Insights into heterogeneous surface induced bubble nucleation mechanisms in cellulose reinforced polylactic acid foams.

As the most abundant natural homo-polymer, cellulose has the potential to enhance polymer properties reducing the cost of raw materials. In this work, the carboxylate cellulose nanofiber (CNF-C) was selected to modify polylactic acid (PLA) foams, and the density functional theory was constructed to help analyze the foaming mechanism quantitatively. The theoretical results showed that the ordered structure, the carboxyl and the hydroxyl of CNF-C were more conducive to providing much stronger CO2 adsorption for bubble nucleation, where the predicted critical bubble size decreased and the cell density increased with the addition of CNF-C. The experimental results revealed that the CNF-C promoted the rheological properties and crystallization behaviors of PLA samples, the PLA/CNF-C foams were characterized with uniform structures, the average cell size decreased from 21.39 µm to 0.19 µm, and the cell number density increased from 2.65×1010cell/cm3 to 2.30×1014cell/cm3. Those improvements resulted in an increase of 394.0 % for the compressive strength of the prepared foams. Generally, the high-performance PLA/CNF-C foams were fabricated successfully without compromising the properties of bio-based and biodegradable, the foaming mechanism was analyzed combining theoretical results with experimental data, and it was believed to provide a guide for cellulose reinforcing biodegradable polymer materials.

Segregation of morphogenetic regulatory function of Shox2 from its cell fate guardian role in sinoatrial node development.

Shox2 plays a vital role in the morphogenesis and physiological function of the sinoatrial node (SAN), the primary cardiac pacemaker, manifested by the formation of a hypoplastic SAN and failed differentiation of pacemaker cells in Shox2 mutants. Shox2 and Nkx2-5 are co-expressed in the developing SAN and regulate the fate of the pacemaker cells through a Shox2-Nkx2-5 antagonistic mechanism. Here we show that simultaneous inactivation of Nkx2-5 in the SAN of Shox2 mutants (dKO) rescued the pacemaking cell fate but not the hypoplastic defects, indicating uncoupling of SAN cell fate determination and morphogenesis. Single-cell RNA-seq revealed that the presumptive SAN cells of Shox2-/- mutants failed to activate pacemaking program but remained in a progenitor state preceding working myocardium, while both wildtype and dKO SAN cells displayed normal pacemaking cell fate with similar cellular state. Shox2 thus acts as a safeguard but not a determinant to ensure the pacemaking cell fate through the Shox2-Nkx2-5 antagonistic mechanism, which is segregated from its morphogenetic regulatory function in SAN development.

Evaluating for unrecognized deficits in perimetry associated with functional upper eyelid malposition.

Objective: To investigate whether functional upper eyelid malposition is associated with unrecognized deficits in automated perimetry among glaucoma patients by examining patients undergoing eyelid surgery who had not been identified as requiring eyelid taping during glaucoma field testing. Methods: In this retrospective pre-post study, an automated database search followed by manual chart review was used to identify eligible patients from January 2012 to March 2020. Included patients had reliable visual field testing within two years before and after functional upper blepharoplasty or ptosis repair and no comorbid ocular diagnoses. As part of routine practice, glaucoma visual field technicians taped patients with pupil-obstructing eyelid malposition; taped examinations were excluded from analysis. Clinical and demographic characteristics, mean deviation, and pattern standard deviation were evaluated within a two year period before and after eyelid surgery. Results: The final analysis included 60 eyes of 38 patients. Change in visual field parameters after eyelid surgery did not reach statistical significance in crude or adjusted analyses. Among patients with ptosis, the margin reflex distance-1 was not associated with change in mean deviation after surgery (Pearson R2 â€‹= â€‹0.0061; P â€‹= â€‹0.700). Five of 17 eyes excluded from analysis due to unreliable pre-operative visual fields demonstrated substantial improvement after surgery. Conclusions: Functional upper eyelid malposition does not appear to cause spurious visual field abnormalities among glaucoma patients with reliable visual fields who were determined not to require eyelid taping at the time of their visual fields. Unreliable visual fields could be a sign of eyelid interference in this population.

Gastrointestinal trichobezoars in the pediatric population: a retrospective study.

BACKGROUND: Trichobezoar is an extremely rare condition characterized by a foreign body in the gastrointestinal tract (GIT) among children. The foreign body may exist in the digestive tract for several years, and it becomes evident if complications develop. The current study aimed to present 21 cases of GIT trichobezoars. METHODS: Retrospective analysis of children who were diagnosed with trichobezoars between August 2012 and December 2022. Patient demographics, clinical presentation, diagnosis, and therapy were collected and analyzed.Twenty-one patients had GIT trichobezoars. Data were collected and analyzed retrospectively. RESULTS: Twenty-one patients were identified. All patients were female. Their mean age at admission was 8.9 ± 1.9 years. Furthermore, 19 (90.5%) patients presented with abdominal pain, 16 (76.2%) with vomiting, and 13 (61.9%) with a palpable mass. Sixteen patients underwent gastroduodenoscopy. Among them, 15 had gastric trichobezoars. Moreover, 12 patients underwent computed tomography scan. Eight patients presented with gastric and small intestinal BZs, one presented with increased small intestinal contents with dilation, and one presented with abundant gastric contents. Then, 20 patients underwent surgery. Among them, five underwent laparoscopic-assisted minilaparotomy (LAML), and the rest underwent laparotomy. The results showed that 10 (50%) patients had gastric trichobezoars; 7 (35%), Rapunzel syndrome; and 3 (15%), small bowel trichobezoars. Two patients developed superficial wound infection postoperatively. One patient had a recurrent gastric trichobezoar. CONCLUSION: Trichobezoar should be considered in young girls with a history of hair eating or those with hair in the vomit or feces. Timely diagnosis and aggressive treatment are the keys to reducing complications and improving prognosis. Laparoscopic-assisted minilaparotomy is a safe, feasible, and effective surgical method for treating trichobezoars.

Effect fraction of Bletilla striata (Thunb.) Reichb.f. alleviates LPS-induced acute lung injury by inhibiting p47phox/NOX2 and promoting the Nrf2/HO-1 signaling pathway.

BACKGROUND & AIMS: The effect fraction of Bletilla striata (Thunb.) Reichb.f. (EFBS), a phenolic-rich extract, has significant protective effects on lipopolysaccharide (LPS)-induced acute lung injury (ALI), but its composition and molecular mechanisms are unclear. This study elucidated its chemical composition and possible protective mechanisms against LPS-induced ALI from an antioxidant perspective. METHODS: EFBS was prepared by ethanol extraction, enriched by polyamide column chromatography, and characterized using ultra-performance liquid chromatography/time-of-flight mass spectrometry. The LPS-induced ALI model and the RAW264.7 model were used to evaluate the regulatory effects of EFBS on oxidative stress, and transcriptome analysis was performed to explore its possible molecular mechanism. Then, the pathway by which EFBS regulates oxidative stress was validated through inhibitor intervention, flow cytometry, quantitative PCR, western blotting, and immunofluorescence techniques. RESULTS: A total of 22 compounds in EFBS were identified. The transcriptome analyses of RAW264.7 cells indicated that EFBS might reduce reactive oxygen species (ROS) production by inhibiting the p47phox/NADPH oxidase 2 (NOX2) pathway and upregulating the nuclear factor erythroid 2-related factor 2 (Nrf2)/heme oxygenase-1 (HO-1) pathway. Both in vitro and in vivo data confirmed that EFBS significantly inhibited the expression and phosphorylation of p47phox protein, thereby weakening the p47phox/NOX2 pathway and reducing ROS production. EFBS significantly increased the expression of Nrf2 in primary peritoneal macrophages and lung tissue and promoted its nuclear translocation, dose-dependent increase in HO-1 levels, and enhancement of antioxidant activity. In vitro, both Nrf2 and HO-1 inhibitors significantly reduced the scavenging effects of EFBS on ROS, further confirming that EFBS exerts antioxidant effects at least partially by upregulating the Nrf2/HO-1 pathway. CONCLUSIONS: EFBS contains abundant phenanthrenes and dibenzyl polyphenols, which can reduce ROS production by inhibiting the p47phox/NOX2 pathway and enhance ROS clearance activity by upregulating the Nrf2/HO-1 pathway, thereby exerting regulatory effects on oxidative stress and improving LPS-induced ALI.

Sketch-Supervised Histopathology Tumour Segmentation: Dual CNN-Transformer With Global Normalised CAM.

Deep learning methods are frequently used in segmenting histopathology images with high-quality annotations nowadays. Compared with well-annotated data, coarse, scribbling-like labelling is more cost-effective and easier to obtain in clinical practice. The coarse annotations provide limited supervision, so employing them directly for segmentation network training remains challenging. We present a sketch-supervised method, called DCTGN-CAM, based on a dual CNN-Transformer network and a modified global normalised class activation map. By modelling global and local tumour features simultaneously, the dual CNN-Transformer network produces accurate patch-based tumour classification probabilities by training only on lightly annotated data. With the global normalised class activation map, more descriptive gradient-based representations of the histopathology images can be obtained, and inference of tumour segmentation can be performed with high accuracy. Additionally, we collect a private skin cancer dataset named BSS, which contains fine and coarse annotations for three types of cancer. To facilitate reproducible performance comparison, experts are also invited to label coarse annotations on the public liver cancer dataset PAIP2019. On the BSS dataset, our DCTGN-CAM segmentation outperforms the state-of-the-art methods and achieves 76.68 % IOU and 86.69 % Dice scores on the sketch-based tumour segmentation task. On the PAIP2019 dataset, our method achieves a Dice gain of 8.37 % compared with U-Net as the baseline network.

Fabrication of bio-based biodegradable poly(lactic acid) (PLA) and poly(3-hydroxybutyrate-co-3-hydroxyvalerate) (PHBV) composite foams for highly efficient oil-water separation.

The open-cell bio-based biodegradable polymer foams show good application prospect in dealing with the serious environmental issue caused by oil spill and organic solvents spills, while the cell structures and hydrophobic properties of the foams limit their performance. In this work, the poly(3-hydroxybutyrate-co-3-hydroxyvalerate) (PHBV) was selected to help prepare bio-based biodegradable poly(lactic acid) (PLA) foams. Based on a two-step foaming method, the crystallization ability of different samples was regulated by the "original crystals" together with PHBV in the foaming process, where skeleton structures were provided to facilitate the open-cell structures and promote their mechanical property. As illustrated, PHBV facilitated the formation of open-cell PLA foams, where the foams displayed superior oil-water separation capacity. The maximum volume expansion ratio of the foams was 80.08, the contact angle of deionized water reached to 134.5°, the adsorption capacity for oil or organic solvents was 10.8 g/g-51.8 g/g, and the adsorption capacity for CCl4 can still maintained 83.5 % of the initial value after 10 adsorption-desorption cycles. This work not only clarified the foaming mechanism of open-cell foams, but also provided a green and simple method for preparing bio-based biodegradable foams possessing excellent oil-water separation performance.

Fiber Sedimentation and Layer-By-Layer Assembly Strategy for Designing Biomimetic Quasi-Ordered Mullite Fiber Aerogels as Extreme Conditions Thermal Insulators.

Ceramic fiber aerogels are attractive thermal insulating materials. In a thermomechanical coupling environment, however, they often show limited mechanical strength and considerably increased heat transfer which can lead to thermal runaway. In this paper, inspired by bird's nest and nacre, we demonstrate a sample strategy combining fiber sedimentation and layer-by-layer assembly to fabricate ultrastrong mullite fiber aerogels (MFAs) with quasi-ordered structures. The fibrous layers and fiber bridges are constructed in a fiber sedimentation self-assembly process. The fiber sedimentation technique optimizes the structure of the MFAs by regulating the fiber orientation. Owing to the quasi-ordered structure, the fabricated MFAs exhibit the integrated properties of high compression fatigue resistance, temperature-invariant compression resilience from -196 to 1300 °C, and low thermal conductivity (0.034 W·m-1·K-1). By deliberately pressing multilayer MFAs into a thin paper, we substantially enhance the load-bearing capacity of the MFAs and achieve large temperature differences (563 °C) between the cold and hot surfaces by using a thin layer of MFAs (3-5 mm) under the simulated high-temperature (685 °C) and high-pressure (0.9 MPa) environment test. The combination of compression resistance, mechanical flexibility, and excellent thermal insulation provides an appealing material for efficient thermal insulation in extreme environments.

DPHL v.2: An updated and comprehensive DIA pan-human assay library for quantifying more than 14,000 proteins.

A comprehensive pan-human spectral library is critical for biomarker discovery using mass spectrometry (MS)-based proteomics. DPHL v.1, a previous pan-human library built from 1,096 data-dependent acquisition (DDA) MS data of 16 human tissue types, allows quantifying of 10,943 proteins. Here, we generated DPHL v.2 from 1,608 DDA-MS data. The data included 586 DDA-MS data acquired from 18 tissue types, while 1,022 files were derived from DPHL v.1. DPHL v.2 thus comprises data from 24 sample types, including several cancer types (lung, breast, kidney, and prostate cancer, among others). We generated four variants of DPHL v.2 to include semi-tryptic peptides and protein isoforms. DPHL v.2 was then applied to two colorectal cancer cohorts. The numbers of identified and significantly dysregulated proteins increased by at least 21.7% and 14.2%, respectively, compared with DPHL v.1. Our findings show that the increased human proteome coverage of DPHL v.2 provides larger pools of potential protein biomarkers.

Metabolic shift and the effect of mitochondrial respiration on the osteogenic differentiation of dental pulp stem cells.

Background: Metabolism shifts from glycolysis to mitochondrial oxidative phosphorylation are vital during the differentiation of stem cells. Mitochondria have a direct function in differentiation. However, the metabolic shift and the effect of mitochondria in regulating the osteogenic differentiation of human dental pulp stem cells (hDPSCs) remain unclear. Methods: Human dental pulp stem cells were collected from five healthy donors. Osteogenic differentiation was induced by osteogenic induction medium. The activities of alkaline phosphatase, hexokinase, pyruvate kinase, and lactate dehydrogenase were analyzed by enzymatic activity kits. The extracellular acidification rate and the mitochondrial oxygen consumption rate were measured. The mRNA levels of COL-1, ALP, TFAM, and NRF1 were analyzed. The protein levels of p-AMPK and AMPK were detected by western blotting. Results: Glycolysis decreased after a slight increase, while mitochondrial oxidative phosphorylation continued to increase when cells grew in osteogenic induction medium. Therefore, the metabolism of differentiating cells switched to mitochondrial respiration. Next, inhibiting mitochondrial respiration with carbonyl cyanide-chlorophenylhydrazone, a mitochondrial uncoupler inhibited hDPSCs differentiation with less ALP activity and decreased ALP and COL-1 mRNA expression. Furthermore, mitochondrial uncoupling led to AMPK activation. 5-Aminoimidazole-4-carboxamide ribonucleotide, an AMPK activator, simulated the effect of mitochondrial uncoupling by inhibiting osteogenic differentiation, mitochondrial biogenesis, and mitochondrial morphology. Mitochondrial uncoupling and activation of AMPK depressed mitochondrial oxidative phosphorylation and inhibited differentiation, suggesting that they may serve as regulators to halt osteogenic differentiation from impaired mitochondrial oxidative phosphorylation.

Deep learning-based semantic segmentation of non-melanocytic skin tumors in whole-slide histopathological images.

Basal cell carcinoma (BCC) and squamous cell carcinoma (SCC) are the two most common skin cancer and impose a huge medical burden on society. Histopathological examination based on whole-slide images (WSIs) remains to be the confirmatory diagnostic method for skin tumors. Accurate segmentation of tumor tissue in WSIs by deep-learning (DL) models can reduce the workload of pathologists and help surgeons ensure the complete removal of tumors. To accurately segment the tumor areas in WSIs of BCC, SCC and squamous cell papilloma (SCP, homologous to SCC) with robust models. We established a data set (ZJU-NMSC) containing 151 WSIs of BCC, SCC and SCP in total. Seven models were utilized to segment WSIs, including the state-of-the-art model, models proposed by us and other models. Dice score, intersection over union, accuracy, sensitivity and specificity were used to evaluate and compare the performance of different models. Heatmaps and tumor tissue masks were generated to reflect the results of the segmentation. The processing times of models are also recorded and compared. While the dice score of most models is higher than 0.85, deeplab v3+ has the best performance and the corresponding tumor tissue mask is more consistent with the ground truth tumor areas even with complex and small lobular lesions. This study broadens the use of DL-based segmentation models in WSIs of skin tumors in terms of tumor types and computational approaches. Segmenting tumor areas can simplify the process of histopathological inspection and benefit the diagnosis and following management of the diseases in practice.

Novel Boron-rich Phosphides with High Hardness, Large Strain, and Magnetism.

Boron-rich compounds have attracted much attention due to their interesting structures and excellent properties. Here, we performed an extensive study on the different B-P stoichiometries under pressure by combining a particle swarm optimization method with first-principles calculations. At 1 atm, BP and B6P are thermodynamically stable, while other stoichiometries are metastable. Under pressure, BP and B6P remain stable relative to constituent pure solids up to 80 GPa, while other stoichiometries become unstable at relatively low pressures. A new Cmca B6P is predicted with the lowest energy at 1 atm and shows higher shear strain than the R3̅m structure, which is known to be more resistant to brittle fracture than B4C. Moreover, the predicted Pm B8P is a magnetic semiconductor with a magnetic moment of 1 µB. All these boron-rich phosphides are hard materials. The present results enrich the B-P phase diagram and promote extensive research on their excellent properties.

In Situ Observation of Fracture along Twin Boundaries in Boron Carbide.

The observation of fracture behaviors in perfect and twinned B4 C crystals via in situ transmission electron microscopy (TEM) mechanical testing is reported. The crystal structure of the synthesized B4 C, composed of B11 C icosahedra connected by boron-deficient C-▫-C chains in a chemical formula of B11 C3 , is determined by state-of-the-art aberration-corrected scanning TEM. The in situ TEM observations reveal that cracking is preferentially initiated at the twin boundaries (TBs) in B4 C under both indentation and tension loading. The cracks then propagate along the TBs, thus resulting in the fracture of B4 C. These results are consistent with the theoretical calculations that show that TBs have a softening effect on B4 C with amorphous bands preferentially nucleated at the TBs. These findings elucidate the atomic arrangement and the role of planar defects in the failure of B4 C. Furthermore, they can guide the design of advanced superhard materials via planar defect control.

Laparoscopic management of benign splenic tumors in children.

Background: The splenic tumor is relatively rare in children. However, diagnosing and managing this disease remain controversial. This study aimed to evaluate the clinical characteristics and pathological features of benign splenic tumors and to explore the feasibility and safety of laparoscopic splenectomy in children. Methods: The clinical data of all patients who were diagnosed with benign splenic tumors and admitted to our center between January 2014 and December 2020 were analyzed retrospectively. Factors that were documented included demographic criteria, clinical manifestations, preoperative examinations, surgical methods, histopathological characteristics, postoperative complications and outcomes. Results: Totally, 24 consecutive patients with a mean age of 10.21±2.59 years were operated laparoscopically. Nine patients presented with recurrent abdominal pain, vomiting or backache. The remaining 15 cases were asymptomatic and were identified incidentally. Laparoscopic total splenectomy (LTS) was performed in 12 cases, laparoscopic partial splenectomy (LPS) in 9 cases and deroofing of splenic cyst in 3 cases. No significant differences were observed in operation time, intraoperative bleeding and postoperative abdominal drainage between the LTS and LPS groups. Histopathology examination reported 12 cases of splenic epithelial cyst, 5 cases of splenic lymphangioma or lymphatic malformation, 3 cases of splenic hamartoma, 3 cases of splenic littoral cell angioma and 1 case of splenic hemangioma. No postoperative complications and tumor recurrence were observed. Conclusions: It is still difficult to precisely diagnose the splenic tumor preoperatively. Laparoscopic splenectomy is safe and feasible in treating the benign splenic tumors, and partial splenectomy is of great significance in preserving the immune function in children.

Chemical Composition, Antitumor Properties, and Mechanism of the Essential Oil from Plagiomnium acutum T. Kop.

Plagiomnium acutum T. Kop. (P. acutum) has been used as a traditional Chinese medicine for thousands of years to treat cancer but lacks evidence. The objective of this work was to reveal the chemical composition of P. acutum essential oil (PEO) and explore its potential antitumor activity and molecular mechanism. PEO was prepared by the simultaneous distillation-extraction method and characterized by gas chromatography/mass spectroscopy. CCK8 assay, flow cytometry, western blot, and immunofluorescence techniques were used to analyze the effects and mechanism of PEO against cancer cells. A total of 74 constituents of PEO were identified, with diterpenes (26.5%), sesquiterpenes (23.89%), and alcohols (21.81%) being the major constituents. Two terpenoids, selina-6-en-4-ol and dolabella-3,7-dien-18-ol, were detected in PEO for the first time. PEO showed significant cell growth inhibitory activity on HepG2 and A549 cells by blocking the G1 phase and inducing apoptosis, which may be attributed to its upregulation of p21Cip1 and p27Kip1 proteins and interference with mitochondrial membrane potential effect. Dolabella-3,7-dien-18-ol accounts for 25.5% of PEO and is one of the main active components of PEO, with IC50 values in HepG2 and A549 cells of (25.820 ± 0.216) µg/mL and (23.597 ± 1.207) µg/mL, respectively. These results confirmed the antitumor medicinal value of P. acutum and showed great application potential in the pharmaceutical industry.

Self-supervised learning mechanism for identification of eyelid malignant melanoma in pathologic slides with limited annotation.

Purpose: The lack of finely annotated pathologic data has limited the application of deep learning systems (DLS) to the automated interpretation of pathologic slides. Therefore, this study develops a robust self-supervised learning (SSL) pathology diagnostic system to automatically detect malignant melanoma (MM) in the eyelid with limited annotation. Design: Development of a self-supervised diagnosis pipeline based on a public dataset, then refined and tested on a private, real-world clinical dataset. Subjects: A. Patchcamelyon (PCam)-a publicly accessible dataset for the classification task of patch-level histopathologic images. B. The Second Affiliated Hospital, Zhejiang University School of Medicine (ZJU-2) dataset - 524,307 patches (small sections cut from pathologic slide images) from 192 H&E-stained whole-slide-images (WSIs); only 72 WSIs were labeled by pathologists. Methods: Patchcamelyon was used to select a convolutional neural network (CNN) as the backbone for our SSL-based model. This model was further developed in the ZJU-2 dataset for patch-level classification with both labeled and unlabeled images to test its diagnosis ability. Then the algorithm retrieved information based on patch-level prediction to generate WSI-level classification results using random forest. A heatmap was computed for visualizing the decision-making process. Main outcome measures: The area under the receiver operating characteristic curve (AUC), accuracy, sensitivity, and specificity were used to evaluate the performance of the algorithm in identifying MM. Results: ResNet50 was selected as the backbone of the SSL-based model using the PCam dataset. This algorithm then achieved an AUC of 0.981 with an accuracy, sensitivity, and specificity of 90.9, 85.2, and 96.3% for the patch-level classification of the ZJU-2 dataset. For WSI-level diagnosis, the AUC, accuracy, sensitivity, and specificity were 0.974, 93.8%, 75.0%, and 100%, separately. For every WSI, a heatmap was generated based on the malignancy probability. Conclusion: Our diagnostic system, which is based on SSL and trained with a dataset of limited annotation, can automatically identify MM in pathologic slides and highlight MM areas in WSIs by a probabilistic heatmap. In addition, this labor-saving and cost-efficient model has the potential to be refined to help diagnose other ophthalmic and non-ophthalmic malignancies.

A Self-Supervised Learning Based Framework for Eyelid Malignant Melanoma Diagnosis in Whole Slide Images.

Eyelid malignant melanoma (MM) is a rare disease with high mortality. Accurate diagnosis of such disease is important but challenging. In clinical practice, the diagnosis of MM is currently performed manually by pathologists, which is subjective and biased. Since the heavy manual annotation workload, most pathological whole slide image (WSI) datasets are only partially labeled (without region annotations), which cannot be directly used in supervised deep learning. For these reasons, it is of great practical significance to design a laborsaving and high data utilization diagnosis method. In this paper, a self-supervised learning (SSL) based framework for automatically detecting eyelid MM is proposed. The framework consists of a self-supervised model for detecting MM areas at the patch-level and a second model for classifying lesion types at the slide level. A squeeze-excitation (SE) attention structure and a feature-projection (FP) structure are integrated to boost learning on details of pathological images and improve model performance. In addition, this framework also provides visual heatmaps with high quality and reliability to highlight the likely areas of the lesion to assist the evaluation and diagnosis of the eyelid MM. Extensive experimental results on different datasets show that our proposed method outperforms other state-of-the-art SSL and fully supervised methods at both patch and slide levels when only a subset of WSIs are annotated. It should be noted that our method is even comparable to supervised methods when all WSIs are fully annotated. To the best of our knowledge, our work is the first SSL method for automatic diagnosis of MM at the eyelid and has a great potential impact on reducing the workload of human annotations in clinical practice.

Artificial intelligence defines protein-based classification of thyroid nodules.

Determination of malignancy in thyroid nodules remains a major diagnostic challenge. Here we report the feasibility and clinical utility of developing an AI-defined protein-based biomarker panel for diagnostic classification of thyroid nodules: based initially on formalin-fixed paraffin-embedded (FFPE), and further refined for fine-needle aspiration (FNA) tissue specimens of minute amounts which pose technical challenges for other methods. We first developed a neural network model of 19 protein biomarkers based on the proteomes of 1724 FFPE thyroid tissue samples from a retrospective cohort. This classifier achieved over 91% accuracy in the discovery set for classifying malignant thyroid nodules. The classifier was externally validated by blinded analyses in a retrospective cohort of 288 nodules (89% accuracy; FFPE) and a prospective cohort of 294 FNA biopsies (85% accuracy) from twelve independent clinical centers. This study shows that integrating high-throughput proteomics and AI technology in multi-center retrospective and prospective clinical cohorts facilitates precise disease diagnosis which is otherwise difficult to achieve by other methods.