Receive a weekly summary and discussion of the top papers of the week by leading researchers in the field.

Dermatology Dermatology

A call for implementing augmented intelligence in pediatric dermatology.

In Pediatric dermatology

Augmented intelligence (AI), the combination of artificial based intelligence with human intelligence from a practitioner, has become an increased focus of clinical interest in the field of dermatology. Technological advancements have led to the development of deep-learning based models to accurately diagnose complex dermatological diseases such as melanoma in adult datasets. Models for pediatric dermatology remain scarce, but recent studies have shown applications in the diagnoses of facial infantile hemangiomas and X-linked hypohidrotic ectodermal dysplasia; however, we see unmet needs in other complex clinical scenarios and rare diseases, such as diagnosing squamous cell carcinoma in patients with epidermolysis bullosa. Given the still limited number of pediatric dermatologists, especially in rural areas, AI has the potential to help overcome health disparities by helping primary care physicians treat or triage patients.

Issa Christopher J, Reimer-Taschenbrecker Antonia, Paller Amy S

2023-Mar-23

artificial intelligence, convolutional neural networks, deep learning, health disparities, pediatric dermatology

General General

A Long-term Consistent Artificial Intelligence and Remote Sensing-based Soil Moisture Dataset.

In Scientific data

The Consistent Artificial Intelligence (AI)-based Soil Moisture (CASM) dataset is a global, consistent, and long-term, remote sensing soil moisture (SM) dataset created using machine learning. It is based on the NASA Soil Moisture Active Passive (SMAP) satellite mission SM data and is aimed at extrapolating SMAP-like quality SM back in time using previous satellite microwave platforms. CASM represents SM in the top soil layer, and it is defined on a global 25 km EASE-2 grid and for 2002-2020 with a 3-day temporal resolution. The seasonal cycle is removed for the neural network training to ensure its skill is targeted at predicting SM extremes. CASM comparison to 367 global in-situ SM monitoring sites shows a SMAP-like median correlation of 0.66. Additionally, the SM product uncertainty was assessed, and both aleatoric and epistemic uncertainties were estimated and included in the dataset. CASM dataset can be used to study a wide range of hydrological, carbon cycle, and energy processes since only a consistent long-term dataset allows assessing changes in water availability and water stress.

Skulovich Olya, Gentine Pierre

2023-Mar-22

Surgery Surgery

Breast Tumor Classification using Short-ResNet with Pixel-based Tumor Probability Map in Ultrasound Images.

In Ultrasonic imaging

Breast cancer is the most common form of cancer and is still the second leading cause of death for women in the world. Early detection and treatment of breast cancer can reduce mortality rates. Breast ultrasound is always used to detect and diagnose breast cancer. The accurate breast segmentation and diagnosis as benign or malignant is still a challenging task in the ultrasound image. In this paper, we proposed a classification model as short-ResNet with DC-UNet to solve the segmentation and diagnosis challenge to find the tumor and classify benign or malignant with breast ultrasonic images. The proposed model has a dice coefficient of 83% for segmentation and achieves an accuracy of 90% for classification with breast tumors. In the experiment, we have compared with segmentation task and classification result in different datasets to prove that the proposed model is more general and demonstrates better results. The deep learning model using short-ResNet to classify tumor whether benign or malignant, that combine DC-UNet of segmentation task to assist in improving the classification results.

Wang You-Wei, Kuo Tsung-Ter, Chou Yi-Hong, Su Yu, Huang Shing-Hwa, Chen Chii-Jen

2023-Mar-23

breast cancer, convolutional neural network, deep learning, tumor classification, ultrasound

General General

Boundary-oriented Network for Automatic Breast Tumor Segmentation in Ultrasound Images.

In Ultrasonic imaging

Breast cancer is considered as the most prevalent cancer. Using ultrasound images is a momentous clinical diagnosis method to locate breast tumors. However, accurate segmentation of breast tumors remains an open problem due to ultrasound artifacts, low contrast, and complicated tumor shapes in ultrasound images. To address this issue, we proposed a boundary-oriented network (BO-Net) for boosting breast tumor segmentation in ultrasound images. The BO-Net boosts tumor segmentation performance from two perspectives. Firstly, a boundary-oriented module (BOM) was designed to capture the weak boundaries of breast tumors by learning additional breast tumor boundary maps. Second, we focus on enhanced feature extraction, which takes advantage of the Atrous Spatial Pyramid Pooling (ASPP) module and Squeeze-and-Excitation (SE) block to obtain multi-scale and efficient feature information. We evaluate our network on two public datasets: Dataset B and BUSI. For the Dataset B, our network achieves 0.8685 in Dice, 0.7846 in Jaccard, 0.8604 in Precision, 0.9078 in Recall, and 0.9928 in Specificity. For the BUSI dataset, our network achieves 0.7954 in Dice, 0.7033 in Jaccard, 0.8275 in Precision, 0.8251 in Recall, and 0.9814 in Specificity. Experimental results show that BO-Net outperforms the state-of-the-art segmentation methods for breast tumor segmentation in ultrasound images. It demonstrates that focusing on boundary and feature enhancement creates more efficient and robust breast tumor segmentation.

Zhang Mengmeng, Huang Aibin, Yang Debiao, Xu Rui

2023-Mar-23

U-Net, breast tumors, deep learning, image segmentation, ultrasound images

General General

ANPELA: Significantly Enhanced Quantification Tool for Cytometry-Based Single-Cell Proteomics.

In Advanced science (Weinheim, Baden-Wurttemberg, Germany)

ANPELA is widely used for quantifying traditional bulk proteomic data. Recently, there is a clear shift from bulk proteomics to the single-cell ones (SCP), for which powerful cytometry techniques demonstrate the fantastic capacity of capturing cellular heterogeneity that is completely overlooked by traditional bulk profiling. However, the in-depth and high-quality quantification of SCP data is still challenging and severely affected by the large numbers of quantification workflows and extreme performance dependence on the studied datasets. In other words, the proper selection of well-performing workflow(s) for any studied dataset is elusory, and it is urgently needed to have a significantly enhanced and accelerated tool to address this issue. However, no such tool is developed yet. Herein, ANPELA is therefore updated to its 2.0 version (https://idrblab.org/anpela/), which is unique in providing the most comprehensive set of quantification alternatives (>1000 workflows) among all existing tools, enabling systematic performance evaluation from multiple perspectives based on machine learning, and identifying the optimal workflow(s) using overall performance ranking together with the parallel computation. Extensive validation on different benchmark datasets and representative application scenarios suggest the great application potential of ANPELA in current SCP research for gaining more accurate and reliable biological insights.

Zhang Ying, Sun Huaicheng, Lian Xichen, Tang Jing, Zhu Feng

2023-Mar-22

cell population identification, comprehensive assessment, parallel computing, protein quantification, single-cell proteomics, trajectory inference

Surgery Surgery

Identification of copper metabolism-related subtypes and establishment of the prognostic model in ovarian cancer.

In Frontiers in endocrinology ; h5-index 55.0

BACKGROUND : Ovarian cancer (OC) is one of the most common and most malignant gynecological malignancies in gynecology. On the other hand, dysregulation of copper metabolism (CM) is closely associated with tumourigenesis and progression. Here, we investigated the impact of genes associated with copper metabolism (CMRGs) on the prognosis of OC, discovered various CM clusters, and built a risk model to evaluate patient prognosis, immunological features, and therapy response.

METHODS : 15 CMRGs affecting the prognosis of OC patients were identified in The Cancer Genome Atlas (TCGA). Consensus Clustering was used to identify two CM clusters. lasso-cox methods were used to establish the copper metabolism-related gene prognostic signature (CMRGPS) based on differentially expressed genes in the two clusters. The GSE63885 cohort was used as an external validation cohort. Expression of CM risk score-associated genes was verified by single-cell sequencing and quantitative real-time PCR (qRT-PCR). Nomograms were used to visually depict the clinical value of CMRGPS. Differences in clinical traits, immune cell infiltration, and tumor mutational load (TMB) between risk groups were also extensively examined. Tumour Immune Dysfunction and Rejection (TIDE) and Immune Phenotype Score (IPS) were used to validate whether CMRGPS could predict response to immunotherapy in OC patients.

RESULTS : In the TCGA and GSE63885 cohorts, we identified two CM clusters that differed significantly in terms of overall survival (OS) and tumor microenvironment. We then created a CMRGPS containing 11 genes to predict overall survival and confirmed its reliable predictive power for OC patients. The expression of CM risk score-related genes was validated by qRT-PCR. Patients with OC were divided into low-risk (LR) and high-risk (HR) groups based on the median CM risk score, with better survival in the LR group. The 5-year AUC value reached 0.74. Enrichment analysis showed that the LR group was associated with tumor immune-related pathways. The results of TIDE and IPS showed a better response to immunotherapy in the LR group.

CONCLUSION : Our study, therefore, provides a valuable tool to further guide clinical management and tailor the treatment of patients with OC, offering new insights into individualized treatment.

Zhao Songyun, Zhang Xin, Gao Feng, Chi Hao, Zhang Jinhao, Xia Zhijia, Cheng Chao, Liu Jinhui

2023

OC, Tumor microenvironment, copper metabolism, immunotherapy, machine learning, risk score signature