Clinical Evaluation of Deep Learning and Atlas-Based Auto-Contouring of Bladder and Rectum for Prostate Radiation Therapy

PurposeAuto-contouring may reduce workload, interobserver variation, and time associated with manual contouring of organs at risk. Manual contouring remains the standard due in part to uncertainty around the time and workload savings after accounting for the review and editing of auto-contours. This preliminary study compares a standard manual contouring workflow with 2 auto-contouring workflows (atlas and deep learning) for contouring the bladder and rectum in patients with prostate cancer.Methods and MaterialsThree contouring workflows were defined based on the initial contour-generation method including manual (MAN), atlas-based auto-contour (ATLAS), and deep-learning auto-contour (DEEP). For each workflow, initial contour generation was retrospectively performed on 15 patients with prostate cancer. Then, radiation oncologists (ROs) edited each contour while blinded to the manner in which the initial contour was generated. Workflows were compared by time (both in initial contour generation and in RO editing), contour similarity, and dosimetric evaluation.ResultsMean durations for initial contour generation were 10.9 min, 1.4 min, and 1.2 min for MAN, DEEP, and ATLAS, respectively. Initial DEEP contours were more geometrically similar to initial MAN contours. Mean durations of the RO editing steps for MAN, DEEP, and ATLAS contours were 4.1 min, 4.7 min, and 10.2 min, respectively. The geometric extent of RO edits was consistently larger for ATLAS contours compared with MAN and DEEP. No differences in clinically relevant dose-volume metrics were observed between workflows.ConclusionAuto-contouring software affords time savings for initial contour generation; however, it is important to also quantify workload changes at the RO editing step. Using deep-learning auto-contouring for bladder and rectum contour generation reduced contouring time without negatively affecting RO editing times, contour geometry, or clinically relevant dose–volume metrics. This work contributes to growing evidence that deep-learning methods are a clinically viable solution for organ-at-risk contouring in radiation therapy.     

First experience with robotic pancreatoduodenectomy in Singapore

INTRODUCTIONRecent studies reported that laparoscopic pancreatoduodenectomy (LPD) is associated with superior perioperative outcomes compared to the open approach. However, concerns have been raised about the safety of LPD, especially during the learning phase. Robotic pancreatoduodenectomy (RPD) has been reported to be associated with a shorter learning curve compared to LPD. We herein present our initial experience with RPD.METHODSA retrospective review of a single-institution prospective robotic hepatopancreaticobiliary (HPB) surgery database of 70 patients identified seven consecutive RPDs performed by a single surgeon in 2016–2017. These were matched at a 1:2 ratio with 14 open pancreatoduodenectomies (OPDs) selected from 77 consecutive pancreatoduodenectomies performed by the same surgeon between 2011 and 2017.RESULTSSeven patients underwent RPD, of which five were hybrid procedures with open reconstruction. There were no open conversions. Median operative time was 710.0 (range 560.0–930.0) minutes. Two major morbidities (> Grade 2) occurred: one gastrojejunostomy bleed requiring endoscopic haemostasis and one delayed gastric emptying requiring feeding tube placement. There were no pancreatic fistulas, reoperations or 90-day/in-hospital mortalities in the RPD group. Comparison between RPD and OPD demonstrated that RPD was associated with a significantly longer operative time. Compared to open surgery, there was no significant difference in estimated blood loss, blood transfusion, postoperative stay, pancreatic fistula rates, morbidity and mortality rates, R0 resection rates, and lymph node harvest rates.CONCLUSIONOur initial experience demonstrates that RPD is feasible and safe in selected patients. It can be safely adopted without any compromise in patient outcomes compared to the open approach.     

STAG2 loss rewires oncogenic and developmental programs to promote metastasis in Ewing sarcoma

1. Download : Download high-res image (172KB)
2. Download : Download full-size image

     

Viral infections in critical care: a narrative review

Viral infections form a substantial part of the intensive care workload, even before the recent and ongoing COVID-19 pandemic. The growing availability of molecular diagnostics for viral infections has led to increased recognition of these pathogens. This additional information, however, provides new challenges for interpretation and management. As the SARS-CoV-2 pandemic has amply demonstrated, the emergence and global spread of novel viruses are likely to provide continued challenges for critical care physicians into the future. This article will provide an overview of viral infections relevant to the critical care physician, discussing the diagnosis and management of respiratory viral infections, blood borne and enteric viruses. We will also discuss herpesviridae complications, commonly seen due to reactivation of latent infections. Further, we explore some rarer and emerging viruses, including recognition of viral haemorrhagic fevers, and briefly discuss post-viral syndromes which may present to the intensive care unit. Finally, we will discuss infection control and its importance in preventing nosocomial viral transmission.     

Viadent, ethanol, and pH effects upon gingival epithelial-like cells, in vitro

Interest has recently been directed towards the use of antiplaque mouthrinses. Most published material concerns the antimicrobial effects of these agents rather than their effects upon oral tissue. This study was conducted to evaluate the effect of a sanguinarine-containing mouthrinse called Viadent upon epithelial-like gingival cells. The cells were grown for 24 hours in supplemented Earle's medium, with and without different Viadent dilutions. Cell counts were made with a hematocytometer. It was found that 50% of the cells were inhibited at 1.2% Viadent. In similar studies, it was found that 70% ethanol and two pH buffers were less toxic than Viadent. Exposure of preformed cell monolayers to Viadent also showed significant inhibition. The relative toxicity of different antiplaque agents may be compared using such cells as a model system. In conclusion, it was observed that Viadent significantly affected gingival cell growth in vitro, that viable cell numbers were greatly reduced by short time exposure, and that the toxic effect of Viadent could only partially be accounted for by ethanol content and/or pH.     

Thrombomodulin Functional Domains Support Osteoblast Differentiation and Bone Healing in Diabetes in Mice

Thrombomodulin (TM) is a transmembrane glycoprotein that contains five functional domains. Soluble TM (sTM), comprising extracellular domains TMD1 (lectin-like), TMD2 (epidermal growth factor [EGF]-like repeat containing), and TMD3 (serine-threonine rich), can be shed from cells by the intramembrane protease rhomboid-like-2 (RHBDL2). TM is expressed by osteoblasts, yet its role there has not been determined. Herein we aimed to investigate the properties of TM and its domains in osteoblast function and bone repair following injury in diabetes. In response to a scratch injury of cultured osteoblast-like MG63 cells, expression of TM and RHBDL2 was enhanced, with increased release of sTM. Conditioned media from the injured cells promoted osteoblast migration, an effect that was lacking with conditioned media from MG63 cells in which TM was silenced by shRNA. Exogenous recombinant TMD1 had no effect on osteoblast activities or on bone repair in vivo. However, TM domains 2 and 3 (TMD2/3), induced MG63 cell migration, proliferation and mineralization in vitro, and when locally administered in mice, improved in vivo healing of injured calvarium. This beneficial effect of TMD2/3, mediated via fibroblast growth factor receptor (FGFR)/ERK signaling pathways, was also observed in vitro under high glucose conditions where endogenous TM expression was reduced, and in vivo in diabetic mice following tibia fracture or calvarium injury, where the osteoblastic response and healing were otherwise dampened. Taken together, osteoblast TM participates in bone healing, and recombinant TMD2/3 holds promise as a novel therapy for diabetic bone defect healing. © 2020 American Society for Bone and Mineral Research.