Artificial intelligence could diagnose rare disorders using just a photo of a face 人工智能只需一张脸的照片就能诊断出罕见的疾病。

时间: 2019-08-30

Rare disorders often show up in someone’s appearance. Individuals with Noonan syndrome—a genetic condition that inhibits the body’s growth and development—can have wide-set eyes, for example, and those with Bain type intellectual disability—caused by a mutated gene on the X chromosome—sport almond-shaped eyes and small chins (see above). Now, researchers have trained artificial intelligence to recognize these features, paving the way for early—and cheap—diagnoses.

罕见的疾病经常出现在某人的外表上。努南综合征患者-抑制身体生长发育的遗传条件-例如,可以有开阔的视野,而那些有贝恩型智力障碍的人-由X染色体上的突变基因引起-运动杏仁形眼睛和小下巴(见上文)。现在,研究人员已经训练了人工智能来识别这些特征,为早期和廉价的诊断铺平了道路。

Scientists built a computer program, DeepGestalt, and trained it on a publicly available data set of more than 17,000 photos of patients affected by more than 200 rare disorders. The program then used deep learning to recognize which patterns of markers were linked to hundreds of different genetic syndromes.

科学家们建立了一个计算机程序DeepGestalt,,并用公开的数据集对它进行了培训,该数据集包含了17000多张病人受到200多种罕见疾病影响的照片。然后,该项目使用深度学习来识别与数百种不同遗传综合征相关的标记模式。

In a test with 502 new images, DeepGestalt successfully placed the correct syndrome in its top 10 list 91% of the time, the researchers report today in Nature Medicine. The program also outperformed doctors in spotting patients with Angelman syndrome and Cornelia de Lange syndrome—an inherited genetic mutation that can cause, among other symptoms, low-set ears and an upturned nose—versus other disorders, and in separating patients with different genetic subtypes of Noonan syndrome.

研究人员今天在《自然医学》杂志上报道,在一项502张新图像的测试中,Deepgestalt成功地将正确的综合症排在了前10位,占91%。该项目在发现Angelman综合征和Cornelia de Lange综合征患者方面也优于医生-一种遗传性基因突变,除其他症状外,还可能导致耳朵下垂和鼻子翘起来-与其他疾病相比,以及分离具有不同努南综合征遗传亚型的患者。

The researchers say the tool could one day be used in combination with genome testing to help doctors search for specific genetic markers and more quickly home in on an accurate diagnosis. This could help reduce the time, cost, and emotional burden of the “diagnostic odyssey” on which millions of families embark each year, seeking care for someone with a rare genetic syndrome.

研究人员说,有朝一日该工具可以与基因组测试结合使用,帮助医生寻找特定的基因标记,更快地找到准确的诊断方法。这有助于减少每年数百万家庭进行的“诊断冒险”的时间、成本和情感负担,寻求对罕见遗传综合征患者的护理。

Given how easy it is to photograph a face, the tool could be abused by employers or insurance providers, the researchers acknowledge. They say proper regulation of the distribution and use of tools like DeepGestalt will be crucial.

研究人员承认,由于拍摄一张脸很容易,这种工具可能会被雇主或保险公司滥用。他们说,对DeepGestalt等工具的分发和使用进行适当的监管至关重要。


英文版:

Rare disorders often show up in someone’s appearance. Individuals with Noonan syndrome—a genetic condition that inhibits the body’s growth and development—can have wide-set eyes, for example, and those with Bain type intellectual disability—caused by a mutated gene on the X chromosome—sport almond-shaped eyes and small chins (see above). Now, researchers have trained artificial intelligence to recognize these features, paving the way for early—and cheap—diagnoses.

罕见的疾病经常出现在某人的外表上。努南综合征患者-抑制身体生长发育的遗传条件-例如,可以有开阔的视野,而那些有贝恩型智力障碍的人-由X染色体上的突变基因引起-运动杏仁形眼睛和小下巴(见上文)。现在,研究人员已经训练了人工智能来识别这些特征,为早期和廉价的诊断铺平了道路。

Scientists built a computer program, DeepGestalt, and trained it on a publicly available data set of more than 17,000 photos of patients affected by more than 200 rare disorders. The program then used deep learning to recognize which patterns of markers were linked to hundreds of different genetic syndromes.

科学家们建立了一个计算机程序DeepGestalt,,并用公开的数据集对它进行了培训,该数据集包含了17000多张病人受到200多种罕见疾病影响的照片。然后,该项目使用深度学习来识别与数百种不同遗传综合征相关的标记模式。

In a test with 502 new images, DeepGestalt successfully placed the correct syndrome in its top 10 list 91% of the time, the researchers report today in Nature Medicine. The program also outperformed doctors in spotting patients with Angelman syndrome and Cornelia de Lange syndrome—an inherited genetic mutation that can cause, among other symptoms, low-set ears and an upturned nose—versus other disorders, and in separating patients with different genetic subtypes of Noonan syndrome.

研究人员今天在《自然医学》杂志上报道,在一项502张新图像的测试中,Deepgestalt成功地将正确的综合症排在了前10位,占91%。该项目在发现Angelman综合征和Cornelia de Lange综合征患者方面也优于医生-一种遗传性基因突变,除其他症状外,还可能导致耳朵下垂和鼻子翘起来-与其他疾病相比,以及分离具有不同努南综合征遗传亚型的患者。

The researchers say the tool could one day be used in combination with genome testing to help doctors search for specific genetic markers and more quickly home in on an accurate diagnosis. This could help reduce the time, cost, and emotional burden of the “diagnostic odyssey” on which millions of families embark each year, seeking care for someone with a rare genetic syndrome.

研究人员说,有朝一日该工具可以与基因组测试结合使用,帮助医生寻找特定的基因标记,更快地找到准确的诊断方法。这有助于减少每年数百万家庭进行的“诊断冒险”的时间、成本和情感负担,寻求对罕见遗传综合征患者的护理。

Given how easy it is to photograph a face, the tool could be abused by employers or insurance providers, the researchers acknowledge. They say proper regulation of the distribution and use of tools like DeepGestalt will be crucial.

研究人员承认,由于拍摄一张脸很容易,这种工具可能会被雇主或保险公司滥用。他们说,对DeepGestalt等工具的分发和使用进行适当的监管至关重要。