Traditional radiomics models mainly rely on explicitly-designed handcrafted features from medical images. This paper aimed to investigate if deep features extracted via transfer learning can generate radiomics signatures for prediction of overall survival (OS) in patients with Glioblastoma Multiforme (GBM). This study comprised a discovery data set of 75 patients and an independent validation data set of 37 patients. A total of 1403 handcrafted features and 98304 deep features were extracted from preoperative multi-modality MR images. After feature selection, a six-deep-feature signature was constructed by using the least absolute shrinkage and selection operator (LASSO) Cox regression model. A radiomics nomogram was further presented by combining the signature and clinical risk factors such as age and Karnofsky Performance Score. Compared with traditional risk factors, the proposed signature achieved better performance for prediction of OS (C-index = 0.710, 95% CI: 0.588, 0.932) and significant stratification of patients into prognostically distinct groups (P < 0.001, HR = 5.128, 95% CI: 2.029, 12.960). The combined model achieved improved predictive performance (C-index = 0.739). Our study demonstrates that transfer learning-based deep features are able to generate prognostic imaging signature for OS prediction and patient stratification for GBM, indicating the potential of deep imaging feature-based biomarker in preoperative care of GBM patients.

No abstract

No abstract

In fully-automatic radiomics model for predicting overall survival (OS) of glioblastoma multiforme (GBM) patients, the effect of image standardization parameters such as voxel size, quantization method and gray level on model reproducibility and prognostic performance are still unclear. In this study, 45792 multiregional radiomics features were automatically extracted from multi-modality MR images with different voxel sizes, quantization methods, and gray levels. The feature reproducibility and prognostic performance were assessed. Multiparametric and fixed-parameter radiomics signatures were constructed based on a training cohort (60 patients). In an independent validation cohort (32 patients), the multiparametric signature achieved better performance for OS prediction (C-Index = 0.705, 95% CI: 0.672, 0.738) and significant stratification of patients into high- and low-risk groups (P = 0.0040, HR = 3.29, 95% CI: 1.40, 7.70), which outperformed the fixed-parameter signatures and conventional factors such as age, Karnofsky Performance Score and tumor volume. This study demonstrated that voxel size, quantization method and gray level had influence on reproducibility and prognosis of radiomics features for GBM OS prediction. An automatic method to determine the optimal parameter settings was provided. It indicated that multiparametric radiomics signature had the potential of offering better prognostic performance than fixed-parameter signatures.

unfolding, [2] and skin wrinkles communicate a change in the state of the body. [3] These exquisite self-wrinkling processes typically originate from heterogeneous growth rates during the growth process, which results in compressive strains on the constrained tissues or organs and leads to mechanical instability. [5,6] Hydrogels possess water-rich structures similar to the abovementioned biological tissues and are regarded as promising scaffolds in biomedical fields. [7][8][9] Given the importance of wrinkled structures, the similarity between hydrogels and biotissue, and the important role of the programmable structure of hydrogels in potential biomedical applications, [10,11] imitating the delicate architecture of biological tissues in a biocompatible, mechanically stable wrinkled hydrogel with programmable patterns can facilitate an understanding of the effects of surface topography on biological properties and will be of benefit to humanity. In fact, motivated by the distinct functionalities of wrinkled structures in biology, material scientists have attempted to mimic this structure in artificial systems. [12][13][14][15][16][17][18][19] Consequently, a series of strategies, including swelling mismatch, [20] prepatterning treatment, [21] or coating of a second layer on the substrate, [5] have been successfully conducted to generate a compressive strain that triggers a controllable wrinkling Wrinkled hydrogels from biomass sources are potential structural biomaterials. However, for biorelated applications, engineering scalable, structure-customized, robust, and biocompatible wrinkled hydrogels with highly oriented nanostructures and controllable intervals is still a challenge. A scalable biomass material, namely cellulose, is reported for customizing anisotropic, all-cellulose, wrinkle-patterned hydrogels (AWHs) through an ultrafast, auxiliary force, acid-induced gradient dual-crosslinking strategy. Direct immersion of a prestretched cellulose alkaline gel in acid and relaxation within seconds allow quick buildup of a consecutive through-thickness modulus gradient with acid-penetration-directed dual-crosslinking, confirmed by visual 3D Raman microscopy imaging, which drives the formation of selfwrinkling structures. Moreover, guided by quantitative mechanics simulations, the structure of AWHs is found to exhibit programmable intervals and aligned nanostructures that differ between ridge and valley regions and can be controlled by tuning the prestretching strain and acid treatment time, and these AWHs successfully induce cell alignment. Thus, a new avenue is opened to fabricate polysaccharide-derived, programmable, anisotropic, wrinkled hydrogels for use as biomedical materials via a bottom-up method.The ORCID identification number(s) for the author(s) of this article can be found under https://doi.

• Multiregional and multiparametric MRI features reliably predicted MGMT methylation in multicentre cohorts. • All-relevant imaging features predicted MGMT methylation better than univariately-predictive and non-redundant features. • Combing clinical factors with radiomics features did not benefit the prediction performance.

The sluggish exciton dissociation and charge transfer and separation due to weak driving force still restrict the further improvement of photocatalytic performance in crystalline carbon nitride (CCN). Here, a series of heptazine-based K+-implanted CCNs (KCCNs) were successfully prepared, while the K+ ion concentration was gradiently inserted by controlling its diffusion from the surface to the bulk in carbon nitride (CN). The built-in electric field (BIEF) induced by the K+ ion concentration gradient and cyano groups has powerful driving force to promote the dissociation of excitons into electrons and holes and accelerate the separation of photogenerated carriers and their transport from the bulk to the surface. Consequently, the KCCN with an optimized BIEF has increased 34 times compared with the original CN for visible light H2 evolution. Our proposed strategy to induce BIEF by constructing concentration gradients through a thermodynamically feasible diffusion-controlled solid-state reaction can be adopted to design other photocatalytic systems.

Here, taking layered polymeric carbon nitride (CN) as an example, the dominant role of in-plane charge transport is demonstrated both theoretically and experimentally. The in-plane “sewing” and interlayer “cutting” effects of three alkaline potassium salts on the CN polymerization process are enhanced with the increase of alkalinity. The in-plane crystallinity-dependent charge transfer and separation and photocatalytic activity are elucidated through the samples of three alkaline potassium salt series. Our findings provide guidance for improving the photoelectric conversion efficiency by optimizing in-plane charge transfer rather than interlayer charge transfer, targeting layered photoelectric conversion materials.