Six artificial intelligence paradigms for tissue characterisation and classification of non-COVID-19 pneumonia against COVID-19 pneumonia in computed tomography lungs.
- 作者列表："Saba L","Agarwal M","Patrick A","Puvvula A","Gupta SK","Carriero A","Laird JR","Kitas GD","Johri AM","Balestrieri A","Falaschi Z","Paschè A","Viswanathan V","El-Baz A","Alam I","Jain A","Naidu S","Oberleitner R","Khanna NN","Bit A","Fatemi M","Alizad A","Suri JS
BACKGROUND:COVID-19 pandemic has currently no vaccines. Thus, the only feasible solution for prevention relies on the detection of COVID-19-positive cases through quick and accurate testing. Since artificial intelligence (AI) offers the powerful mechanism to automatically extract the tissue features and characterise the disease, we therefore hypothesise that AI-based strategies can provide quick detection and classification, especially for radiological computed tomography (CT) lung scans. METHODOLOGY:Six models, two traditional machine learning (ML)-based (k-NN and RF), two transfer learning (TL)-based (VGG19 and InceptionV3), and the last two were our custom-designed deep learning (DL) models (CNN and iCNN), were developed for classification between COVID pneumonia (CoP) and non-COVID pneumonia (NCoP). K10 cross-validation (90% training: 10% testing) protocol on an Italian cohort of 100 CoP and 30 NCoP patients was used for performance evaluation and bispectrum analysis for CT lung characterisation. RESULTS:Using K10 protocol, our results showed the accuracy in the order of DL > TL > ML, ranging the six accuracies for k-NN, RF, VGG19, IV3, CNN, iCNN as 74.58 ± 2.44%, 96.84 ± 2.6, 94.84 ± 2.85%, 99.53 ± 0.75%, 99.53 ± 1.05%, and 99.69 ± 0.66%, respectively. The corresponding AUCs were 0.74, 0.94, 0.96, 0.99, 0.99, and 0.99 (p-values < 0.0001), respectively. Our Bispectrum-based characterisation system suggested CoP can be separated against NCoP using AI models. COVID risk severity stratification also showed a high correlation of 0.7270 (p < 0.0001) with clinical scores such as ground-glass opacities (GGO), further validating our AI models. CONCLUSIONS:We prove our hypothesis by demonstrating that all the six AI models successfully classified CoP against NCoP due to the strong presence of contrasting features such as ground-glass opacities (GGO), consolidations, and pleural effusion in CoP patients. Further, our online system takes < 2 s for inference.
背景: 新型冠状病毒肺炎大流行目前没有疫苗。因此，唯一可行的预防方案依赖于通过快速准确的检测来检测COVID-19-positive个病例。由于人工智能 (AI) 提供了自动提取组织特征和表征疾病的强大机制，因此我们假设基于AI的策略可以提供快速检测和分类，特别是对于放射计算机断层扫描 (CT) 肺扫描。 方法: 六个模型，两个基于传统机器学习 (ML) (k-NN和RF)，两个基于迁移学习 (TL) (VGG19和InceptionV3)，最后两个是我们的定制设计的深度学习 (DL) 模型 (CNN和iCNN)，用于COVID肺炎 (CoP) 和非COVID肺炎 (NCoP) 之间的分类。对90% 例CoP和30例NCoP患者的意大利队列的K10交叉验证 (10% 训练: 100测试) 方案用于性能评估和CT肺表征的双谱分析。 结果: 使用K10协议，我们的结果显示准确度为dl > tltl > ml ml，范围k-NN，RF，VGG19，IV3，CNN，iCNN的六个准确度为74.58 ± 2.44% 96.84，2.6 ± 94.84，2.85% ± ，分别为99.53 ± 0.75% 、99.53 ± 1.05% 和99.69 ± 0.66%。相应的auc分别为0.74、0.94、0.96、0.99和0.99 (p值 <0.0001)。我们的基于双谱的表征系统建议使用AI模型可以将CoP与NCoP分开。COVID风险严重程度分层也显示与临床评分 (如磨玻璃密度 (GGO)) 的高度相关性为0.7270 (p <0.0001)，进一步验证了我们的AI模型。 结论: 我们通过证明所有六个AI模型成功地将CoP分类为NCoP，证明了我们的假设，这是因为在CoP患者中存在强烈的对比特征，如磨玻璃影 (GGO) 、固结和胸腔积液。此外，我们的在线系统需要 < 2 s进行推理。
METHODS::The ATP binding-cassette superfamily corresponds the mostly transmembrane transporters family found in humans. These proteins actively transport endogenous and exogenous substrates through biological membranes in body tissues, so they have an important role in the regulation of many physiological functions necessary for human homeostasis, as well as in response regulation to several pharmacological substrates. The development of multidrug resistance has become one of the main troubles in conventional chemotherapy in different illnesses including cancer, being the increased efflux of antineoplastic drugs the main reason for this multidrug resistance, with a key role of the ABC superfamily. Likely, the interindividual variability in the pharmacological response among patients is well known, and may be due to intrinsically factors of the disease, genetic and environmental ones. Thus, the understanding of this variability, especially the genetic variability associated with the efficacy and toxicity of drugs, can provide a safer and more effective pharmacological treatment, so ABC genes are considered as important regulators due to their relationship with the reduction in pharmacological response. In this review, updated information about transporters belonging to this superfamily was collected, the possible role of these transporters in cancer, the role of genetic variability in their genes, as well as some therapeutic tools that have been tried to raise against main transporters associated with chemoresistance in cancer.
METHODS:BACKGROUND:Cholinergic neurotransmission regulates neuroinflammation in Parkinson disease (PD). RESEARCH DESIGN AND METHODS:The authors conducted a delayed-start study of donepezil for cognitive decline in non-demented PD patients. The study consisted of a 96-week randomized placebo-controlled double-blind phase 1, followed by a 24-week donepezil extension phase 2. The primary outcome measure was a change in the Mini-Mental State Examination (MMSE) at week 120. RESULTS:A total of 98 patients were randomly allocated to the early-start (donepezil-to-donepezil) and delayed-start (placebo-to-donepezil) groups. Mean (SD) of the baseline MMSE was 27.6 (2.0) and 28.0 (2.1), respectively. MMSE change at week 120 was better in the early-start group than in the delayed-start group, but the difference was not significant. The MMSE declined in apolipoprotein ε4 carriers, but not in non-carriers, and the factor interaction (intervention × ε4 genotype) was highly significant (P < 0.001). Analyzed with the interaction, the difference was significant (group difference 1.95 [0.33 to 3.57], P = 0.018). The MMSE decline slope in phase 1 was significantly better in the early-start group than in the delayed-start group (P = 0.048). CONCLUSIONS:Cognitive function deteriorated in ε4 carriers, but not in non-carriers, and early-start donepezil may postpone cognitive decline in the former.
METHODS::Since the discovery of dental pulp stem cells, a lot of teams have expressed an interest in dental pulp regeneration. Many approaches, experimental models and biological explorations have been developed, each including the use of stem cells and scaffolds with the final goal being clinical application in humans. In this review, the authors' objective was to compare the experimental models and strategies used for the development of biomaterials for tissue engineering of dental pulp with stem cells. Electronic queries were conducted on PubMed using the following terms: pulp regeneration, scaffold, stem cells, tissue engineering and biomaterial. The extracted data included the following information: the strategy envisaged, the type of stem cells, the experimental models, the exploration or analysis methods, the cytotoxicity or viability or proliferation cellular tests, the tests of scaffold antibacterial properties and take into account the vascularization of the regenerated dental pulp. From the 71 selected articles, 59% focused on the "cell-transplantation" strategy, 82% used in vitro experimentation, 58% in vivo animal models and only one described an in vivo in situ human clinical study. 87% used dental pulp stem cells. A majority of the studies reported histology (75%) and immunohistochemistry explorations (66%). 73% mentioned the use of cytotoxicity, proliferation or viability tests. 48% took vascularization into account but only 6% studied the antibacterial properties of the scaffolds. This article gives an overview of the methods used to regenerate dental pulp from stem cells and should help researchers create the best development strategies for research in this field.