Deep Learning-Based Classification and Semantic Segmentation of Lung Tuberculosis Lesions in Chest X-ray Images.

We present a deep learning (DL) network-based approach for detecting and semantically segmenting two specific types of tuberculosis (TB) lesions in chest X-ray (CXR) images. In the proposed method, we use a basic U-Net model and its enhanced versions to detect, classify, and segment TB lesions in CXR images. The model architectures used in this study are U-Net, Attention U-Net, U-Net++, Attention U-Net++, and pyramid spatial pooling (PSP) Attention U-Net++, which are optimized and compared based on the test results of each model to find the best parameters. Finally, we use four ensemble approaches which combine the top five models to further improve lesion classification and segmentation results. In the training stage, we use data augmentation and preprocessing methods to increase the number and strength of lesion features in CXR images, respectively. Our dataset consists of 110 training, 14 validation, and 98 test images. The experimental results show that the proposed ensemble model achieves a maximum mean intersection-over-union (MIoU) of 0.70, a mean precision rate of 0.88, a mean recall rate of 0.75, a mean F1-score of 0.81, and an accuracy of 1.0, which are all better than those of only using a single-network model. The proposed method can be used by clinicians as a diagnostic tool assisting in the examination of TB lesions in CXR images.

Clinical value and application of a novel nomogram containing inflammatory, nutritional and clinical markers in predicting overall survival of stage II/III gastric cancer patients after radical resection: a bi-centered retrospective study of 2,443 patients.

OBJECTIVES: We aimed to identify nutritional, inflammatory and clinical indicators associated with stage II/III gastric cancer in patients, and construct a nomogram model for accurate prediction of prognosis of patients. METHODS: We retrospectively recruited stage II/III gastric cancer (GC) patients who underwent radical gastrectomy at Fudan University Shanghai Cancer Center, from 2012 to 2019. The patients were randomly divided into training and internal validation sets, and then the Maximum log-rank statistic method was used to determine the optimal cut-off value. Next, we performed univariate and multivariate Cox regression analyses to identify independent risk factors associated with overall survival (OS). These were subsequently used to develop a nomogram model. We validated this model in patients with stage II/III gastric cancer (from 2010 to 2019) at Guangxi Medical University Affiliated Tumor Hospital. RESULTS: A total of 2,443 patients met our inclusion criteria and were therefore included in our study. Patients from Fudan University Shanghai Cancer Center were randomly divided into training (n=1725) and internal validation (n=430) sets, while those from Guangxi Medical University Affiliated Tumor Hospital were used as the external validation set (n=288). Results from univariate and multivariate Cox regression analyses revealed that age (adjusted HR, 1.23; 95% CI, 1.05-1.44; P=0.012), TNM stage (adjusted HR, 3.62; 95% CI, 2.79-4.68; P<0.001), CEA (adjusted HR, 1.40; 95% CI, 1.14-1.71; P<0.001), CA199 (adjusted HR, 1.47; 95% CI, 1.21-1.79; P<0.001), and Prognostic Nutritional Index (PNI, adjusted HR, 0.81; 95% CI, 0.67-0.98; P=0.026) were independent prognostic factors for OS in the training set. The established nomogram model, with a C-index of 0.67, had 3- and 5-year Area under Curve (AUC) values of 0.719 and 0.714, respectively. Notably, the model effectively distinguished patients' OS in both the internal (P<0.001) and external (P<0.001) datasets. CONCLUSIONS: PNI is an independent prognostic factor for stage II/III GC patients after radical resection. The established novel nomogram model, based on nutritional, inflammatory and clinical indicators, can accurately and efficiently predict prognosis of stage II/III GC patients.

Preoperative Predictors of Lymph Node Metastasis in Colon Cancer.

BACKGROUND: Lymph node metastasis (LNM) is a well-established prognostic factor for colon cancer. Preoperative LNM evaluation is relevant for planning colon cancer treatment. The aim of this study was to construct and evaluate a nomogram for predicting LNM in primary colon cancer according to pathological features. PATIENTS AND METHODS: Six-hundred patients with clinicopathologically confirmed colon cancer (481 cases in the training set and 119 cases in the validation set) were enrolled in the Affiliated Cancer Hospital of Guangxi Medical University from January 2010 to December 2019. The expression of molecular markers (p53 and ß-catenin) was determined by immunohistochemistry. Multivariate logistic regression was used to screen out independent risk factors, and a nomogram was established. The accuracy and discriminability of the nomogram were evaluated by consistency index and calibration curve. RESULTS: Univariate logistic analysis revealed that LNM in colon cancer is significantly correlated (P <0.05) with tumor size, grading, stage, preoperative carcinoembryonic antigen (CEA) level, and peripheral nerve infiltration (PNI). Multivariate logistic regression analysis confirmed that CEA, grading, and PNI were independent prognostic factors of LNM (P <0.05). The nomogram for predicting LNM risk showed acceptable consistency and calibration capability in the training and validation sets. CONCLUSIONS: Preoperative CEA level, grading, and PNI were independent risk factor for LNM. Based on the present parameters, the constructed prediction model of LNM has potential application value.

A review of current and emerging approaches for Quaternary marine sediment dating.

Dating methodologies for Quaternary marine sediments play increasingly important roles in the reconstruction of paleoenvironments and paleoclimate in (paleo)oceanography. Previous reviews or studies have focused mainly on one or two methodologies, and their applications in one specific environment. With the continuing technological and methodological advances in different methods over the past few decades, an up-to-date comparison of the pros and cons of each dating methodology is needed to clearly understand their applications in marine geoscience research. In this review, we first briefly summarized the common methods of absolute dating and relative dating. These are (1) radioisotope dating with different half-lives using natural nuclides of 234Th, 210Pb, 230Th, and 226Ra, cosmogenic nuclides of 7Be, 14C, 10Be, 32Si, 26Al, 36Cl and 21Ne, and the artificial radionuclides of 137Cs, 239, 240Pu, 241Am and 129I that have been induced by atmospheric nuclear tests, accidents in nuclear plants, and discharges of radioactive wastes; (2) radiation exposure dating of luminescence and electron paramagnetic resonance (ESR) dating; and (3) stratigraphic dating of δ18O and paleomagnetic sequence. Applications and limitations from the marine terraces, estuaries, to hadal trenches have been summarized to each technique in the study of Quaternary marine geoscience extending from the Anthropocene through the Pleistocene. Finally, we introduced some emerging event dating methods, namely the arrivals of microplastics, mercury isotopes, and organic pollutant deposition that all appeared after the industrial resolution in our now changing ocean influenced by acidification, global warming, and anthropogenic activities. We ended by discussing future perspectives for reliable and high-resolution chronology by interdisciplinary methods including computer programming to better understand the natural geological evolution and predict the future changes in earth science.

Evaluation of Red Cell Distribution Width to Lymphocyte Ratio as Potential Biomarker for Detection of Colorectal Cancer.

BACKGROUND AND AIM: Colorectal cancer (CRC) is the third most lethal cancer globally. This study sought to determine the feasibility of using red cell distribution width-to-lymphocyte ratio (RLR) as a tool to facilitate CRC detection. METHODS: Seventy-eight healthy controls, 162 patients diagnosed with CRC, and 94 patients with colorectal polyps (CP) from June 2017 to October 2018 were retrospectively reviewed. Clinical data were obtained to analyze preoperative RLR level, and receiver operating characteristic (ROC) curve analysis was performed to estimate the potential role of RLR as a CRC biomarker. RESULTS: RLR was higher in patients with CRC than in healthy participants (P < 0.05). ROC analysis indicated that combined detection of RLR and CEA appears to be a more effective marker to distinguish among controls, CP, and CRC patients, yielding 56% sensitivity and 90% specificity. RLR levels were significantly greater in those who had more advanced TNM stages (P < 0.05) and patients with distant metastasis stages (P < 0.05). CONCLUSIONS: RLR might serve as a potential biomarker for CRC diagnosis.