Machine learning-based response assessment in patients with rectal cancer after neoadjuvant chemoradiotherapy: radiomics analysis for assessing tumor regression grade using T2-weighted magnetic resonance images.

PURPOSE: This study aimed to assess tumor regression grade (TRG) in patients with rectal cancer after neoadjuvant chemoradiotherapy (NCRT) through a machine learning-based radiomics analysis using baseline T2-weighted magnetic resonance (MR) images. MATERIALS AND METHODS: In total, 148 patients with locally advanced rectal cancer(T2-4 or N+) who underwent MR imaging at baseline and after chemoradiotherapy between January 2010 and May 2021 were included. A region of interest for each tumor mass was drawn by a radiologist on oblique axial T2-weighted images, and main features were selected using principal component analysis after dimension reduction among 116 radiomics and three clinical features. Among eight learning models that were used for prediction model development, the model showing best performance was selected. Treatment responses were classified as either good or poor based on the MR-assessed TRG (mrTRG) and pathologic TRG (pTRG). The model performance was assessed using the area under the receiver operating curve (AUROC) to classify the response group. RESULTS: Approximately 49% of the patients were in the good response (GR) group based on mrTRG (73/148) and 26.9% based on pTRG (28/104). The AUCs of clinical data, radiomics models, and combined radiomics with clinical data model for predicting mrTRG were 0.80 (95% confidence interval [CI] 0.73, 0.87), 0.74 (95% CI 0.66, 0.81), and 0.75(95% CI 0.68, 0.82), and those for predicting pTRG was 0.62 (95% CI 0.52, 0.71), 0.74 (95% CI 0.65, 0.82), and 0.79 (95% CI 0.71, 0.87). CONCLUSION: Radiomics combined with clinical data model using baseline T2-weighted MR images demonstrated feasible diagnostic performance in predicting both MR-assessed and pathologic treatment response in patients with rectal cancer after NCRT.

Machine Learning-Assisted Identification of Single-Layer Graphene via Color Variation Analysis.

Techniques such as using an optical microscope and Raman spectroscopy are common methods for detecting single-layer graphene. Instead of relying on these laborious and expensive methods, we suggest a novel approach inspired by skilled human researchers who can detect single-layer graphene by simply observing color differences between graphene flakes and the background substrate in optical microscope images. This approach implemented the human cognitive process by emulating it through our data extraction process and machine learning algorithm. We obtained approximately 300,000 pixel-level color difference data from 140 graphene flakes from 45 optical microscope images. We utilized the average and standard deviation of the color difference data for each flake for machine learning. As a result, we achieved F1-Scores of over 0.90 and 0.92 in identifying 60 and 50 flakes from green and pink substrate images, respectively. Our machine learning-assisted computing system offers a cost-effective and universal solution for detecting the number of graphene layers in diverse experimental environments, saving both time and resources. We anticipate that this approach can be extended to classify the properties of other 2D materials.

Accurate and Precise Determination of Mechanical Properties of Silicon Nitride Beam Nanoelectromechanical Devices.

Accurate and precise determination of mechanical properties of nanoscale materials is mandatory since device performances of nanoelectromechanical systems (NEMS) are closely related to the flexural properties of the materials. In this study, the intrinsic mechanical properties of highly stressed silicon nitride (SiN) beams of varying lengths are investigated using two different techniques: Dynamic flexural measurement using optical interferometry and quasi-static flexural measurement using atomic force microscopy. The resonance frequencies of the doubly clamped, highly stressed beams are found to be inversely proportional to their length, which is not usually observed from a beam but is expected from a string-like structure. The mass density of the SiN beams can be precisely determined from the dynamic flexural measurements by using the values for internal stress and Young's modulus determined from the quasi-static measurements. As a result, the mass resolution of the SiN beam resonators was predicted to be a few attograms, which was found to be in excellent agreement with the experimental results. This work suggests that accurate and precise determination of mechanical properties can be achieved through combined flexural measurement techniques, which is a crucial key for designing practical NEMS applications such as biomolecular sensors and gas detectors.

Carbon nanotube field effect transistors with suspended graphene gates.

Novel field effect transistors with suspended graphene gates are demonstrated. By incorporating mechanical motion of the gate electrode, it is possible to improve the switching characteristics compared to a static gate, as shown by a combination of experimental measurements and numerical simulations. The mechanical motion of the graphene gate is confirmed by using atomic force microscopy to directly measure the electrostatic deflection. The device geometry investigated here can also provide a sensitive measurement technique for detecting high-frequency motion of suspended membranes as required, e.g., for mass sensing.