Biomedical Engineering Letters最新文献

Multiclass classification of brain tumors from magnetic resonance (MR) images is challenging due to high inter-class similarities. To this end, convolution neural networks (CNN) have been widely adopted in recent studies. However, conventional CNN architectures fail to capture the small lesion patterns of brain tumors. To tackle this issue, in this paper, we propose a global transformer network dubbed GT-Net for multiclass brain tumor classification. The GT-Net mainly comprises a global transformer module (GTM), which is introduced on the top of a backbone network. A generalized self-attention block (GSB) is proposed to capture the feature inter-dependencies not only across spatial dimension but also channel dimension, thereby facilitating the extraction of the detailed tumor lesion information while ignoring less important information. Further, multiple GSB heads are used in GTM to leverage global feature dependencies. We evaluate our GT-Net on a benchmark dataset by adopting several backbone networks, and the results demonstrate the effectiveness of GTM. Further, comparison with state-of-the-art methods validates the superiority of our model.

Practical application of surface-enhanced Raman spectroscopy (SERS) has suffered from several limitations by heterogeneous distribution of hot-spots, such as high signal fluctuation and the resulting low reliability in detection. Herein, we develop a strategy of more sensitive and reliable SERS platform through designing spatially homogeneous gold nanoparticles (GNPs) on a uniform gold nanoisland (GNI) pattern. The proposed SERS substrate is successfully fabricated by combining two non-lithographic techniques of electron beam evaporation and convective self-assembly. These bottom-up methods allow a simple, cost-effective, and large-area fabrication. Compared to the SERS substrates obtained from two separate nanofabrication methods, Raman spectra measured by the samples with both GNPs and GNIs present a significant increase in the signal intensity as well as a notable improvement in signal fluctuation. The simulated near-field analyses demonstrate the formation of highly amplified plasmon modes within and at the gaps of the GNP-GNI interfaces. Moreover, the suggested SERS sensor is evaluated to detect the glucose concentration, exhibiting that the detection sensitivity is improved by more than 10 times compared to the sample with only GNI patterns and a fairly good spatial reproducibility of 7% is accomplished. It is believed that our suggestion could provide a potential for highly sensitive, low-cost, and reliable SERS biosensing platforms that include many advantages for healthcare devices.

Supplementary information: The online version contains supplementary material available at 10.1007/s13534-024-00381-4.

Supramalleolar osteotomy (SMO) is a representative procedure to restore a malalignment in the varus ankle deformity by shifting the concentrated pressure on the medial ankle joint to the lateral area. Additionally, fibula osteotomy (FO) is selectively selected and performed according to the surgeon's preference. However, it is controversial whether FO is effective in shifting the abnormal pressure from the medial to the lateral area on the ankle joint. Some cadaveric studies have been performed to prove this. However, it is difficult to consistently reconstruct amount of the varus ankle deformities angle in cadavers and to guarantee reliable contact pressure between the ankle joint. Thus, the aim of this study was predicted and quantitatively compared a peak pressure between single SMO and SMO with FO procedure by using a finite element analysis as a powerful biomechanical tool to those limitations of cadaveric study. This study reconstructed total 4 3D foot and ankle models including a normal and pre-op model and 2 post-op models. The pre-op model was modified by assigning 10° varus tilting corresponding to stage 3b in the classification of varus ankle osteoarthritis based on the validated normal model. Also, the post-op models were reconstructed by applying single SMO and SMO with FO, respectively. All of the models were assumed as one-leg standing position and to mimic smooth ankle joint motion. Peak contact pressure change was predicted at the medial ankle joint by using computational simulation. As a result, 2 post-op models showed a remarkably peak pressure reduction by up to 5.5 times on the medial tibiotalar joint. However, a comparison between single SMO and SMO with FO model showed no appreciable differences. In conclusion, this study predicted that single SMO may be as effective as SMO with FO in reducing peak contact pressure on the medial tibiotalar joint in varus ankle osteoarthritis.