Background: Inflammatory bowel disease (IBD) is an immune-mediated disorder characterized by intestinal inflammation and includes two subtypes: Crohn's disease (CD) and ulcerative colitis (UC). The computed tomography manifestations of colonic CD (cCD) and UC are similar, and differential diagnosis is challenging. Our study aimed to investigate the feasibility of using a modified YOLOv5 algorithm for differentiating between cCD and UC on computed tomography enterography (CTE) images. Methods: This multicenter retrospective study analyzed data from a total of 29 cCD patients and 29 UC patients. Five submodels (YOLOv5n, YOLOv5s, YOLOv5m, YOLOv5l, and YOLOv5x) of YOLOv5 were trained and evaluated on the datasets. The CTE images of the cCD group and UC group were divided into a training set, validation set, and test set at a ratio of 8:1:1. Finally, the precision (Pr), recall rate (Rc), and mean average precision (mAP_0.5 and mAP_0.5:0.95) of the models were compared. Results: The YOLOv5x model showed the best performance among the five submodels, with mAP_0.5 of 0.97 and mAP_0.5:0.95 of 0.97 and 0.84 in the validation set and mAP_0.5 and mAP_0.5:0.95 of 0.97 and 0.83 in the test set, respectively. These results demonstrated similar diagnostic accuracy to the two radiologists (84.5%). Conclusion: The modified YOLOv5 algorithm is a feasible approach to distinguish between cCD and UC on CTE images. These findings may facilitate the early detection and differential diagnosis of IBD.

Various deep learning models have been developed and employed for medical image classification. This study conducted comprehensive experiments on 12 models, aiming to establish reliable benchmarks for research on breast dynamic contrast-enhanced magnetic resonance imaging image classification. Twelve deep learning models were systematically compared by analyzing variations in 4 key hyperparameters: optimizer (Op), learning rate, batch size (BS), and data augmentation. The evaluation criteria encompassed a comprehensive set of metrics including accuracy (Ac), loss value, precision, recall rate, F1-score, and area under the receiver operating characteristic curve. Furthermore, the training times and model parameter counts were assessed for holistic performance comparison. Adjustments in the BS within Adam Op had a minimal impact on Ac in the convolutional neural network models. However, altering the Op and learning rate while maintaining the same BS significantly affected the Ac. The ResNet152 network model exhibited the lowest Ac. Both the recall rate and area under the receiver operating characteristic curve for the ResNet152 and Vision transformer-base (ViT) models were inferior compared to the others. Data augmentation unexpectedly reduced the Ac of ResNet50, ResNet152, VGG16, VGG19, and ViT models. The VGG16 model boasted the shortest training duration, whereas the ViT model, before data augmentation, had the longest training time and smallest model weight. The ResNet152 and ViT models were not well suited for image classification tasks involving small breast dynamic contrast-enhanced magnetic resonance imaging datasets. Although data augmentation is typically beneficial, its application should be approached cautiously. These findings provide important insights to inform and refine future research in this domain.

OBJECTIVE:Acute aortic dissection (AD) is a life threatening condition that poses considerable challenges for timely diagnosis. Non-contrast computed tomography (CT) is frequently used to diagnose AD in certain clinical settings, but its diagnostic accuracy can vary among radiologists. This study aimed to develop and validate an interpretable You only look once version 8 (YOLOv8) deep learning model based on non-contrast CT to detect AD. METHODS:This retrospective study included patients from five institutions, divided into training, internal validation, and external validation cohorts. The YOLOv8 deep learning model was trained on annotated non-contrast CT images. Its performance was evaluated using area under the curve (AUC), sensitivity, specificity, and inference time compared with findings from vascular interventional radiologists, general radiologists, and radiology residents. In addition, gradient weighted class activation mapping (Grad-CAM) saliency map analysis was performed. RESULTS:A total of 1 138 CT scans were assessed (569 with AD, 569 controls). The YOLOv8s submodel achieved an AUC of 0.964 (95% confidence interval [CI] 0.939 - 0.988) in the internal validation cohort and 0.970 (95% CI 0.946 - 0.990) in the external validation cohort. In the external validation cohort, the performance of the three groups of radiologists in detecting AD was inferior to that of the YOLOv8s model. The model's sensitivity (0.976) was slightly higher than that of vascular interventional specialists (0.965; p = .18), and its specificity (0.935) was superior to that of general radiologists (0.835; p < .001). The model's inference time was 3.47 seconds, statistically significantly shorter than the radiologists' mean interpretation time of 25.32 seconds (p < .001). Grad-CAM analysis confirmed that the model focused on anatomically and clinically relevant regions, supporting its interpretability. CONCLUSION:The YOLOv8s deep learning model reliably detected AD on non-contrast CT and outperformed radiologists, particularly in time efficiency and diagnostic accuracy. Its implementation could enhance AD screening in specific settings, support clinical decision making, and improve diagnostic quality.

To develop and validate a nomogram combining radiomics and pathology features to distinguish between aldosterone-producing adenomas (APAs) and nonfunctional adrenal adenomas (NF-AAs). Consecutive patients diagnosed with adrenal adenomas via computed tomography (CT) or pathologic analysis between January 2011 and November 2022 were eligible for inclusion in this retrospective study. CT images and hematoxylin eosin–stained slides were used for annotation and feature extraction. The selected radiomics and pathology features were used to develop a risk model using various machine learning models, and the area under the receiver operating characteristic curve (AUC) was determined to evaluate diagnostic performance. The predicted results from radiomics and pathology features were combined and visualized using a nomogram. A total of 211 patients (APAs, n = 59; NF-AAs, n = 152) were included in this study, with patients randomly divided into either the training set or the testing set at a ratio of 8:2. The ExtraTrees model yielded a sensitivity of 0.818, a specificity of 0.733, and an accuracy of 0.756 (AUC = 0.817; 95

Objectives: Atypical presentations, lack of biomarkers, and low sensitivity of plain CT can delay the diagnosis of superior mesenteric artery (SMA) abnormalities, resulting in poor clinical outcomes. Our study aims to develop a deep learning (DL) model for detecting SMA abnormalities in plain CT and evaluate its performance in comparison with a clinical model and radiologist assessment. Materials and methods: A total of 1048 patients comprised the internal (474 patients with SMA abnormalities, 474 controls) and external testing (50 patients with SMA abnormalities, 50 controls) cohorts. The internal cohort was divided into the training cohort (n = 776), validation cohort (n = 86), and internal testing cohort (n = 86). A total of 5 You Only Look Once version 8 (YOLOv8)-based DL submodels were developed, and the performance of the optimal submodel was compared with that of a clinical model and of experienced radiologists. Results: Of the submodels, YOLOv8x had the best performance. The area under the curve (AUC) of the YOLOv8x submodel was higher than that of the clinical model (internal test set: 0.990 vs 0.878, P =.002; external test set: 0.967 vs 0.912, P =.140) and that of all radiologists (P <.001). The YOLOv8x submodel, when compared with radiologist assessment, demonstrated higher sensitivity (internal test set: 100.0 % vs 70.7 %, P =.002; external test set: 96.0 % vs 68.8 %, P <.001) and specificity (internal test set: 90.7 % vs 66.0 %, P =.025; external test set: = 88.0 % vs 66.0 %, P <.001). Conclusion: Using plain CT images, YOLOv8x was able to efficiently identify cases of SMA abnormalities. This could potentially improve early diagnosis accuracy and thus improve clinical outcomes.

Background You Only Look Once version 5 (YOLOv5), a one-stage deep-learning (DL) algorithm for object detection and classification, offers high speed and accuracy for identifying targets.Purpose To investigate the feasibility of using the YOLOv5 algorithm to non-invasively distinguish between aldosterone-producing adenomas (APAs) and non-functional adrenocortical adenomas (NF-ACAs) on computed tomography (CT) images.Material and Methods A total of 235 patients who were diagnosed with ACAs between January 2011 and July 2022 were included in this study. Of the 215 patients, 81 (37.7%) had APAs and 134 (62.3%) had NF-ACAs' they were randomly divided into either the training set or the validation set at a ratio of 9:1. Another 20 patients, including 8 (40.0%) with APA and 12 (60.0%) with NF-ACA, were collected for the testing set. Five submodels (YOLOv5n, YOLOv5s, YOLOv5m, YOLOv5l, and YOLOv5x) of YOLOv5 were trained and evaluated on the datasets.Results In the testing set, the mAP_0.5 value for YOLOv5x (0.988) was higher than the values for YOLOv5n (0.969), YOLOv5s (0.965), YOLOv5m (0.974), and YOLOv5l (0.983). The mAP_0.5:0.95 value for YOLOv5x (0.711) was also higher than the values for YOLOv5n (0.587), YOLOv5s (0.674), YOLOv5m (0.671), and YOLOv5l (0.698) in the testing set. The inference speed of YOLOv5n was 2.4 ms in the testing set, which was the fastest among the five submodels.Conclusion The YOLOv5 algorithm can accurately and efficiently distinguish between APAs and NF-ACAs on CT images, especially YOLOv5x has the best identification performance.

BackgroundThis study aims to explore the accuracy of Convolutional Neural Network (CNN) models in predicting malignancy in Dynamic Contrast-Enhanced Breast Magnetic Resonance Imaging (DCE-BMRI).MethodsA total of 273 benign lesions (benign group) and 274 malignant lesions (malignant group) were collected and randomly divided into a training set (246 benign and 245 malignant lesions) and a testing set (28 benign and 28 malignant lesions) in a 9:1 ratio. An additional 53 lesions from 53 patients were designated as the validation set. Five models-VGG16, VGG19, DenseNet201, ResNet50, and MobileNetV2-were evaluated. Model performance was assessed using accuracy (Ac) in the training and testing sets, and precision (Pr), recall (Rc), F1 score (F1), and area under the receiver operating characteristic curve (AUC) in the validation set.ResultsThe accuracy of VGG19 on the test set (0.96) is higher than that of VGG16 (0.91), DenseNet201 (0.91), ResNet50 (0.67), and MobileNetV2 (0.88). For the validation set, VGG19 achieved higher performance metrics (Pr 0.75, Rc 0.76, F1 0.73, AUC 0.76) compared to the other models, specifically VGG16 (Pr 0.73, Rc 0.75, F1 0.70, AUC 0.73), DenseNet201 (Pr 0.71, Rc 0.74, F1 0.69, AUC 0.71), ResNet50 (Pr 0.65, Rc 0.68, F1 0.60, AUC 0.65), and MobileNetV2 (Pr 0.73, Rc 0.75, F1 0.71, AUC 0.73). S4 model achieved higher performance metrics (Pr 0.89, Rc 0.88, F1 0.87, AUC 0.89) compared to the other four fine-tuned models, specifically S1 (Pr 0.75, Rc 0.76, F1 0.74, AUC 0.75), S2 (Pr 0.77, Rc 0.79, F1 0.75, AUC 0.77), S3 (Pr 0.76, Rc 0.76, F1 0.73, AUC 0.75), and S5 (Pr 0.77, Rc 0.79, F1 0.75, AUC 0.77). Additionally, S4 model showed the lowest loss value in the testing set. Notably, the AUC of S4 for BI-RADS 3 was 0.90 and for BI-RADS 4 was 0.86, both significantly higher than the 0.65 AUC for BI-RADS 5.ConclusionsThe S4 model we propose has demonstrated superior performance in predicting the likelihood of malignancy in DCE-BMRI, making it a promising candidate for clinical application in patients with breast diseases. However, further validation is essential, highlighting the need for additional data to confirm its efficacy.

Purpose: To evaluate the capability of deep transfer learning (DTL) and fine-tuning methods in differentiating malignant from benign lesions in breast dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI). Methods: The diagnostic efficiencies of the VGG19, ResNet50, and DenseNet201 models were tested under the same dataset. The model with the highest performance was selected and modified utilizing three fine-tuning strategies (S1-3). Fifty additional lesions were selected to form the validation set to verify the generalization abilities of these models. The accuracy (Ac) of the different models in the training and test sets, as well as the precision (Pr), recall rate (Rc), F1 score (), and area under the receiver operating characteristic curve (AUC), were primary performance indicators. Finally, the kappa test was used to compare the degree of agreement between the DTL models and pathological diagnosis in differentiating malignant from benign breast lesions. Results: The Pr, Rc, f1, and AUC of VGG19 (86.0%, 0.81, 0.81, and 0.81, respectively) were higher than those of DenseNet201 (70.0%, 0.61, 0.63, and 0.61, respectively) and ResNet50 (61.0%, 0.59, 0.59, and 0.59). After fine-tuning, the Pr, Rc, f1, and AUC of S1 (87.0%, 0.86, 0.86, and 0.86, respectively) were higher than those of VGG19. Notably, the degree of agreement between S1 and pathological diagnosis in differentiating malignant from benign breast lesions was 0.720 (κ = 0.720), which was higher than that of DenseNet201 (κ = 0.440), VGG19 (κ = 0.640), and ResNet50 (κ = 0.280). Conclusion: The VGG19 model is an effective method for identifying benign and malignant breast lesions on DCE-MRI, and its performance can be further improved via fine-tuning. Overall, our findings insinuate that this technique holds potential clinical application value.

The origin of metastatic liver tumours (arising from gastric or colorectal sources) is closely linked to treatment choices and survival prospects. However, in some instances, the primary lesion remains elusive even after an exhaustive diagnostic investigation. Consequently, we have devised and validated a radiomics nomogram for ascertaining the primary origin of liver metastases stemming from gastric cancer (GCLMs) and colorectal cancer (CCLMs). This retrospective study encompassed patients diagnosed with either GCLMs or CCLMs, comprising a total of 277 GCLM cases and 278 CCLM cases. Radiomic characteristics were derived from venous phase computed tomography (CT) scans, and a radiomics signature (RS) was computed. Multivariable regression analysis demonstrated that gender (OR = 3.457; 95% CI: 2.102-5.684; p < 0.001), haemoglobin levels (OR = 0.976; 95% CI: 0.967-0.986; p < 0.001), carcinoembryonic antigen (CEA) levels (OR = 0.500; 95% CI: 0.307-0.814; p = 0.005), and RS (OR = 2.147; 95% CI: 1.127-4.091; p = 0.020) exhibited independent associations with GCLMs as compared to CCLMs. The nomogram, combining RS with clinical variables, demonstrated strong discriminatory power in both the training (AUC = 0.71) and validation (AUC = 0.78) cohorts. The calibration curve, decision curve analysis, and clinical impact curves revealed the clinical utility of this nomogram and substantiated its enhanced diagnostic performance.

Background In clinical practice, reducing unnecessary biopsies for mammographic BI-RADS 4 lesions is crucial. The objective of this study was to explore the potential value of deep transfer learning (DTL) based on the different fine-tuning strategies for Inception V3 to reduce the number of unnecessary biopsies that residents need to perform for mammographic BI-RADS 4 lesions. Methods A total of 1980 patients with breast lesions were included, including 1473 benign lesions (185 women with bilateral breast lesions), and 692 malignant lesions collected and confirmed by clinical pathology or biopsy. The breast mammography images were randomly divided into three subsets, a training set, testing set, and validation set 1, at a ratio of 8:1:1. We constructed a DTL model for the classification of breast lesions based on Inception V3 and attempted to improve its performance with 11 fine-tuning strategies. The mammography images from 362 patients with pathologically confirmed BI-RADS 4 breast lesions were employed as validation set 2. Two images from each lesion were tested, and trials were categorized as correct if the judgement (≥ 1 image) was correct. We used precision (Pr), recall rate (Rc), F1 score (F1), and the area under the receiver operating characteristic curve (AUROC) as the performance metrics of the DTL model with validation set 2. Results The S5 model achieved the best fit for the data. The Pr, Rc, F1 and AUROC of S5 were 0.90, 0.90, 0.90, and 0.86, respectively, for Category 4. The proportions of lesions downgraded by S5 were 90.73%, 84.76%, and 80.19% for categories 4 A, 4B, and 4 C, respectively. The overall proportion of BI-RADS 4 lesions downgraded by S5 was 85.91%. There was no significant difference between the classification results of the S5 model and pathological diagnosis ( P = 0.110). Conclusion The S5 model we proposed here can be used as an effective approach for reducing the number of unnecessary biopsies that residents need to conduct for mammographic BI-RADS 4 lesions and may have other important clinical uses.

目的 探讨基于DenseNet 201深度迁移学习(DTL)模型在改善乳腺MRI乳腺影像报告和数据系统(BI-RADS)4类病变的细分类潜力.方法 以DenseNet 201神经网络为基础,对11 256幅良性组和5 448幅恶性组乳腺动态对比增强磁共振成像(DCE-MRI)图像建立DTL模型,将良性组、恶性组图像分别按照9:1随机分为训练集(良性组10 146幅,恶性组4 908幅)和测试集(良性组1 110幅,恶性组540幅).选取BI-RADS 4类患者81例作为验证集,所有患者乳腺病变均经病理证实,其中60例恶性,21例良性.每例患者选择10幅DCE-MRI增强图像进行验证,如果8幅图像归类正确则认定为此例患者归类正确.以验证集准确率、召回率、F1评分及受试者工作特征(ROC)曲线的曲线下面积(AUC)作为性能指标.结果 DTL模型在训练集和测试集的准确率均为100.00％.验证集准确率、召回率、F1评分及AUC分别为98.00％、0.98、0.98和0.97.21例良性病变中,DTL模型预测正确20例(占95.24％),预测恶性概率为3.50％～27.60％.60例恶性病变中,DTL模型预测正确58例(占96.67％),预测恶性概率为51.50％～93.60％.基于DenseNet 201的DTL模型与病理组织学在对乳腺MRI良恶性病变的分类诊断结果差异无统计学意义(P=0.859).结论 基于DenseNet 201的DTL模型可作为乳腺MRI BI-RADS 4类病变细分的有效方法.

Objective:To investigate the value of a deep learning method based on MobileNet in classification of bedside chest radiograph and improvement of the work efficiency.Methods:A total of 6, 320 bedside chest radiographs from January 2017 to December 2022 in the Second Peoples′ Hospital of Changzhou were retrospectively collected. The included cases were divided into normal group (885 images), pneumonia group (1 927 images), pleural effusion group (373 images), and pneumonia with pleural effusion group (3 135 images). Three hundred and fifty images were selected as a validation set, while the remaining images were divided into a train set (4 775 images) and a test set (1 195 images) using simple randomization, by 8∶2 ratio. Two lightweight convolutional neural network models (MobileNetV1 and MobileNetV2) were used to construct a bedside chest radiograph classification model, based on which two fine-tuning strategies were designed. Four models were generated namely MobileNetV1_False (V1_False), MobileNetV1_True (V1_True), MobileNetV2_False (V2_False) and MobileNetV2_True (V2_True). In the first stage, a binary classification model was established to divide the images into normal and lesion groups; then a four-class classification model was established in the second stage, with which the images were divided into four groups: normal, pneumonia, pleural effusion and pneumonia with pleural effusion. Metrics for model performance evaluation including accuracy (Ac), precision (Pr), recall rate (Rc), F1 score (F1) and area under the receiver operating characteristic curve (AUC) were calculated.Results:In both the first and second stages, V1_True and V2_True had higher Ac, Pr, Rc, and F1 than V1_False and V2_False in both the training set and validation set; and the V1_True model outperformed the other three models in classification. The classification Ac of the V1_True model in the validation set was higher than that of radiologists in the first stage [95.71% (335/350) vs. 90.29% (316/350)] and in the second stage [93.43% (327/350) vs. 87.14% (305/350)]. The recognition time of V1_True model′s in the validation set of 350 bedside chest radiographs was significantly less than that of the radiologists (mean: 17 s vs. 300 min).Conclusions:V1_True is an optimal MobileNet model for classifying bedside chest radiographs. The application of this model in clinical practice may help to accurately identify the information of lung lesions from bedside chest radiographs in time, and may improve the work efficiency in the radiology department.

Inflammatory bowel disease (IBD) is a complex, immune-mediated gastrointestinal disorder with ill-defined etiology, multifaceted diagnostic criteria, and unpredictable treatment response. Innovations in IBD diagnostics include developments in genomic sequencing, molecular analytics and gastrointestinal imaging, and have leverage above large data platforms to develop clinically meaningful artificial intelligence (AI) tools. Artificial intelligence, through machine learning, facilitates the interpretation of large arrays of data, and provides insight to improve IBD outcomes, especially with the application of deep learning. This review aims to summarize the recent progress of artificial intelligence in inflammatory bowel disease and its applications.

This study aimed to evaluate the predictive value of computed tomography (CT) texture features in the treatment response of patients with advanced pancreatic cancer (APC) receiving palliative chemotherapy. This study enrolled 84 patients with APC treated with first-line chemotherapy and conducted texture analysis on primary pancreatic tumors. 59 patients and 25 were randomly assigned to the training and validation cohorts at a ratio of 7:3. The treatment response to chemotherapy was evaluated according to the Response Evaluation Criteria in Solid Tumors (RECIST1.1). The patients were divided into progressive and non-progressive groups. The least absolute shrinkage selection operator (LASSO) was applied for feature selection in the training cohort and a radiomics signature (RS) was calculated. A nomogram was developed based on a multivariate logistic regression model incorporating the RS and carbohydrate antigen 19-9 (CA19-9), and was internally validated using the C-index and calibration plot. We performed the decision curve analysis (DCA) and clinical impact curve analysis to reflect the clinical utility of the nomogram. The nomogram was further externally confirmed in the validation cohort. The multivariate logistic regression analysis indicated that the RS and CA19-9 were independent predictors (P < 0.05), and a trend was found for chemotherapy between progressive and non-progressive groups. The nomogram incorporating RS, CA19-9 and chemotherapy showed favorable discriminative ability in the training (C-index = 0.802) and validation (C-index = 0.920) cohorts. The nomogram demonstrated favorable clinical utility. The RS of significant texture features was significantly associated with the early treatment effect of patients with APC treated with chemotherapy. Based on the RS, CA19-9 and chemotherapy, the nomogram provided a promising way to predict chemotherapeutic effects for APC patients.

It is crucial to diagnose breast cancer early and accurately to optimize treatment. Presently, most deep learning models used for breast cancer detection cannot be used on mobile phones or low-power devices. This study intended to evaluate the capabilities of MobileNetV1 and MobileNetV2 and their fine-tuned models to differentiate malignant lesions from benign lesions in breast dynamic contrast-enhanced magnetic resonance images (DCE-MRI).

目的 探讨基于MobileNetV2深度迁移学习(DTL)对乳腺X线摄影乳腺影像报告和数据系统(BI-RADS)4类病变降级分类的价值.方法 将良性组、恶性组图像分别按照9:1随机分为训练集(良性组9346幅,恶性组4421幅)和测试集(良性组1038幅,恶性组491幅).通过模型微调构建基于MobileNetV2的DTL模型,并对9346幅良性组和4421幅恶性组乳腺X线图像进行学习,另外搜集由5位影像科医师报告的乳腺X线BI-RADS 4类病变患者共382例作验证集.每个病变均选择头尾位(CC位)和内外斜位(MLO位)两幅图像进行验证,如有1幅图像归类正确,则判断为该例归类正确.以验证集准确率、召回率、F1评分及受试者工作特征曲线(ROC)曲线下面积(AUC)作为DTL模型的性能指标.结果 模型在训练集和测试集准确率分别为100％、98％.在验证集准确率、召回率、F1评分及AUC分别为0.91、0.91、0.91和0.91.模型对BI-RADS 4A、4B、4C类病变降级比例分别为87.7％、80.2％和75.2％.对BI-RADS 4类病变总体降级比例为81.9％,且DTL模型与病理组织学在对乳腺X线摄影良恶性病变的分类诊断结果差异无统计学意义(P=0.206).结论 基于MobileNetV2的DTL模型是乳腺X线摄影BI-RADS 4类病变降级的有效方法.

目的:探讨基于DenseNet201深度迁移学习(DTL)在改善乳腺MRI BI-RADS 3类病变分类诊断的潜力.方法:采用基于DenseNet201的DTL模型对11 256幅良性组和5 448幅恶性组乳腺DCE-MRI图像进行学习.将良性组、恶性组图像按照9:1随机分为训练集(良性组:10 146幅;恶性组4908幅)和测试集(良性组:1110幅;恶性组:540幅).收集乳腺MRI报告BI-RADS 3类患者201例作验证集,其中197例良性,4例恶性.以验证集准确度、召回率、F1评分及ROC曲线下面积作为性能指标.结果:训练集和测试集最高准确度分别为100.00％和99.52％.验证集平均准确度、平均召回率、平均F1评分及ROC曲线下面积分别为98.00％、0.98、0.98和0.98.DTL模型对201例病变归类正确199例,归类准确度99.00％.结论:基于DenseNet201的DTL模型是提高乳腺MRIBI-RADS 3类病变良恶性诊断准确性的有效方法.

Abstract Background Dispensing Error occurs frequently in outpatient pharmacies. The aim of this article is to explore the value of the YOLO-V5 deep learning algorithm for object detection in packaging drugs from outpatient pharmacies and to develop an artificial intelligence assist pharmacist drug dispensation system (AI-APDDS) capable of completely reducing or eliminating dispensing errors. Methods A total of 1784 images from 136 different packaging drugs were collected and labeled to form a deep learning dataset. The dataset was split into training and validation sets at a ratio of 3:1. The YOLO-V5 deep-learning algorithm was trained using images from our dataset (training epochs:1000, batch size:4, learning rate:0.01). The values of precision (Pr) and mean average precision(mAP) were used as measures for model performance evaluation. Results Pr in the training set was, for all models, equal to 1.00. The mAP_0.5 of YOLO-V5x was 0.992, which was higher than that of YOLO-V5n (0.975), YOLO-V5s (0.976), YOLO-V5m (0.978), and YOLO-V5l (0.983). The mAP_0.5:0.95 of YOLO-V5x was 0.955, which was higher than those of YOLO-V5n (0.916), YOLO-V5s (0.920), YOLO-V5m (0.921) and YOLO-V5l (0.935). The training time and model size were 92.67 hours and 465MB, respectively, for YOLO-V5x, which were the highest among the four models. The speed of detection for one image was 6.0ms for YOLO-V5n, which was the fastest among the four models. Conclusions It can identify the packaging drugs accurately for five submodels of YOLO-V5, it is feasible to implement the artificial-intelligence assisted drug dispensation system for pharmacist to achieve “zero” dispensing error.

In order to achieve better performance, artificial intelligence is used in breast cancer diagnosis. In this study, we evaluated the efficacy of different fine-tuning strategies of deep transfer learning (DTL) based on the DenseNet201 model to differentiate malignant from benign lesions on breast dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI). We collected 4260 images of benign lesions and 4140 images of malignant lesions of the breast pertaining to pathologically confirmed cases. The benign and malignant groups was randomly divided into a training set and a testing set at a ratio of 9:1. A DTL model based on the DenseNet201 model was established, and the effectiveness of 4 fine-tuning strategies (S0: strategy 0, S1: strategy; S2: strategy; and S3: strategy) was compared. Additionally, DCE-MRI images of 48 breast lesions were selected to verify the robustness of the model. Ten images were obtained for each lesion. The classification was considered correct if more than 5 images were correctly classified. The metrics for model performance evaluation included accuracy (Ac) in the training and testing sets, precision (Pr), recall rate (Rc), f1 score (f1), and area under the receiver operating characteristic curve (AUROC) in the validation set. The Ac of the 4 fine-tuning strategies reached 100.00% in the training set. The S2 strategy exhibited good convergence in the testing set. The Ac of S2 was 98.01% in the testing set, which was higher than those of S0 (93.10%), S1 (90.45%), and S3 (93.90%). The average classification Pr, Rc, f1, and AUROC of S2 in the validation set were (89.00%, 80.00%, 0.81, and 0.79, respectively) higher than those of S0 (76.00%, 67.00%, 0.69, and 0.65, respectively), S1 (60.00%, 60.00%, 0.60, 0.66, and respectively), and S3 (77.00%, 73.00%, 0.74, 0.72, respectively). The degree of coincidence between S2 and the histopathological method for differentiating between benign and malignant breast lesions was high (κ = 0.749). The S2 strategy can improve the robustness of the DenseNet201 model in relatively small breast DCE-MRI datasets, and this is a reliable method to increase the Ac of discriminating benign from malignant breast lesions on DCE-MRI.

Abstract Background To develop and validate a radiomics nomogram to determine the primary site of liver metastases from gastric and colorectal cancer based on texture analysis. Methods We enrolled 555 patients with liver metastases, comprising 277 with gastric cancer (GC) and 278 with colorectal cancer (CRC), and randomly divided them into the training and validation cohorts at a ratio of 7:3. Radiomics features were extracted from venous phase computed tomography (CT) scans. Univariate analysis revealed three texture features potentially correlated with the identification model (p < 0.1). The selected features were combined with their coefficients to construct the radiomics signature (RS). A nomogram was developed with the RS (p = 0.02) and clinical features (p < 0.05). Nomogram performance was determined by its discriminative ability and clinical utility. Results The multivariable logistic regression model included gender, blood hemoglobin (HGB), carcinoembryonic antigen (CEA), and RS. The nomogram showed great discrimination in the training cohort (AUC = 0.71) and in the validation cohort (AUC = 0.78). The nomogram also demonstrated favorable clinical consistency. Conclusion This study presents a radiomics nomogram incorporating RS and clinical features with great discrimination and high clinical value for the differentiation of liver metastases originating from gastric or colorectal cancer.