Deep Learning-Guided Dosimetry for Mitigating Local Failure of Patients With Non-Small Cell Lung Cancer Receiving Stereotactic Body Radiation Therapy.

PURPOSE: Non-small cell lung cancer (NSCLC) stereotactic body radiation therapy with 50 Gy/5 fractions is sometimes considered controversial, as the nominal biologically effective dose (BED) of 100 Gy is felt by some to be insufficient for long-term local control of some lesions. In this study, we analyzed such patients using explainable deep learning techniques and consequently proposed appropriate treatment planning criteria. These novel criteria could help planners achieve optimized treatment plans for maximal local control. METHODS AND MATERIALS: A total of 535 patients treated with 50 Gy/5 fractions were used to develop a novel deep learning local response model. A multimodality approach, incorporating computed tomography images, 3-dimensional dose distribution, and patient demographics, combined with a discrete-time survival model, was applied to predict time to failure and the probability of local control. Subsequently, an integrated gradient-weighted class activation mapping method was used to identify the most significant dose-volume metrics predictive of local failure and their optimal cut-points. RESULTS: The model was cross-validated, showing an acceptable performance (c-index: 0.72, 95% CI, 0.68-0.75); the testing c-index was 0.69. The model's spatial attention was concentrated mostly in the tumors' periphery (planning target volume [PTV] - internal gross target volume [IGTV]) region. Statistically significant dose-volume metrics in improved local control were BED Dnear-min ≥ 103.8 Gy in IGTV (hazard ratio [HR], 0.31; 95% CI, 015-0.63), V104 ≥ 98% in IGTV (HR, 0.30; 95% CI, 0.15-0.60), gEUD ≥ 103.8 Gy in PTV-IGTV (HR, 0.25; 95% CI, 0.12-0.50), and Dmean ≥ 104.5 Gy in PTV-IGTV (HR, 0.25; 95% CI, 0.12-0.51). CONCLUSIONS: Deep learning-identified dose-volume metrics have shown significant prognostic power (log-rank, P = .003) and could be used as additional actionable criteria for treatment planning in NSCLC stereotactic body radiation therapy patients receiving 50 Gy in 5 fractions. Although our data do not confirm or refute that a significantly higher BED for the prescription dose is necessary for tumor control in NSCLC, it might be clinically effective to escalate the nominal prescribed dose from BED 100 to 105 Gy.

Stereotactic Ablative Radiotherapy of Ventricular Tachycardia Using Tracking: Optimized Target Definition Workflow.

Background and Purpose: Stereotactic arrhythmia radioablation (STAR) has been suggested as a promising therapeutic alternative in cases of failed catheter ablation for recurrent ventricular tachycardias in patients with structural heart disease. Cyberknife® robotic radiosurgery system utilizing target tracking technology is one of the available STAR treatment platforms. Tracking using implantable cardioverter-defibrillator lead tip as target surrogate marker is affected by the deformation of marker-target geometry. A simple method to account for the deformation in the target definition process is proposed. Methods: Radiotherapy planning CT series include scans at expiration and inspiration breath hold, and three free-breathing scans. All secondary series are triple registered to the primary CT: 6D/spine + 3D translation/marker + 3D translation/target surrogate-a heterogeneous structure around the left main coronary artery. The 3D translation difference between the last two registrations reflects the deformation between the marker and the target (surrogate) for the respective respiratory phase. Maximum translation differences in each direction form an anisotropic geometry deformation margin (GDM) to expand the initial single-phase clinical target volume (CTV) to create an internal target volume (ITV) in the dynamic coordinates of the marker. Alternative GDM-based target volumes were created for seven recent STAR patients and compared to the original treated planning target volumes (PTVs) as well as to analogical volumes created using deformable image registration (DIR) by MIM® and Velocity® software. Intra- and inter-observer variabilities of the triple registration process were tested as components of the final ITV to PTV margin. Results: A margin of 2 mm has been found to cover the image registration observer variability. GDM-based target volumes are larger and shifted toward the inspiration phase relative to the original clinical volumes based on a 3-mm isotropic margin without deformation consideration. GDM-based targets are similar (mean DICE similarity coefficient range 0.80-0.87) to their equivalents based on the DIR of the primary target volume delineated by dedicated software. Conclusion: The proposed GDM method is a simple way to account for marker-target deformation-related uncertainty for tracking with Cyberknife® and better control of the risk of target underdose. The principle applies to general radiotherapy as well.

The use of Pantherpix pixel detector in radiotherapy QA.

There are various different detectors, which can be used for radiotherapy measurements, and more are about to be adopted. Hybrid pixel detectors (HPD) have been originally developed for the high energy physics. However, over the last few years they also expanded in the medical physics. Novel 2D detector Pantherpix is a HPD designed specifically for the radiotherapy. In this article, its properties are characterised and an assessment of its use in radiotherapy photon beams is provided. Properties such as response stability, response linearity, angular dependence and energy dependence were studied. In order to prove sufficient clinical quality for relative dosimetry, further measurements were undertaken (i.e. dose profiles and collimator scatter factors). Acquired results were compared with ion chamber and gafchromic film results. Namely the applicability of PhPix for cobalt beam therapy, which is still widely used (and will be used in near future) in economically less developed countries, is considered.