Geometric uncertainty analysis of MLC tracking for lung SABR

Rapid distortion correction enables accurate magnetic resonance imaging-guided real-time adaptive radiotherapy

Background and purpose

Magnetic resonance imaging (MRI)-Linac systems combine simultaneous MRI with radiation delivery, allowing treatments to be guided by anatomically detailed, real-time images. However, MRI can be degraded by geometric distortions that cause uncertainty between imaged and actual anatomy. In this work, we develop and integrate a real-time distortion correction method that enables accurate real-time adaptive radiotherapy.

Materials and methods

The method was based on the pre-treatment calculation of distortion and the rapid correction of intrafraction images. A motion phantom was set up in an MRI-Linac at isocentre (P0 ), the edge (P 1) and just outside (P 2) the imaging volume. The target was irradiated and tracked during real-time adaptive radiotherapy with and without the distortion correction. The geometric tracking error and latency were derived from the measurements of the beam and target positions in the EPID images.

Results

Without distortion correction, the mean geometric tracking error was 1.3 mm at P 1 and 3.1 mm at P 2. When distortion correction was applied, the error was reduced to 1.0 mm at P 1 and 1.1 mm at P 2. The corrected error was similar to an error of 0.9 mm at P0 where the target was unaffected by distortion indicating that this method has accurately accounted for distortion during tracking. The latency was 319 ± 12 ms without distortion correction and 335 ± 34 ms with distortion correction.

Conclusions

We have demonstrated a real-time distortion correction method that maintains accurate radiation delivery to the target, even at treatment locations with large distortion.

Application of an automatic, uncertainty model-guided, target-generating algorithm to lung stereotactic body radiotherapy

PURPOSE

This work evaluated a new radiotherapy target-generating framework (the αTarget algorithm) for creating internal target volumes for lung SBRT.

METHODS

Nineteen patients previously treated with definitive intent SBRT to the lung were identified from a clinical database. For each patient's 4DCT simulation scan, deformable image registration was used between phases of the scan in order to generate voxelized models of motion for 35 individual gross tumor volumes. These motion models were then used with a new implementation of a previously described target-generating algorithm to create new internal target volumes (αITVs). The resulting αITVs were analyzed with respect to their volume and the coverage they provided each tumor voxel per that voxel's motion model. The clinically used ITVs were similarly analyzed, and were then compared to the αITVs using paired Student's t-tests. In addition, isotropic margins were added to the αITVs in order to determine the largest margin magnitude that could be added without exceeding the volume of the clinical ITVs.

RESULTS

The αITVs increased the target coverage provided to each tumor's 5th-percentile-most-covered-voxel an average of 50.3% compared to the clinical ITVs (p < 0.0001). At the same time, the αITVs had volumes that were, on average, 31.4% smaller (p < 0.0001). The differences in volume were large enough that, on average, an extra 2 mm isotropic margin could be added to the αITV before it had a volume greater than the clinical ITV.

CONCLUSIONS

The αTarget algorithm can generate more effective lung SBRT internal target volumes that provide greater coverage with smaller volumes. In combination with numerous other advantages of the framework, this effectiveness makes the αTarget algorithm a powerful new method for advanced IGRT or adaptive radiotherapy techniques.

Quantification of the geometric uncertainty when using implanted markers as a surrogate for lung tumor motion

BACKGROUND

Fiducial markers are used as surrogates for tumor location during radiation therapy treatment. Developments in lung fiducial marker and implantation technology have provided a means to insert markers endobronchially for tracking of lung tumors. This study quantifies the surrogacy uncertainty (SU) when using endobronchially implanted markers as a surrogate for lung tumor position.

METHODS

We evaluated SU for 17 patients treated in a prospective electromagnetic-guided MLC tracking trial. Tumor and markers were segmented on all phases of treatment planning 4DCTs and all frames of pretreatment kilovoltage fluoroscopy acquired from lateral and frontal views. The difference in tumor and marker position relative to end-exhale position was calculated as the SU for both imaging methods and the distributions of uncertainties analyzed.

RESULTS

The mean (range) tumor motion amplitude in the 4DCT scan was 5.9 mm (1.7-11.7 mm) in the superior-inferior (SI) direction, 2.2 mm (0.9-5.5 mm) in the left-right (LR) direction, and 3.9 mm (1.2-12.9 mm) in the anterior-posterior (AP) direction. Population-based analysis indicated symmetric SU centered close to 0 mm, with maximum 5th/95th percentile values over all axes of -2.0 mm/2.1 mm with 4DCT, and -2.3/1.3 mm for fluoroscopy. There was poor correlation between the SU measured with 4DCT and that measured with fluoroscopy on a per-patient basis. We observed increasing SU with increasing surrogate motion. Based on fluoroscopy analysis, the mean (95% CI) SU was 5% (2%-8%) of the motion magnitude in the SI direction, 16% (6%-26%) of the motion magnitude in the LR direction, and 33% (23%-42%) of the motion magnitude in the AP direction. There was no dependence of SU on marker distance from the tumor.

CONCLUSION

We have quantified SU due to use of implanted markers as surrogates for lung tumor motion. Population 95th percentile range are up to 2.3 mm, indicating the approximate contribution of SU to total geometric uncertainty. SU was relatively small compared with the SI motion, but substantial compared with LR and AP motion. Due to uncertainty in estimations of patient-specific SU, it is recommended that population-based margins are used to account for this component of the total geometric uncertainty.

A Sentence Classification Framework to Identify Geometric Errors in Radiation Therapy from Relevant Literature

The objective of systematic reviews is to address a research question by summarizing relevant studies following a detailed, comprehensive, and transparent plan and search protocol to reduce bias. Systematic reviews are very useful in the biomedical and healthcare domain; however, the data extraction phase of the systematic review process necessitates substantive expertise and is labour-intensive and time-consuming. The aim of this work is to partially automate the process of building systematic radiotherapy treatment literature reviews by summarizing the required data elements of geometric errors of radiotherapy from relevant literature using machine learning and natural language processing (NLP) approaches. A framework is developed in this study that initially builds a training corpus by extracting sentences containing different types of geometric errors of radiotherapy from relevant publications. The publications are retrieved from PubMed following a given set of rules defined by a domain expert. Subsequently, the method develops a training corpus by extracting relevant sentences using a sentence similarity measure. A support vector machine (SVM) classifier is then trained on this training corpus to extract the sentences from new publications which contain relevant geometric errors. To demonstrate the proposed approach, we have used 60 publications containing geometric errors in radiotherapy to automatically extract the sentences stating the mean and standard deviation of different types of errors between planned and executed radiotherapy. The experimental results show that the recall and precision of the proposed framework are, respectively, 97% and 72%. The results clearly show that the framework is able to extract almost all sentences containing required data of geometric errors.

Dose-based optimisation for multi-leaf collimator tracking during radiation therapy

Motion in the patient anatomy causes a reduction in dose delivered to the target, while increasing dose to healthy tissue. Multi-leaf collimator (MLC) tracking has been clinically implemented to adapt dose delivery to account for intrafraction motion. Current methods shift the planned MLC aperture in the direction of motion, then optimise the new aperture based on the difference in fluence. The drawback of these methods is that 3D dose, a function of patient anatomy and MLC aperture sequence, is not properly accounted for. To overcome the drawback of current fluence-based methods, we have developed and investigated real-time adaptive MLC tracking based on dose optimisation. A novel MLC tracking algorithm, dose optimisation, has been developed which accounts for the moving patient anatomy by optimising the MLC based on the dose delivered during treatment, simulated using a simplified dose calculation algorithm. The MLC tracking with dose optimisation method was applied in silico to a prostate cancer VMAT treatment dataset with observed intrafraction motion. Its performance was compared to MLC tracking with fluence optimisation and, as a baseline, without MLC tracking. To quantitatively assess performance, we computed the dose error and 3D γ failure rate (2 mm/2%) for each fraction and method. Dose optimisation achieved a γ failure rate of (4.7 ± 1.2)% (mean and standard deviation) over all fractions, which was significantly lower than fluence optimisation (7.5 ± 2.9)% (Wilcoxon sign-rank test p < 0.01). Without MLC tracking, a γ failure rate of (15.3 ± 12.9)% was achieved. By considering the accumulation of dose in the moving anatomy during treatment, dose optimisation is able to optimise the aperture to actively target regions of underdose while avoiding overdose.