Search

Radiation doses to the heart are potentially high in patients undergoing radiotherapy for thymoma or thymic carcinoma because of their origin site and propensity for pericardial invasion. We investigated potential relationships between radiation pneumonitis (RP) and the dosimetric parameters of lung and heart substructures in patients with thymic epithelial tumors. This retrospective study included 70 consecutive patients who received definitive or postoperative radiotherapy at a median dose of 58.3 Gy. Heart substructures were delineated according to a published atlas. The primary end point of ≥ grade 2 RP was observed in 13 patients (19%) despite a low lung dose; median lung V20 (i.e. percentage of the volume receiving at least 20 Gy) was only 16.6%. In a univariate analysis, four lung parameters, heart V35, three pulmonary artery (PA) parameters, two left ventricle parameters, and left atrium V35 were associated with the development of RP. In a multivariate analysis, only PA V35 remained significant (hazard ratio 1.04; 95% CI 1.01–1.07, p = 0.007). PA V35 of the RP versus non-RP groups were 84.2% versus 60.0% (p = 0.003). The moderate dose sparing of PA could be a candidate as a planning constraint for reducing the risk of RP in thoracic radiotherapy.

Embedded systems such as self-driving systems require a computing platform with high computing power and low power consumption. Multi-/many-core platforms satisfy exactly these requirements. However, for hard real-time applications, multiple demands on shared resources can hinder real-time performance. Memory is among the resources that can most dramatically impair the desired performance. Therefore, we addressed contentions induced by the shared memory. We improve the predictability of contentions by dividing tasks into the memory access phase and the execution phase using a Directed Acyclic Graph (DAG). Existing methods are able to make accurate contention estimations for one Compute Cluster (CC) of a Clustered many-core processor. Our method is able to do the same for multiple CCs, thereby doubling the scalability in consideration of contentions. Using an Integer Linear Programming (ILP) formulation, we produced a static, non-preemptive, partitioned, time-triggered schedule. We also conducted an experiment in order to minimize the makespan. The evaluation confirmed that our new method reduced the makespan by increasing the number of CCs.

The safety and efficacy of dose-escalated radiotherapy with intensity-modulated radiotherapy (IMRT) and image-guided radiotherapy (IGRT) remain unclear in salvage radiotherapy (SRT) after radical prostatectomy. We examined the impact of these advanced radiotherapy techniques and dose intensification on the toxicity of SRT. This multi-institutional retrospective study included 421 patients who underwent SRT at the median dose of 66 Gy in 2-Gy fractions. IMRT and IGRT were used for 225 (53%) and 321 (76%) patients, respectively. At the median follow-up of 50 months, the cumulative incidence of late grade 2 or higher gastrointestinal (GI) and genitourinary (GU) toxicities was 4.8% and 24%, respectively. Multivariate analysis revealed that the non-use of either IMRT or IGRT, or both (hazard ratio [HR] 3.1, 95% confidence interval [CI] 1.8–5.4, p p = 0.048) were associated with late GI toxicity, whereas a higher dose ≥68 Gy was the only factor associated with GU toxicities (HR 3.1, CI 1.3–7.4, p = 0.012). This study suggested that the incidence of GI toxicities can be reduced by IMRT and IGRT in SRT, whereas dose intensification may increase GU toxicity even with these advanced techniques.

Computing platforms for embedded systems are increasingly being transformed into multi/many-core platforms because embedded systems have become extensive, complex, and automated. In the case of an autonomous driving system, various applications are simultaneously running, and low power consumption and large-scale calculation are required. Many-core processors with a multiple instruction, multiple data (MIMD) architecture can meet these requirements. This paper proposes a scheduling algorithm for an automotive driving system expressed in a directed acyclic graph (DAG) and we use Kalray MPPA-256 as the target many-core processor. On the basis of the architecture of Kalray MPPA-256, task processing that requires large-scale calculation and intercore communication is performed while avoiding communication contention by using a proposed grouping computational resource. In addition, we propose a scheduling method for a multi-rate DAG which is a DAG with multiple periods. This method generates a DAG task in a hyperperiod and schedules the DAG with dependency on tasks that have been released closely. The formulas for prioritization and processor selection are proposed for various generated tasks in a hyperperiod. Evaluation results show that the proposed algorithm is superior to existing DAG scheduling algorithms with regard to schedulability and deadline miss ratio.

MR-only simulations provide pseudo-CT images which are segmented into 5 kinds of tissues from DIXON-based images. However, it is difficult to register pseudo-CT images to cone-beam CT (CBCT) images collected for image-guided radiation therapy (IGRT), because of the lack of contrasts among tissues. We validated gaps of IGRT between pseudo-CT or planning CT and CBCT for patients without implanted markers. We also propose calcification-assisted registration for MR-only simulation. We conducted retrospective analyses to verify the registration accuracy in 15 patients who underwent volumetric modulated arc therapy (VMAT) for prostate cancer. They underwent planning CT and pseudo-CT. Pseudo-CT images after deformable image registration (DIR) to planning CT images were rendered automatic pelvic bone matching to CBCT images. Patient positions on the pseudo-CT images after DIR were shifted on the basis of tissues around the prostate. We compared registration gaps between the images of planning CT and pseudo-CT with DIR, assuming that the tissue-based matching between the planning CT and CBCT was the gold standard. To the pseudo-CT images with DIR, calcifications detected on planning CT were added. We validated IGRT accuracy for a calcification-assisted registration. The absolute registration errors of the pseudo-CT, in comparison with the planning CT, were 0.34 ± 0.50 (lateral), 1.3 ± 1.3 (longitudinal), and 1.1 ± 1.0 mm (vertical). The absolute registration errors of the pseudo-CT with calcification contouring, in comparison with the planning CT, were 0.41 ± 1.0 (lateral), 0.87 ± 0.92 (longitudinal), and 0.74 ± 0.64 mm (vertical). Reduced absolute registration errors were observed in the proposed approach in the longitudinal (P

This paper proposes an anomaly prediction method based on k-means clustering that assumes embedded devices with memory constraints to predict control system anomalies. With this method, by checking control system behavior, it is possible to predict anomalies. However, continuing clustering is difficult because data accumulate in memory similar to existing k-means clustering method, which is problematic for embedded devices with low memory capacity. Therefore, we also propose k-means clustering to continue clustering for infinite stream data. The proposed k-means clustering method is based on online k-means clustering of sequential processing. The proposed k-means clustering method only stores data required for anomaly prediction and releases other data from memory. Experimental results show that anomalies can be predicted by k-means clustering, and the proposed method can predict anomalies similar to standard k-means clustering while reducing memory consumption. Moreover, the proposed k-means clustering demonstrates better results of anomaly prediction than existing online k-means clustering.