Search

Background and purpose: Proton therapy is expected to outperform photon-based treatment regarding organs at risk (OAR) sparing but to date there is no method to practically measure clinical benefit. Here, we introduce the novel ROCOCO Performance Scoring System (RPSS) translating dose differences into clinically relevant endpoints and apply this to a treatment plan comparison of volumetric modulated arc therapy (VMAT) and intensity modulated proton therapy (IMPT) in 20 pilocytic astrocytoma patients. Material and methods: The RPSS was developed on the basis of expert-based weighting factors and toxicity scores per OAR. The imaging datasets of 20 pilocytic astrocytoma patients having undergone radiotherapy were included in this in silico dosimetric comparison trial as proof of principle. For each of these patients, treatment plans to a total dose of 54 Gy (RBE) were generated for VMAT and IMPT and these were compared regarding radiation dose to the clinical target volume (CTV) and OARs. The RPSS was calculated for each treatment plan comparing VMAT and IMPT. Results: In 40 analysed treatment plans, the average and low dose volumes to various OARs were significantly reduced when using IMPT compared to VMAT (p < 0.05). Using the RPSS, a significant difference between both treatment modalities was found, with 85% of the patients having a lower RPSS in favour of the IMPT plan. Conclusion: There are dosimetric differences between IMPT and VMAT in pilocytic astrocytoma patients. In absence of clinically validated NTCP models we introduce the RPSS model in order to objectively compare treatment modalities by translating dosimetric differences in potential clinical differences.

Introduction We present results of the worldwide first systematic study on the sensitivity of prompt-gamma-imaging (PGI) to detect anatomical changes in proton therapy for the ongoing evaluation in prostate-cancer treatments. Materials&Methods Spot-wise range shifts were monitored with a PGI-slit-camera during 40 fractions of hypo-fractionated prostate-cancer treatments (5 patients, 2 fields, each 1.5GyE). In-room CTs were acquired for these fractions and range shifts of spot-wise integrated depth-dose (IDD) profiles serve as ground-truth. For both PGI and IDD data, spots were clustered based on Bragg-peak position and proton number to mitigate statistical uncertainty in the PGI measurement using a low-dose spot cut-off at 5e7 protons, a minimum number of 3e9 protons per cluster, and a minimum/maximum cluster volume of 1cm3/8cm3. Clusters with absolute range shift ≥5mm were classified as relevant anatomical changes. Results A strong correlation (rPearson=0.72) was found between ground-truth IDD and PGI range shifts per cluster with an average absolute deviation of 1.3mm over all fractions. In total, 245/7143 (3.4%) clusters (found within 24/72 fields) contained relevant IDD-based range shifts. PGI detected these changes with a sensitivity of 68%, specificity of 96%, and accuracy of 95%. The results might be affected by potential intra-fractional changes between in-room CT acquisition and treatment delivery. A higher sensitivity is also expected for a gantry-mounted camera system with decreased positioning uncertainty. Conclusion Our systematic investigation on the sensitivity of a PGI-slit-camera with a first quantitative comparison of range shifts from PGI and IDD profiles demonstrates the capability to locally detect relevant anatomical changes in patients.

Introduction Within a clinical study, we investigate the potential benefit of prompt-gamma-imaging (PGI) based range verification in proton therapy. As the manual interpretation of detected spot-wise range-shift information is time-consuming and complex, we aim to automatically detect and classify treatment deviations from realistic PGI data using convolutional neural networks (CNNs). Materials & Methods For 12 head-and-neck cancer patients and an anthropomorphic head phantom, monitoring of single fields from pencil-beam-scanning plans with the IBA slit camera was considered. In total, 386 treatment deviations were simulated on planning and control CTs and manually classified into 7 classes: non-relevant changes (NRC) and relevant changes triggering treatment intervention due to range-prediction errors (±RPE), setup errors in beam direction (±SE), anatomical changes (AC), or a combination of such errors (CE). The spatial maps of filtered PGI-determined range deviations were converted to 16x16x16 voxel grids. Three complexity levels were investigated using 3D-CNNs [training cohort (n=9), test cohort (n=4), Fig.1]: (A) optimal PGI data, (B) realistic PGI data with simulated Poisson noise based on the locally delivered proton number, (C) realistic PGI data with additional positioning uncertainty of the slit camera. Results During validation on the independent test data, the 3D-CNNs achieved multi-class accuracies of 81%, 77%, 76% and binary accuracies of 97%, 95%, 93% for the respective complexity levels (A,B,C) (Fig.2). In the most realistic scenario (C), relevant treatment deviations were detected with 97% sensitivity and 82% specificity. Misclassifications of the AC class were caused by similar PGI characteristics of the CE class. Conclusion CNNs can reliably detect and classify relevant treatment deviations from realistically simulated PGI data. While validation on measured patient data is needed, our study highlights the potential of automated PGI interpretation, which is desired for broad clinical application and a prerequisite for including PGI in an automated feedback loop for online adaptation.

Purpose & Objective Prompt-gamma imaging (PGI) based range verification has been utilized in first pencil-beam scanning (PBS) proton therapy treatments and is under systematic investigation concerning its potential benefit in a clinical study at our institution. Manual interpretation of the detected spot-wise range shift information is time-consuming, highly complex, and therefore not feasible in a broad routine application. Here, we present an approach to automatically detect and classify treatment deviations in realistically simulated PGI data for head and neck cancer treatments using convolutional neural networks (CNNs). Materials & Methods For 12 patients and an anthropomorphic head phantom, PBS treatment plans were generated and one field per plan was assumed to be monitored with the IBA slit camera. In total, 386 scenarios resembling different relevant or non-relevant treatment deviations were simulated on planning and control CTs and manually classified into 7 classes: non-relevant changes (NRC) and relevant changes triggering treatment intervention due to range prediction errors (±RPE), setup errors in beam direction (±SE), anatomical changes (AC), or a combination of such errors (CE). After filtering of PBS spots with reliable PGI information, the 3D spatial maps of PGI-determined range deviations (reference vs. change scenario) were converted to 16x16x16 voxel grids. Three complexity levels of simulated PGI data were investigated: (A) optimal PGI data, (B) realistic PGI data with simulated Poisson noise based on the locally delivered proton number, (C) realistic PGI data with an additional positioning uncertainty of the slit camera following an experimentally determined distribution. For each complexity level, 3D-CNNs (6 convolutional & 2 downsampling layers) were trained on a subset of 8 patients and the phantom dataset using patient-specific leave-one-out cross-validation and tested on an independent test cohort of 4 patients. Results On the test data, the CNN ensemble achieved an accuracy of 0.81, 0.77, and 0.76 for the complexity levels (A), (B), and (C), respectively. Similarly, for the task to solely differentiate relevant from non-relevant changes, the binary accuracy was 0.97, 0.95, and 0.93. The trained ensemble provided fast (

Across organisations data exchange is a topic of increasing interest. Blockchain technology is a promising technology for this management problem. Understanding the technology requires an understanding of the vocabulary related to blockchain systems and blockchain technology. The aim of this paper is to develop vocabulary and definitions for blockchain system, blockchain system type, blockchain system functional characteristics, blockchain subtype, blockchain system non-functional characteristics, blockchain system application area and governance of blockchain systems. This research is done by a systematic literature review and a systematic mapping study to identify vocabulary for defining a proposed conceptual framework for blockchain systems. None of the definitions have yet been used. As such, more work has to be done before actual implementations of putting such technology in a real life setting.

Purpose & Objective Currently, the uncertainty in CT-based range prediction is substantially impairing the accuracy of particle therapy. Direct determination of stopping-power ratio (SPR) from dual-energy CT (DECT) has been proposed (DirectSPR) and initial validation studies in phantoms and biological tissues have proven a superior accuracy. However, a validation of range prediction in patients has not been achieved by any means. Here, we present the first verification of CT based proton range prediction in patients, using prompt-gamma imaging (PGI). Materials & Methods A PGI slit camera system of improved positioning accuracy, using a floor-based docking station, was developed. Its accuracy and positioning reproducibility were determined with x-ray and PGI measurements. The PGI system was clinically applied to monitor absolute proton ranges for a 1.5 GyE field during hypo-fractionated treatment of 3 prostate-cancer patients using pencil beam scanning (PBS) (Fig. 1). Per patient 3 fractions were monitored. For all monitored fractions, in-room control-CT (cCT) scans were acquired in treatment position enabling PGI-based spot-by-spot range analysis for the actual patient anatomy: The PGI measurements were compared to simulations of the expected PGI signal based on the respective cCT. Three different SPR prediction models were applied in the simulation: A standard CT-number-to-SPR conversion (Std-HLUT), a HLUT optimized with DECT-derived SPR information (Adapt-HLUT), and the directly voxel-wise calculated SPR based on the input from DECT (DirectSPR). To verify range prediction in patients, the histogram of PGI-derived range shifts from all PBS spots was analyzed concerning its Gaussian mean – acting as surrogate for the accuracy of the respective range prediction method. It is independent from random uncertainty contributions (e.g. positioning, statistical uncertainty in shift determination). Results The accuracy and precision for global PGI range verification (averaging over multiple spots) was determined to be 0.7 mm (2σ) and 1.3 mm (2σ), respectively. The precision is limited by remaining uncertainties in image registration and positioning reproducibility (1.1 mm, 2σ). Hence, the absolute verification uncertainty of the cumulative mean shift (for 9 monitored fractions) is 0.8 mm (2σ), which is smaller than the range prediction uncertainty for deep-seated tumors (about 10 mm for prostate treatments). The comparison of the PGI-measured and predicted spot-wise ranges for in total 12000 PBS spots from the 9 analyzed fractions resulted in an range prediction offset of 0.6 mm, 1.3 mm and 4.4 mm, for the DirectSPR, Adapt-HLUT and Std-HLUT approaches, respectively. Conclusion The accuracy of PGI-based range verification was improved to enable the worldwide first in-man validation of CT-based stopping-power prediction. The evaluation of the first clinical PGI data for prostate-cancer treatments, systematically acquired within a clinical study, confirms the superiority of DECT-based range prediction in patients.

Purpose & Objective Anatomical changes during the course of proton therapy treatment can result in relevant changes in proton range, potentially causing severe under- or overdosage. Verifying the proton treatment, ideally in real-time, is thus highly desirable. Here, we present the first systematic evaluation of the sensitivity of a prompt-gamma-imaging (PGI) based range verification system to detect anatomical changes in prostate-cancer treatments. Materials & Methods A PGI slit-camera system was clinically applied to monitor spot-wise proton range deviations during 7 and 9 fractions of hypo-fractionated pencil beam scanning (PBS) treatment for 2 prostate-cancer patients, respectively (2 opposing fields, 1.5 GyE each). For all monitored fractions, in-room control CT scans (cCT) were acquired in treatment position, serving as ground-truth reference. Based on the evaluation of planning CT (pCT) and cCT data on the level of CT images, dose distributions and derived line-dose profiles, anatomical changes were identified and scored concerning cause and magnitude. The detectability of these changes with PGI was determined by manually comparing expected range shifts from line-dose profiles (pCT vs. cCT) with PGI-derived spot-wise range shifts for distal PBS spots (Fig.1). This evaluation was performed for both, measured as well as simulated PGI data based on cCT (no statistical uncertainty). Furthermore, the sensitivity for a binary differentiation between relevant (strong/moderate) and no relevant anatomical changes within a fraction was determined. Working towards an automated classification of treatment deviations for real-time treatment verification, a simple two-parametric model was established to classify each monitored field into global, local and not clinically relevant anatomical changes. Results From 64 detected anatomical changes in 32 monitored treatment fields, in total 66% (84%) were also identified by measured (simulated) PGI data (Fig.2a). All strong changes (14/64) were identified correctly. For the differentiation between relevant from non-relevant changes, a sensitivity of 69% (95%) was achieved for measured (simulated) PGI data. The first attempt for automated classification was able to reliably differentiate global from local changes (Fig.2b). However, it was more difficult to distinguish treatments with no relevant from local anatomical changes. Also in the ground-truth classification, this decision was sometimes also borderline. Conclusion In the first systematic investigation of the sensitivity of PGI-based treatment verification in clinical prostate-cancer treatments, its capability to detect strong anatomical changes has been clearly demonstrated. In clinical PGI application the sensitivity is a bit smaller than for idealized PGI simulations, still severe changes were detected for all cases. The next step is to establish a reliable automated interpretation of PGI data. In a first trial, we established a two-parametric prediction model with already encouraging results.

Introduction: Anatomical changes during proton therapy can cause severe dosimetric deviation. Treatment verification is thus highly desirable. Here, we present the first systematic evaluation of the sensitivity of a Prompt-Gamma-Imaging (PGI) based range verification system to detect anatomical changes in prostate-cancer treatments. Materials and Methods: Spot-wise range deviations were monitored with a PGI slit camera during in total 16 fractions of hypo-fractionated Pencil-Beam-Scanning (PBS) prostate-cancer treatments (2 patients, 2 fields, each 1.5GyE). For all monitored fractions, in-room control-CT scans were acquired, serving as ground-truth reference for the identification and scoring of anatomical changes (strong/moderate/light). The sensitivity to detect these changes was determined for both, clinically measured and simulated PGI-data, respectively: For distal PBS spots, expected shifts, determined from line-dose profiles (planning-CT vs. control-CT), were manually compared with PGI-derived spot-wise shifts (Fig.1). Furthermore, a simple two-parametric model was established to classify each monitored field into scenarios of global, local and no-clinically-relevant anatomical changes. Results: Overall 66% (84%) of the 64 detected anatomical changes were identified from measured (simulated) PGI-data (Fig.2a). All strong changes (14/64) were identified correctly. The first attempt for automated field-wise classification was able to correctly classify most global changes (9/11). However, differentiation between non-relevant from local changes seemed more difficult (4/6 and 7/14 fields classified correctly, respectively); but even ground-truth classification was often borderline in those cases (Fig.2b). Conclusion: In the first systematic investigation of the sensitivity of clinical PGI-based treatment verification, its capability to detect strong anatomical changes has been clearly demonstrated. Moving towards automated interpretation of PGI-data, a simple two-parametric model already showed encouraging results.