A recent study in The Lancet Oncology showed that AI found 20 percent more cancers on mammogram screens than radiologists alone.
Published in The Lancet Oncology, the study analyzed the scans of 80,000 Swedish women with a moderate hereditary risk of breast cancer who had undergone a mammogram between April 2021 and July 2022. Half of these scans were read by AI and then a radiologist to double-check the findings. The second group of scans was read by two researchers without the help of AI. (Currently, the standard of care across Europe is to have two radiologists analyze a scan before diagnosing a patient with breast cancer.)
The study showed that the AI group detected cancer in 6 out of every 1,000 scans, while the radiologists detected cancer in 5 per 1,000 scans. In other words, AI found 20 percent more cancers than the highly-trained radiologists.
Scientists have been using MRI images (like the ones pictured here) to train artificial intelligence to detect cancers earlier and with more accuracy. Here, MIT's AI system, MIRAI, looks for patterns in a patient's mammograms to detect breast cancer earlier than ever before. news.mit.edu
But even though the AI was better able to pinpoint cancer on an image, it doesn’t mean radiologists will soon be out of a job. Dr. Laura Heacock, a breast radiologist at NYU, said in an interview with CNN that radiologists do much more than simply screening mammograms, and that even well-trained technology can make errors. “These tools work best when paired with highly-trained radiologists who make the final call on your mammogram. Think of it as a tool like a stethoscope for a cardiologist.”
AI is still an emerging technology, but more and more doctors are using them to detect different cancers. For example, researchers at MIT have developed a program called MIRAI, which looks at patterns in patient mammograms across a series of scans and uses an algorithm to model a patient's risk of developing breast cancer over time. The program was "trained" with more than 200,000 breast imaging scans from Massachusetts General Hospital and has been tested on over 100,000 women in different hospitals across the world. According to MIT, MIRAI "has been shown to be more accurate in predicting the risk for developing breast cancer in the short term (over a 3-year period) compared to traditional tools." It has also been able to detect breast cancer up to five years before a patient receives a diagnosis.
The challenges for cancer-detecting AI tools now is not just accuracy. AI tools are also being challenged to perform consistently well across different ages, races, and breast density profiles, particularly given the increased risks that different women face. For example, Black women are 42 percent more likely than white women to die from breast cancer, despite having nearly the same rates of breast cancer as white women. Recently, an FDA-approved AI device for screening breast cancer has come under fire for wrongly detecting cancer in Black patients significantly more often than white patients.
As AI technology improves, radiologists will be able to accurately scan a more diverse set of patients at a larger volume than ever before, potentially saving more lives than ever.
Vanessa Colimorio (left) and Sharon Kochlany (right) at a farm with their four-year-old twin daughters and one-year-old son. The kids share the same sperm donor.
Sharon Kochlany and Vanessa Colimorio's four-year-old twin girls had a classic school assignment recently: make a family tree. They drew themselves and their one-year-old brother branching off from their moms, with aunts, uncles, and grandparents forking off to the sides.
The recently-gained sovereignty of queer families stands to be lost if a consumer DNA test brings a stranger's identity out of the woodwork.
What you don't see in the invisible space between Kochlany and Colimorio, however, is the sperm donor they used to conceive all three children.
To look at a family tree like this is to see in its purest form that kinship can supersede biology—the boundaries of where this family starts and stops are clear to everyone in it, in spite of a third party's genetic involvement. This kind of self-definition has always been synonymous with LGBTQ families, especially those that rely on donor gametes (sperm or eggs) to exist.
But the world around them has changed quite suddenly: The recent consumer DNA testing boom has made it more complicated than ever for families built through reproductive technology—openly, not secretively—to maintain the strong sense of autonomy and privacy that can be crucial for their emotional security. Prospective parents and cryobanks are now mulling how best to bring a new generation of donor-conceived people into this world in a way that leaves open the choice to know more about their ancestry without obliterating an equally important choice: the right not to know about biological relatives.
For queer parents who have long fought for social acceptance, having a biological relationship to their children has been revolutionary, and using an unknown donor as a means to this end especially so. Getting help from a friend often comes with the expectation that the friend will also have social involvement in the family, which some people are comfortable with, but being able to access sperm from an unknown donor—which queer parents have only been able to openly do since the early 1980s—grants them the reproductive autonomy to create families seemingly on their own. That recently-gained sovereignty stands to be lost if a consumer DNA test brings a stranger's identity out of the woodwork.
At the same time, it's natural for donor-conceived people to want to know more about where they come from ethnically, even if they don't want to know the identity of their donor. As a donor-conceived person myself, I know my donor's self-reported ethnicity, but have often wondered how accurate it is.
Opening the Pandora's box of a consumer DNA test as a way to find out has always felt profoundly unappealing to me, however. Many people have accidentally learned they're donor-conceived by unwittingly using these tools, but I already know that about myself going in, and subsequently know I'll be connected to a large web of people whose existence I'm not interested in learning about. In addition to possibly identifying my anonymous donor, his family could also show up, along with any donor-siblings—other people with whom I share a donor. My single lesbian mom is enough for me, and the trade off to learn more about my ethnic ancestry has never seemed worth it.
In 1992, when I was born, no one was planning for how consumer DNA tests might upend or illuminate one's sense of self. But the donor community has always had to stay nimble with balancing privacy concerns and psychological well-being, so it should come as no surprise that figuring out how to do so in 2020 includes finding a way to offer ancestry insight while circumventing consumer DNA tests.
A New Paradigm
This is the rationale behind unprecedented industry news that LeapsMag can exclusively break: Within the next few weeks, California Cryobank, the largest cryobank in the country, will begin offering genetically-verified ancestry information on the free public part of every donor's anonymous profile in its database, something no other cryobanks yet offer (an exact launch date was not available at the time of publication). Currently, California Cryobank's donor profiles include a short self-reported list that might merely say, "Ancestry: German, Lebanese, Scottish."
The new information will be a report in pie chart form that details exactly what percentages of a donor's DNA come from up to 26 ethnicities—it's analogous to, but on a smaller scale than, the format offered by consumer DNA testing companies, and uses the same base technology that looks for single nucleotide polymorphisms in DNA that are associated with specific ethnicities. But crucially, because the donor takes the DNA test through California Cryobank, not a consumer-facing service, the information is not connected in a network to anyone else's DNA test. It's also taken before any offspring exist so there's no chance of revealing a donor-conceived person's identity this way.
Later, when a donor-conceived person is born, grows up, and wants information about their ethnicity from the donor side, all they need is their donor's anonymous ID number to look it up. The donor-conceived person never takes a genetic test, and therefore also can't accidentally find donor siblings this way. People who want to be connected to donor siblings can use a sibling registry where other people who want to be found share donor ID numbers and look for matches (this is something that's been available for decades, and remains so).
"With genetic testing, you have no control over who reaches out to you, and at what point in your life."
California Cryobank will require all new donors to consent to this extra level of genetic testing, setting a new standard for what information prospective parents and donor-conceived people can expect to have. In the immediate, this information will be most useful for prospective parents looking for donors with specific backgrounds, possibly ones similar to their own.
It's a solution that was actually hiding in plain sight. Two years ago, California Cryobank's partner Sema4, the company handling the genetic carrier testing that's used to screen for heritable diseases, started analyzing ethnic data in its samples. That extra information was being collected because it can help calculate a more accurate assessment of genetic risks that run in certain populations—like Ashkenazi Jews and Tay Sachs disease—than relying on oral family histories. Shortly after a plan to start collecting these extra data, Jamie Shamonki, chief medical officer of California Cryobank, realized the companies would be sitting on a goldmine for a different reason.
"I didn't want to use one of these genetic testing companies like Ancestry to accomplish this," says Shamonki. "The whole thing we're trying to accomplish is also privacy."
Consumer-facing DNA testing companies are not HIPAA compliant (whereas Sema4, which isn't direct-to-consumer, is HIPAA compliant), which means there are no legal privacy protections covering people who add their DNA to these databases. Although some companies, like 23andMe, allow users to opt-out of being connected with genetic relatives, the language can be confusing to navigate, requires a high level of knowledge and self-advocacy on the user's part, and, as an opt-out system, is not set up to protect the user from unwanted information by default; many unwittingly walk right into such information as a result.
Additionally, because consumer-facing DNA testing companies operate outside the legal purview that applies to other health care entities, like hospitals, even a person who does opt-out of being linked to genetic relatives is not protected in perpetuity from being re-identified in the future by a change in company policy. The safest option for people with privacy concerns is to stay out of these databases altogether.
For California Cryobank, the new information about donor heritage won't retroactively be added to older profiles in the system, so donor-conceived people who already exist won't benefit from the ancestry tool, but it'll be the new standard going forward. The company has about 500 available donors right now, many of which have been in their registry for a while; about 100 of those donors, all new, will have this ancestry data on their profiles.
Shamonki says it has taken about two years to get to the point of publicly including ancestry information on a donor's profile because it takes about nine months of medical and psychological screening for a donor to go from walking through the door to being added to their registry. The company wanted to wait to launch until it could offer this information for a significant number of donors. As more new donors come online under the new protocol, the number with ancestry information on their profiles will go up.
For Parents: An Unexpected Complication
While this change will no doubt be welcome progress for LGBTQ families contemplating parenthood, it'll never be possible to put this entire new order back in the box. What are such families who already have donor-conceived children losing in today's world of widespread consumer genetic testing?
Kochlany and Colimorio's twins aren't themselves much older than the moment at-home DNA testing really started to take off. They were born in 2015, and two years later the industry saw its most significant spike. By now, more than 26 million people's DNA is in databases like 23andMe and Ancestry; as a result, it's estimated that within a year, 90 percent of Americans of European descent will be identifiable through these consumer databases, by way of genetic third cousins, even if they didn't want to be found and never took the test themselves. This was the principle behind solving the Golden State Killer cold case.
The waning of privacy through consumer DNA testing fundamentally clashes with the priorities of the cyrobank industry, which has long sought to protect the privacy of donor-conceived people, even as open identification became standard. Since the 1980s, donors have been able to allow their identity to be released to any offspring who is at least 18 and wants the information. Lesbian moms pushed for this option early on so their children—who would obviously know they couldn't possibly be the biological product of both parents—would never feel cut off from the chance to know more about themselves. But importantly, the openness is not a two-way street: the donors can't ever ask for the identities of their offspring. It's the latter that consumer DNA testing really puts at stake.
"23andMe basically created the possibility that there will be donors who will have contact with their donor-conceived children, and that's not something that I think the donor community is comfortable with," says I. Glenn Cohen, director of Harvard Law School's Center for Health Law Policy, Biotechnology & Bioethics. "That's about the donor's autonomy, not the rearing parents' autonomy, or the donor-conceived child's autonomy."
Kochlany and Colimorio have an open identification donor and fully support their children reaching out to California Cryobank to get more information about him if they want to when they're 18, but having a singular name revealed isn't the same thing as having contact, nor is it the same thing as revealing a web of dozens of extended genetic relations. Their concern now is that if their kids participate in genetic testing, a stranger—someone they're careful to refer to as only "the donor" and never "dad"—will reach out to the children to begin some kind of relationship. They know other people who are contemplating giving their children DNA tests, and feel staunchly that it wouldn't be right for their family.
"With genetic testing, you have no control over who reaches out to you, and at what point in your life," Kochlany says. "[People] reaching out and trying to say, 'Hey I know who your dad is' throws a curveball. It's like, 'Wait, I never thought I had a dad.' It might put insecurities in their minds."
"We want them to have the opportunity to choose whether or not they want to reach out," Colimorio adds.
Kochlany says that when their twins are old enough to start asking questions, she and Colimorio plan to frame it like this: "The donor was kind of like a technology that helped us make you a person, and make sure that you exist," she says, role playing a conversation with their kids. "But it's not necessarily that you're looking to this person [for] support or love, or because you're missing a piece."
It's a line in the sand that's present even for couples still far off from conceiving. When Mallory Schwartz, a film and TV producer in Los Angeles, and Lauren Pietra, a marriage and family therapy associate (and Shamonki's step-daughter), talk about getting married someday, it's a package deal with talking about how they'll approach having kids. They feel there are too many variables and choices to make around family planning as a same-sex couple these days to not have those conversations simultaneously. Consumer DNA databases are already on their minds.
"It frustrates me that the DNA databases are just totally unregulated," says Schwartz. "I hope they are by the time we do this. I think everyone deserves a right to privacy when making your family [using a sperm donor]."
"I wouldn't want to create a world where people who are donor-conceived feel like they can't participate in this technology because they're trying to shut out [other] information."
On the prospect of having a donor relation pop up non-consensually for a future child, Pietra says, "I don't like it. It would be really disappointing if the child didn't want [contact], and unfortunately they're on the receiving end."
You can see how important preserving the right to keep this door closed is when you look at what's going on at The Sperm Bank of California. This pioneering cryobank was the first in the world to openly serve LGBTQ people and single women, and also the first to offer the open identification option when it opened in 1982, but not as many people are asking for their donor's identity as expected.
"We're finding a third of young people are coming forward for their donor's identity," says Alice Ruby, executive director. "We thought it would be a higher number." Viewed the other way, two-thirds of the donor-conceived people who could ethically get their donor's identity through The Sperm Bank of California are not asking the cryobank for it.
Ruby says that part of what historically made an open identification program appealing, rather than invasive or nerve-wracking, is how rigidly it's always been formatted around mutual consent, and protects against surprises for all parties. Those [donor-conceived people] who wanted more information were never barred from it, while those who wanted to remain in the dark could. No one group's wish eclipsed the other's. The potential breakdown of a system built around consent, expectations, and respect for privacy is why unregulated consumer DNA testing is most concerning to her as a path for connecting with genetic relatives.
For the last few decades in cryobanks around the world, the largest cohort of people seeking out donor sperm has been lesbian couples, followed by single women. For infertile heterosexual couples, the smallest client demographic, Ruby says donor sperm offers a solution to a medical problem, but in contrast, it historically "provided the ability for [lesbian] couples and single moms to have some reproductive autonomy." Yes, it was still a solution to a biological problem, but it was also a solution to a social one.
The Sperm Bank of California updated its registration forms to include language urging parents, donor-conceived people, and donors not to use consumer DNA tests, and to go through the cryobank if they, understandably, want to learn more about who they're connected to. But truthfully, there's not much else cryobanks can do to protect clients on any side of the donor transaction from surprise contact right now—especially not from relatives of the donor who may not even know someone in their family has donated sperm.
A Tricky Position
Personally, I've known I was donor-conceived from day one. It has never been a source of confusion, angst, or curiosity, and in fact has never loomed particularly large for me in any way. I see it merely as a type of reproductive technology—on par with in vitro fertilization—that enabled me to exist, and, now that I do exist, is irrelevant. Being confronted with my donor's identity or any donor siblings would make this fact of my conception bigger than I need it to be, as an adult with a full-blown identity derived from all of my other life experiences. But I still wonder about the minutiae of my ethnicity in much the same way as anyone else who wonders, and feel there's no safe way for me to find out without relinquishing some of my existential independence.
"People obviously want to participate in 23andMe and Ancestry because they're interested in knowing more about themselves," says Shamonki. "I wouldn't want to create a world where people who are donor-conceived feel like they can't participate in this technology because they're trying to shut out [other] information."
After all, it was the allure of that exact conceit—knowing more about oneself—that seemed to magnetically draw in millions of people to these tools in the first place. It's an experience that clearly taps into a population-wide psychic need, even—perhaps especially—if one's origins are a mystery.
In this illustration of immune defense, Y-shaped proteins called antibodies attack a virus.
When we get sick, our immune system sends its soldier cells to the battlefield. Called B-cells, they "examine" the foreign particles that shouldn't be in our bloodstream—and start producing the antibodies, the proteins to neutralize the invaders.
To screen the antibodies, scientists have developed a proprietary way to make the effective ones glow – like a literal "lightbulb" moment.
The better these antibodies are at neutralizing the pathogen, the faster we recover.
The antibodies acquired from COVID-19 survivors already showed promise in treating other patients, but because they must be obtained from people, generating a regular supply is not feasible. To close the gap, researchers are trying to identify the B-cells that make the best antibodies—and then farm them in laboratories at scale.
Scientists at Berkley Lights, a biotechnology company in California, have been screening B-cells from recovered patients and testing the antibodies they release for virus-neutralizing abilities. To screen the antibodies, scientists there have developed a proprietary way to make the effective ones glow – like a literal "lightbulb" moment.
So how does it work? First, the individual B-cells are placed into microscopic chambers called nano-pens, where they secrete the antibodies. Once released, the antibodies are flushed over tiny beads that have bits of the viral particles attached to them, along with special molecules that can emit fluorescent light.
"When an antibody binds to the bead, we see a bright light on the bead," explains John Proctor, the company's senior vice president of antibody therapeutics. "So we can identify which cells are making the antibodies."
Then the antibodies are tested for their ability to counteract the coronavirus's spike proteins, which the virus uses to break into our cells. Not all antibodies are equally good at this crucial defense move—some can block only parts of the virus's machinery, while others can neutralize it completely. Proctor and his colleagues are looking for the latter.
Once scientists identify the best performing B-cells, they crack the cells open—or in scientific terms "lyse" them—and extract the genetic instructions for making the antibodies. As it turns out, B-cells aren't very efficient at pumping out massive amounts of antibodies, so scientists insert these genetic instructions into a different, more prolific type of cell.
Named Chinese Hamster Ovary Cells or CHO, these cells are commonly used in the pharmaceutical industry because they can generate therapeutic proteins en masse. Under the right nutrient conditions, which include a lot of sugar, CHO cells can keep making the antibodies at commercial levels. "They are engineered to operate in very large bioreactors," Proctor explains.
While traditional antibody screening can take three months, the Beacon System can do it in less than 24 hours.
Berkeley Lights' technology has already been used to screen the antibodies of recovered patients from Vanderbilt University Medical Center. In another example, a biotech company GenScript ProBio used the platform to screen mice engineered to have human antibodies for the coronavirus.
In addition to its small, lab-on-a-chip size, Berkeley Lights' system allows scientists to greatly speed up the screening process. While traditional antibody screening can take three months, the Beacon System can do it in less than 24 hours. "We only need one B-cell per pen and a couple of beads to see that fluorescent signal," Proctor says. "It is a more advanced way to process and analyze cells, and that level of sensitivity is unique to our technology."
B-cells secreting antibodies inside the Berkeley Lights Beacon System Nano-Pens.
While vaccines are likely to take months to develop and test, antibodies might arrive to the battleground sooner. With the extremely limited treatment options for COVID-19, antibody-based therapeutics can potentially bridge this gap.