Thinking About Ob/Gyn
A fresh and evidence-based perspective of all things related to obstetrics and gynecology. Follow us on Instagram @thinkingaboutobgyn or visit thinkingaboutobgyn.com for show notes and more.
Thinking About Ob/Gyn
Episode 10.5 The Dense Breast Dilemma
Howard and Antonia explore the safety of medications during pregnancy and the controversial reporting requirements for breast density on mammograms, examining how science is being overshadowed by fear-mongering in healthcare decision-making.
• Examining the evidence behind avoiding fluconazole (Diflucan) in first trimester, finding that short courses likely pose minimal risk
• Discussing the important distinction between possibility and probability when evaluating medication safety in pregnancy
• Analyzing the wide variation in cesarean delivery rates across US counties, from 5.4% to over 53% for low-risk patients
• Critiquing politically-motivated FDA actions on SSRIs, food dyes, and other health policies not supported by scientific evidence
• Explaining why the FDA's requirement to notify women of dense breasts on mammograms may cause more harm than good
• Demonstrating how supplemental testing for women with dense breasts leads to false positives and unnecessary procedures
• Reviewing the historical development of prenatal diagnosis from early ultrasound to cell-free DNA testing
00:00:33 Evidence for Diflucan in Pregnancy
00:12:12 Cesarean Delivery Rates Across US Counties
00:16:39 FDA's Position on SSRIs and Food Dyes
00:28:46 Managing Dense Breasts in Mammography
00:44:46 History of Prenatal Abnormality Diagnosis
Follow us on Instagram @thinkingaboutobgyn.
Welcome to Thinking About OBGYN. Today's episode features Howard Harrell and Antonia Roberts discussing breast density and more.
Speaker 2:Howard.
Speaker 3:Antonia.
Speaker 2:What are we thinking about on today's episode?
Speaker 3:Well, we're going to talk about the management of dense breasts when we encounter those on mammography and for cancer screening and some other new quick articles. But first, what's the thing we do without evidence?
Speaker 2:Well, here's one I've heard about, although I don't think I've personally seen or even thought about doing this myself, and that is avoiding short courses of diflucan or fluconazole in the first trimester out of concern for birth defects.
Speaker 3:Okay, sure. Well, this is a reminder that every medication that we use, even prenatal vitamins, has not undergone the normal sort of clinical trials for safety and efficacy that we would normally do with medications, because of ethical concerns about randomizing pregnant patients to medications to any medication versus some placebo, and then carefully measuring what the differences in outcomes are. So all of our knowledge about medication safety is based upon what may be low quality retrospective epidemiologic data, and much of that data ends up being mixed because no two data sets look exactly the same, and it's a difficult problem. If any of it does tend to show a trend towards, say, more malformations with the medication, then people become very concerned, though often the difficulty is determining noise in the data from true signal, and this gets down to confounding and recall bias and all the problems. With retrospective data like this and I think that's the case here, where there are lots of little studies, little retrospective studies, the data is very mixed. Some of this data has also been written about by people who specialize in lawsuits and fear-mongering and things like that. So, in particular, I've written before on Howardisms about a Canadian researcher named Berard who seems to universally find that all medicines cause significant problems. Based upon what I would call faulty research methodologies also what judges call them that and then testifies on behalf, seemingly always, of a plaintiff's problems based upon what I would call faulty research methodologies Also, a judge has called them that and then testifies on behalf, seemingly always, of a plaintiff's attorney against manufacturers. So it takes a lot of discipline to look through some of this literature and think about what it really means and what inferences we can really draw from it. And we're living in this moment now where everyone is seemingly scared of everything due to so much fear mongering on social media and from some of our new government officials who also are trial attorneys who like filing lawsuits against manufacturers like vaccine manufacturers.
Speaker 3:But one key mistake is that people confuse possibility with probability. So just because something could possibly cause a birth defect doesn't mean, for example, that it's likely that it does or that it's probable. Now Burrard, in particular, did author a paper about fluconazole in 2019. And this paper concluded that fluconazole increased the risk of spontaneous abortion and that doses higher than 150 milligrams might increase the risk of cardiac defects. Of course, a paper like that will create a firestorm in the lay media.
Speaker 3:Studies of this nature will often involve a million patients and the mining of their electronic data and their records, and with relatively rare events like newborn cardiac anomalies. These sorts of approaches are very sensitive to underreporting or overreporting and they're notorious for producing false positive results. There was a paper in 2020 in the British Medical Journal that concluded that fluconazole did not cause oral clefts or conral-troncal malformations, but that there might be an association with musculoskeletal malformations. So different malformations there's all these different malformations you could look at in some of these mixed data. Some will show no association with one type of malformation and then another one will show an association with a different malformation, but not the one that the previous paper didn't find. This is mixed data. This is usually noise when you see it.
Speaker 2:Yeah. So two different papers finding two very, very unrelated types of findings that are not consistent with each other. So that should make us question whether either paper is correct because of that amount of inconsistency. And it's also true that if a study that's looking at especially millions of people and it finds a few dozen abnormalities within those millions of people millions of people and it finds a few dozen abnormalities within those millions of people, it's likely that those few dozen patients with the abnormalities probably after the fact probably thought about all of their possible medical or dietary exposures a lot more than the rest of the millions of patients that didn't have abnormalities.
Speaker 2:Because this is a retrospective thing, they're being asked after these things were identified and diagnosed. So it's like a recall bias thing. Usually, if anyone has any kind of adverse or unexplained outcome, they're going to tend to really rack their brains thinking was it because of that one time I had that one infection or that fever, or that vaccine or that antibiotic or, in this case, that one treatment for yeast? So when you go back years later and just look at the charts of these mothers whose children had abnormalities, you're probably just going to find more information just for that reason, especially relating to medical or environmental exposures, than you will in the charts of the healthy controls. Essentially.
Speaker 2:So that is a fundamental methodological limitation of retrospective data like this. It's not like they have a prospective collection of data for this or that. They called each person in the study and asked them specifically whether they were exposed to something, or especially when or how much. So common medications that women take while pregnant will include things like antibiotics for UTIs, medications for yeast infections or bacterial vaginosis, and Tylenol for various aches and pains. So you see all sorts of studies on all those different kinds of medications that have this same fundamental flaw that attributes abnormalities to those medications. But the issue is those patients are much more likely to have their use of those medications recorded in the chart when they come up with some kind of fetal abnormality, compared to their peers who probably used the same drug but had a healthy baby.
Speaker 3:Yeah, the phenomenon happens too where the general OBGYN does a screening ultrasound, find something that they're worried about, sends it then to the maternal fetal medicine doctor and then eventually there's an anomaly diagnosed with ultrasound. But if you look in the charts of women who have had extra ultrasounds or fetal MRIs or things like that in maternal fetal medicine offices, you'll see CPT codes like exposure to certain medications.
Speaker 3:It might be an antidepressant or an antipsychotic or an antiepileptic or whatever. There'll be all sorts of codes that are detailing these exposures that they've put in there that the general OBGYN doesn't have on all the patients who have normal charts and or normal ultrasounds in those charts. So you end up with overreporting of medication exposures and underreporting in the normal cases. And then there is that recall bias. So if your child has autism and someone comes to you and takes it, maybe a pediatrician takes a detailed developmental history and they ask you what you were exposed to or what you might have taken during pregnancy, you're going to think about all these things.
Speaker 3:You might have already thought about them because you're already primed by social media and things like this, as you said, to rack your brain and think about what you did wrong to cause your child to have a problem. So you think about the Tylenol that you took and the food coloring, because you heard that was bad, or the hair dye or hairspray or the fluconazole or the epidural that you received or whatever else that you've been primed to think of due to social media, influencers and other folks. But then in the control sample where you pull out hundreds or thousands or a million electronic files of women who had no abnormalities reported for their children. Well, those patients didn't give the degree of specificity about all these things because it didn't matter to anyone. It was never reported in the chart.
Speaker 2:Yeah, the only thing really I can think of. I have two kids. One has autism, one doesn't. I drank a lot of green tea with the it must be the green tea. Yeah, tiktok, better pick up on that now. So what's our best answer on the fluconazole then?
Speaker 3:Is there any cause for concern? Well, we'll put a link to a 2024 meta-analysis in the European Journal of Epidemiology that concluded that fluconazole used during the first trimester does not significantly increase the risk of major congenital malformations, either overall or by subtype. When you adjust the data correctly, they say there still might be a risk and again, don't confuse possibility with probability. There might be a risk of major congenital malformations with sustained doses or doses greater than 150 milligrams. So you know something other than for a yeast infection. Somebody has a serious fungal infection and they're on sustained long-term doses, for example. So I think the takeaway is that giving a Diflucan or two in the first trimester is very unlikely to have any negative effect when you look at the whole of the data. But if you had a patient who needed a 10-day course or a prolonged course or higher doses than that, then there's still a possibility not a probability, but a possibility of an increased risk of major congenital malformations.
Speaker 2:Yeah, but possibility is not the same as probability, right?
Speaker 3:And I think, yes and I think that's the moment we're living in A lot of things are possible, and sometimes we continue to say that a thing's possible until we have sufficient evidence to say that it's not possible. But in pregnancy in particular, in these sorts of exposures, we don't ever have those trials, and so it takes a pretty high bar for us to say we don't think there's a connection, just as we say there's no connection between vaccines and autism. We have that compelling data that says it. It's a scientific impossibility. But just because something's possible doesn't mean that it's probable. So we err on the side of minimal risk exposure while you're pregnant. So we try to avoid anything unnecessary medications or anything during pregnancy. And even when we say to avoid something, that doesn't mean we know that they're dangerous or harmful. And I do think, though, that this idea of minimal risk and saying that something's possible but not necessarily probable, it gets us confused, and so we're living in an age where this difference is important, I think.
Speaker 2:So avoid high dose or prolonged dosing of fluconazole in the first trimester.
Speaker 3:Yeah, and then Bactrim and Coumadin and lisinopril and losartan and lithium and methimazole some of the anti-seizure medications, if you can control them without them. But remember, even with the seizure medications we're weighing risk versus benefit. In fact, with many of those medications we're weighing risk versus benefit. It's not that we're not aware that there may be an increased risk associated with the use of some of the anti-seizure medications. It's just that not using them is associated with more risk. So we're trying to minimize risk and we'd prefer that the person didn't have a seizure disorder. But here we are and we have to treat them. For Bactrim we do have good alternatives and it's one of the reasons why I tend to always just prefer Macrobid as my first line medication for urinary tract infections during pregnancy. And obviously we don't use ACE inhibitors or angiotensin receptor blockers and we switch those medicines off of those patients in the first trimester when we see them.
Speaker 2:Yeah, because they have good alternatives too, right, all right, well, let's move on. Did you see the research letter in the Green Journal August 7th of 2025 that looked at the risk of cesarean delivery in individual counties across the US?
Speaker 3:delivery in individual counties across the US. I did. I was a little upset that they didn't give us the actual raw data set in a spreadsheet to look at, because I had to zoom way in on their maps to find my county. I had to actually cross-reference this with a map to figure this out and see whether it was green or not.
Speaker 2:And green is good right.
Speaker 3:It is.
Speaker 2:All right, well, was your county green?
Speaker 3:It was, I was in a green county. Right, it is All right.
Speaker 2:Well, was your county green.
Speaker 3:It was. I was in a green county.
Speaker 2:Literally in a green county. Yeah, nice pun there, well, okay, well, my county was not green NTSV. The nulliparous term singleton vertex cesarean rate above 23.6, but less than 30%. So they picked 23.6% as a cutoff for their the little healthy people 2030 campaign target. There's also the next color up is a dark blue where the rate is greater than 30. So at least it could have been worse in my county.
Speaker 3:Yeah Well, it's not a competition. There aren't that many counties that are dark blue, but unfortunately they happen to be in a lot of the more populous areas of the United States. The whole Southern tip of Florida looks pretty bad, honestly.
Speaker 2:And then, on the opposite extreme, they recorded 5.4% NTSV rate and this was in a county in Alaska. And then the highest rate in any US county was 53.4% and that was in Seminole County, Georgia. And again, that's just NTSV rate, not their total cesarean rate. That would include repeats, Repeats, yeah. And so in theory, NTSV is supposed to represent the simplest, lowest risk patients for cesarean.
Speaker 3:Yeah, well, in general the South had the highest rate when you look at this, while the West had the lowest rate. And I think one of the interesting things about studies like this one other than finding your county on the map, if you want to is that it shows how wide the variation is for the same metric in different parts of the country. Yeah, of course, those patients in that small area in Alaska they are not the same patients as in other places. You can imagine all the high-risk folks aren't staying there while they're pregnant or these are a different patient population. But still, the general trend of low single digits in some counties in the United States to 30 or 40 or even 50 plus percent is not fully accounted for just by patient differences, not even close. So that much variation is unnecessary and it shows a quality of care issue is unnecessary and it shows a quality of care issue.
Speaker 3:I say they need to put all that data online and let folks look at it hospital by hospital. Until they do, you won't see many real attempts to improve it. Now, what's the right number? That's a whole other debate. It's possible that some counties have too low of a rate, but clearly, if you look at the averages. There's some very wide variation in outlying counties and that number should be well less than 22%. I think it should be in the low teens.
Speaker 2:I'm trying to look again to see if they talked anything about BMI. There's some regional variation there, but it's not to the same scale.
Speaker 3:They didn't account for it.
Speaker 2:Okay, they didn't have that level of data. It's just the reporting data. Okay, they didn't have that level of data. Fair.
Speaker 3:But I think, yes, that's why it's higher in the South in general than it is in the West is. But I think, yes, that's why it's higher in the South in general than it is in the West is because BMIs are higher in the South than they are in the West. I have seen individual hospitals posting individual NTSC cesarean rates just in their physician workspaces, where it'll list the rate for each individual provider.
Speaker 2:Yeah, and California has shown that that will lead to some competition in lowering rates. Yeah, yeah, and California has shown that that will issued for SSRIs in pregnancy. It's being considered by the FDA after their scientific committee was replaced by a whole new panel of people by Mr Kennedy.
Speaker 3:Not doctor, right, mr Kennedy?
Speaker 2:Mr.
Speaker 3:Kennedy, jr Kennedy, not doctor right.
Speaker 2:Mr Kennedy, mr Kennedy Jr.
Speaker 3:Yeah, I'm afraid we're going to fill several episodes with Mr Kennedy Jr's actions. But yeah, there was. In July 21st there was a meeting in which several Kennedy appointees these are all interesting. You can find this information and drill down on who these folks are. They spoke about potential birth defects associated with SSRIs, but none of them were able to show any scientific evidence that indicated that these medicines are bad for pregnant women or that they cause birth defects. We discussed the safety of SSRIs back in episode 6.1, and we did a deep dive there if people want to listen to some of that. But it should suffice to say that early preliminary data about possible issues with SSRI and their use during pregnancy has not been borne out by larger studies and we continue to conclude that they are safe to use as any other medication we use during pregnancy. So this FDA panel which got together and made these unscientific assertions, like a lot of other activity that the FDA is doing now, will likely lead to patients suffering and maybe even dying unnecessarily out of fear-mongering.
Speaker 2:Yeah, it's not an exaggeration to say that treating depression saves lives and antidepressants save lives. We should remind listeners that suicide is a leading cause of death for pregnant and postpartum women in the United States and in fact, when you count suicide and substance abuse in the umbrella of mental health conditions, then this category of mental health conditions is the number one cause of death for new and expecting mothers in the US. So I'm really struggling to understand what is the agenda for Mr Kennedy and this panel to try to tear down one of the few things that actually helps with this number one cause of maternal death.
Speaker 3:Yeah, well, or add in vaccination. In the same way, if you look at the spike in maternal mortality related to COVID, you can infer from that how many women's lives were saved from the COVID vaccine, and at least during the pandemic, the intense couple of years of the pandemic that was the leading cause of maternal death, and we could prevent 90-something percent of it 97% of it with vaccines. And the women who chose to get vaccinated, they didn't die from it, and the increased mortality we saw was in the group of women who chose not to. So, yes, what is the agenda here? I would encourage folks to listen to that prior episode because their patients are going to come in and I'm seeing it on a daily basis, where people are coming in and parroting to me the things that the FDA and RFK Jr are saying, and also the American College of Obstetricians and Gynecologists and the Society for Internal Fetal Medicine have recently released statements reiterating that these medications are safe to use during pregnancy, and we can put a link to that statement from the SMFM. More and more we're going to have to start telling our patients that information that comes from the CDC and the FDA is unreliable and unscientific. This includes, obviously, information about vaccines, but food safety and a lot of other items that we're seeing.
Speaker 3:I mentioned food safety because a lot of folks may be excited about actions that the current administration has taken against things like high fructose corn syrup and food dyes. Both of these have been popular as fringe health concepts for a while and lots of doctors, including well-intentioned OBGYNs and others, parrot misinformation about these types of items. But, just like SSRIs during pregnancy, there's no evidence that any of the food dyes that have been recently banned or high fructose corn syrup are uniquely unhealthy. We have no evidence that any of the banned food diets are associated with any negative health outcomes. And, of course, we do have a problem with people consuming too much sugar in their diets. That's for sure. But high fructose corn syrup is no worse than cane sugar and, in fact, by making the claim that it is that cane sugar is somehow healthier, we might see people switch to over-consuming cane sugar because they feel like it's natural and healthy compared to high fructose corn syrup.
Speaker 2:Yeah, Just going from one bad thing to another rather than just actually reducing excessive sugar intake. And yeah, these ideas are popular across political parties. So if you go to a lot of those higher end fancier grocery stores which I love going to, it's always such a treat. Some of that stuff, you see, you have to take with a grain of salt. Like they don't sell the typical American Coca-Cola because it has the high fructose corn syrup but then they sell the Mexican Coca-Cola in the glass bottles because it has the natural cane sugar in it. The implication is that somehow the cane sugar is healthier or better for you, even though it has the same amount of simple added carbs in there. So a lot of people who would oppose the current administration's attacks, maybe on the vaccines and medicine in general, at the same time are supporting these nonsensical actions in the food industry?
Speaker 3:Yeah, and they shouldn't. The science is science here. And if you're attacking those items and you're supporting the type of science that would make a link between vaccines and autism, for example. So if we're going to be scientific, let's be scientific and let's have a high standard of evidence for the decisions that we make. The availability of high fructose corn syrup in the United States merely represents economic interest here and the availability of cane sugar in Mexico represents the economic issues there. It has nothing to do with health in either country.
Speaker 3:Cane sugar, which is a sucrose, is a disaccharide that has one molecule of glucose and one molecule of fructose, so it's half glucose and half fructose. Now high fructose corn syrup, that evil industry-laden curse upon scourge upon mankind. Well, it comes in two varieties. One is HFCS42, and that's 42% fructose and 53% glucose, and the other is HFCS55, and that's 55% fructose and 42% glucose. The numbers aren't up to 100 because the missing percentages are just a few intermediate chains of glucose and other oligosaccharide sugar chains that are intermediate between the two. But the point is high fructose corn syrup and cane sugar are essentially the same thing. They're half glucose and half fructose. They're the same thing.
Speaker 3:Which one you buy has more to do with where you live and which group of farmers are subsidized to grow what. Cane sugar has a history of being grown closer to the equator. It also has a very long and troubling history of slavery and abuse. High fructose corn syrup, on the other hand, has a history of being manufactured from corn, which we grow more robustly in northern altitudes than we do sugar cane, and that's it. We process them to end up with roughly the same flavoring and sweetening capacities. There's nothing else to it, and anyone who believes that high fructose corn syrup is less healthy for you than cane sugar believes in an elaborate conspiracy theory and lies that they've seen online in bogus documentaries. And some of the same folks that talk about this are the ones that believe that vaccines are evil as well. So let's be careful not to support any of the administration's missteps while opposing others. A lot of these really fall into the same category, by which I mean not being able to interpret scientific evidence correctly.
Speaker 2:It can be a difficult thing when you see a recommendation that you think on the surface looks good.
Speaker 2:For example, our FDA has recently endorsed removing the black box warning on hormone treatments for menopause and a lot of our OBGYN colleagues have really celebrated that action.
Speaker 2:But we've talked on this podcast for years now about the prior misunderstandings and continued misunderstandings about hormone safety and continued misunderstandings about hormone safety, and there have been black box warnings issued for SSRIs in the past that were then removed. So this is a similar thing right now. I think that the current time will reinforce to many physicians that the FDA is one of many sources of information about medications and they serve an important regulatory step. But it's never been true that we should use a medicine just because the FDA approved it, or that we should not use a medicine because or if the FDA hasn't approved it. Evidence, of course, evolves over time and OBGYNs in particular use a lot of medications off label, meaning the FDA hasn't approved it, but the studies are still very convincing and robust, so new data and new understanding is coming out continuously. So it is nice that the FDA removed this black box wording for menopausal hormone treatment, but I don't know that they did it for the right reasons.
Speaker 3:Yeah, no, I think they did it for bad reasons. Honestly, a lot of the MAHA movement and the functional medicine movement centers around just a huge industry of selling supplements and hormones. This industry is bigger than big pharma, by the way, in terms of financial dollars and completely unregulated. They sell the fountain of youth in hormone replacement therapy for men and women and it's a huge, booming business. This warning is being removed because it's a very pro-hormone industry that supports the players that are there in DC right now. They overuse these hormones, testosterone in particular, both in men and women, and they sell these medicines in routes of administration that we know are unsafe, like testosterone pellets, which ACOG recommends against using, even though many of our colleagues use them.
Speaker 3:So, yes, sometimes Bobby Kennedy's FDA may get something right almost by accident, but directionally they're still wrong. Their reasons are wrong and in past administrations of the FDA they've gotten some things wrong, though directionally they were right. Their intention was good, but we know the science wasn't there or they overreacted. Let's say I think the difference is important. Directionally, the current Kennedy administration is anti-scientific and anti-health. At least in past administrations, for all the problems and corruptions with industries and other things, there was at least a scientific approach and at least a desire to try to generate scientific support for the policies. And so the direction is wrong that we're moving in and that looks like cutting funding for important scientific research and moving that funding to well new studies on the relationship between vaccines and autism, for example, when we already know definitively that there isn't a link between the two. But it does seem like Mr Kennedy is going to announce this month that vaccines are the cause of autism.
Speaker 2:And in the meantime, just today, the latest one I saw was that they cut out funding for research on preventing drowning or reducing drowning deaths in kids. They just cut that. So every day almost, it seems like there's some new, whatever the opposite of good news from there. Well, do you think we've just lost any listeners, Any major RFK groupies? I hope not.
Speaker 3:And I doubt it. I think our listeners are on point here and we're going to have an infectious disease doctor on later this season and we're going to go through some of the well, we'll go through some normal infectious disease stuff too, but we're going to be talking about vaccines.
Speaker 2:So stay tuned listeners that are still here. All right, I guess we'll keep checking our inboxes over the next couple of weeks and see what comes in. Well, let's move on again. Let's talk about breast density on mammography. The number of mammograms being reported with the result of dense breast tissue has increased in the last year, and that's not because women are growing denser breasts, it's because the FDA in September 2024 required all women undergoing mammography to be told if they have dense breasts or not, whether or not any suspicious masses were found, and about 50% of women do have dense breast by the criteria that's been set out. And this already was happening in some states that had passed laws requiring that kind of reporting. So a lot of us in states like that have been dealing with dense mammogram reports for years.
Speaker 3:But it will be new, probably at least to some of our listeners, but it will be new probably, at least to some of our listeners yeah, the changes to the reporting requirements were also never based on scientific consensus or literature, but rather they've come about from advocacy groups, In some cases even just single politicians who had personal stories of missed breast cancers in their family. But large advocacy groups have focused on arguing to federal and state administrators for years now that this extra testing for dense breasts was a necessary step to deal with cancers which were potentially obscured by breast density. But there's no evidence that identifying women with breast density and then doing some additional testing is beneficial to them.
Speaker 2:Yeah, so there's an article in the May 13th issue 2025 of JAMA that goes over this and reviews related literature, so I thought we could talk about their conclusions and their summary. They acknowledge that, yes, half of women have dense breasts, but there's significant inter-observer variability among radiologists about which women have dense breasts and which mammograms fit this criteria. They cite a study in this JAMA paper where 17% of women were labeled as having breast density that did not have breast density according to a different radiologist who looked at the same film.
Speaker 3:Right, and there's some different definitions about breast density too. I also thought it was interesting that the AI reads tend to do this better, but they're not reimbursed by insurance, so they're still relying upon the human radiologist to read this. But it's clear that the density of breast tissue decreases the sensitivity of mammography slightly for detection of cancer. So 3D mammography, which I think is what most people are getting now, has a sensitivity of 77% and a specificity of 88% for detection of breast cancer in women with dense breasts. So, by comparison, for women who are not labeled as having dense breasts, this goes up to 88% sensitivity and 92% specificity. So 77 compared to 88% sensitivity and 88 compared to 92% specificity.
Speaker 2:Okay. So yeah, definitely, on the surface it looks like the numbers are a little better. So two questions come to mind. The first is whether having dense breasts by itself increases the chance of having breast cancer. And secondly, do these women with dense breasts then need something more than mammography, since, as you said, the numbers, the mammography doesn't perform quite as well in them?
Speaker 3:Yeah, well, I think the question of whether or not breast density alone increases your risk of breast cancer is a difficult one. One of the problems with all of this data is that there are so many false positives and a significant number of women who are ultimately labeled as having breast cancer. Well, they don't. They're false positives. Even they don't. They're false positives even though they might receive treatment. So anytime you do more screening, you will have more detection of breast cancer.
Speaker 3:So some of the models do indicate that women with dense breasts are at a slightly higher risk for breast cancer than women who don't have dense breasts, but it doesn't put them into what we would call a high risk category just because they have dense breasts.
Speaker 3:So typically high risk women are those who are considered to have a greater than 20% lifetime risk for developing breast cancer based upon their risk factors. Breast density doesn't contribute that much of an increased risk, and it's debatable whether it increases the risk at all, because it might just increase the risk of extra screening and more false positive diagnoses of breast cancer, and so it's always difficult to tease that out. Ideally, you'd want to know whether or not an absolute outcome like breast cancer mortality is higher in women who have dense breasts compared to those with not dense breasts after you've accounted for all other confounders. But that's very difficult to understand and we just don't have robust data for that. So our best answer is dense breasts might increase your risk a little bit, but not significantly and not enough that you would be considered high risk.
Speaker 2:Okay, but if we let's pretend that it does, let's assume that it does slightly increase the risk of breast cancer and that mammography performs less well in that population of women with dense breasts. So then does it make sense to add on additional testing Because that's what people are advocating for For example, reflexing to an ultrasound if the mammogram says dense breasts?
Speaker 3:Well, the bottom line answer, and the answer that the authors of this paper give, is no, it doesn't make sense. All women, including those who have dense breasts, should have their lifetime risk of breast cancer calculated. That's how we know if they're high risk or not that 20% threshold. Now there are several tools available for this, and there's problems with many of these tools. They do tend to overestimate risk. So we can talk about which tool might be best at another time. There's a lot there. We likely have new tools or things like AI models in the future that will come out that will do this better, and so this is an evolving idea. But for now, you could use the breast cancer risk assessment tool or the breast cancer surveillance consortium risk calculator or the Tyra Kuzik model and then determine what the patient's risk of breast cancer is, and two of those models account for breast density. Now, using the breast cancer risk assessment tool, I calculated that your lifetime risk of developing breast cancer is 13.6%, which compares to an average risk of 12.4%, but obviously is below 20%, so you're not high risk. Then, using a Tire-Kuzik model, I calculated your risk at 10.5%, compared to a population lifetime risk of 10.8%, and in that questionnaire I presumed just to make you higher risk of breast densities that you had dense breasts, because that model incorporates that question. And then I put your information with the same assumptions into the consortium calculator and that one doesn't give a lifetime risk but it gives five and 10 year risk. And so your 10 year risk is 1.3%, which compares to an average risk of 1.59. So again lower than average risk. So you're-year risk is 1.3%, which compares to an average risk of 1.59. So again lower than average risk. So you're not high risk. And in that one I also indicated that you had dense breasts. So it does lower your risk. This is a way of thinking about this. It lowers your risks if I put into the models that you have the lowest category of density, or basically fatty breasts, the lowest category of density or basically fatty breasts. So if that were the case, your 10-year risk becomes only 0.31% in the consortium calculator. And that similar risk reduction occurs with the Tyra Kuzik model.
Speaker 3:If I do the same manipulation so you can play around with these numbers yourself.
Speaker 3:Listeners can pull up the calculators and do this and put in your information and just see what would happen if you change the breast density alone.
Speaker 3:But since half of women basically have dense breasts, then the average woman has dense breasts and therefore the average breast cancer rates already largely account for women having dense breasts. So you're essentially lowering your risk of breast cancer if you don't have dense breasts and I think that's the reframe here and another way of thinking about it the average number incorporates all the women with dense breasts and so if you don't have dense breasts, you're at a lower than average risk, at least according to these models, and that may be a good reframe. And instead of thinking that dense breasts increase your risk of breast cancer, maybe we should think of not having dense breasts lowers your risk of breast cancer. In fact, dense breasts are the norm and if you don't have dense breasts, congratulations. And I think that's the important part about understanding how this affects your risk in these calculators. You need dense breasts to basically be at average risk or maybe slightly above average risk in these calculators, but it doesn't get you anywhere near that 20% lifetime risk of breast cancer, which we would consider high risk.
Speaker 2:Okay. Well, now I understand why you're asking me all those questions about my family history.
Speaker 3:They ask some detailed questions in that tire cruising model.
Speaker 2:So yeah, but it's true, none at all in my family.
Speaker 2:So that seems to be a lot more important than breast density. So, yeah, that is an interesting reframe, and perhaps we should tell patients who don't have dense breasts that you're at a lower risk of breast cancer and mammography is going to work better for you than for the average patient, and maybe you can get by with less frequent screening. It's interesting, really, how a reframe can change your perception, and we might also tell women with larger breasts that they are at a little bit increased risk of breast cancer. Or, on the flip side, we could tell women with smaller breasts that they're at decreased risk of breast cancer, or, on the flip side, we could tell women with smaller breasts that they're at decreased risk of breast cancer. Depending on which approach you take, you could find yourself advocating for more or less screening. So the real question is what does the science say on it? And that's the problem we have when health policy is dictated by non-scientific advocacy groups that we've been seeing more of now than ever, as we were just discussing.
Speaker 3:Yeah, I think this is why you can't just do things, because it makes sense to you. And there are obviously stories where someone went in you and I have obviously seen these in practice. They had a normal mammogram last year and now they've come in a few months later not even time, not even a year later and they felt a tumor and they've got breast cancer. And then it's an easy thing to say, well, why didn't they see that on my mammogram? Well, it's not super sensitive, it doesn't pick up every cancer. And then the radiologist might say, well, you've got denser breasts, so it was probably obscured by the densities. And then you think, well, why didn't we do something about that? But even when you do the extra testing, it's not perfect. So we don't need to rationalize why we didn't find it with dense breasts. It's just a limitation of the screening and our screening processes take this into account.
Speaker 3:So the authors of this article they review the evidence for the question that we've posed. You could do ultrasound. They review the evidence for the question that we've posed. You could do ultrasound, which is what's commonly being recommended, or you could do MRI. So, as of now, there are no randomized controlled trials which have shown that women with dense breasts benefit from either ultrasound or MRI in reducing breast cancer deaths and overall, the performance of these additional tests likely leads to more harm than good. This is what everyone always forgets about tests, and about mammography in particular, is that there are negatives associated with doing screening tests, and the more tests you do, the more of those negative outcomes you're going to see. And I'm not just talking about cost and anxiety and discomfort though those things shouldn't be discounted.
Speaker 3:I'm talking about biopsies and lumpectomies and over-diagnosis of DCIS that was not going to be progressive. And misdiagnosis of breast cancers that are treated and receive full treatment even though they weren't cancers. And extra radiation exposure, with all the potentially increased imaging and mammography in particular. Mri doesn't have radiation, but the contrast itself is radioactive and overall has about the same net negative effects in terms of radiation exposure as mammograms do. One model that the authors refer to suggested that biennial implementation of ultrasonography for women with dense breasts between the ages of 50 to 74 who had a prior negative mammogram would result in an additional 354 biopsy recommendations per 1,000 women, with potential benefit of perhaps preventing one cancer death per 3,000. This is a model, not a trial.
Speaker 3:Mri, though, is worse.
Speaker 3:If you add an MRI to women with dense breasts who underwent biennial screening from age 40 to 74, then over that 35-year time period, the model suggests that you might save one life per 1,000 women.
Speaker 3:But the number of false positives will go up from 1,392 per 1,000 to 1,850, almost a 50% increase and the number of benign biopsies will triple nearly from 221 per 1,000 to 628 per 1,000. So there's a reason why we don't recommend MRI for breast cancer screening, and it's because of the high rate of false positives and no difference in radiation exposure, and so it's really just justified. These extra tests are justified when your lifetime risk of breast cancer is above 20%. So the conclusion of these data is that there's more harm than good for a woman with dense breast to undergo supplemental testing, either with ultrasound or MRI, and that we should calculate their lifetime risk and if it's above 20%, then they should be getting an MRI screening, and that's what we've recommended anyway. She's supposed to come back with a report of the dense breast and then you're supposed to calculate her risk and talk to her about it, but of course, what actually happens is people just order the extra testing.
Speaker 2:The other interesting thing in all that modeling is that it assumes a person is getting biannual screening, so once every two years, which is what the USPSTF recommends. Those numbers are going to get even worse, like the false positives, when you are looking at women getting yearly screening and then potentially yearly follow up with the ultrasound or MRI. Since so many patients are getting yearly mammograms anyway, then the yearly mammogram itself is going to be more effective at finding occult breast cancers than every other year, followed by those follow-up ultrasounds.
Speaker 3:statistically, yeah, that's a good point, and the points raised in the article really don't even apply to the vast majority of the women in the United States receiving mammograms who have dense breasts, because the vast majority of the 40 million women in the US receiving mammograms are getting yearly mammograms, not biennial screening as most of our professional societies and the science tends to recommend.
Speaker 3:So in many cases the harms are going to be even more magnified when you add that extra testing on than we've highlighted. But I do think it's helpful to reframe this for our patients. When they come in with this concern because they're being scared by it, honestly, they get a letter and certified letter that their breasts are dense and they're required by statute or by FDA to tell them this information, as if there's a strong scientific basis for this extra testing and not to disparage our radiology colleagues, but they're not fighting against this too hard. It's more revenue, it's more business for them to have half of women get ultrasounds after their mammography. So they're not really contextualizing this that well. But maybe we can reframe this for folks and let them know that not having dense breasts lowers their risk of breast cancer and that's all that is. And having dense breasts puts them at the average risk.
Speaker 2:Yeah, another reason that health policy needs to be dictated by science and not special interest groups. Okay, let's try to get to the history segment this time. So you told Jacqueline on your last episode about her son Noah, that we would discuss the history of antenatal diagnosis of fetal abnormalities, and you just didn't get to it last time because you ran out of time.
Speaker 3:We were ready, but she's a talker. She just took the whole time.
Speaker 2:It was good stuff, so I guess I'm glad you didn't get to it, because we got to hear more about more of what she was saying.
Speaker 3:Yeah, all right. Well, let's do it then.
Speaker 2:Okay, so you and Scott Guthrie, the neonatologist, had talked about previously the history of listening to heart tones and seeing meconium fluid as the very beginnings of understanding fetal well-being, and there's a lot that could be added to that story as fetal monitoring was developed with other new technologies. But let's talk about specifically diagnosing structural or genetic or chromosomal abnormalities. So this starts with ultrasound.
Speaker 3:Yeah, we trace obstetric ultrasound back. This may surprise folks to know how old this is, but there's a landmark paper in 1958 that was published in the Lancet by Professor Ian Donald and colleagues in Glasgow, scotland. Published in the Lancet by Professor Ian Donald and colleagues in Glasgow, scotland, and building on earlier physics such as the piezoelectric effect that was discovered way back in 1880, along with sonar technology that was developed to detect submarines during World War I, their team demonstrated the use of pulsed ultrasound to investigate abdominal masses, and this article, which is really quite interesting, actually included the first image of a fetal head that was published at least, and we can put that picture on the Instagram if people are interested to see it what a fetal head on ultrasound looked like in 1958.
Speaker 2:Yeah, I'm looking at that picture and it's three fuzzy lines basically. Maybe if you tell me it's a head, I'll say sure, but I wouldn't be able to tell that just by looking at it. It's not the greatest ultrasound image I've ever seen.
Speaker 3:Yeah, there's a little printed out ultrasound picture in my baby book from when I was in utero and it looks. I can't tell you what it is at all. I'm sure it's also faded a little bit over time, but who knows what they were looking at with it. But yeah, so it took a long time for image quality to be good enough to do anything with really. But these were very rudimentary assessments back then and a lot of big, complex machinery too, nothing portable. So obviously you could use even this rudimentary ultrasound though to determine things like whether the fetus was head down Although people back then routinely used x-ray for that if that was in question.
Speaker 3:And it wasn't until 1968 that there was real-time imaging. So these were static images prior to 68. And then that would allow, for example, the detection of cardiac activity. You could actually see the heart beating and moving. And then in the few years after that, they introduced grayscale imaging. This picture we're looking at is really just black and white. Honestly, there's very little tonality to it. But when you had more grayscale, that improved the amount of visible detail and the clarity of the images, and then of course it had to become way less expensive and way more portable. So in the early days it was really just used to detect position in places that had it and maybe the presence of twins or triplets and also placenta location became something they could do in the 60s pretty easily. By 1972, it was useful enough that you could use ultrasound to guide the amniocentesis that we were already doing, but that made amniocentesis a lot safer. You could find the pocket of fluid, you could see where the placenta was, and so that became important too.
Speaker 2:Okay, how about fetal measurements like estimating the weight?
Speaker 3:Yeah Well, it wasn't widespread, but fetal biometry became more and more available as people learned how to do it. The biparietal diameter itself was first described in 1968, but it took time to build up the databases of measurements and the nomograms today that we use to accurately determine gestational age and fetal sizes.
Speaker 2:Okay, and you just mentioned the amniocentesis, right?
Speaker 3:Well, amniocentesis, right. Well, amniocentesis was done way, way long before any of this, originally to relieve excess amniotic fluid. So imagine severe poly and a woman who can't breathe and is uncomfortable. We were doing that a long time ago and at least by 1877, there's reports of what we might today call a therapeutic amniocentesis. In 1930, a group of doctors injected contrast dye and then they took x-rays. So they put the dye in utero in the amnion and then they would take an x-ray to use the contrast to outline the amniotic cavity on x-ray cavity on x-ray. And also by the 1950s it was used to determine the delta OD 450 levels for management of RH disease, where they would draw some of the fluid and look for the refraction of light at 450 nanometers. And this corresponded with the breakdown of hemoglobin.
Speaker 3:But then in 1956, a paper was published by Fritz Fuchs and Pavel Ries in Nature showing that you could determine the fetal sex by looking for the presence of bar bodies or the inactive X chromosome of fetal cells. And this immediately had some clinical applications for sex-linked disorders like hemophilia. By 1960, it was being used in the management of prenatal diagnosis of hemophilia and by 1964 for Duchesne's muscular dystrophy. Then in 1966, steele and Breg were able to culture fetal cells and do a full chromosomal analysis, or what today we'd call a karyotype. So now you could diagnose at least major chromosomal abnormalities. Another important paper in 1970 by Nadler and Derby appeared in the New England Journal of Medicine that described the role of amniocentesis for the diagnosis of genetic defects, and this really led to the creation of genetic laboratories around the country and wider clinical adoption. So this happened at around the same time as Roe v Wade, so there was an interest then in possible terminations for chromosomal abnormalities. And then how do you determine those abnormalities during the pregnancy?
Speaker 2:Yeah, and we talked before on a previous episode about the whole term advanced maternal age and how that's become a diagnosis, and that was created at that time to indicate patients who would be a candidate for amniocentesis, because the risk of Down syndrome was 1 in 270 at a maternal age of 35. And the fetal loss rate associated with amniocentesis, also at the time was thought to be 1 in 270. So the decision was made that when those two numbers crossed and then beyond age 35, the risk of Down syndrome increases, more so than the risk of fetal loss. So then at that crossing point you should be offered an amniocentesis.
Speaker 3:Yeah, the advance to the geriatric. What is a geriatric primograph that all these women hate when they see that on their billing levels. It's an archaic diagnosis, but in 1972, it had been discovered by Brock and Sutcliffe that elevated levels of alpha-fetoprotein in amniotic fluid indicated a possible neural tube defect. So one of the side effects of taking amniotic fluid out was what else can we find in there that's useful for birth defects? So they discovered alpha-fetoprotein.
Speaker 2:Yeah, and it was also discovered that some of that elevated alpha-fetoprotein leaked into the maternal bloodstream, and so that elevated maternal serum AFP levels could also indicate possible neural tube defect. And so this was early in the 1970s, before ultrasound was really anywhere near capable of seeing those neural tube defects, so this became yet another thing to screen for.
Speaker 3:Right and over time as we checked more of these maternal serum alpha-fetoprotein levels we noticed that low alpha-fetop fetal protein was associated with an increased risk of Down syndrome and trisomy 18. So this led to AFP screening becoming fairly standard by about 1983 when it was FDA approved and really widely available after 1985 when ACOG recommended that we make it available to every pregnant woman for screening. So it became a standard of care after that. But of course the sensitivity and specificity of AFP is pretty poor, so folks look for more biomarkers.
Speaker 3:In the 1980s AFP was combined with HCG levels and unconjugated estriol levels and that gave us the triple screen for Down syndrome. But that test only discovered about 70% of cases and of course one in 20 women who took the test would be told that their test was abnormal. So a 5% false positive rate. But this was better than AFP alone. And eventually inhibin was added to that screen to make the quad screen or the quadruple test. This got the sensitivity up to 81%. Quad screen or the quadruple test this got the sensitivity up to 81% and ultrasound quality improved greatly over time and anatomic surveys came to be focused largely just on what we call soft markers for aneuploidy.
Speaker 3:But these screening tests and just the age-based screens themselves. They still led to the need for definitive diagnostic tests, which would be the amniocentesis and then eventually chorionic villus sampling, which was first performed in 1983, but became widely available in the early 1990s. Also in the 1990s we discovered that a thickened nuchal translucency between 11 and 14 weeks of pregnancy was strongly associated with Down syndrome and other chromosomal aneuploidies, and this was eventually combined with pregnancy associated plasma protein A, which is typically decreased in Down syndrome pregnancies, and HCG, which is typically increased, and we got the combined first trimester screen. That was done either as an integrated or sequential screen and with the integrated screen we found sensitivities as high as 94% for detection of Down syndrome.
Speaker 2:All right, that's pretty good. And then after that we got to cell-free DNA.
Speaker 3:Yeah, in 1997, we discovered that small fragments of DNA from the placenta circulate in a pregnant person's bloodstream. We couldn't do anything with them. We knew they were there. But we needed what we now call massively parallel shotgun sequencing to be able to clinically analyze these fragments. And that was first demonstrated in 2008. And then the first commercially available non-invasive prenatal screens as we know them today came out in 2011.
Speaker 3:So millions of these fragments, both maternal and placental, are sequenced and they're mapped to their chromosomes of origin. And placental are sequenced and they're mapped to their chromosomes of origin, and then we can determine the relative proportion of those, so that if you have, say, a three to two ratio of sequences associated with chromosome 21, for example, then you can assume that there's a trisomy 21 pregnancy or do that for 18 or 13 or anything else. And all of a sudden that became dramatically more sensitive, with far fewer false positives than anything we'd done up to that point. But eventually we were also able to do chromosomal microarrays and now whole exome sequencing. So with chromosomal microarray testing we can detect small submicroscopic, missing or duplicated segments of DNA, and this became a standard of care by 2016. And now with whole exome sequencing we essentially can test for any genetic abnormality or genetic characteristic if we want to, and that with testing for RH, the presence of the blood type, for example, for RH screening.
Speaker 2:All right, that was a nice little historical segment. I think we're probably out of time for today.
Speaker 3:There's lots of things we could talk about in the future of what we do with all this massive amounts of DNA evidence. It's a whole new ethical landscape that people are going to live in.
Speaker 3:And what do we do even with sequencing our adult patients? I think in 10 years we'll have pretty standard offerings that your adult doctor will have your whole exome sequenced and be able to probe that for different genetic information. But then of course, do you want to know if you're at increased risk for Alzheimer's, if there's nothing we can do about it, or do you want to know if you're going to have Huntington's disease? There's a lot of ethical things that we're going to deal with in the next 10 years related to this massive amount of genetic information. Well, we'll have to do some DNA ethics episode in the future, but otherwise we'll be back in a couple of weeks.
Speaker 1:Sounds good Thanks for listening. Be sure to check out thinkingaboutobgyncom for more information and be sure to follow us on Instagram. We'll be back in two weeks.