Astronomy Tool Can Now Detect COVID in Breath

2023-05-18 02:15:51
关注

Astronomers and physicists have long used a laser-based sensor called an “optical frequency comb” to study the material makeup of the cosmos and to make timekeeping more accurate. But the COVID pandemic has pushed this versatile tool from the world of space and physics into health care.

Optical frequency combs are lasers that simultaneously shoot pulses of light at multiple frequencies. Because these superfast pulses are precisely spaced along the light spectrum—from infrared through the visible colors to ultraviolet—they form a series of peaks on a graph of the frequencies that look like the teeth of a comb. This “comb” can be used in a variety of ways. For instance, different types of molecules absorb different colors of light; by detecting which colors of light are absorbed near specific frequencies, the comb can identify specific molecules in an air sample. In a recent study, scientists proved this tool can detect COVID from Breathalyzer-type tests in which subjects simply blow into a tube—potentially paving the way for fast, noninvasive diagnostic tests for a multitude of diseases.

Every time humans exhale, we expel more than 1,000 kinds of trace molecules called volatile organic compounds, or VOCs. “Changes in VOC profiles can be linked to specific health conditions,” says Cristina Davis, associate vice chancellor for interdisciplinary research and strategic initiatives at the University of California, Davis. Scholars have known for millennia that certain breath odors are associated with clinical diseases or disorders. References in ancient Greek and Chinese medical literature indicate that doctors used the nose as a diagnostic tool, Davis says. More recently, dogs have been trained to identify some diseases in humans—and laser comb detectors also need training, says physicist Jun Ye, co-author of the new study. “We are training our frequency comb nose using machine learning, and once it’s trained, it becomes an electronic dog—with much greater sensitivity,” Ye says.

This powerful artificial nose has the advantage of being able to sniff out disease in a way that’s quick and noninvasive, says the study’s lead author, Qizhong Liang, a graduate student at the University of Colorado Boulder. For maximum accuracy, Liang adds, positive results from the new technology’s COVID tests should be followed up with a more reliable PCR test. But for quick screening at airports, concert venues or hospitals, it could beat other methods, such as body temperature scans, that are used to assess potential COVID infection without requiring an invasive nose swab.

Frequency combs can do more than identify molecules. They were originally developed in the 1990s to make more accurate optical atomic clocks, for which the inventors won the 2005 Nobel Prize in Physics. The combs can measure the natural oscillation of atoms so precisely that they have become an indispensable component of atomic clocks, which keep time incredibly well by counting these oscillations. In astronomy, researchers use optical frequency combs to measure the frequencies of light coming from distant stars; disruptions can indicate a star has an exoplanet. In atmospheric science, they have been used to study greenhouse gases. And in 2008 Ye and his colleagues at JILA (a joint institute of the National Institute of Standards and Technology and the University of Colorado Boulder) first proved that the technology could be used as a breath test for disease biomarkers. They set this finding aside, but global events suddenly gave them a very good reason to revive the work in April 2020. “Using breath as a diagnostic tool has been around for a while,” Davis says, “but I think it took a pandemic for research interest to really to really be moved forward.”

Shortly after the COVID pandemic began, Ye got a call from his colleagues at the National Academies of Sciences, Engineering, and Medicine and the Air Force Office of Scientific Research. They wanted to know whether his early frequency comb “Breathalyzer” research might help develop a noninvasive COVID test.

Ye and his team started by updating their 2008 technology. The researchers extended the laser comb’s frequency range from the near-infrared region of the spectrum into the mid-infrared part—where molecules absorb light two to three times more strongly. That signal boost allowed the researchers to improve the tool’s detection sensitivity by 1,000-fold, letting them identify molecules at extremely low concentrations on the scale of hundreds of parts per trillion.

Next, Ye’s team collected breath samples from 170 University of Colorado Boulder students and staff from May 2021 to January 2022. Each participant received conventional PCR nasal swab COVID tests, and about half were positive. The researchers then used the frequency comb to analyze light-absorption patterns among molecules in the participants’ breath. Applying machine learning to the frequency comb data, combined with the already-known PCR data on who was positive or negative, they found six “discriminating molecules” that indicated COVID infection. The work was described in a paper published in April in the Journal of Breath Research.

Liang says AI was key to the project’s success because of the vast amount of information the frequency comb gathers when analyzing breath. “Machine learning can analyze all of this data simultaneously and can automatically figure out the best way of utilizing all of that discriminating information to make a prediction model,” he says.

Frequency combs aren’t the only way to test human breath for COVID or other diseases. Other methods include gas chromatography/mass spectrometry systems such as the InspectIR test, which received emergency use authorization from the U.S. Food and Drug Administration in April 2022. In such chemistry-based techniques, the gas molecules to be analyzed are separated by an inert gas, broken down into fragments and then measured. Davis calls these types of tests the “gold standard,” but they require time, specialized training and bulky equipment that limits their use to the lab. Davis has been working on a smaller, portable type of test, an ion mobility spectrometer, which identifies substances based on the mobility of their molecules in an electric field. Other options use chemicals that bind to VOCs to isolate and test them. “There are more than 15 companies working on a variety of these kinds of tests,” Davis says.

The frequency comb technology is different, Liang says, because it uses laser spectroscopy and therefore “detects the molecules in breath in a nondestructive way.” By this he means the comb does not cause a sample to degrade or create any unwanted by-products, as breath tests that rely on chemical reactions can. Frequency comb technology also has the potential to be really, really fast: it could potentially eventually provide results in seconds, compared with minutes in other breath tests.

That said, you probably won’t see laser combs the next time you catch a flight. “Breath tests, in general, have not reached prime time yet,” says Wilbur Lam, a COVID test expert, pediatrician and biomedical engineer at Emory University and the Georgia Institute of Technology. With the frequency comb method, he says, “you get an optical signal, and whether that optical signal is truly indicative of a COVID infection really has to be proven. Right now, they’re showing some correlation. But how does it correlate with other types of conditions that could affect the breath?”

If frequency comb “Breathalyzers” do prove themselves in further research, they could make a huge difference in many clinical settings beyond rapid testing for COVID. Study co-author Kristen Bjorkman, director of interdisciplinary research at the BioFrontiers Institute, suggests this technology might one day be used to detect chronic obstructive pulmonary disease, kidney failure, lung and pancreatic cancers and even Alzheimer’s disease. Multiple early studies have provided preliminary evidence that the contents of exhaled breaths can be used for these diagnoses.

Breathalyzer-style tests could also be ideal for diagnosing children, and Ye says some pediatricians have already approached him about a frequency comb test for asthma in kids. When a child shows up sick at the emergency room, Ye explains, a lot of invasive tests are required to determine if the symptoms are caused by a bacterial or viral illness or asthma. He says one Denver-based pediatrician told him, “‘Imagine you can do a breath analysis on children, which is totally noninvasive. Kids won’t cry if they have to just donate a breath.’”

参考译文
天文学工具现可检测呼气中的新冠病毒
天文学家和物理学家长期以来一直使用一种称为“光频梳”的激光传感器,用于研究宇宙物质的组成,并提高计时的精确度。但新冠疫情的爆发,使得这一多功能工具从太空和物理领域进入医疗领域。光频梳是能同时以多种频率发射脉冲光的激光器。由于这些超快脉冲在光谱上排列得非常精确——从红外线到可见光再到紫外线——它们在频率图谱上形成了一系列尖峰,看起来就像梳子的梳齿。这个“光频梳”可以用于多种用途。例如,不同的分子会吸收不同颜色的光;通过检测特定频率下哪些颜色的光被吸收,光频梳可以识别空气样本中的特定分子。在最近的一项研究中,科学家证明,这种工具可以通过呼吸测试检测出新冠——只需受试者对着一个管子吹气,就有望为多种疾病的快速、无创诊断测试开辟道路。每次人类呼气时,我们都会呼出超过1000种称为挥发性有机化合物(VOCs)的微量分子。“VOC模式的变化可以和特定的健康状况联系起来,”加州大学戴维斯分校的跨学科研究及战略倡议副主管克里斯蒂娜·戴维斯说。学者们早已知道某些呼吸气味与临床疾病或紊乱有关。戴维斯指出,古代希腊和中国医学文献中的记载表明,医生曾以鼻子作为诊断工具。最近,狗已被训练用于识别人体中的某些疾病,物理学家、这篇新研究的共同作者荣·叶说,光频梳探测器也需要“训练”。“我们正在用机器学习训练我们的光频梳‘鼻子’,一旦训练完成,它就变成了一只‘电子狗’,具有更高的灵敏度,”叶说。这项强大的“人工鼻”优势在于,它能够以快速、无创的方式检测疾病,新研究的主要作者、科罗拉多大学博尔德分校的研究生梁启忠指出。为了达到最大准确性,梁补充说,新技术的新冠检测结果应随后进行更可靠的PCR测试加以确认。然而在机场、演唱会场地或医院等进行快速筛查时,它可能比其他方法,比如体温扫描,要更胜一筹,后者用于在不需要侵入性鼻拭子的情况下评估潜在的新冠感染。光频梳的功能不仅限于识别分子。它们最初是在20世纪90年代开发的,用于制造更精确的光学原子钟,其发明者因此获得了2005年的诺贝尔物理学奖。光频梳可以测量原子的自然振荡如此精确,以致它们已成为原子钟中必不可少的组成部分,原子钟通过计算这些振荡来保持极高的计时精度。在天文学中,研究人员使用光频梳来测量来自遥远恒星的光的频率;扰动可以表明恒星拥有系外行星。在大气科学中,它们被用来研究温室气体。2008年,叶和他的同事们(JILA研究所,系美国国家标准与技术研究院和科罗拉多大学博尔德分校的联合研究所)首次证明了这一技术可以用于疾病生物标志物的呼吸测试。他们将这一发现暂搁一旁,但全球事件突然给他们一个极好的理由,在2020年4月重新启动这项研究。“将呼吸作为诊断工具已经有一段时间了,”戴维斯说,“但我认为,这次疫情推动了人们对这项研究的兴趣真正迅速前进。”新冠疫情刚刚开始后不久,叶就接到美国国家科学院、工程院和医学院以及空军科学办公室的同事打来的电话,他们想知道他的早期光频梳“呼吸检测器”研究是否能帮助开发一种无创的新冠检测。叶和他的团队从更新2008年的技术开始入手。研究人员将激光梳的频率范围从光谱的近红外区域扩展到中红外区域——在那里,分子对光的吸收强度要高出两到三倍。这种信号增强使得研究人员能够将工具的检测灵敏度提高1000倍,使他们能在极低浓度下识别分子,浓度低至每万亿分之几百的水平。接下来,叶的团队从2021年5月到2022年1月,收集了170名科罗拉多大学博尔德分校学生和员工的呼吸样本。每位参与者都接受了传统的PCR鼻拭子新冠检测,大约一半人呈阳性。研究人员随后使用光频梳分析参与者呼吸中分子的光吸收模式。他们将机器学习应用于光频梳数据,并与已知的PCR检测数据(谁阳性或阴性)结合,发现了六个“辨别分子”,这些分子可以指示新冠感染。这项工作在2023年4月发表于《呼吸研究杂志》。梁说,人工智能是该项目成功的关键,因为光频梳在分析呼吸时能收集到大量信息。“机器学习可以同时分析所有这些数据,并能自动找出最佳方式,利用所有这些辨别信息来构建预测模型,”他说。光频梳并不是检测人类呼吸中新冠或其他疾病的方法。其他方法包括气相色谱-质谱系统,例如在2022年4月获得了美国食品药品监督管理局(FDA)紧急使用授权的InspectIR检测。在这些基于化学的技术中,要分析的气体分子首先由惰性气体分离,然后被打碎成碎片并进行测量。戴维斯称这些测试为“黄金标准”,但它们需要时间、专门的培训以及笨重的设备,这限制了它们仅能在实验室中使用。戴维斯正在研究一种更小、便携的测试,称为离子迁移谱仪,它通过测量分子在电场中的迁移速度来识别物质。其他方法使用能与VOCs结合的化学试剂以分离和检测它们。“目前有超过15家公司正在研究各种这样的测试,”戴维斯说。梁认为,频率梳技术不同,因为它使用激光光谱技术,因此“以一种非破坏性方式检测呼吸中的分子”。他的意思是,频率梳不会导致样本降解,也不会像依赖化学反应的呼吸测试一样产生任何不想要的副产物。频率梳技术还可能非常非常快:它最终可能在几秒钟内提供结果,而其他呼吸检测需要几分钟。尽管如此,当你下次坐飞机时,你可能不会看到激光频率梳的应用。“一般来说,呼吸测试尚未进入主流,”埃默里大学和佐治亚理工学院的新冠测试专家、儿科医生和生物医学工程师威尔伯·兰说。他说,对于频率梳方法,“你得到了一个光学信号,这个光学信号是否真的能代表新冠感染,还需要被证明。目前,他们只显示出一些相关性。但这种相关性是否与可能影响呼吸的其他情况相关?”如果频率梳“呼吸检测器”能在进一步的研究中证明其有效性,它们可能在许多临床环境中带来重大变革,远不止新冠的快速检测。这项研究的共同作者、BioFrontiers研究所跨学科研究主任克里斯汀·比约克曼建议,这项技术有一天可能会用于检测慢性阻塞性肺病、肾衰竭、肺癌和胰腺癌,甚至阿尔茨海默病。多项早期研究已提供了初步证据,呼出的呼吸成分可用于这些诊断。呼吸检测风格的测试也可能特别适合儿童的诊断,叶说,一些儿科医生已经联系他,希望使用频率梳测试儿童哮喘。当一个孩子因病来到急诊室时,叶解释说,需要很多侵入性测试,以确定症状是由于细菌或病毒感染还是哮喘引起的。他说,一位来自丹佛的儿科医生告诉他:“‘想象一下,你可以对儿童进行呼吸分析,这完全是非侵入性的。如果他们只需捐献一口气,孩子们就不会哭。’”
您觉得本篇内容如何
评分

评论

您需要登录才可以回复|注册

提交评论

提取码
复制提取码
点击跳转至百度网盘