We conducted a meta-analysis of published retrospective studies and compared the effectiveness of pars plana vitrectomy with and without internal limiting membrane (ILM) peeling for idiopathic epiretinal membrane (IERM). The results revealed that patients in the IERM+ILM peeling group had better BCVA after surgery within 12 months than those in IERM peeling group. But patients in the IERM peeling group showed better BCVA in the 18th month. More retrospective studies or randomized controlled trials are required to investigate and compare the long-term effect of IERM removal with and without ILM peeling.
Though deep learning has shown successful performance in classifying the label and severity stage of certain disease, most of them give few evidence on how to make prediction. Here, we propose to exploit the interpretability of deep learning application in medical diagnosis. Inspired by Koch's Postulates, a well-known strategy in medical research to identify the property of pathogen, we define a pathological descriptor that can be extracted from the activated neurons of a diabetic retinopathy detector. To visualize the symptom and feature encoded in this descriptor, we propose a GAN based method to synthesize pathological retinal image given the descriptor and a binary vessel segmentation. Besides, with this descriptor, we can arbitrarily manipulate the position and quantity of lesions. As verified by a panel of 5 licensed ophthalmologists, our synthesized images carry the symptoms that are directly related to diabetic retinopathy diagnosis. The panel survey also shows that our generated images is both qualitatively and quantitatively superior to existing methods.
Purpose: To evaluate the performance of Delta4DVH Anatomy in patient-specific intensity-modulated radiotherapy quality assurance. Materials and Methods: Dose comparisons were performed between Anatomy doses calculated with treatment plan dose measured modification and pencil beam algorithms, treatment planning system doses, film doses, and ion chamber measured doses in homogeneous and inhomogeneous geometries. The sensitivity of Anatomy doses to machine errors and output calibration errors was also investigated. Results: For a Volumetric Modulated Arc Therapy (VMAT) plan evaluated on the Delta4 geometry, the conventional gamma passing rate was 99.6%. For a water-equivalent slab geometry, good agreements were found between dose profiles in film, treatment planning system, and Anatomy treatment plan dose measured modification and pencil beam calculations. Gamma passing rate for Anatomy treatment plan dose measured modification and pencil beam doses versus treatment planning system doses was 100%. However, gamma passing rate dropped to 97.2% and 96% for treatment plan dose measured modification and pencil beam calculations in inhomogeneous head & neck phantom, respectively. For the 10 patients’ quality assurance plans, good agreements were found between ion chamber measured doses and the planned ones (deviation: 0.09% ± 1.17%). The averaged gamma passing rate for conventional and Anatomy treatment plan dose measured modification and pencil beam gamma analyses in Delta4 geometry was 99.6% ± 0.89%, 98.54% ± 1.60%, and 98.95% ± 1.27%, respectively, higher than averaged gamma passing rate of 97.75% ± 1.23% and 93.04% ± 2.69% for treatment plan dose measured modification and pencil beam in patients’ geometries, respectively. Anatomy treatment plan dose measured modification dose profiles agreed well with those in treatment planning system for both Delta4 and patients’ geometries, while pencil beam doses demonstrated substantial disagreement in patients’ geometries when compared to treatment planning system doses. Both treatment planning system doses are sensitive to multileaf collimator and monitor unit (MU) errors for high and medium dose metrics but not sensitive to the gantry and collimator rotation error smaller than 3°. Conclusions: The new Delta4DVH Anatomy with treatment plan dose measured modification algorithm is a useful tool for the anatomy-based patient-specific quality assurance. Cautions should be taken when using pencil beam algorithm due to its limitations in handling heterogeneity and in high-dose gradient regions.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
hi@scite.ai
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.