orientation optimization for intensity-modulated proton therapy

PURPOSE

In intensity-modulated proton therapy (IMPT), unaccounted-for variation in biological effectiveness contributes to the discrepancy between the constant relative biological effectiveness (RBE) model prediction and experimental observation. It is desirable to incorporate biological doses in treatment planning to improve modeling accuracy and consequently achieve a higher therapeutic ratio. This study addresses this demand by developing a method to incorporate linear energy transfer (LET) into beam orientation optimization (BOO).

METHODS

Instead of RBE-weighted dose, this LET weighted BOO (LETwBOO) framework uses the dose and LET product (LET  ×  D) as the biological surrogate. The problem is formulated with a physical dose fidelity term, a LET  ×  D constraint term, and a group sparsity term. The LET  ×  D of organs at risks is penalized for minimizing the biological effect while maintaining the physical dose objectives. Group sparsity is used to reduce the number of active beams from 600-800 non-coplanar candidate beams to between 2 and 4. This LETwBOO method was tested on three skull base tumor (SBT) patients and three bilateral head-and-neck (H&N) patients. The LETwBOO plans were compared with IMPT plans using manually selected beams with only physical dose constraint (MAN) and the initial MAN plan reoptimized with additional LET  ×  D constraint (LETwMAN).

RESULTS

The LETwBOO plans show superior physical dose and LET  ×  D sparing. On average, the [mean, maximal] doses of organs at risks (OARs) in LETwBOO are reduced by [2.85, 4.6] GyRBE from the MAN plans in the SBT cases and reduced by [0.9, 2.5] GyRBE in the H&N cases, while LETwMAN is comparable to MAN. cLET × Ds of PTVs are comparable in LETwBOO and LETwMAN, where c is a scaling factor of 0.04 μm/keV. On average, in the SBT cases, LETwBOO reduces the OAR [mean, maximal] cLET × D by [1.1, 2.9] Gy from the MAN plans, compared to the reduction by LETwMAN from MAN of [0.7, 1.7] Gy. In the H&N cases, LETwBOO reduces the OAR [mean, maximal] cLET × D by [0.8, 2.6] Gy from the MAN plans, compared to the reduction by LETwMAN from MAN of [0.3, 1.2] Gy.

CONCLUSION

We developed a novel LET weighted BOO method for IMPT to generate plans with improved physical and biological OAR sparing compared with the plans unaccounted for biological effects from BOO.

Quantifying vascular invasion in pancreatic cancer-a contrast CT based method for surgical resectability evaluation

Pancreatic cancer (PC) is one of the most lethal cancers, with frequent local therapy resistance and dismal 5-year survival rate. To date, surgical resection remains to be the only treatment option offering potential cure. Unfortunately, at diagnosis, the majority of patients demonstrate varying levels of vascular infiltration, which can contraindicate surgical resection. Patients unsuitable for immediate resection are further divided into locally advanced (LA) and borderline resectable (BR), with different treatment goals and therapeutic designs. Accurate definition of resectability is thus critical for PC patients, yet the existing methods to determine resectability rely on descriptive abutment to surrounding vessels rather than quantitative geometric characterization. Here, we aim to introduce a novel intra-subject object-space support-vector-machine (OsSVM) method to quantitatively characterize the degree of vascular involvement -- the main factor determining the PC resectability. Intra-subject OsSVMs were applied on 107 contrast CT scans (56 LA, BR and 26 resectable (RE) PC cases) for optimized tumor-vessel separations. Nine metrics derived from OsSVM margins were calculated as indicators of the overall vascular infiltration. The combined sets of matrics selected by the elastic net yielded high classification capability between LA and BR (AUC=0.95), as well as BR and RE (AUC=0.98). The proposed OsSVM method may provide an improved quantitative imaging guideline to refine the PC resectability grading system.

Image domain multi-material decomposition using single energy CT

Multi-material decomposition (MMD) technique decomposes the CT images into basis material images and has been promising in clinical practice for material composition quantification within the human body. MMD could be implemented using the image data acquired from spectral CT or its special case, dual-energy CT (DECT) while the spectral CT data acquisition usually requires a hardware modification. In this paper, we propose an image domain MMD method using single energy CT (SECT). The proposed objective function applies a least square data fidelity term to enforce the minimization between the linear combination of decomposed material image and the measured SECT image, and an edge-preserving (EP) regularization term to meet the piecewise constant property of the material image. We apply the optimization transfer principle to form a pixel-wise separable quadratic surrogate (PWSQS) function in each iteration to decrease the objective function. The pixelwise direct inversion method assisted by the two-material assumption (TMA) is applied to obtain a good initial value. The proposed method is evaluated using a digital phantom, a Catphan phantom and the clinical data. A low-pass filtration method is implemented for a comparison purpose. In the phantom study, the proposed TMA method achieves high volume fraction accuracy (VFA) of 79.64% and the proposed EP method further increases the VFA by 15.56% and decreases the decomposition standard deviation (STD) by 81.51% compared with the TMA method. At the comparable noise level, the proposed EP method increases spatial resolution by an overall factor of 1.01 when the modulation transfer function magnitude is decreased to 50% compared with the low-pass filtration method. In clinical data study, the virtual non-contrast image generated by the proposed method achieves the root-mean-squared-relative error of 2.93% compared with the contrast-free ground-truth image.

A novel energy layer optimization framework for spot-scanning proton arc therapy

PURPOSE

Spot-scanning proton arc therapy (SPAT) is an emerging modality to improve plan conformality and delivery efficiency. A greedy and heuristic method is proposed in the existing SPAT algorithm to select energy layers and sequence energy switching with gantry rotation, which does not promise optimality in either dosimetry or efficiency. We aim to develop a method to solve the energy layer switching and dosimetry optimization problems in an integrated framework for SPAT.

METHODS

In an integrated approach, energy layer optimization for spot-scanning proton arc therapy (ELO-SPAT) is formulated with a dose fidelity term, a group sparsity regularization, a log barrier regularization, and an energy sequencing (ES) penalty. The combination of L2,1/2-norm group sparsity regularization and log barrier function allows one energy layer being selected per control point. The ES regularization term sorts the delivery sequence from high energy to low energy to reduce the total energy layer switching time (ELST) and subsequently the total delivery time. Within the ES penalty, the gradient of layer weights between adjacent beams is first calculated along beam direction and then along energy direction. The gradients indicate energy switch patterns between two adjacent beams. The time-wise costly energy switch-up is more heavily penalized in the ES term. This ELO-SPAT method was tested on one frontal base-of-skull (BOS) patient, one chordoma (CHDM) patient with a simultaneous integrated boost, one bilateral head-and-neck (H&N) patient, and one lung (LNG) patient. We compared ELO-SPAT with intensity-modulated proton therapy (IMPT) using discrete beams and SPArc by Ding et al. For the two arc algorithms, both the plans with and without energy sequencing were created and compared.

RESULTS

Energy layer optimization for spot-scanning proton arc therapy reduced the runtime of optimization by 84% on average compared with the greedy SPArc method. In both the ELO-SPAT plans with and without ES, one energy layer per control point was selected. Without ES regularization, the energy sequence was arbitrary, with around 40-60 switch-up for the tested cases. After adding ES regularization, the number of energy switch-up was reduced to less than 20. Compared with the energy sequenced SPArc plans, the ELO-SPAT plans with ES led to 24% less total ELST for synchrotron plans and 14% less for cyclotron plans. Both the ELO-SPAT and SPArc plans achieved better sparing compared with the IMPT plans for most Organs-at-risks (OARs), with or without ES. Without ES, the ELO-SPAT plans achieved further improvement of the OARs compared with the SPArc plans, with an averaged reduction of OAR [Dmean, Dmax] by [1.57, 3.34] GyRBE. Adding the ES regularization degraded the plan quality, but the ELO-SPAT plans still had comparable or slightly better sparing than the SPArc plans with ES, with an averaged reduction of OAR [Dmean, Dmax] by [1.42, 2.34] GyRBE.

CONCLUSION

We developed a computationally efficient spot-scanning proton arc optimization method, which solved energy layer selection and sequencing in an integrated framework, generating plans with good dosimetry and high delivery efficiency.

Multimodality image registration in the head-and-neck using a deep learning-derived synthetic CT as a bridge

PURPOSE

To develop and demonstrate the efficacy of a novel head-and-neck multimodality image registration technique using deep-learning-based cross-modality synthesis.

METHODS AND MATERIALS

Twenty-five head-and-neck patients received magnetic resonance (MR) and computed tomography (CT) (CTaligned ) scans on the same day with the same immobilization. Fivefold cross validation was used with all of the MR-CT pairs to train a neural network to generate synthetic CTs from MR images. Twenty-four of 25 patients also had a separate CT without immobilization (CTnon-aligned ) and were used for testing. CTnon-aligned 's were deformed to the synthetic CT, and compared to CTnon-aligned registered to MR. The same registrations were performed from MR to CTnon-aligned and from synthetic CT to CTnon-aligned . All registrations used B-splines for modeling the deformation, and mutual information for the objective. Results were evaluated using the 95% Hausdorff distance among spinal cord contours, landmark error, inverse consistency, and Jacobian determinant of the estimated deformation fields.

RESULTS

When large initial rigid misalignment is present, registering CT to MRI-derived synthetic CT aligns the cord better than a direct registration. The average landmark error decreased from 9.8 ± 3.1 mm in MR→CTnon-aligned to 6.0 ± 2.1 mm in CTsynth →CTnon-aligned deformable registrations. In the CT to MR direction, the landmark error decreased from 10.0 ± 4.3 mm in CTnon-aligned →MR deformable registrations to 6.6 ± 2.0 mm in CTnon-aligned →CTsynth deformable registrations. The Jacobian determinant had an average value of 0.98. The proposed method also demonstrated improved inverse consistency over the direct method.

CONCLUSIONS

We showed that using a deep learning-derived synthetic CT in lieu of an MR for MR→CT and CT→MR deformable registration offers superior results to direct multimodal registration.

Many-isocenter optimization for robotic radiotherapy

Despite significant dosimetric gains, clinical implementation of the 4π non-coplanar radiotherapy on the widely available C-arm gantry system is hindered by limited clearance, and the need to perform complex coordinated gantry and couch motion. A robotic radiotherapy platform would be conducive to such treatment but a new conflict between field size and MLC modulation resolution needs to be managed for versatile applications. This study investigates the dosimetry and delivery efficiency of purposefully creating many isocenters to achieve simultaneously high MLC modulation resolution and large tumor coverage. An integrated optimization framework was proposed for simultaneous beam orientation optimization (BOO), isocenter selection, and fluence map optimization (FMO). The framework includes a least-square dose fidelity objective, a total variation term for regularizing the fluence smoothness, and a group sparsity term for beam selection. A minimal number of isocenters were identified for efficient target coverage. Colliding beams excluded, high-resolution small-field 4π intensity-modulated radiotherapy (IMRT) treatment plans with 50 cm source-to-isocenter distance (SID-50) on 10 Head and Neck (H&N) cancer patients were compared with low-resolution large-field plans with 100 cm SID (SID-100). With the same or better target coverage, the average reduction of [Dmean, Dmax] of 20-beam SID-50 plans from 20-beam SID-100 plans were [2.09 Gy, 1.19 Gy] for organs at risk (OARs) overall, [3.05 Gy, 0.04 Gy] for parotid gland, [3.62 Gy, 5.19 Gy] for larynx, and [3.27 Gy, 1.10 Gy] for mandible. R50 and integral dose were reduced by 5.3% and 9.6%, respectively. Wilcoxon signed-rank test showed significant difference (p   <  0.05) in planning target volume (PTV) homogeneity, PTV Dmax, R50, Integral dose, and OAR Dmean and Dmax. The estimated delivery time of 20-beam [SID-50, SID-100] plans were [19, 18] min and [14, 9] min, assuming 5 fractions and 30 fractions, respectively. With clinically acceptable delivery efficiency, many-isocenter optimization is dosimetrically desirable for treating large targets with high modulation resolution on the robotic platform.

Technical Note: Robust individual thermoluminescence dosimeter tracking using optical fingerprinting

PURPOSE

The thermoluminescence dosimeter (TLD) has desirable features including low cost, reusability, small size, and relatively low energy dependence. However, the commonly available poly-crystal TLDs (e.g., TLD-100) exhibit high interdetector variability that requires individual calibration for high detection accuracy. To improve individual TLD tracking robustness, we developed an optical fingerprinting method to identify the TLD-100 chips.

METHODS

Seven hundred and fifty-two images were initially captured using a digital microscope camera to build a feature library for both facets of 376 TLD-100 chips. A median intensity thresholding method was used to segment images into foreground and background. The affine transformation was used to register the segmented images to the same position. The fingerprint of each image was calculated from its registered image. All fingerprints were then recorded in an Elasticsearch® search database. The TLD fingerprint match was tested three times when the library was established and repeated once 20 months later. All chips were irradiated at 0, 1, 4, and 8 Gy on a calibrated clinical MV linac to establish the individual calibration curve.

RESULTS

The true positive rate of identifying TLDs based on their optical fingerprints was 100% at initialization of the inventory. After 20 months and multiple deployments for characterization, calibration, and dose measurement, the true positive match rate dropped to 99% with zero false-positive matches. The TLDs exhibited high self-consistency in the dose-response test with R² between 0.988 and 1 with linear regression.

CONCLUSIONS

The TLD-100 chips surface textures are unique and sufficient to support accurate identification based on the optical fingerprinting. This method provides inexpensive and robust management of the TLDs for individual calibration and dosimetry.

A sparse orthogonal collimator for small animal intensity-modulated radiation therapy. Part II: hardware development and commissioning

PURPOSE

A dose-modulation device for small animal radiotherapy is required to use clinically analogous treatment techniques, which will likely increase the translatability of preclinical research results. Because the clinically used multileaf collimator (MLC) is impractical for miniaturization, we have developed a simpler, better-suited sparse orthogonal collimator (SOC) for delivering small animal intensity-modulated radiation therapy (IMRT) using a rectangular aperture optimization (RAO) treatment planning system.

METHODS

The SOC system was modeled in computer-aided design software and fabricated with machined tungsten leaves and three-dimensional (3D) printed leaf housing. A graphical user interface was developed for controlling and calibrating the SOC leaves, which are driven by Arduino-controlled stepper motors. A Winston-Lutz test was performed to assess mechanical alignment, and abutting field and grid dose patterns were created to analyze intra- and intercalibration leaf positioning error. Leaf transmission and penumbra were measured over the full range of gantry angles and leaf positions, respectively. Three SOC test plans were delivered, and film measurements were compared to the intended dose distributions. The differences in maximum, mean, and minimum, as well as pixelwise absolute dose differences, were compared for each structure, and a gamma analysis was performed for the target structures using criteria of 4% dose difference and 0.3 mm distance to agreement.

RESULTS

The Winston-Lutz test revealed maximum directional offsets between the SOC and primary collimator axes of 0.53 mm at 0° and 0.68 mm over the full 360°. Upper and lower abutting field patterns had maximum dose deviations of 18.8 ± 3.1% and 15.5 ± 2.9%, respectively, and grid patterns showed intra- and intercalibration repeatability of 93% and 91%, respectively. Extremely low midleaf (0.15 ± 0.05%) and interleaf (0.27 ± 0.22%) transmission was measured, with no significant rotational variation. The average penumbra was ~0.8 mm for all leaves at field center, with a range of 0.17 mm for all leaf positions. A highly concave test plan was delivered with a ~ 95% gamma analysis pass rate, and a realistic mouse phantom liver irradiation plan achieved a pass rate of ~98%. A highly complex dose distribution was also created with 551 SOC apertures averaging 2.4 mm in size.

CONCLUSIONS

A sparse orthogonal collimator was developed and commissioned, with promising preliminary dosimetry results. The SOC design, with its limited moving components and high dose-modulation resolution, is ideal for delivering high-quality small animal IMRT with our RAO-based treatment planning system.

A sparse orthogonal collimator for small animal intensity-modulated radiation therapy part I: Planning system development and commissioning

PURPOSE

To achieve more translatable preclinical research results, small animal irradiation needs to more closely simulate human radiotherapy. Although the clinical gold standard is intensity-modulated radiation therapy (IMRT), the direct translation of this method for small animals is impractical. In this study we describe the treatment planning system for a novel dose modulation device to address this challenge.

METHODS

Using delineated target and avoidance structures, a rectangular aperture optimization (RAO) problem was formulated to penalize deviations from a desired dose distribution and limit the number of selected rectangular apertures. RAO was used to create IMRT plans with highly concave targets in the mouse brain, and the plan quality was compared to that using a hypothetical miniaturized multileaf collimator (MLC). RAO plans were also created for a realistic application of mouse whole liver irradiation and for a highly complex two-dimensional (2D) dose distribution as a proof-of-principle. Beam commissioning data, including output and off-axis factors and percent depth dose (PDD) curves, were acquired for our small animal irradiation system and incorporated into the treatment planning system. A plan post-processing step was implemented for aperture size-specific dose recalculation and aperture weighting reoptimization.

RESULTS

The first RAO test case achieved highly conformal doses to concave targets in the brain, with substantially better dose gradient, conformity, and target dose homogeneity than the hypothetical miniaturized MLC plans. In the second test case, a highly conformal dose to the liver was achieved with significant sparing of the kidneys. RAO also successfully replicated a complex 2D dose distribution with three prescription dose levels. Energy spectra for field sizes 1 to 20 mm were calculated to match the measured PDD curves, with maximum and mean dose deviations of 4.47 ± 0.30% and 1.71 ± 0.18%. The final reoptimization of aperture weightings for the complex RAO test plan was able to reduce the maximum and mean dose deviations between the optimized and recalculated dose distributions from 10.3% to 6.6% and 4.0% to 2.8%, respectively.

CONCLUSIONS

Using the advanced optimization techniques, complex IMRT plans were achieved using a simple dose modulation device. Beam commissioning data were incorporated into the treatment planning process to more accurately predict the resulting dose distribution. This platform substantially reduces the gap in treatment plan quality between clinical and preclinical radiotherapy, potentially increasing the value and flexibility of small animal studies.

Motion Compensated Dynamic MRI Reconstruction With Local Affine Optical Flow Estimation

This paper proposes a novel framework to reconstruct dynamic magnetic resonance imaging (DMRI) with motion compensation (MC). Specifically, by combining the intensity-based optical flow constraint with the traditional compressed sensing scheme, we are able to jointly reconstruct the DMRI sequences and estimate the interframe motion vectors. Then, the DMRI reconstruction can be refined through MC with the estimated motion field. By employing the coarse-to-fine multi-scale resolution strategy, we are able to update the motion field in different spatial scales. The estimated motion vectors need to be interpolated to the finest resolution scale to compensate the DMRI reconstruction. Moreover, the proposed framework is capable of handling a wide class of prior information (regularizations) for DMRI reconstruction, such as sparsity, low rank, and total variation. The formulated optimization problem is solved by a primal-dual algorithm with linesearch due to its efficiency when dealing with non-differentiable problems. Experiments on various DMRI datasets validate the reconstruction quality improvement using the proposed scheme in comparison to several state-of-the-art algorithms.

Image-domain multimaterial decomposition for dual-energy computed tomography with nonconvex sparsity regularization

Dual-energy computed tomography (CT) has the potential to decompose tissues into different materials. However, the classic direct inversion (DI) method for multimaterial decomposition (MMD) cannot accurately separate more than two basis materials due to the ill-posed problem and amplified image noise. We propose an integrated MMD method that addresses the piecewise smoothness and intrinsic sparsity property of the decomposition image. The proposed MMD was formulated as an optimization problem including a quadratic data fidelity term, an isotropic total variation term that encourages image smoothness, and a nonconvex penalty function that promotes decomposition image sparseness. The mass and volume conservation rule was formulated as the probability simplex constraint. An accelerated primal-dual splitting approach with line search was applied to solve the optimization problem. The proposed method with different penalty functions was compared against DI on a digital phantom, a Catphan^® 600 phantom, a quantitative imaging phantom, and a pelvis patient. The proposed framework distinctly separated the CT image up to 12 basis materials plus air with high decomposition accuracy. The cross talks between two different materials are substantially reduced, as shown by the decreased nondiagonal elements of the normalized cross correlation (NCC) matrix. The mean square error of the measured electron densities was reduced by 72.6%. Across all datasets, the proposed method improved the average volume fraction accuracy from 61.2% to 99.9% and increased the diagonality of the NCC matrix from 0.73 to 0.96. Compared with DI, the proposed MMD framework improved decomposition accuracy and material separation.

Automated 4π radiotherapy treatment planning with evolving knowledge-base

PURPOSE

Non-coplanar 4π radiotherapy generalizes intensity modulated radiation therapy (IMRT) to automate beam geometry selection but requires complicated hyperparameter tuning to attain superior plan quality, which can be tedious and inconsistent. In this study, a fully automated 4π treatment planning was developed using evolving knowledge-base (EKB) planning guided by dose prediction.

METHODS

Twenty 4π lung and twenty 4π head and neck (HN) cases were included. A statistical voxel dose learning model was initially trained on low-quality plans created using generic hyperparameter templates without manual tuning. To improve the automated plan quality without being limited by the training data quality, a new 4π optimization problem was formulated to include a one-sided penalty on the organ-at-risk (OAR) dose deviation from the predicted dose. This directional OAR penalty encourages superior OAR sparing. The fast iterative shrinkage-thresholding algorithm (FISTA) was used to solve the large-scale beam orientation optimization problem. With the improved plans, new predictions were created to guide the next loop of EKB planning for a total of 10 loops. Plan quality was evaluated using a plan quality metric (PQM) points system based on clinical dose constraints and compared with automated planning approaches guided by manual high-quality plans using all non-coplanar beams, automated plans using individually evolved targeted dose, and manually created 4π plans.

RESULTS

For the lung cases, the final EKB plans had significantly higher PQM than manually created 4π (+2.60%). The improvements plateaued after the third loop. The final HN EKB plans and manually created 4π plans had comparable PQMs, but had lower PQM compared to automated plans using a high-quality training set (-3.00% and -4.44%, respectively). The PQM consistently increased up to the sixth loop. Individually evolved plans were able to improve the plan quality from initial condition due to the one-sided cost function but the 60% of them were trapped in undesired local minima that were substantially worse than their corresponding EKB plans.

CONCLUSION

Evolving knowledge-base planning is a novel automated planning technique guided by the predicted three-dimensional dose distribution, which can evolve from low-quality plans. EKB allows new beams to be used in the automated planning workflow for superior plan quality.