Clinical association between plan complexity and the local-recurrence-free-survival of non-small-cell lung cancer patients receiving stereotactic body radiation therapy

PURPOSE

To investigate the clinical impact of plan complexity on the local recurrence-free survival (LRFS) of non-small cell lung cancer (NSCLC) patients treated with stereotactic body radiation therapy (SBRT).

METHODS

Data from 123 treatment plans for 113 NSCLC patients were analyzed. Plan-averaged beam modulation (PM), plan beam irregularity (PI), monitor unit/Gy (MU/Gy) and spherical disproportion (SD) were calculated. The γ passing rates (GPR) were measured using ArcCHECK 3D phantom with 2 %/2mm criteria. High complexity (HC) and low complexity (LC) groups were statistically stratified based on the aforementioned metrics, using cutoffs determined by their significance in correlation with survival time, as calculated using the R-3.6.1 packages. Kaplan-Meier analysis, Cox regression, and Random Survival Forest (RSF) models were employed for the analysis of local recurrence-free survival (LRFS). Propensity-score-matched pairs were generated to minimize bias in the analysis.

RESULTS

The median follow-up time for all patients was 25.5 months (interquartile range 13.4-41.2). The prognostic capacity of PM was suggested using RSF, based on Variable Importance and Minimal Depth methods. The 1-, 2-, and 3-year LRFS rates in the HC group were significantly lower than those in the LC group (p = 0.023), when plan complexity was defined by PM. However, no significant difference was observed between the HC and LC groups when defined by other metrics (p > 0.05). All γ passing rates exceeded 90.5 %.

CONCLUSIONS

This study revealed a significant association between higher PM and worse LRFS in NSCLC patients treated with SBRT. This finding offers additional clinical evidence supporting the potential optimization of pre-treatment quality assurance protocols.

Towards Automation in Radiotherapy Planning: A Deep Learning Approach for the Delineation of Parotid Glands in Head and Neck Cancer

The delineation of parotid glands in head and neck (HN) carcinoma is critical to assess radiotherapy (RT) planning. Segmentation processes ensure precise target position and treatment precision, facilitate monitoring of anatomical changes, enable plan adaptation, and enhance overall patient safety. In this context, artificial intelligence (AI) and deep learning (DL) have proven exceedingly effective in precisely outlining tumor tissues and, by extension, the organs at risk. This paper introduces a DL framework using the AttentionUNet neural network for automatic parotid gland segmentation in HN cancer. Extensive evaluation of the model is performed in two public and one private dataset, while segmentation accuracy is compared with other state-of-the-art DL segmentation schemas. To assess replanning necessity during treatment, an additional registration method is implemented on the segmentation output, aligning images of different modalities (Computed Tomography (CT) and Cone Beam CT (CBCT)). AttentionUNet outperforms similar DL methods (Dice Similarity Coefficient: 82.65% ± 1.03, Hausdorff Distance: 6.24 mm ± 2.47), confirming its effectiveness. Moreover, the subsequent registration procedure displays increased similarity, providing insights into the effects of RT procedures for treatment planning adaptations. The implementation of the proposed methods indicates the effectiveness of DL not only for automatic delineation of the anatomical structures, but also for the provision of information for adaptive RT support.

A management method for the statistical results of patient-specific quality assurance for intensity-modulated radiation therapy

There are many reports concerning patient-specific quality assurance (QA) for intensity-modulated radiation therapy (IMRT). However, reports about the statistical results of QA are lacking. Management methods for the results of the QA are needed, even though we have the ESTRO group recommendation that a tolerance limit of 1.96 standard deviation (SD) be established in each institution. The purpose of this study was to establish a management method for determining the tolerance limit and to report the statistical results of patient-specific QA. From April 2006 to March 2015, five linacs in the National Cancer Center, Tokyo, Japan, were used to treat 1185 patients with IMRT. Patient-specific QA was performed using an ion chamber, films, and some detectors. To establish a management method for the results, differences between the measured and calculated doses in the ion chamber were analyzed for each linac, each phantom, and each treatment site. The overall mean dose difference was 0.5 ± 1.3%, and the mean dose difference in each linac was 0.6 ± 1.2%, 0.9 ± 1.3%, -0.4 ± 1.4%, -0.1 ± 1.2% and -0.1 ± 0.9%. The difference between linacs and between treatment sites was significant (P < 0.001 and 0.01, respectively). The proportion of the dose difference within ±3% was 97.7%, and that was improved from 2006 to 2014. The results of the patient-specific QA should be managed for each linac and each treatment site in order to decide the suitable tolerance limit. Reports of statistical results will be helped if a new tolerance limit and action level will be considered.

A machine learning approach to the accurate prediction of multi-leaf collimator positional errors

Discrepancies between planned and delivered movements of multi-leaf collimators (MLCs) are an important source of errors in dose distributions during radiotherapy. In this work we used machine learning techniques to train models to predict these discrepancies, assessed the accuracy of the model predictions, and examined the impact these errors have on quality assurance (QA) procedures and dosimetry. Predictive leaf motion parameters for the models were calculated from the plan files, such as leaf position and velocity, whether the leaf was moving towards or away from the isocenter of the MLC, and many others. Differences in positions between synchronized DICOM-RT planning files and DynaLog files reported during QA delivery were used as a target response for training of the models. The final model is capable of predicting MLC positions during delivery to a high degree of accuracy. For moving MLC leaves, predicted positions were shown to be significantly closer to delivered positions than were planned positions. By incorporating predicted positions into dose calculations in the TPS, increases were shown in gamma passing rates against measured dose distributions recorded during QA delivery. For instance, head and neck plans with 1%/2 mm gamma criteria had an average increase in passing rate of 4.17% (SD = 1.54%). This indicates that the inclusion of predictions during dose calculation leads to a more realistic representation of plan delivery. To assess impact on the patient, dose volumetric histograms (DVH) using delivered positions were calculated for comparison with planned and predicted DVHs. In all cases, predicted dose volumetric parameters were in closer agreement to the delivered parameters than were the planned parameters, particularly for organs at risk on the periphery of the treatment area. By incorporating the predicted positions into the TPS, the treatment planner is given a more realistic view of the dose distribution as it will truly be delivered to the patient.