Delayed haemolytic transfusion reaction due to Kidd antibodies

A delayed haemolytic transfusion reaction (DHTR) encompasses a positive direct antiglobulin test (DAT) developed anytime between 24 hours to 28 days after cessation of transfusion, a positive eluate or a newly identified alloantibody in the plasma or serum along with features of haemolysis in the patient. Routinely, it is expected that with the transfusion of one unit of packed red cells in a patient of average height and weight, the haemoglobin level and hematocrit increase by 1 g/dL and 3% respectively. However, in a patient with DHTR, an inadequate rise of post-transfusion haemoglobin (< 1 g/dL) or rapid fall in haemoglobin back to pre-transfusion levels is observed. Kidd antibodies are particularly known to cause DHTR, maybe alone or in unison with other antibodies. Detection of these alloantibodies is consequential in providing good transfusion support to these patients. These events may be difficult to detect as they may present as varied clinical features or immunological nuisances. In this case series, we have presented three cases of DHTR with special emphasis on their clinical presentations, immunohaematological evaluations, laboratory parameters and the role of proper transfusion support in these patients to avoid further complications.     

Dietary patterns and the risk of abnormal blood lipids among young adults: A prospective cohort study

Background and aimsThe extent to which dietary patterns influence the risk of abnormal blood lipids throughout young adulthood remains unclear. The aim was to investigate whether early young adulthood dietary patterns predict the risk of abnormal blood lipids during later young adulthood.Methods and resultsWe used data from a long running birth cohort study in Australia. Western dietary pattern rich in meats, processed foods and high-fat dairy products and prudent pattern rich in fruit, vegetables, fish, nuts, whole grains and low-fat dairy products were derived using principal component analysis at the 21-year follow-up from dietary data obtained using a food frequency questionnaire. After 9-years, fasting blood samples of all participants were collected and their total, low-density lipoprotein (LDL) and high-density lipoprotein (HDL) cholesterols and triglyceride (TG) levels were measured. Abnormal blood lipids were based on clinical cut-offs for total, LDL and HDL cholesterols, and TG and relative distributions for total:HDL and TG:HDL cholesterols ratios. Log-binomial models were used to estimate risk of each outcome in relation to dietary patterns. Greater adherence to the Western pattern predicted increased risks of high LDL (RR: 1.47; 95%CI: 1.06, 2.03) and TG (1.90; 1.25, 2.86), and high ratios of total:HDL (1.48; 1.00, 2.19) and TG:HDL (1.78; 1.18, 2.70) cholesterols in fully adjusted models. Conversely, a prudent pattern predicted reduced risks of low HDL (0.58; 0.42, 0.78) and high TG (0.66; 0.47, 0.92) and high total:HDL (0.71; 0.51, 0.98) and TG:HDL (0.61; 0.45, 0.84) cholesterols ratios.ConclusionThis is the first prospective study to show greater adherence to unhealthy Western diet predicted increased risks of abnormal blood lipids, whereas healthy prudent diet predicted lower such risks in young adults. Addressing diets in early course may improve cardiovascular health of young adults.     

Registries,Databases and Repositories for Developing Artificial Intelligence in Cancer Care

Modern artificial intelligence techniques have solved some previously intractable problems and produced impressive results in selected medical domains. One of their drawbacks is that they often need very large amounts of data. Pre-existing datasets in the form of national cancer registries, image/genetic depositories and clinical datasets already exist and have been used for research. In theory, the combination of healthcare Big Data with modern, data-hungry artificial intelligence techniques should offer significant opportunities for artificial intelligence development, but this has not yet happened. Here we discuss some of the structural reasons for this, barriers preventing artificial intelligence from making full use of existing datasets, and make suggestions as to enable progress. To do this, we use the framework of the 6Vs of Big Data and the FAIR criteria for data sharing and availability (Findability, Accessibility, Interoperability, and Reuse). We share our experience in navigating these barriers through The Brain Tumour Data Accelerator, a Brain Tumour Charity-supported initiative to integrate fragmented patient data into an enriched dataset. We conclude with some comments as to the limits of such approaches.     

下一代诊断病理学

随着人类基因组测序、生物大数据信息分析、分子病理检测和人工智能辅助病理诊断等技术进步及其应用, 临床医学发展迈向精准诊疗时代。这一时代背景下, 传统诊断病理学迎来前所未有的历史机遇, 正在向"下一代诊断病理学(next-generation diagnostic pathology)"迈进。下一代诊断病理学以病理形态和临床信息为诊断基础, 以分子检测与生物信息分析、智慧制样与流程质控、智能诊断与远程会诊、病灶活体可视化与"无创"病理诊断等创新前沿交叉技术为主要特征, 以多组学和跨尺度整合诊断为病理报告内容, 实现对疾病的"最后诊断", 并预测疾病演进和结局、建议治疗方案和评估治疗反应, 形成新的疾病诊断"金标准"。未来, 需要激发病理学科创新活力, 加快下一代诊断病理学成熟和应用, 重塑病理学科理论和技术体系, 发挥诊断病理学在疾病"防、诊、治、养"等过程中的重要作用, 促进临床医学进一步发展, 服务健康中国战略。     

基于超声影像组学的乳腺癌预测模型的诊断性能研究

目的评估并比较基于不同机器学习算法建立的乳腺癌超声影像组学预测模型的诊断性能。 方法回顾性收集2017年1月至2019年4月就诊皖南医学院第一附属医院、有明确病理结果的乳腺肿块病例828例，以2018年8月31日为节点将其分为训练集（526例）和验证集（302例），提取肿块的超声影像组学特征并进行特征筛选，运用k最近邻（kNN）、逻辑回归（LR）、朴素贝叶斯（NB）、随机森林（RF）和支持向量机（SVM）5种机器学习算法分别建立预测模型，使用重复交叉验证方法做内部验证，计算比较各模型的敏感度、特异度、阳性预测值（PPV）和阴性预测值（NPV），并实施外部验证，绘制ROC曲线并比较ROC曲线下面积（AUC）以评价模型的鉴别诊断性能，绘制校准曲线评价模型校准度。 结果从提取的109个影像组学特征中筛选出19个特征建立了5种机器学习算法模型。在内部验证中，5种模型的敏感度、特异度、PPV、NPV比较，总体差异均有统计学意义（P均＜0.001）。LR模型的特异度、PPV、NPV中位数分别为0.769、0.816、0.778，3项指标均高于其他4种模型；敏感度中位数为0.824，高于kNN、RF和SVM模型。此外，SVM模型的特异度、PPV、NPV中位数分别为0.706、0.774、0.759，虽均低于LR模型，但均高于其他3种模型。在外部验证中，LR、SVM、RF、kNN和NB的AUC依次为0.890、0.832、0.821、0.746和0.703，其中LR与SVM的AUC差异有统计学意义（P=0.012）；此外，各模型在校准性能上表现并不一致，LR和SVM模型的校准曲线显示乳腺癌实际概率与预测概率之间的一致性较好。 结论以超声影像组学特征为基础，运用不同机器学习算法建立的乳腺癌超声预测模型，均表现出较高的诊断性能，其中LR模型表现最为突出；选择合适的机器学习算法有助于进一步提高预测模型的诊断性能，提供更加准确的量化预测结果。