Tuesday, July 1, 2014
Guest Blogger Professor Seema Mohapatra: Facial Recognition Technology, Genetic Diseases, and Privacy
Recently scientists at Oxford University revealed that they have developed a facial-recognition program that uses ordinary family photos to help diagnose rare genetic conditions. The computer software uses facial feature recognition to look for similarities from a bank of publically available computer-based photos of facial structures that have similar genetic conditions, such as Down Syndrome and Angelman Syndrome. The new computer software was developed by Christoffer Nellåker and Andrew Zisserman of the University of Oxford, along with their colleagues. It was designed to assist doctors in making a preliminary diagnosis of rare diseases. Currently, it correctly predicts a genetic disorder on average 93 percent of the time.
The concept of using facial feature recognition to diagnose diseases is not new, as many rare disorders do not have an accompanying genetic test, and specialists rely on analysis of facial features to help in diagnosis. However, doctors with the requisite skill set for such diagnoses are scarce, and thus, until this point, many families have had to wait years for their child’s prognosis.
There are upwards of 17,000 genetic disorders that have been diagnosed, of which about 700 can be diagnosed with the assistance of abnormal facial characteristic recognition. The program developed by the scientists at Oxford currently can help identify 90 individual disorders, and the hope is that it will be used where specialists are unavailable. The software is not currently used as a sole test for diagnosis, but is used for assisting pediatricians in the process. According to Nellåker, "it's not sufficiently accurate to provide a rock-solid diagnosis, but it helps narrow down the possibilities." The researchers submitted an image of Abraham Lincoln, who has long been suspected of having Marfan Syndrome, through the program. Sure enough, Marfan was one of the suspected diseases based on his facial features.
The researchers hope that in the future, this program could be used to identify those born with the disorder, allowing for the option of early treatment. Additionally, “[a] diagnosis of a rare genetic disorder can be a very important step” and “can provide parents with some certainty and help with genetic counseling on risks for other children or how likely a condition is to be passed on,” according to lead researcher, Nellåker.
Although the researchers may have the best intentions, there are some serious privacy concerns that may arise should this technology be used for other purposes. Last year, the American College of Medical Genetics and Genomics (ACMG) published its recommendation regarding the disclosure of information found during genome sequencing in clinical settings. The recommendation included disclosure of “incidental findings,” meaning that secondary predispositions unrelated to the reasons for undergoing the sequencing should also be disclosed to the patient. However, on April 1, 2014, a new policy was announced by the ACMG, effectively reversing its earlier position, and stating that a patient should be able to decide if they want to know about any unexpected test results. The new recommendation includes a push for physician consultation regarding the question if the tests should be done at all. Therefore, the decision making point is before the sample is sent to the lab, rather than when the results are received by the clinician. So, imagine a scenario where a patient is being tested for disease A, but then shows signs of Disease B in the facial recognition screening? In the clinical setting, the physician should discuss this possibility with a patient prior to deciding whether to undergo such screening.
What if this technology becomes the standard of care as a part of a pediatric well visit? Could parents decide for their children that their children should be evaluated “just in case”? Could children opt out of such screening?
What is even more potentially worrisome about this technology is that it will not necessarily be limited to the doctor-patient realm and may become more readily available. Let’s say anyone could use the picture of me on this blog or on my Facebook page to screen me for genetic ailments. What laws would prevent that from happening right now? Would that be considered be an invasion of privacy tort? It is not clear I currently have a reasonable expectation of privacy of my facial features when I have a profile picture or headshot on my law school website.
Additionally, should facial recognition technology be added as a sort of genetic information that should be covered under the Genetic Information Nondiscrimination Act (GINA)? Because the software just narrows down a possibility and is not necessarily revealing genetic information, it is not clear GINA applies to such technology. So, could employers run pictures of candidates through the software to ensure “healthy” candidates? Some internet dating sites are already using facial recognition technology to help match one with a date that looks like his or her ex, by request. Could E-harmony lead to E-ugenics by using the facial recognition screening software to exclude individuals who may show signs of genetic ailments? I am not sure of the answers to these questions, but ,any of these issues need to be explored as technology is moving at a faster pace than our laws.
 Chris Weller, Rare Genetic Disorders Could Be Diagnosed With Facial Recognition Computer Software, MEDICAL DAILY, Jun. 24, 2014. Available here.
 Josh Clark, Does the shape of my face show that I have a genetic disorder? Available here.
 Diagnose Rare Disorders Using Photos, DECCAN CHRONICLES, Jun. 25, 2014. Available here.