Radiological imaging techniques, including magnetic resonance imaging (MRI) and positron emission tomography (PET), are the standard-of-care non-invasive diagnostic approaches widely applied in neuro-oncology. Unfortunately, accurate interpretation of radiological imaging data is constantly challenged by the indistinguishable radiological image features shared by different pathological changes associated with tumor progression and/or various therapeutic interventions. In recent years, machine learning (ML)-based artificial intelligence (AI) technology has been widely applied in medical image processing and bioinformatics due to its advantages in implicit image feature extraction and integrative data analysis. Despite its recent rapid development, ML technology still faces many hurdles for its broader applications in neuro-oncological radiomic analysis, such as lack of large accessible standardized real patient radiomic brain tumor data of all kinds and reliable predictions on tumor response upon various treatments. Therefore, understanding ML-based AI technologies is critically important to help us address the skyrocketing demands of neuro-oncology clinical deployments. Here, we provide an overview on the latest advancements in ML techniques for brain tumor radiomic analysis, emphasizing proprietary and public dataset preparation and state-of-the-art ML models for brain tumor diagnosis, classifications (e.g., primary and secondary tumors), discriminations between treatment effects (pseudoprogression, radiation necrosis) and true progression, survival prediction, inflammation, and identification of brain tumor biomarkers. We also compare the key features of ML models in the realm of neuroradiology with ML models employed in other medical imaging fields and discuss open research challenges and directions for future work in this nascent precision medicine area.

Accurate handwriting recognition has been a challenging computer vision problem, because static feature analysis of the text pictures is often inadequate to account for high variance in handwriting styles across people and poor image quality of the handwritten text. Recently, by introducing machine learning, especially convolutional neural networks (CNNs), the recognition accuracy of various handwriting patterns is steadily improved. In this paper, a deep CNN model is developed to further improve the recognition rate of the MNIST handwritten digit dataset with a fast-converging rate in training. The proposed model comes with a multi-layer deep arrange structure, including 3 convolution and activation layers for feature extraction and 2 fully connected layers (i.e., dense layers) for classification. The model’s hyperparameters, such as the batch sizes, kernel sizes, batch normalization, activation function, and learning rate are optimized to enhance the recognition performance. The average classification accuracy of the proposed methodology is found to reach 99.82% on the training dataset and 99.40% on the testing dataset, making it a nearly error-free system for MNIST recognition.

Simultaneous high transparency and high haze are necessary for high-efficiency optical, photonic, and optoelectronic applications. However, a typical highly transparent film lacks high optical haze or vice versa. Here, we report a silk fibroin-based optical film that exhibits both ultrahigh optical transparency (>93%) and ultrahigh optical transmission haze (>65%). Also, in combination with the soft lithography method, different nanostructured silk fibroin films are presented and their optical properties are characterized as well. To demonstrate its exceptional performance in both high transmission and high optical haze, we combine the silk fibroin with the silicon photodiode and show that the efficiency can be increased by 6.96% with the silk fibroin film without patterns and 14.9% with the nanopatterned silk fibroin film. Silk provides excellent mechanical, optical, and electrical properties, and the reported high-performance silk fibroin can enable the development of next-generation biocompatible eco-friendly flexible electronic and optical devices.

Micron and sub-micron particles tend to be self-organized into well-defined arrays during conventional convective assembly, however this technique hardly works for the assembly of tiny nanoparticles smaller than 10 nm into monolayered arrays. The intrinsic difficulty is mainly related to the evaporation and convections occurring at a much larger volume than nanoparticle size. As a result, multilayered nanoparticle arrays are usually created. In this paper, we report that solvent evaporation and nanoparticle convection are artificially confined in nanoscale grooves near a pipette's periphery and the grooves can help assemble parallel strips because of the uni-directional translation of the pipette. The stacking layer of each strip is correlated with the grooves' diameter, thus the nanoparticle monolayer is fabricated in the tiny grooves. The narrowest strip width observed experimentally is geometrically limited by the effective diameter of the charged nanoparticles in gold colloid because nanoparticle convections from the inner chamber to the pipette's periphery will not occur within the grooves finer than the effective diameters.

By combining spin coating and inverse nanotransfer printing, silica-coated gold nanoparticles are patterned onto polydimethylsiloxane (PDMS) superhydrophobic surfaces to form a hierarchical structure. A layer of nanoparticles is spin-coated on a flat silicon substrate serving as the stamp, which is then transferred to the raised regions of PDMS surfaces. Our inverse nanotransfer printing is in contrast to the standard nanotransfer printing, which transfers metal from the raised regions of a stamp to a flat PDMS surface. The fabricated hierarchical surface exhibits a higher contact angle and delays the Cassie-Wenzel transition during evaporation of a sessile droplet, indicating an improvement of superhydrophobicity. Finally, we demonstrate that the fabricated nanoparticle-based superhydrophobic surfaces can enhance the Raman intensity and significantly decrease the surface-enhanced Raman scattering detection limit.