Image Analysis-Based Machine Learning for the Diagnosis of Retinopathy of Prematurity: A Meta-analysis and Systematic Review.

TOPIC: To evaluate the performance of machine learning (ML) in the diagnosis of retinopathy of prematurity (ROP) and to assess whether it can be an effective automated diagnostic tool for clinical applications. CLINICAL RELEVANCE: Early detection of ROP is crucial for preventing tractional retinal detachment and blindness in preterm infants, which has significant clinical relevance. METHODS: Web of Science, PubMed, Embase, IEEE Xplore, and Cochrane Library were searched for published studies on image-based ML for diagnosis of ROP or classification of clinical subtypes from inception to October 1, 2022. The quality assessment tool for artificial intelligence-centered diagnostic test accuracy studies was used to determine the risk of bias (RoB) of the included original studies. A bivariate mixed effects model was used for quantitative analysis of the data, and the Deek's test was used for calculating publication bias. Quality of evidence was assessed using Grading of Recommendations Assessment, Development and Evaluation. RESULTS: Twenty-two studies were included in the systematic review; 4 studies had high or unclear RoB. In the area of indicator test items, only 2 studies had high or unclear RoB because they did not establish predefined thresholds. In the area of reference standards, 3 studies had high or unclear RoB. Regarding applicability, only 1 study was considered to have high or unclear applicability in terms of patient selection. The sensitivity and specificity of image-based ML for the diagnosis of ROP were 93% (95% confidence interval [CI]: 0.90-0.94) and 95% (95% CI: 0.94-0.97), respectively. The area under the receiver operating characteristic curve (AUC) was 0.98 (95% CI: 0.97-0.99). For the classification of clinical subtypes of ROP, the sensitivity and specificity were 93% (95% CI: 0.89-0.96) and 93% (95% CI: 0.89-0.95), respectively, and the AUC was 0.97 (95% CI: 0.96-0.98). The classification results were highly similar to those of clinical experts (Spearman's R = 0.879). CONCLUSIONS: Machine learning algorithms are no less accurate than human experts and hold considerable potential as automated diagnostic tools for ROP. However, given the quality and high heterogeneity of the available evidence, these algorithms should be considered as supplementary tools to assist clinicians in diagnosing ROP. FINANCIAL DISCLOSURE(S): Proprietary or commercial disclosure may be found in the Footnotes and Disclosures at the end of this article.

Modification of Frictional Properties of Hydrogel Surface via Laser Ablated Topographical Micro-Textures.

Hydrogels and biological cartilage tissues are highly similar in structure and composition due to their unique characteristics such as high-water content and low friction coefficients. The introduction of hydrogel cartilage can effectively reduce the friction coefficient and wear coefficient of the original bone joint and the implanted metal bone joint (generally titanium alloy or stainless steel), which is considered as a perfect replacement material for artificial articular cartilage. How to accurately regulate the local tribological characteristics of hydrogel artificial cartilage according to patient weight and bone shape is one of the important challenges in the current clinical application field of medical hydrogels. In this study, the mechanism by which micro-pits improve the surface friction properties was studied. Ultraviolet lasers were used to efficiently construct micro-pits with different shapes on a polyvinyl alcohol hydrogel in one step. It was shown that by using such a maskless laser processing, the performance of each part of the artificial cartilage can be customized flexibly and effectively. We envision that the approach demonstrated in this article will provide an important idea for the development of a high-performance, continuous and accurate method for controlling surface friction properties of artificial cartilage.

Distinguish the Value of the Benign Nevus and Melanomas Using Machine Learning: A Meta-Analysis and Systematic Review.

Background: Melanomas, the most common human malignancy, are primarily diagnosed visually, beginning with an initial clinical screening and followed potentially by dermoscopic analysis, a biopsy, and histopathological examination. We aimed to systematically review the performance and quality of machine learning-based methods in distinguishing melanoma and benign nevus in the relevant literature. Method: Four databases (Web of Science, PubMed, Embase, and the Cochrane library) were searched to retrieve the relevant studies published until March 26, 2022. The Predictive model Deviation Risk Assessment tool (PROBAST) was used to assess the deviation risk of opposing law. Result: This systematic review included thirty researches with 114007 subjects and 71 machine learning models. The convolutional neural network was the main machine learning method. The pooled sensitivity was 85% (95% CI 82-87%), the specificity was 86% (82-88%), and the C-index was 0.87 (0.84-0.90). Conclusion: The findings of our study showed that ML algorithms had high sensitivity and specificity for distinguishing between melanoma and benign nevi. This suggests that state-of-the-art ML-based algorithms for distinguishing melanoma from benign nevi may be ready for clinical use. However, a large proportion of the earlier published studies had methodological flaws, such as lack of external validation and lack of clinician comparisons. The results of these studies should be interpreted with caution.

Establishment and validation of a nomogram for predicting immune-related prognostic features in trunk melanoma-specific death.

Background: Trunk melanoma is one of the most common and deadly types of melanomas. Multiple factors are associated with the prognosis of patients with trunk melanoma. Currently, direct, and reliable clinical tools for early assessment of individual specific risk of death are limited, and most of them are prediction models for all-cause death. Their accuracy in predicting competitiveness events, which make up a relatively large portion, may be substantially compromised. Hence, we conducted this study to investigate the risk factors of trunk melanoma-specific death to establish a comprehensive prediction model suitable for clinical application. Methods: Patients with trunk melanoma analyzed in this study were from the SEER program [2010-2015]. The random sampling method was used to split the included cases into the training and validation cohorts at a ratio of 7:3. Univariate and multivariate competing risk models were used to screen the independent influencing factors of specific death, and then a nomogram covering these independent predictors was constructed. The concordance index (C-index) and a calibration curve were used to evaluate the calibration degree and accuracy of the nomogram. Results: We identified 21,198 patients with trunk melanoma from the SEER database, and 3,814 of them died (17.99%). Among the death cases, deaths from other causes accounted for 66.50%The prognostic nomogram included 8 variables and 16 independent influencing factors. The overall C-index in the training set was 0.89, and the receiver operating characteristic (ROC) curve for predicting 1-, 3-, and 5-year survival was 0.928 [95% confidence interval (CI): 0.911-0.945], 0.907 (95% CI: 0.895-0.918), and 0.891 (95% CI: 0.879-0.902), respectively. The C-index of the model in the validation set was 0.89, and the area under the ROC curve (AUC) for predicting 1-, 3-, and 5-year cancer-specific death (CSD) was 0.927 (95% CI: 0.899-0.955), 0.916 (95% CI: 0.901-0.930), and 0.905 (95% CI: 0.899-0.921). Both the training set and the validation set showed the ideal calibration degree. Conclusions: This model can be used as a potential tool for prognostic risk management of trunk melanoma in the presence of many competing events.

Printed Diodes: Materials Processing, Fabrication, and Applications.

Printing techniques for the fabrication of diodes have received increasing attention over the last decade due to their great potential as alternatives for high-throughput and cost-effective manufacturing approaches compatible with both flexible and rigid substrates. Here, the progress achieved and the challenges faced in the fabrication of printed diodes are discussed and highlighted, with a focus on the materials of significance (silicon, metal oxides, nanomaterials, and organics), the techniques utilized for ink deposition (gravure printing, screen printing, inkjet printing, aerosol jet printing, etc.), and the process through which the printed layers of diode are sintered after printing. Special attention is also given to the device applications within which the printed diodes have been successfully incorporated, particularly in the fields of rectification, light emission, energy harvesting, and displays. Considering the unmatched production scalability of printed diodes and their intrinsic suitability for flexible and wearable applications, significant improvement in performance and intensive research in development and applications of the printed diodes will continuously progress in the future.

Recent Advances in Stretchable Supercapacitors Enabled by Low-Dimensional Nanomaterials.

Supercapacitors (SCs) have shown great potential for mobile energy storage technology owing to their long-term durability, electrochemical stability, structural simplicity, as well as exceptional power density without much compromise in the energy density and cycle life parameters. As a result, stretchable SC devices have been incorporated in a variety of emerging electronics applications ranging from wearable electronic textiles to microrobots to integrated energy systems. In this review, the recent progress and achievements in the field of stretchable SCs enabled by low-dimensional nanomaterials such as polypyrrole, carbon nanotubes, and graphene are presented. First, the three major categories of stretchable supercapacitors are discussed: double-layer supercapacitors, pseudo-supercapacitors, and hybrid supercapacitors. Then, the representative progress in developing stretchable electrodes with low-dimensional (0D, 1D, and 2D) nanomaterials is described. Next, the design strategies enabling the stretchability of the devices, including the wavy-shape design, wire-shape design, textile-shape design, kirigami-shape design, origami-shape design, and serpentine bridge-island design are emphasized, with the aim of improving the electrochemical performance under the complex stretchability conditions that may be encountered in practical applications. Finally, the newest developments, major challenges, and outlook in the field of stretchable SC development and manufacturing are discussed.