Search

PURPOSE The purpose of this study is to evaluate the prediction and classification performances of the gamma passing rate (GPR) for different machine learning models and to select the best model for achieving machine learning-based patient-specific quality assurance (PSQA). METHODS The measurement verification of 356 head-and-neck volumetric modulated arc therapy plans was performed using a diode array phantom (Delta4 Phantom), and GPR values at 2%/2 mm with global normalization and 3%/2 mm with local normalization were calculated. Machine learning models, including ridge regression (RIDGE), random forest (RF), support vector regression (SVR), and stacked generalization (STACKING), were used to predict the GPR. Each machine learning model was trained using 260 plans, and the prediction accuracy was evaluated using the remaining 96 plans. The prediction error between the measured and predicted GPR was evaluated. For the classification evaluation, the lower control limit for the measured GPR and lower control limit for predicted GPR (LCLp ) was defined to identify whether the GPR values represent a "pass" or a "fail." LCLp values with 99% and 99.9% confidence levels were calculated as the upper prediction limits for the GPR estimated from the linear regression between the measured and predicted GPR. RESULTS There was an overestimation trend of the low measured GPR. The maximum prediction errors for RIDGE, RF, SVR, and STACKING were 3.2%, 2.9%, 2.3%, and 2.2% at global 2%/2 mm and 6.3%, 6.6%, 6.1%, and 5.5% at the local 3%/2 mm, respectively. In the global 2%/2 mm, the sensitivity was 100% for all the machine learning models except RIDGE when using 99% LCLp . The specificity was 76.1% for RIDGE, RF, and SVR and 66.3% for STACKING, however, the specificity decreased dramatically when 99.9% LCLp was used. In the local 3%/2 mm, however, only STACKING showed 100% sensitivity when using 99% LCLp . The decrease in the specificity using 99.9% LCLp was smaller than that in the global 2%/2mm, and the specificity for RIDGE, RF, SVR, and STACKING was 61.3%, 61.3%, 72.0%, and 66.8%, respectively. CONCLUSIONS STACKING had better prediction accuracy for low GPR values than other machine learning models. Applying LCLp to a regression model enabled the consistent evaluation of quantitative and qualitative GPR predictions. Adjusting the confidence level of the LCLp helped improve the balance between the sensitivity and specificity. We suggest that STACKING can assist the safe and efficient operation of PSQA. This article is protected by copyright. All rights reserved.

Purpose We analyzed interfractional robustness of scanning carbon ion radiotherapy (CIRT) for prostate cancer based on the dose distribution using daily in‐room computed tomography (CT) images. Materials and Methods We analyzed 11 consecutive patients treated with scanning CIRT for localized prostate cancer in our hospital between December 2015 and January 2016. In‐room CT images were taken under treatment conditions in every treatment session. The dose distribution on each in‐room CT image was recalculated, while retaining the pencil beam arrangement of the initial treatment plan. Then, the dose–volume histogram (DVH) parameters including the percentage of the clinical target volume (CTV) with 95% and 90% of the prescribed dose area (V95% of CTV, V90% of CTV) and V80% of rectum were calculated. The acceptance criteria for the CTV and rectum were set at V95% of CTV ≥95%, V90% of CTV ≥98%, and V80% of rectum < 10 ml. Results V95% of CTV, V90% of CTV, and V80% of rectum for the reproduced plans were 98.8 ± 3.49%, 99.5 ± 2.15%, and 4.39 ± 3.96 ml, respectively. Acceptance of V95% of CTV, V90% of CTV, and V80% of rectum was obtained in 123 (94%), 125 (95%) and 117 sessions (89%), respectively. Acceptance of the mean dose of V95% of CTV, V90% of CTV, and V80% of rectum for each patient was obtained in 10 (91%), 10 (91%), and 11 patients (100%), respectively. Conclusions We demonstrated acceptable interfractional robustness based on the dose distribution in scanning CIRT for prostate cancer.

Background/aim The present study aimed to compare the radiation dose distribution of carbon-ion radiotherapy (CIRT) for stage I esophageal cancer with three-dimensional conformal radiotherapy (3DCRT) and volumetric modulated arc therapy (VMAT). Patients and methods Fifteen patients with cT1bN0M0 esophageal cancer who received 3DCRT at Kanagawa Cancer Center between January 2014 and April 2019 were enrolled. The dose-volume histogram parameters of the target volume and normal organs planned with CIRT, 3DCRT, and VMAT were evaluated. Results The homogeneity index for the target volume of CIRT was significantly lower than that of 3DCRT and VMAT. In addition, the radiation dose of CIRT to the heart, lungs, spinal cord, and skin was significantly lower than that of 3DCRT and VMAT. Conclusion Favorable dose distributions with CIRT were demonstrated compared with 3DCRT and VMAT for esophageal cancer.

Background/Aim The efficacy and safety of carbon-ion radiotherapy (CIRT) for prostate cancer have already been demonstrated. The number of hemodialysis (HD) patients is increasing. Although the toxicity of CIRT in HD patients may be more severe, it has been insufficiently investigated. Therefore, we retrospectively analyzed the safety of CIRT for HD patients with prostate cancer in the present study. Materials and Methods Five HD patients with prostate cancer who underwent CIRT at the Kanagawa Cancer Center during November 2015–2020 were included in this study. CIRT was delivered by the raster scanning method (sCIRT). Adverse events were assessed using the CTCAE v5.0. The dose-volume histogram (DVH) parameters of the target volume and normal organs were evaluated between initial planning computed tomography (CT) and in-room CT images. Results In the acute phase, Grade 1 genitourinary toxicity was recorded in 1 patient. In the late phase, Grade 1 genitourinary toxicity was recorded in 2 patients. No gastrointestinal toxicities were noted during the follow-up period. In-room CT analysis revealed no significant differences among all DVH parameters of the target volume and normal organs when compared with the treatment plan dose. Conclusions The safety of sCIRT for prostate cancer in HD patients was investigated in the present study. In-room CT analysis suggested the robustness of the treatment plan. According to the present results, sCIRT for prostate cancer can be safely performed in HD patients.

From August 2019 to August 2020, we inserted polyglycolic acid (PGA) spacers and administered carbon ion radiotherapy (CIRT) to three cases of retroperitoneal sarcoma at our hospital. We aimed to investigate its utility and safety for retroperitoneal sarcoma. We analyzed changes in PGA spacer volume and corresponding computed tomography (CT) values in addition to the dose distribution using in-room CT images that were obtained during treatment. We assessed adverse events and investigated the suitability, safety and effectivity of PGA spacer insertion. During treatment, changes in PGA spacer volumes and CT values were confirmed. Volumes increased in patients with a folded PGA spacer, and it increased 1.6-fold by the end of irradiation compared with planning CT. The CT values decreased by 20–50 Hounsfield units at the end of irradiation compared to the planning CT. Dose distribution evaluation showed that the dose to the gastrointestinal tract adjacent to the tumor was maintained below the tolerable dose, and a sufficient dose was delivered to the target by PGA spacer insertion. One case of subileus caused during abdominal surgery for PGA spacer insertion occurred. No other adverse events, such as digestive disorders, were observed. CIRT with PGA spacer insertion for retroperitoneal sarcomas is safe and effective. For cases in which there is no option but to perform irradiation using a PGA spacer, precautionary measures such as verification of dose distributions using CT images are necessary.

Background/aim The purpose of this study was to compare the dose distribution between scanning carbon-ion radiotherapy (sCIRT) and volumetric-modulated arc therapy with stereotactic body radiation therapy (VMAT-SBRT) for stage I non-small cell lung cancer (NSCLC). Patients and methods Fifteen patients with early-stage NSCLC who underwent sCIRT at Kanagawa Cancer Center between 2018-2020 were enrolled. Dose-volume histogram parameters of the planned target volume and normal organs for sCIRT and VMAT-SBRT were evaluated. Results The homogeneity index of the target volume of sCIRT was significantly lower than that of VMAT-SBRT. The dose of sCIRT was significantly lower than that of VMAT-SBRT at low volumes in the lung, heart, spinal cord, and esophagus. Conclusion The dose distribution of sCIRT for early-stage NSCLC was better than that of VMAT-SBRT.

The Kanagawa Cancer Center (KCC) is the core facility for cancer care in Kanagawa Prefecture. It serves a population of more than 9 million people. Planning for the ionbeam Radiation Oncology Center in Kanagawa (i-ROCK) started in 2015. The basic framework and design of i-ROCK was established in 2010, and it is the first cancercenter-based carbon ion radiation therapy facility in the world and the fifth carbon ion radiation therapy facility in Japan. The building was completed in August 2014 (Figure 1). The installation of beam delivery equipment has been completed. The start of clinical operation is planned for December 2015. An overview of i-ROCK is presented in this report.