Deep-learning based fast and accurate 3D CT deformable image registration in lung cancer

BACKGROUND

Deformable Image Registration (DIR) is an essential technique required in many applications of radiation oncology. However, conventional DIR approaches typically take several minutes to register one pair of 3D CT images and the resulting deformable vector fields (DVFs) are only specific to the pair of images used, making it less appealing for clinical application.

PURPOSE

A deep-learning-based DIR method using CT images is proposed for lung cancer patients to address the common drawbacks of the conventional DIR approaches and in turn can accelerate the speed of related applications, such as contour propagation, dose deformation, adaptive radiotherapy (ART), etc. METHODS: A deep neural network based on VoxelMorph was developed to generate DVFs using CT images collected from 114 lung cancer patients. Two models were trained with the weighted mean absolute error (wMAE) loss and structural similarity index matrix (SSIM) loss (optional) (i.e., the MAE model and the M+S model). In total, 192 pairs of initial CT (iCT) and verification CT (vCT) were included as a training dataset and the other independent 10 pairs of CTs were included as a testing dataset. The vCTs usually were taken 2 weeks after the iCTs. The synthetic CTs (sCTs) were generated by warping the vCTs according to the DVFs generated by the pre-trained model. The image quality of the synthetic CTs was evaluated by measuring the similarity between the iCTs and the sCTs generated by the proposed methods and the conventional DIR approaches, respectively. Per-voxel absolute CT-number-difference volume histogram (CDVH) and MAE were used as the evaluation metrics. The time to generate the sCTs was also recorded and compared quantitatively. Contours were propagated using the derived DVFs and evaluated with SSIM. Forward dose calculations were done on the sCTs and the corresponding iCTs. Dose volume histograms (DVHs) were generated based on dose distributions on both iCTs and sCTs generated by two models, respectively. The clinically relevant DVH indices were derived for comparison. The resulted dose distributions were also compared using 3D Gamma analysis with thresholds of 3 mm/3%/10% and 2 mm/2%/10%, respectively.

RESULTS

The two models (wMAE and M+S) achieved a speed of 263.7±163 / 265.8±190 ms and a MAE of 13.15±3.8 / 17.52±5.8 HU for the testing dataset, respectively. The average SSIM scores of 0.987±0.006 and 0.988±0.004 were achieved by the two proposed models, respectively. For both models, CDVH of a typical patient showed that less than 5% of the voxels had a per-voxel absolute CT-number-difference larger than 55 HU. The dose distribution calculated based on a typical sCT showed differences of ≤2cGy[RBE] for clinical target volume (CTV) D95 and D5 , within ±0.06% for total lung V5 , ≤1.5cGy[RBE] for heart and esophagus Dmean , and ≤6cGy[RBE] for cord Dmax compared to the dose distribution calculated based on the iCT. The good average 3D Gamma passing rates (> 96% for 3 mm/3%/10% and > 94% for 2 mm/2%/10%, respectively) were also observed.

CONCLUSION

A deep neural network-based DIR approach was proposed and has been shown to be reasonably accurate and efficient to register the initial CTs and verification CTs in lung cancer.

Technical note: Investigation of dose and LETd effect to rectum and bladder by using non-straight laterals in prostate cancer receiving proton therapy

BACKGROUND

Parallel-opposed lateral beams are the conventional beam arrangements in proton therapy for prostate cancer. However, when considering linear energy transfer (LET) and RBE effects, alternative beam arrangements should be investigated.

PURPOSE

To investigate the dose and dose averaged LET (LET_d ) impact of using new beam arrangements rotating beams 5°-15° posteriorly to the laterals in prostate cancer treated with pencil-beam-scanning (PBS) proton therapy.

METHODS

Twenty patients with localized prostate cancer were included in this study. Four proton treatment plans for each patient were generated utilizing 0°, 5°, 10°, and 15° posterior oblique beam pairs relative to parallel-opposed lateral beams. Dose-volume histograms (DVHs) from posterior oblique beams were analyzed. Dose-LET_d -volume histogram (DLVH) was employed to study the difference in dose and LET_d with each beam arrangement. DLVH indices, V ( d , l ) $V( {d,l} )$ , defined as the cumulative absolute volume that has a dose of at least d (Gy[RBE]) and a LET_d of at least l (keV/µm), were calculated for both the rectum and bladder to the whole group of patients and two-sub groups with and without hydrogel spacer. These metrics were tested using Wilcoxon signed-rank test.

RESULTS

Rotating beam angles from laterals to slightly posterior by 5°-15° reduced high LET_d volumes while it increased the dose volume in the rectum and increased LET_d in bladders. Beam angles rotated five degrees posteriorly from laterals (i.e., gantry in 95° and 265°) are proposed since they achieved the optimal balance of better LET_d sparing and minimal dose increase in the rectum. A reduction of V(50 Gy[RBE], 2.6 keV/µm) from 7.41 to 3.96 cc (p < 0.01), and a slight increase of V(50 Gy[RBE], 0 keV/µm) from 20.1 to 21.6 cc (p < 0.01) were observed for the group without hydrogel spacer. The LET_d sparing was less effective for the group with hydrogel spacer, which achieved the reduction of V(50 Gy[RBE], 2.6 keV/µm) from 4.28 to 2.10 cc (p < 0.01).

CONCLUSIONS

Posterior oblique angle plans improved LET_d sparing of the rectum while sacrificing LET_d sparing in the bladder in the treatment of prostate cancer with PBS. Beam angle modification from laterals to slightly posterior may be a strategy to redistribute LET_d and perhaps reduce rectal toxicity risks in prostate cancer patients treated with PBS. However, the effect is reduced for patients with hydrogel spacer.

Virtual particle Monte Carlo: A new concept to avoid simulating secondary particles in proton therapy dose calculation

BACKGROUND

In proton therapy dose calculation, Monte Carlo (MC) simulations are superior in accuracy but more time consuming, compared to analytical calculations. Graphic processing units (GPUs) are effective in accelerating MC simulations but may suffer thread divergence and racing condition in GPU threads that degrades the computing performance due to the generation of secondary particles during nuclear reactions.

PURPOSE

A novel concept of virtual particle (VP) MC (VPMC) is proposed to avoid simulating secondary particles in GPU-accelerated proton MC dose calculation and take full advantage of the computing power of GPU.

METHODS

Neutrons and gamma rays were ignored as escaping from the human body; doses of electrons, heavy ions, and nuclear fragments were locally deposited; the tracks of deuterons were converted into tracks of protons. These particles, together with primary and secondary protons, are considered to be the realistic particles. Histories of primary and secondary protons were replaced by histories of multiple VPs. Each VP corresponded to one proton (either primary or secondary). A continuous-slowing-down-approximation model, an ionization model, and a large angle scattering event model corresponding to nuclear interactions were developed for VPs by generating probability distribution functions (PDFs) based on simulation results of realistic particles using MCsquare. For efficient calculations, these PDFs were stored in the Compute Unified Device Architecture textures. VPMC was benchmarked with TOPAS and MCsquare in phantoms and with MCsquare in 13 representative patient geometries. Comparisons between the VPMC calculated dose and dose measured in water during patient-specific quality assurance (PSQA) of the selected 13 patients were also carried out. Gamma analysis was used to compare the doses derived from different methods and calculation efficiencies were also compared.

RESULTS

Integrated depth dose and lateral dose profiles in both homogeneous and inhomogeneous phantoms all matched well among VPMC, TOPAS, and MCsquare calculations. The 3D-3D gamma passing rates with a criterion of 2%/2 mm and a threshold of 10% was 98.49% between MCsquare and TOPAS and 98.31% between VPMC and TOPAS in homogeneous phantoms, and 99.18% between MCsquare and TOPAS and 98.49% between VPMC and TOPAS in inhomogeneous phantoms, respectively. In patient geometries, the 3D-3D gamma passing rates with 2%/2 mm/10% between dose distributions from VPMC and MCsquare were 98.56 ± 1.09% in patient geometries. The 2D-3D gamma analysis with 3%/2 mm/10% between the VPMC calculated dose distributions and the 2D measured planar dose distributions during PSQA was 98.91 ± 0.88%. VPMC calculation was highly efficient and took 2.84 ± 2.44 s to finish for the selected 13 patients running on four NVIDIA Ampere GPUs in patient geometries.

CONCLUSION

VPMC was found to achieve high accuracy and efficiency in proton therapy dose calculation.

Exploratory study of seed spots analysis to characterize dose and linear energy transfer effect in adverse event initialization of pencil beam scanning proton therapy

BACKGROUND

Both dose and linear-energy-transfer (LET) could play a substantial role in adverse event (AE) initialization of cancer patients treated with pencil-beam-scanning proton therapy (PBS). However, not all the voxels within the AE regions are directly induced from the dose and LET effect. It is important to study the synergistic effect of dose and LET in AE initialization by only including a subset of voxels that are dosimetrically important.

PURPOSE

To perform exploratory investigation of the dose and LET effects upon AE initialization in PBS using seed spots analysis.

METHODS

113 head and neck (H&N) cancer patients receiving curative PBS were included. Among them, 20 patients experienced unanticipated CTCAEv4.0 grade≥3 AEs (AE group) and 93 patients did not (control group). Within the AE group, 13 AE patients were included in the seed spot analysis to derive the descriptive features of AE initialization and the remaining 7 mandible osteoradionecrosis patients and 93 control patients were used to derive the feature-based volume constraint of mandible osteoradionecrosis. The AE regions were contoured and the corresponding dose-LET volume histograms (DLVHs) of AE regions were generated for all patients in the AE group. We selected high LET voxels (the highest 5% of each dose bin) with a range of moderate to high dose (≥∼40 Gy[RBE]) as critical voxels. Critical voxels which were contiguous with each other were grouped into clusters. Each cluster was considered as a potential independent seed spot for AE initialization. Seed spots were displayed in a 2D dose-LET plane based on their mean dose and LET to derive the descriptive features of AE initialization. A volume constraint of mandible osteoradionecrosis was then established based on the extracted features using a receiver operating characteristic curve.

RESULTS

The product of dose and LET (xBD) was found to be a descriptive feature of seed spots leading to AE initialization in this preliminary study. The derived xBD volume constraint for mandible osteoradionecrosis showed good performance with an area-under-curve of 0.87 (sensitivity of 0.714 and specificity of 0.807 in the leave-one-out cross validation) for the very limited patient data included in this study.

CONCLUSION

Our exploratory study showed that both dose and LET were observed to be important in AE initializations. The derived xBD volume constraint could predict mandible osteoradionecrosis reasonably well in the very limited H&N cancer patient data treated with PBS included in this study. This article is protected by copyright. All rights reserved.

A novel and individualized robust optimization method using normalized dose interval volume constraints (NDIVC) for intensity-modulated proton radiotherapy

PURPOSE

Intensity-modulated proton therapy (IMPT) is known to be sensitive to patient setup and range uncertainty issues. Multiple robust optimization methods have been developed to mitigate the impact of these uncertainties. Here, we propose a new robust optimization method, which provides an alternative way of robust optimization in IMPT, and is clinically practical, which will enable users to control the balance between nominal plan quality and plan robustness in a user-defined fashion.

METHOD

We calculated nine individual dose distributions which corresponded to one nominal and eight extreme scenarios caused by patient setup and proton beam's range uncertainties. For each voxel, the normalized dose interval (NDI) is defined as the full dose range variation divided by the maximum dose in all uncertainty scenarios (NDI = [max - min dose]/max dose), which was then used to calculate the normalized dose interval volume histogram (NDIVH) curves. The areas under the NDIVH curves were used to quantify plan robustness. A normalized dose interval volume constraint (NDIVC) applied to the target was incorporated to specify the desired robustness which was user-defined. Users could then explore the trade-off between nominal plan quality and plan robustness by adjusting the position of the NDIVCs on the NDIVH curves freely. We benchmarked our method using one lung, five head and neck (H&N), and three prostate cases by comparing our results to those derived using the voxel-wise worst-case robust optimization.

RESULTS

Using the benchmark cases, our new method achieved quality IMPT plans comparable to those derived from the voxel-wise worst-case robust optimization for both nominal plan quality and plan robustness in general; even more conformal and more homogeneous target dose distributions in some cases, if proper NDIVCs were applied. The AUC under NDIVH, as a precise quantitative index of plan robustness, was consistent with DVH bandwidths. Additionally, we demonstrated the feasibility of adjusting the position of NDIVCs in the NDIVH curves which allowed users to explore the trade-off between nominal plan quality and plan robustness.

CONCLUSIONS

The NDIVH-based robust optimization method provided a novel and individualized way of robust optimization in IMPT, and enables users to adjust the balance between nominal plan quality and plan robustness in a user-defined fashion. This method is applicable for continued improvement and developing the next generation of IMPT planning algorithms in the future.

Assessing the robustness of passive scattering proton therapy with regard to local recurrence in stage III non-small cell lung cancer: a secondary analysis of a phase II trial

Background

We assessed the robustness of passive scattering proton therapy (PSPT) plans for patients in a phase II trial of PSPT for stage III non-small cell lung cancer (NSCLC) by using the worst-case scenario method, and compared the worst-case dose distributions with the appearance of locally recurrent lesions.

Methods

Worst-case dose distributions were generated for each of 9 patients who experienced recurrence after concurrent chemotherapy and PSPT to 74 Gy(RBE) for stage III NSCLC by simulating and incorporating uncertainties associated with set-up, respiration-induced organ motion, and proton range in the planning process. The worst-case CT scans were then fused with the positron emission tomography (PET) scans to locate the recurrence.

Results

Although the volumes enclosed by the prescription isodose lines in the worst-case dose distributions were consistently smaller than enclosed volumes in the nominal plans, the target dose coverage was not significantly affected: only one patient had a recurrence outside the prescription isodose lines in the worst-case plan.

Conclusions

PSPT is a relatively robust technique. Local recurrence was not associated with target underdosage resulting from estimated uncertainties in 8 of 9 cases.

Preliminary evaluation of multifield and single-field optimization for the treatment planning of spot-scanning proton therapy of head and neck cancer

PURPOSE

Spot-scanning proton therapy (SSPT) using multifield optimization (MFO) can generate highly conformal dose distributions, but it is more sensitive to setup and range uncertainties than SSPT using single-field optimization (SFO). The authors compared the two optimization methods for the treatment of head and neck cancer with bilateral targets and determined the superior method on the basis of both the plan quality and the plan robustness in the face of setup and range uncertainties.

METHODS

Four patients with head and neck cancer with bilateral targets who received SSPT treatment in the authors' institution were studied. The patients had each been treated with a MFO plan using three fields. A three-field SFO plan (3F-SFO) and a two-field SFO plan (2F-SFO) with the use of a range shifter in the beam line were retrospectively generated for each patient. The authors compared the plan quality and robustness to uncertainties of the SFO plans with the MFO plans. Robustness analysis of each plan was performed to generate the two dose distributions consisting of the highest and the lowest possible doses (worst-case doses) from the spatial and range perturbations at every voxel. Dosimetric indices from the nominal and worst-case plans were compared.

RESULTS

The 3F-SFO plans generally yielded D95 and D5 values in the targets that were similar to those of the MFO plans. 3F-SFO resulted in a lower dose to the oral cavity than MFO in all four patients by an average of 9.9 Gy, but the dose to the two parotids was on average 6.7 Gy higher for 3F-SFO than for MFO. 3F-SFO plans reduced the variations of dosimetric indices under uncertainties in the targets by 22.8% compared to the MFO plans. Variations of dosimetric indices under uncertainties in the organs at risk (OARs) varied between organs and between patients, although they were on average 9.2% less for the 3F-SFO plans than for the MFO plans. Compared with the MFO plans, the 2F-SFO plans showed a reduced dose to the parotids for both the nominal dose and in the worst-case scenario, but the plan robustness in the target of the 2F-SFO plans was not notably greater than that of the MFO plans.

CONCLUSIONS

Compared with MFO, 3F-SFO improves plan robustness in the targets but degrades dose sparing in the parotids in both the nominal and worst-case scenarios. Although 2F-SFO improves parotid sparing compared with MFO, it produces little improvement in plan robustness. Therefore, considering its tolerable target coverage and sparing of OARs in worst-case scenarios, the authors recommend MFO as the planning method for the treatment of head and neck cancer with bilateral targets.

Influence of robust optimization in intensity-modulated proton therapy with different dose delivery techniques

PURPOSE

The distal edge tracking (DET) technique in intensity-modulated proton therapy (IMPT) allows for high energy efficiency, fast and simple delivery, and simple inverse treatment planning; however, it is highly sensitive to uncertainties. In this study, the authors explored the application of DET in IMPT (IMPT-DET) and conducted robust optimization of IMPT-DET to see if the planning technique's sensitivity to uncertainties was reduced. They also compared conventional and robust optimization of IMPT-DET with three-dimensional IMPT (IMPT-3D) to gain understanding about how plan robustness is achieved.

METHODS

They compared the robustness of IMPT-DET and IMPT-3D plans to uncertainties by analyzing plans created for a typical prostate cancer case and a base of skull (BOS) cancer case (using data for patients who had undergone proton therapy at our institution). Spots with the highest and second highest energy layers were chosen so that the Bragg peak would be at the distal edge of the targets in IMPT-DET using 36 equally spaced angle beams; in IMPT-3D, 3 beams with angles chosen by a beam angle optimization algorithm were planned. Dose contributions for a number of range and setup uncertainties were calculated, and a worst-case robust optimization was performed. A robust quantification technique was used to evaluate the plans' sensitivity to uncertainties.

RESULTS

With no uncertainties considered, the DET is less robust to uncertainties than is the 3D method but offers better normal tissue protection. With robust optimization to account for range and setup uncertainties, robust optimization can improve the robustness of IMPT plans to uncertainties; however, our findings show the extent of improvement varies.

CONCLUSIONS

IMPT's sensitivity to uncertainties can be improved by using robust optimization. They found two possible mechanisms that made improvements possible: (1) a localized single-field uniform dose distribution (LSFUD) mechanism, in which the optimization algorithm attempts to produce a single-field uniform dose distribution while minimizing the patching field as much as possible; and (2) perturbed dose distribution, which follows the change in anatomical geometry. Multiple-instance optimization has more knowledge of the influence matrices; this greater knowledge improves IMPT plans' ability to retain robustness despite the presence of uncertainties.