脑动脉瘤的不规则形态，尤其是子瘤的存在，是动脉瘤破裂风险的关键因素．临床上对子瘤的评估主要是通过时间飞跃法磁共振血管造影（Time of Flight-Magnetic Resonance Angiography，TOF-MRA）进行图像重建及基于医生视觉和经验的判断，这限制了诊断的效率和准确性．本文提出了一种基于3D ResNet50改进的并行多尺度注意力融合网络（Parallel Multiscale Attention Fusion Networks，PMAF-Net）的子瘤自动分类方法，PMAF-Net采用多尺度卷积并加权融合通道和空间注意力权重以提高特征提取能力．实验所用TOF-MRA数据291例，其中训练集128例，验证集32例，测试集131例．与其他分类网络比较，PMAF-Net在测试集上表现最好，准确率为83.97%，召回率为84.48%，精确率为80.33%，F1分数为0.823 5，受试者工作特征曲线（ROC）也显示出模型最佳的分类性能（AUC为0.900 8）．实验结果表明该网络能更准确地识别出子瘤型动脉瘤，有望对动脉瘤破裂风险评估和量化提供支持．

胰腺因其解剖结构复杂多变、周围环境复杂等特点，始终是医学图像分割中最具挑战性的任务之一．针对以上问题，提出一种融合双解码和全局注意力上采样模块的深度学习分割模型（Combining Dual Decoding and Global Attention Upsampling Modules Network，DGANet）．模型由一个编码器和两个解码器构成，两个解码器实现了对不同深度特征信息的充分利用；模型采用全局注意力上采样模块（Global Attention Upsampling，GAU），利用高层丰富的语义信息来引导低层选择更为精准的特征信息．利用长海医院提供的数据集进行实验，结果表明平均Dice相似系数为86.28%，交并比（Intersection-over-Union，IoU）为0.77，豪斯多夫距离（Hausdorff Distance，HD）为7.7 mm，数据证实了该模型在胰腺囊性肿瘤分割中具有一定的临床意义和价值．

随着老龄化加剧，心血管疾病患病人数逐年增加，借助医学图像实现心脏功能的评估在诊疗过程中起着重要作用．心脏分割是评估心脏功能的前提，一直受到临床医生和科学研究者的密切关注．本文从传统方法和深度学习方法角度梳理了近十年以来关于心脏分割研究的文献．重点介绍了基于主动轮廓和图谱模型的传统分割方法，以及基于U-Net和全卷积神经网络（FCN）的深度学习算法．其中针对通过增加局部模块、优化损失函数、强化网络结构等方式改进深度学习网络以实现心脏特定区域精准分割这一主题进行了详细展开，并从心脏磁共振、X射线计算机断层扫描（CT）和超声3种成像模态对上述方法进行总结．最后总结了该领域目前的研究现状并对未来研究方向进行了展望．

术前精准预测胰腺囊性肿瘤的类型，对制定个体化诊疗方案具有重要的临床价值．针对胰腺浆液性和黏液性囊性肿瘤的分类鉴别问题，本文探讨了基于深度学习的多源特征分类模型在胰腺囊性肿瘤的术前辅助诊断中的应用．首先，通过深度学习和影像组学技术从分割图像中提取深度学习特征和影像组学特征，并对病例的临床特征进行评估和量化，然后采用最小绝对收缩选择算子（LASSO）及交叉验证的方法筛选特征，随之构建出两个多源特征模型，即影像组学联合深度学习（RAD_DL）模型、临床特征联合RAD_DL（Clinical_RAD_DL）模型，把传统的影像组学（RAD）模型和深度学习（DL）模型作为对照，最后选用支持向量机（SVM）、自适应提升算法（ADAboost）、随机森林（Random Forest）以及逻辑回归（Logistic）进行分类．采用准确率、召回率、精确率、曲线下面积（AUC）值以及精确率和召回率的调和平均数（F1值）作为评价指标，比较上述4种不同特征模型的分类效能，用校准曲线和决策曲线来评估其临床应用价值．结果显示Clinical_RAD_DL特征模型的分类效能表现最佳，准确率是0.923 1，召回率是0.882 4，精确率是0.882 0，F1是0.882 2，AUC是0.912 6，校准曲线和决策曲线显示出Clinical_RAD_DL特征模型的临床应用价值是最高的．实验表明基于深度学习的多源特征分类模型，对胰腺黏液性和浆液性囊性肿瘤具有较好的分类效果，可以为临床上精准诊疗提供帮助．

医生根据磁共振影像征象对患者的乳腺病变程度进行BI-RADS分类评估时存在一定的主观性，且 BI-RADS 3-5类病变的良恶性存在交叉，在临床诊断时极易发生因诊断类别较高而造成不必要的有创治疗．针对这些问题，本文应用影像组学技术对乳腺的T₁加权（T1W）和动态对比增强（DCE）磁共振图像进行特征提取和融合，采用最小绝对收缩和选择算子（LASSO）算法筛选出各特征集的最优特征集，并分别使用支持向量机（SVM）、随机森林（RF）、K最近邻（KNN）及逻辑回归（LR）算法进行BI-RADS 3-5类乳腺病变三分类，并且在此基础上实现乳腺良恶性分类．结果显示基于特征融合的四个影像组学模型对乳腺病变BI-RADS 3-5类的分类准确率分别为81.25%、87.50%、78.38%、81.25%；对乳腺病变良恶性鉴别的准确率分别为90.91%、93.55%、92.73%、94.55%. 这表明MRI影像组学结合机器学习的算法对乳腺病变BI-RADS分类效果及良恶性鉴别效果均较好，且特征融合可进一步提高分类预测的准确率．

Classification of brain tumors is one of the daunting tasks in medical imaging and incorrect decisions during the diagnosis process may lead to increased human fatality. Latest advances in artificial intelligence and deep learning have opened the path for the success of numerous medical image analysis tasks, including the recognition of brain tumors. In this paper, we propose a simple architecture based on deep learning that results in strong generalization without needing much preprocessing. The proposed Multi-modal Squeeze and Excitation model (MSENet) receives multiple representations of a given tumor image, learns end-to-end and effectively predicts the severity level of tumor. Convolution feature descriptors from multiple deep pre-trained models are used to effectively describe the tumor images and are supplied as input to the proposed MSENet. The squeeze and excitation blocks of the MSENet allow the model to prioritize tumor regions while giving less emphasis to the rest of the image, serving as an attention mechanism in the model. The proposed model is evaluated on the benchmark brain tumor dataset that is publicly accessible from the Figshare repository. Experimental studies reveal that in terms of model parameters, the proposed approach is simple and leads to competitive performance compared to those of existing complex models. By increasing complexity, the proposed model leads to generalize better and achieves state-of-the-art accuracy of 96.05% on Figshare dataset. Compared to the existing models, the proposed model neither uses segmentation nor uses augmentation techniques and without much pre-processing achieves competitive performance.

With the advancement in technology, machine learning can be applied to diagnose the mass/tumor in the brain using magnetic resonance imaging (MRI). This work proposes a novel developed transfer deep-learning model for the early diagnosis of brain tumors into their subclasses, such as pituitary, meningioma, and glioma. First, various layers of isolated convolutional-neural-network (CNN) models are built from scratch to check their performances for brain MRI images. Then, the 22-layer, binary-classification (tumor or no tumor) isolated-CNN model is re-utilized to re-adjust the neurons’ weights for classifying brain MRI images into tumor subclasses using the transfer-learning concept. As a result, the developed transfer-learned model has a high accuracy of 95.75% for the MRI images of the same MRI machine. Furthermore, the developed transfer-learned model has also been tested using the brain MRI images of another machine to validate its adaptability, general capability, and reliability for real-time application in the future. The results showed that the proposed model has a high accuracy of 96.89% for an unseen brain MRI dataset. Thus, the proposed deep-learning framework can help doctors and radiologists diagnose brain tumors early.

Brain tumor analysis is essential to the timely diagnosis and effective treatment of patients. Tumor analysis is challenging because of tumor morphology factors like size, location, texture, and heteromorphic appearance in medical images. In this regard, a novel two-phase deep learning-based framework is proposed to detect and categorize brain tumors in magnetic resonance images (MRIs). In the first phase, a novel deep-boosted features space and ensemble classifiers (DBFS-EC) scheme is proposed to effectively detect tumor MRI images from healthy individuals. The deep-boosted feature space is achieved through customized and well-performing deep convolutional neural networks (CNNs), and consequently, fed into the ensemble of machine learning (ML) classifiers. While in the second phase, a new hybrid features fusion-based brain-tumor classification approach is proposed, comprised of both static and dynamic features with an ML classifier to categorize different tumor types. The dynamic features are extracted from the proposed brain region-edge net (BRAIN-RENet) CNN, which is able to learn the heteromorphic and inconsistent behavior of various tumors. In contrast, the static features are extracted by using a histogram of gradients (HOG) feature descriptor. The effectiveness of the proposed two-phase brain tumor analysis framework is validated on two standard benchmark datasets, which were collected from Kaggle and Figshare and contain different types of tumors, including glioma, meningioma, pituitary, and normal images. Experimental results suggest that the proposed DBFS-EC detection scheme outperforms the standard and achieved accuracy (99.56%), precision (0.9991), recall (0.9899), F1-Score (0.9945), MCC (0.9892), and AUC-PR (0.9990). The classification scheme, based on the fusion of feature spaces of proposed BRAIN-RENet and HOG, outperform state-of-the-art methods significantly in terms of recall (0.9913), precision (0.9906), accuracy (99.20%), and F1-Score (0.9909) in the CE-MRI dataset.

Microsatellite instability (MSI) has been approved as a pan-cancer biomarker for immune checkpoint blockade (ICB) therapy. However, current MSI identification methods are not available for all patients. We proposed an ensemble multiple instance deep learning model to predict microsatellite status based on histopathology images, and interpreted the pathomics-based model with multi-omics correlation. Two cohorts of patients were collected, including 429 from The Cancer Genome Atlas (TCGA-COAD) and 785 from an Asian colorectal cancer (CRC) cohort (Asian-CRC). We established the pathomics model, named Ensembled Patch Likelihood Aggregation (EPLA), based on two consecutive stages: patch-level prediction and WSI-level prediction. The initial model was developed and validated in TCGA-COAD, and then generalized in Asian-CRC through transfer learning. The pathological signatures extracted from the model were analyzed with genomic and transcriptomic profiles for model interpretation. The EPLA model achieved an area-under-the-curve (AUC) of 0.8848 (95% CI: 0.8185-0.9512) in the TCGA-COAD test set and an AUC of 0.8504 (95% CI: 0.7591-0.9323) in the external validation set Asian-CRC after transfer learning. Notably, EPLA captured the relationship between pathological phenotype of poor differentiation and MSI (< 0.001). Furthermore, the five pathological imaging signatures identified from the EPLA model were associated with mutation burden and DNA damage repair related genotype in the genomic profiles, and antitumor immunity activated pathway in the transcriptomic profiles. Our pathomics-based deep learning model can effectively predict MSI from histopathology images and is transferable to a new patient cohort. The interpretability of our model by association with pathological, genomic and transcriptomic phenotypes lays the foundation for prospective clinical trials of the application of this artificial intelligence (AI) platform in ICB therapy.© The author(s).

The grain size of metal materials has a significant impact on their macroscopic properties. However, original metallographic images often suffer from issues such as substantial noise, missing grain boundaries, low contrast, and blurred edges. These challenges hinder the accurate extraction of complete grain boundaries, limiting the precision of grain size measurement and material performance prediction. Therefore, effectively reconstructing incomplete grain boundaries is particularly crucial. This paper proposes a grain boundary reconstruction and grain size measurement method based on an improved channel attention mechanism. A generative adversarial network (GAN) serves as the backbone, with a custom-designed channel attention module embedded in the generator. Combined with a global context attention mechanism, the method captures the global contextual information of the image, enhancing the network’s semantic understanding and reconstruction accuracy for regions with missing grain boundaries. During the image reconstruction process, the method effectively leverages long-range feature correlations within the image, significantly improving network performance. To address the Mode Collapse observed during experiments, the loss function is optimized using Focal Loss, balancing the ratio of positive and negative samples and improving network robustness. Compared with other attention modules, the improved channel attention module significantly enhances the performance of the generative network. Experimental results demonstrate that the generative network based on this module outperforms comparable modules in terms of MIoU (86.25%), Accuracy (95.06%), and Precision (86.54%). The grain boundary reconstruction method based on the improved channel attention mechanism not only effectively improves the accuracy of grain boundary reconstruction but also significantly enhances the generalization ability of the network. This provides reliable technical support for the characterization of the microstructure and the performance prediction of metal materials.