欢迎来到艾兰博曼,请 登录 | 立即注册
查看: 1048|回复: 1

【资讯翻译】In Newborn Screening, Big Data Saves Little People

[复制链接]

10

听众

10

收听

1954

积分

管理员

Rank: 9Rank: 9Rank: 9

积分
1954
发表于 2014-8-1 10:08:06 | 显示全部楼层 |阅读模式
In Newborn Screening, Big Data Saves Little People

By Amy Saenger, PhD

Over the past 50 years newborn screening for inborn errors of metabolism has evolved into complex algorithms and tiered testing. It is now largely consolidated and highly automated, thanks to innovations in tandem mass spectrometry. At Tuesday’s plenary session, Piero Rinaldo, MD, PhD,shared the podium with Kaitlyn Bloom,PhD, who— in an AACC first—discussed laboratory medicine from a patient’s perspective,shedding light on the important contributions laboratory professionals make every single day.

Bloom is currently a molecular postdoctoral fellow at the Children’s Hospital of Philadelphia. However, this would not have been the case had she not been screened at birth. Diagnosed with phenylketonuria (PKU) at 11 days old, Bloom shared her personal story about the power of newborn screening and the testing of just one dried blood spot. “I consider myself extremely lucky to be born in an era where technology detected my PKU as a newborn,” Bloom said. “I feel blessed by PKU because, over time, it has created a passion within me to learn more about the genetics of human metabolism and establish a career in that field.”

A single dried blood spot yields more than 100 markers and ratios and ultimately can aid in the diagnosis of numerous inherited disorders. But despite the great strides in technology in this area, many questions remain. For example, what should be done with all the data? How can false-positive results be prevented? And how can results be interpreted easily by the laboratory and provide value in an output which is clear and actionable?

Rinaldo’s groundbreaking research has found some solutions to these challenges.One is an online database used by an international collaborative group of newborn screening laboratories. The Region 4 Stork Project (R4S) involves 200 labs in 60 countries that have more than 1.2 million data points on at least 16,000 true-positive cases detected with newborn screening. This innovative approach uses post-analytical interpretive tools and multivariate pattern recognition software, allowing for greater differentiation and interpretation of values based not on normal versus abnormal, but from one disease state to another. “As the number of tests included in newborn screening grows, performance metrics must exceed the historic standard,”Rinaldo said. The current false positive rate for newborn screening is excessive,he added.

The interpretive tools look specifically for a single condition or help differentiate between multiple conditions, as long as there are a sufficient number of true-positive cases documented in the database. All of the patient’s results are merged into a single score reported as a percentile rank among all cases with the same condition and compared to uniform interpretation guidelines. The power of this system derives from the sheer number of users who are able to input their cases into the database.

Complex disease profiles like those encountered in newborn screening require condition-specific disease ranges rather than traditional analyte-based cutoffs. R4S post-analytical tools take clinical validation from the conventional static process that’s performed early during test development to a constantly evolving, dynamic refinement of disease range—one that continues to improve throughout the entire process. Rinaldo has pioneered efforts in this area, and although the short-term goal of R4S is to reduce the number of false-positive newborn screening results, when asked if it will one day be the demise of strict analyte-based cutoffs he responded,“you betcha.”

2

听众

0

收听

594

积分

高级会员

Rank: 4

积分
594
发表于 2014-8-1 18:16:46 | 显示全部楼层
在过去的50年,新生儿先天性代谢缺陷筛查已经逐渐演变成复杂的算法和分层检测。它现在已经极大的综合化和高度自动化,这些都多亏了新的串联质谱法的运用。在周二的2014 AACC年会的全体会议上,Piero Rinaldo 博士和Kaitlyn Bloom 博士在讲台上与我们共同分享了他们的经验,Kaitlyn Bloom 博士是第一个在AACC年会上从患者角度来讨论医学检验的人,从而揭示了检验专业人员每一天的重要贡献。

  Bloom 目前是一名费城儿童医院从事分子方向研究的博士后。然而,这不会一直是她在出生时没有被筛查出来的原因。在出生11天后就被确诊为患有苯丙酮尿症(PKU),Bloom 分享了她个人关于新生儿筛查和只用一个干血点检测的故事。“我认为自己非常幸运,出生在可在新生儿期检测苯丙酮尿症(PKU)该项技术的区域,”Bloom 说。“因为苯丙酮尿症,我觉得很幸福,因为,随着时间的推移,它激发了我的激情去更多的了解关于人体新陈代谢的基因并在该领域建立我的职业。”

  一个单一的干血点可以检测出超过100种标记物和比值,并最终可以在众多的遗传性疾病的诊断中提供帮助。但是,尽管在这方面的技术上已经获得了巨大进步,但仍然存在许多问题。例如,应如何处理这些所有的数据?如何预防出现假阳性结果?以及结果如何能较容易的被实验室解释并且使得这些价值可清晰的和具有可操作性的被输入和输出?

  对于这些挑战,Rinaldo 的开创性研究已经发现了一些解决方案,其中一个是使用新生儿筛查实验室的国际合作小组的在线数据库,这是一项被称为R4S(The Region 4 Stork Project)的计划,该项目涉及60个国家的200个实验室,拥有超过120万个数据点、至少16000例使用新生儿筛查显示为真阳性病例。这种创新的方法使用后分析解释工具和多变量模式识别软件,允许不基于正常与非正常价值的较大差异和解释,除了从一种疾病状态到另一种状态。“随着在新生儿筛查中被包含的检测项目的增多,性能指标也必须超过以往的历史水平,”Rinaldo说道。同时,他补充说道,目前新生儿筛查中的假阳性率是过高的。

  该解释性工具为单一条件寻找特异性,或为在多条件中寻找特异性提供支持,只要数据库中有足够的真阳性病例。所有患者的结果都合并成一份单一的分数报告,作为一个百分等级在所有条件相似的病例中并且跟单一的解释指引作对照。该系统的能量由可向该系统中输入他们的病例的绝对用户数量驱动。

  像那些在新生儿筛查中遇到的复杂状况的疾病,需要特定条件疾病范围更胜以传统分析为基础的临界值。R4S 后分析工具(R4S post-analytical tools)从传统的静态过程中获取临床验证,那是在测试发展到一个持续进化期间的早期表现,疾病范围的动态细化——它会随着整个进程的不断推进而持续改进。Rinaldo 已经在这个领域率先发力,虽然 R4S 计划的短期目标是在新生儿筛查结果中减少假阳性的数量,当被问及是否将来有一天消除以严格分析为基础的临界值时,他的回应是“那还用说”。
您需要登录后才可以回帖 登录 | 立即注册

本版积分规则

© 2013 艾兰博曼 All Rights Reserved.
( 浙ICP备07020270号-8 )
 
快速回复 返回顶部 返回列表