BackgroundAccurate and standardized descriptions of organs at risk (OARs) are essential in radiation therapy for treatment planning and evaluation. Traditionally, physicians have contoured patient images manually, which, is time-consuming and subject to inter-observer variability.This study aims to a) investigate whether customized, deep-learning-based auto-segmentation could overcome the limitations of manual contouring and b) compare its performance against a typical, atlas-based auto-segmentation method organ structures in liver cancer.MethodsOn-contrast computer tomography image sets of 70 liver cancer patients were used, and four OARs (heart, liver, kidney, and stomach) were manually delineated by three experienced physicians as reference structures. Atlas and deep learning auto-segmentations were respectively performed with MIM Maestro 6.5 (MIM Software Inc., Cleveland, OH) and, with a deep convolution neural network (DCNN). The Hausdorff distance (HD) and, dice similarity coefficient (DSC), volume overlap error (VOE), and relative volume difference (RVD) were used to quantitatively evaluate the four different methods in the case of the reference set of the four OAR structures.ResultsThe atlas-based method yielded the following average DSC and standard deviation values (SD) for the heart, liver, right kidney, left kidney, and stomach: 0.92 ± 0.04 (DSC ± SD), 0.93 ± 0.02, 0.86 ± 0.07, 0.85 ± 0.11, and 0.60 ± 0.13 respectively. The deep-learning-based method yielded corresponding values for the OARs of 0.94 ± 0.01, 0.93 ± 0.01, 0.88 ± 0.03, 0.86 ± 0.03, and 0.73 ± 0.09. The segmentation results show that the deep learning framework is superior to the atlas-based framwork except in the case of the liver. Specifically, in the case of the stomach, the DSC, VOE, and RVD showed a maximum difference of 21.67, 25.11, 28.80% respectively.ConclusionsIn this study, we demonstrated that a deep learning framework could be used more effectively and efficiently compared to atlas-based auto-segmentation for most OARs in human liver cancer. Extended use of the deep-learning-based framework is anticipated for auto-segmentations of other body sites.

Background: Proton beam has an excellent depth dose distribution due to its unique physical properties, and thus proton beam therapy (PBT) has been tried and showed promising outcomes in hepatocellular carcinoma (HCC). The purpose of this phase II study is to evaluate the efficacy of hypofractionated PBT in HCC. Methods:The eligibility criteria for this study were as follows: patients with HCC lesion(s) who were failed after, were difficult to treat with, or refused to other local treatments; tumor size and number of ≤7 and ≤2 cm, respectively, and HCC lesion(s) of ≥2 cm from gastrointestinal organs; Child-Pugh score of ≤7; Eastern Cooperative Oncology Group performance status ≤1; and age ≥18 years. The prescribed dose of PBT was 70 Gy equivalent in 10 fractions. The primary endpoint was 3-year local progression-free survival (LPFS) rate.Results: Forty-five patients were prospectively enrolled, and there were 35 men and 10 women with a median age of 63 years (range, 46-78 years). Thirty-seven patients had recurrent and/or residual disease, and eight patients had treatment-naive disease. All patients received the planned treatments without treatment interruption, and grade ≥3 acute toxicity did not occur. The median follow-up duration was 35.1 months (range, 11.2-56.3 months) and local progression occurred in two patients (8.7%). The 3-year rates of LPFS and overall survival (OS) were 95.2% (95% confidence interval [CI], 89.1%−100%) and 86.4% (95% CI, 72.9-99.9%), respectively. Conclusion:Hypofractionated PBT showed promising LPFS and OS, and further studies are warranted to compare PBT with other local modalities.

To evaluate the clinical efficacy and feasibility of proton beam radiotherapy (PBT) using the simultaneous integrated boost (SIB) technique in locally advanced pancreatic cancer (LAPC), 81 LAPC patients receiving PBT using SIB technique were analyzed. The prescribed doses to planning target volume (PTV)1 and PTV2 were 45 or 50 GyE and 30 GyE in 10 fractions, respectively. Of 81 patients, 18 patients received PBT without upfront and maintenance chemotherapy (group I), 44 received PBT followed by maintenance chemotherapy (group II), and 19 received PBT after upfront chemotherapy followed by maintenance chemotherapy (n = 16) (group III). The median follow-up time was 19.6 months (range 2.3–57.6 months), and the median overall survival (OS) times of all patients and of those in groups I, II, and III were 19.3 months (95% confidence interval [CI] 16.8–21.7 months), 15.3 months (95% CI 12.9–17.7 months), 18.3 months (95% CI 15.9–20.7 months), and 26.1 months (95% CI 17.8–34.3 months), respectively (p = 0.043). Acute and late grade ≥ 3 toxicities related to PBT were not observed. PBT with the SIB technique showed promising OS for LAPC patients with a safe toxicity profile, and intensive combinations of PBT and chemotherapy could improve OS in these patients.

To evaluate the benefit of adjuvant treatments, such as chemoradiotherapy (CRT) and chemotherapy (CTx), compared with no adjuvant treatment (No-AT) in resected gallbladder (GB) cancer patients, 151 patients were analyzed: 98 (64.9%) patients received adjuvant treatment with CRT (n = 59, 39.1%) or CTx (n = 39, 25.8%), and the remaining 53 (35.1%) did not (No-AT). The clinicopathological factors, patterns of failure, locoregional recurrence-free survival (LRFS), recurrence-free survival (RFS) and overall survival (OS) were compared among the three groups according to tumor stage. In patients with T2-3N0M0 stage disease, the incidences of locoregional recurrence and distant recurrence and 5-year LRFS, RFS and OS rates were not significantly different among the No-AT, CTx, and CRT groups ( p > 0.05 each). In those with T2-3N1-2M0 stage disease, the incidences of locoregional recurrence (11.4%, 78.1%, and 68.4%, respectively) and distant recurrence (42.8%, 73.9% and 66.7%, respectively) in the CRT group were significantly lower than those in the No-AT and CTx groups ( p < 0.05), and the CRT group had significantly higher 5-year LRFS (82,1%, 26.8%, and 19.0%), RFS (53.3%, 11.6% and 16.7%) and OS rates (64.0%, 22.7% and 4.3%) than the CTx and No-AT groups ( p < 0.05 each). Therefore, adjuvant CRT may improve the LRFS and RFS and subsequently improve OS in lymph node-positive resected GB cancer.

Objective To assess prognostic factors of patients with operable oral cavity squamous cell carcinoma (OSCC), focusing on the associations with smoking/alcohol exposure and age. Materials and Methods A total of 247 patients with OSCC who received curative surgery ± adjuvant radiotherapy were analyzed. The patient subgroups were divided according to pretreatment smoking/alcohol exposure. Individuals aged 45 years or less were classified as younger patients. Results The median follow‐up was 52.2 months. The 5‐year locoregional progression‐free survival (LRFFS), distant metastasis‐free survival (DMFS), overall survival (OS), and cancer‐specific survival (CSS) rates were 85.2%, 88.3%, 78.1%, and 83.5%, respectively. An advanced stage, differentiation, and lympho‐vascular space invasion were significantly associated with lower OS and CSS. In a subgroup analysis of younger patients (n = 49), more smoking/alcohol exposure was significantly associated with better OS (hazard ratio [HR]: 0.21, 95% confidence interval [CI]: 0.05–0.95, p = .043). With increasing age, the HR for smoking/alcohol exposure with respect to OS increased up to 11.59 (95% CI: 1.49–89.84, p = .019) in older patients. Conclusion Younger OSCC patients with non‐ or less smoking/alcohol exposure showed unfavorable outcomes. The prognostic significance of pretreatment smoking/alcohol exposure changed from favorable to detrimental with increasing age in operable OSCC.