We introduce MedMNIST v2, a large-scale MNIST-like dataset collection of standardized biomedical images, including 12 datasets for 2D and 6 datasets for 3D. All images are pre-processed into a small size of 28 × 28 (2D) or 28 × 28 × 28 (3D) with the corresponding classification labels so that no background knowledge is required for users. Covering primary data modalities in biomedical images, MedMNIST v2 is designed to perform classification on lightweight 2D and 3D images with various dataset scales (from 100 to 100,000) and diverse tasks (binary/multi-class, ordinal regression, and multi-label). The resulting dataset, consisting of 708,069 2D images and 9,998 3D images in total, could support numerous research/educational purposes in biomedical image analysis, computer vision, and machine learning. We benchmark several baseline methods on MedMNIST v2, including 2D/3D neural networks and open-source/commercial AutoML tools. The data and code are publicly available at https://medmnist.com/.

There have been considerable debates over 2D and 3D representation learning on 3D medical images. 2D approaches could benefit from large-scale 2D pretraining, whereas they are generally weak in capturing large 3D contexts. 3D approaches are natively strong in 3D contexts, however few publicly available 3D medical dataset is large and diverse enough for universal 3D pretraining. Even for hybrid (2D + 3D) approaches, the intrinsic disadvantages within the 2D/3D parts still exist. In this study, we bridge the gap between 2D and 3D convolutions by reinventing the 2D convolutions. We propose ACS (axial-coronal-sagittal) convolutions to perform natively 3D representation learning, while utilizing the pretrained weights on 2D datasets. In ACS convolutions, 2D convolution kernels are split by channel into three parts, and convoluted separately on the three views (axial, coronal and sagittal) of 3D representations. Theoretically, ANY 2D CNN (ResNet, DenseNet, or DeepLab) is able to be converted into a 3D ACS CNN, with pretrained weight of a same parameter size. Extensive experiments validate the consistent superiority of the pretrained ACS CNNs, over the 2D/3D CNN counterparts with/without pretraining. Even without pretraining, the ACS convolution can be used as a plug-and-play replacement of standard 3D convolution, with smaller model size and less computation.

Iron is one of the essential mineral elements for the human body and this nutrient deficiency is a worldwide public health problem. Iron is essential in oxygen transport, participates in many enzyme systems in the body, and is an important trace element in maintaining basic cellular life activities. Iron also plays an important role in collagen synthesis and vitamin D metabolism. Therefore, decrease in intracellular iron can lead to disturbance in the activity and function of osteoblasts and osteoclasts, resulting in imbalance in bone homeostasis and ultimately bone loss. Indeed, iron deficiency, with or without anemia, leads to osteopenia or osteoporosis, which has been revealed by numerous clinical observations and animal studies. This review presents current knowledge on iron metabolism under iron deficiency states and the diagnosis and prevention of iron deficiency and iron deficiency anemia (IDA). With emphasis, studies related to iron deficiency and bone loss are discussed, and the potential mechanisms of iron deficiency leading to bone loss are analyzed. Finally, several measures to promote complete recovery and prevention of iron deficiency are listed to improve quality of life, including bone health.

Many studies have indicated that static magnetic fields (SMFs) have positive effects on bone tissue, including bone formation and bone healing process. Evaluating the effects of SMFs on bone cell (especially osteoblast) function and exploring the mechanism, which is critical for understanding the possible risks or benefits from SMFs to the balance of bone remodeling. Iron and magnetic fields have the natural relationship, and iron is an essential element for normal bone metabolism. Iron overload or deficiency can cause severe bone disorders including osteoporosis. However, there are few reports regarding the role of iron in the regulation of bone formation under SMFs. In this study, hypomagnetic field (HyMF) of 500 nT, moderate SMF (MMF) of 0.2 T, and high SMF (HiMF) of 16 T were used to investigate how osteoblast (MC3T3-E1) responses to SMFs and iron metabolism of osteoblast under SMFs. The results showed that SMFs did not pose severe toxic effects on osteoblast growth. During cell proliferation, iron content of osteoblast MC3T3-E1 cells was decreased in HyMF, but was increased in MMF and HiMF after exposure for 48 h. Compared to untreated control (i.e., geomagnetic field, GMF), HyMF and MMF exerted deleterious effects on osteoblast differentiation by simultaneously retarding alkaline phosphatase (ALP) activity, mineralization and calcium deposition. However, when exposed to HiMF of 16 T, the differentiation potential showed the opposite tendency with enhanced mineralization. Iron level was increased in HyMF, constant in MMF and decreased in HiMF during cell differentiation. In addition, the mRNA expression of transferrin receptor 1 (TFR1) was promoted by HyMF but was inhibited by HiMF. At the same time, HiMF of 16 T and MMF of 0.2 T increased the expression of ferroportin 1 (FPN1). In conclusion, these results indicated that osteoblast differentiation can be regulated by altering the strength of the SMF, and iron is possibly involved in this process.

This article reviews the current development of nanoparticles on bone tissue engineering & bone cells and potential mechanisms of the effects of NPs on bone cell to potentially reveal new therapeutic strategies to improve the effectiveness of bone regeneration therapy.