We collected breast cancer ultrasound images diagnosed breast cancer in radiology of Renmin Hospital of Wuhan University from December 2020 to December 2022. The dataset contains 927 images including benign and malignant cancer from patients with the range of age 17 to 79. Meanwhile, the ethical approval of this study was approved by the ethics committee of Renmin Hospital of Wuhan University (WDRY2022-K217). Each image contains tumor regions. At the same time, the dataset has different tumor area and morphology features, including contrast, brightness, and fuzzy. In summary, we collected a breast ultrasound image for the segmentation task.

We collected breast cancer ultrasound images diagnosed breast cancer in radiology of Renmin Hospital of Wuhan University from December 2020 to December 2022. The dataset contains 927 images including benign and malignant cancer from patients with the range of age 17 to 79. Meanwhile, the ethical approval of this study was approved by the ethics committee of Renmin Hospital of Wuhan University (WDRY2022-K217). Each image contains tumor regions. At the same time, the dataset has different tumor area and morphology features, including contrast, brightness, and fuzzy. In summary, we collected a breast ultrasound image for the segmentation task.

We used a weak coherent light source with a central wavelength of 840 nm and a full width at half maximum (FWHM) of 25.429 nm to generate the weak coherent light. A 50:50 fiber coupler split the weak coherent light and directed it toward the optical delay line and simulated sample. A photodetector finally detected the optical signals. A transparent glass plate simulated the sample and generated the weak coherent interference heterodyne detector signal.

We used a weak coherent light source with a central wavelength of 840 nm and a full width at half maximum (FWHM) of 25.429 nm to generate the weak coherent light. A 50:50 fiber coupler split the weak coherent light and directed it toward the optical delay line and simulated sample. A photodetector finally detected the optical signals. A transparent glass plate simulated the sample and generated the weak coherent interference heterodyne detector signal.

The dataset includes the weak coherent interference heterdyne detection signal in the thesis "Weak coherent light interference heterodyne detection based on time-domain signal analysis". The weak coherent light-generated photodetection signals were incident on the photodetectors, which were further converted into analog voltage signals. A data acquisition card sampled the analog signals at a rate of 250 kSa/s and converted them into digital signals

The dataset includes the weak coherent interference heterdyne detection signal in the thesis "Weak coherent light interference heterodyne detection based on time-domain signal analysis". The weak coherent light-generated photodetection signals were incident on the photodetectors, which were further converted into analog voltage signals. A data acquisition card sampled the analog signals at a rate of 250 kSa/s and converted them into digital signals

BackgroundOsteoporosis presents a significant global health challenge, compromising bone quality and elevating fracture susceptibility. While dual-energy x-ray absorptiometry (DXA) stands as the gold standard for bone mineral density (BMD) assessment and osteoporosis diagnosis, its costliness and complexity impede widespread screening adoption. Predictive modeling of BMD, leveraging genetic and clinical data, emerges as a more viable and cost-effective approach for osteoporosis and fracture risk evaluation.Methods and FindingsWe developed BMD prediction models for the femoral neck (FNK) and lumbar spine (SPN) using various methods within a UK Biobank (UKBB) training set comprising 17,964 individuals from the British white population. Models based on Regression with Least Absolute Shrinkage and Selection Operator (LASSO), selected based on the coefficient of determination (R2) from a model selection subset of 5,973 individuals from the British white population, underwent testing on five UKBB test sets and 12 independent cohorts of diverse ancestries, totaling over 15,000 individuals. Furthermore, we assessed the correlation of predicted BMDs with fragility fractures in a distinct case-control set of over 287,000 participants lacking DXA-BMDs in the UKBB of the European white population.Incorporating genetic factors marginally improved predictions, capturing an additional 2.3% variation for FNK-BMD and 3% for SPN-BMD over clinical factors alone. Predicted BMDs exhibited significant associations with fragility fracture risk in the European white population. Nonetheless, the predictive model's performance varied between the UKBB population of other ethnic groups and the independent cohorts.ConclusionsOur study yields novel insights into predicting osteoporosis and fracture risk. Genetic factors enhance BMD predictive performance beyond clinical factors alone. Adjusting inclusion thresholds for genetic variants (e.g., 5×10^(-6) or 5×10^(-7)) rather than solely considering genome-wide association study (GWAS)-significant variants may further refine the model's explanatory power for BMD variations. This study also underscores the imperative for training models on diverse population to bolster predictive performance across various ethnic and geographical populations.

BackgroundOsteoporosis presents a significant global health challenge, compromising bone quality and elevating fracture susceptibility. While dual-energy x-ray absorptiometry (DXA) stands as the gold standard for bone mineral density (BMD) assessment and osteoporosis diagnosis, its costliness and complexity impede widespread screening adoption. Predictive modeling of BMD, leveraging genetic and clinical data, emerges as a more viable and cost-effective approach for osteoporosis and fracture risk evaluation.Methods and FindingsWe developed BMD prediction models for the femoral neck (FNK) and lumbar spine (SPN) using various methods within a UK Biobank (UKBB) training set comprising 17,964 individuals from the British white population. Models based on Regression with Least Absolute Shrinkage and Selection Operator (LASSO), selected based on the coefficient of determination (R2) from a model selection subset of 5,973 individuals from the British white population, underwent testing on five UKBB test sets and 12 independent cohorts of diverse ancestries, totaling over 15,000 individuals. Furthermore, we assessed the correlation of predicted BMDs with fragility fractures in a distinct case-control set of over 287,000 participants lacking DXA-BMDs in the UKBB of the European white population.Incorporating genetic factors marginally improved predictions, capturing an additional 2.3% variation for FNK-BMD and 3% for SPN-BMD over clinical factors alone. Predicted BMDs exhibited significant associations with fragility fracture risk in the European white population. Nonetheless, the predictive model's performance varied between the UKBB population of other ethnic groups and the independent cohorts.ConclusionsOur study yields novel insights into predicting osteoporosis and fracture risk. Genetic factors enhance BMD predictive performance beyond clinical factors alone. Adjusting inclusion thresholds for genetic variants (e.g., 5×10^(-6) or 5×10^(-7)) rather than solely considering genome-wide association study (GWAS)-significant variants may further refine the model's explanatory power for BMD variations. This study also underscores the imperative for training models on diverse population to bolster predictive performance across various ethnic and geographical populations.

We collected breast cancer ultrasound images diagnosed breast cancer in radiology of Renmin Hospital of Wuhan University from December 2020 to December 2022. The dataset contains 927 images including benign and malignant cancer from patients with the range of age 17 to 79. Meanwhile, the ethical approval of this study was approved by the ethics committee of Renmin Hospital of Wuhan University (WDRY2022-K217). Each image contains tumor regions. At the same time, the dataset has different tumor area and morphology features, including contrast, brightness, and fuzzy. In summary, we collected a breast ultrasound image for the segmentation task.

We collected breast cancer ultrasound images diagnosed breast cancer in radiology of Renmin Hospital of Wuhan University from December 2020 to December 2022. The dataset contains 927 images including benign and malignant cancer from patients with the range of age 17 to 79. Meanwhile, the ethical approval of this study was approved by the ethics committee of Renmin Hospital of Wuhan University (WDRY2022-K217). Each image contains tumor regions. At the same time, the dataset has different tumor area and morphology features, including contrast, brightness, and fuzzy. In summary, we collected a breast ultrasound image for the segmentation task.