Episode 59 - Stefan Enroth
Self-sampling for Better Screening and Diagnosis of Gynecological Cancers
Theme: Women and Medicine
Published: 29 May 2024
Summary
The research of Stefan Enroth, Associate Professor of Computational Genomics at Uppsala University, focuses on finding biomarkers for better screening and diagnosis of gynecological cancers, particularly ovarian cancer. In this podcast episode, Stefan Enroth emphasizes the importance of early detection for improving survival rates and highlights the challenges of finding reliable biomarkers for a disease that's often discovered late. He discusses his work on self-sampling methods, which allow women to collect their own samples at home, and the potential benefits of this approach for cost-effectiveness and accessibility. Enroth also touches upon the ethical considerations and potential limitations of broader self-sampling programs, highlighting the need for further research and careful implementation in the healthcare system.
Keywords
Gynecological cancers, biomarkers, self-screening, diagnosis, healthcare system
Suggested Link/s
Personal website: https://www.uu.se/en/contact-and-organisation/staff?query=N4-698 External link, opens in new window.
Publication: https://uu.diva-portal.org/smash/get/diva2:1657110/FULLTEXT01.pdf External link, opens in new window.
Transcript of the Episode
Stefan Enroth 0:09
We try to find biomarkers for ovarian cancer. Ovarian cancer is a terrible disease, usually discovered late, meaning that it has really poor survival numbers, so five years survival is very low, but if we were able to find it early, the survival rates are fairly high. It would be really good if we could find the cancer early. And doing so, I think we need to have some sort of self sampling methodology for providing cost efficiency to the whole project.
Natalie von der Lehr 0:42
Welcome, to SCAS Talks, a podcast by the Swedish Collegium for Advanced Study. My name is Natalie von der Lehr, and in this episode, I talk to Stefan Enroth, Associate Professor of Computational Genomics at Uppsala University and fellow within the Natural Sciences Program at SCAS during the spring of 2022. The research of Stefan Enroth aims at developing models for biomarker discovery in non-communicable diseases, especially gynecological cancers. He investigates the possibilities and limitations of self collected sampling methods in particular. And this is the second episode in our theme "Women in Medicine". With us in the studio is also Keiko Snarberg, who is currently studying journalism at Södertörn University.
Very welcome to SCAS Talks. Would you like to say a few more words about yourself, Stefan.
Stefan Enroth 1:40
Thank you, Natalie and Keiko. So my name is Stefan Enroth. I work here at Uppsala University, and have been doing so for - well, I did my undergraduate studies here as well. So I've been here for more than 25 years now. So quite a while, but in the current department for 10-12, years now.
Natalie von der Lehr 1:57
How come you got into this research topic?
Stefan Enroth 2:00
Well, I think for myself and for many researchers, the path is not completely straight to begin with. So I started my studies here in Uppsala in Information Technology Engineering, and finished my degree here, and then worked in the industry for a little while employing machine learning methods to in wearable computers and a company based in Stockholm, but then retrained as a bioinformatician. And took my PhD in bioinformatics in 2010 and then moved to employing those tools and knowledges to more medical problems in the Department of Immunology, Genetics and Pathology, especially medical genetics, as a research track. And in the group I joined - we had a long experience of investigating efficient methods for cervical cancer screening, based on home sampling and HPV testing, and then we expanded from that field into other types of cancers as well, but always with the idea of using self sampling and home sampling as an efficient way and cost efficient way of collecting samples. We try to find biomarkers for ovarian cancer. Ovarian cancer is a terrible disease, usually discovered late, meaning that it has really poor survival numbers. So five years survival is very low. But if we were able to find it early, the survival rates are fairly high. It would be really good if we could find the cancer early, and doing so, I think we need to have some sort of self sampling methodology for providing cost efficiency to the whole project.
Natalie von der Lehr 3:31
Maybe we need a little bit of definitions for the continuation of this discussion. First of all, what is meant by gynecological cancers?
Stefan Enroth 3:40
That's a broad term that's sort of collecting five cancers, and those are cervical cancer. So we talked a little bit about ovarian cancer. We have endometrial cancer, vaginal cancer and vulva cancer. So those are cancers in the reproductive organs of women. I'm really not that big of a fan of the term gynecological cancer, because these five cancers are really different. They have different etiology, they have different symptoms, they have different treatments and different ways of being discovered. But we use them still as a collective term, gynological cancer.
Natalie von der Lehr 4:12
What kind of knowledge gaps are there in this group of cancers?
Stefan Enroth 4:16
Well, there are lots of unknowns in terms of why these cancers occur. Some of them we know, like cervical cancer, is caused by an infection of HPV, so that's easy to test for, but it's really not that known. For instance, for ovarian cancer, the sort of underlying reasons to how and why the cancer is developed, which also makes it a little bit harder to find good biomarkers, because can't really make any pre-assumptions on what we are looking for. So there is a lot of unknown factors there in the background of the cancers.
Natalie von der Lehr 4:45
Since you mentioned biomarker again, what is a biomarker?
Stefan Enroth 4:49
In very broad terms, a biomarker is something that we can measure, that it's indicative of something. So that's something in the first part could be, for instance a genetic mutation that we can measure in our bodies, or it could be a specific molecule, such as a protein, that we can measure in blood. And the process that we are observing could be a cancer, or it could be some other type of disease that sort of changes either the expression levels of these proteins in relation to the disease, it could go either go up or it could go down. It's a difference compared to some other states.
Natalie von der Lehr 5:24
You already mentioned the importance of early diagnosis, but just to get the overall picture in the beginning, why is this so important, and what can you do to improve it?
Stefan Enroth 5:37
So for ovarian cancer, for instance, we can broadly classify the cancer into four stages, from stage one to stage four, with a higher stage has a poorer prognosis in terms of survival and treatability of the disease. And already in stage three, the cancer is usually spread. So it's not defined to a single location, but it could have spread throughout the body and the abdomen. Well, if the cancer is discovered in stage one, and there is a survivability rate for almost 90% for five years, while for it's discovered later in stages three or four, it's really low, maybe only 20% and there's a big increase in survival and treatability if the cancer is discovered in the early stages.
Natalie von der Lehr 6:18
But let's get into some details of your research. Can you tell us more about what you're doing? You mentioned this screening program. What are you doing there?
Stefan Enroth 6:29
So we have three main focus areas for our biomarker studies in ovarian cancer. There are three different applications areas, we can say. And of course, a screening program would be the Holy Grail. That would be fantastic, if we had a population screening program that would pick up cancers early. That would save a lot of lives each year. That's really difficult. So apart from the screening program, we're also interested in finding biomarkers that could separate benign from malign tumors in symptomatic women. In Sweden today, about 80% of the women that do seek health care for symptoms that could be ovarian cancer are having benign conditions, but they are still operated upon. So that could be a way of reducing surgery to separate the benign from the malign conditions. And the third area is just try and find biomarkers for detecting relapse. That's also a major problem area in ovarian cancer, because after treatment, around 50% of women will have relapse within two years. So three different application areas of the biomarker, and it doesn't have to be the same biomarkers in all of them, they could be different. They can be representative of different processes, and it could also be different sample types that are most appropriate for these different applications areas.
Natalie von der Lehr 7:47
And how do you go about them to find these biomarkers and all that?
Stefan Enroth 7:52
Maybe a fortunate thing is that ovarian cancer is fairly rare. So in Sweden, it's about one in 10,000 women that will develop the cancer. But that's also a challenge for us, because it's difficult to find samples, especially before the diagnosis, which would be the samples that you need to find the biomarker for early discovery. So what we have been focusing on is to find biomarkers for separating benign versus malign conditions, and that we can do with great biobanks that we have here in Upsala, for instance, that collect samples at the hospital regularly that you can use as a researcher for your research projects. And so we get these samples, and we try to get as many as we can, and then we try to measure as many biomarkers as we can, and then we use computational methods to find good predictive biomarkers, either single value biomarkers, sort of a single protein that you can measure, or combinations. And this is where the computational methods and the machine learning comes in, because it's not possible, really, to do it by yourself. So you need a computational algorithm to find these things for you.
Natalie von der Lehr 9:00
Yes, because nowadays, these kind of methods that you use, they generate a huge amount of data.
Stefan Enroth 9:06
Yes, they do. The latest version of the protein assays that we've used measures over 5000 proteins in each sample. And if you have hundreds of samples, which is always necessary to have rather large cohorts, to find things that are common between cases and not something that's specific to a smaller set of women, or a single woman, or something like that, then you need, you need to have many samples to look at, and then the data quickly becomes really big.
Natalie von der Lehr 9:33
If I've understood it correctly. You have done a study recently using self sampling. Can you tell us a little bit more about that?
Stefan Enroth 9:40
So I have to trace that back a little bit first. So the projects were done in cervical cancer screening, because that was done with self sampling of cervical vaginal fluid that the women could collect themselves at home using a small brush, which is inserted into the vagina, and then you take out some of the fluid, and you put it on a paper card, and it dries. So it's a dry format, and you can send it through ordinary mail back to the lab, and then we could do testing on that for screening of HPV, human papilloma virus, which is the cause for cervical cancer. And this is a really efficient way of collecting samples. And we figured that it would be fantastic if we can use the same sample type to also scan for additional cancers, but maybe based on other molecules than just DNA. So we started looking at using that sample, and in collaboration with clinicians in Gothenburg and in Uppsala, we've collected cervical vaginal fluid from women at time of diagnosis. So we have both benign tumors and malign conditions in all stages. And then we started working from those samples.
Natalie von der Lehr 10:42
Keiko, you look like you want to ask a question.
Keiko Snarberg 10:45
Actually, I do. Don't most women think it's unpleasant to take the samples themselves?
Stefan Enroth 10:49
We've actually conducted those studies when we did the cervical cancer screening projects and asked women if they preferred to do it that way, or the sort of standard way of meaning that you have to make an appointment and go to a gynecologist and have a cell sample. And the actually, the self sampling method was preferred by a majority of the women that were participating in the study. One of the aspects is that you can do it when it's a good time for you. You don't have to book a time and take time off work and go somewhere, and you can do it yourself at home.
Keiko Snarberg 11:26
But will the testing be trustworthy? I mean, you're not a professional, you don't know exactly how it should be done, maybe?
Stefan Enroth 11:35
The absolute majority of samples were perfectly fine. We also set up studies where we had women sample themselves, and also a sample taken by a midwife, the same type of sample, but in a hospital setting, and then compared. And the results were the same. So it's a fairly robust way of sampling. It's not so difficult. It's very rare that you don't get any usable samples in this way, and you don't need much material for proteomics research. It's enough to take a punch, a small circle, from the card that's about three millimeters in diameter, and that's enough to measure 1000s of proteins or or do DNA analysis,
Natalie von der Lehr 12:17
And then you can use the same card again for another test.
Stefan Enroth 12:20
Yes, that's also a major advantage, and they're really easy to store. You can basically store them in a box on a shelf in room temperature, so they're stable.
Natalie von der Lehr 12:30
So you've tried this self sampling on other types of cancers, then what are your results so far on that?
Stefan Enroth 12:39
Well, the results from cervical cancers were really, really good compared to the previous cell sample investigation or screening program that we had. We found twice as many pre stages of cervical cancer, but actually the half the cost and quicker. But for ovarian cancer, there's no implemented screening program today. So these are research studies, but what we have found is protein biomarkers that could be indicative. They're not precisely accurate enough, maybe to justify screening, because there's an issue with false positives, meaning that although the test can discover the majority of the cases, quite a few women that don't have the cancer have a different protein profile than what's considered to be normal, and these will have to be investigated somehow in the healthcare system. So one of the points needed to be considered when starting a screening program is, what will we do with the results? So how will we sort of process all the results in the normal healthcare system? And if we have too many false positives, all of these women will also have to have investigations, further investigations, examinations, and so on. And that's also a big thing to think about before we can start the program.
Natalie von der Lehr 13:50
Also for the women themselves, it's of course worrying.
Stefan Enroth 13:53
Absolutely, that could be very worrying, knowing that maybe you have a cancer that's also a factor that needs to be considered.
Natalie von der Lehr 14:00
But what would it take then, apart from eliminating these false positives or taking them down, what would it take to implement such a screening system in the actual healthcare system?
Stefan Enroth 14:11
In Sweden, today, we have two running screening programs. It's breast cancer and cervical cancer, and recently, we've also started with colorectal cancer screening, but that's not implemented completely yet, but it's on its way. So those are the only screening programs that we have. And the Swedish welfare board has, I think it's 15 criteria that you need to fulfill before a screening program can be implemented. And one of them is, of course, that the test has to be good enough. But there are also other issues. There are ethical issues and there are legal issues. There has to be a time frame where the test can make a difference in relation to the disease, because there's no point in finding out that you have the disease if you can't act on the information. So there are a lot of different parameters that need to be fulfilled. But the one that is closest to our research is, of course, finding a good enough test with really high specificity and sensitivity for finding all the cancers and only the cancers.
Natalie von der Lehr 15:12
Do you think that's possible?
Stefan Enroth 15:13
So I think maybe a population screening test will be a little bit further in the future, but I think we can definitely already, maybe say with some confidence that we can separate benign from malign conditions in symptomatic women. And I think we also have a good opportunity for finding relapses early. Mainly because that's a lot smaller group, and it's women that's already within the healthcare system. So it's easy to sort of track the women that are at risk, as compared to finding the one in 10,000 cancers in the general population.
Natalie von der Lehr 15:44
In the previous episode within the same theme, I talked to Anat Biegon, also about Women in Medicine, and we talked a lot about how women have been understudied, both the diseases, but also how women have not been included in clinical trials and so on. How was it in this kind of cancers - they are specific to women. What does it look like? How is the research and is there like a misrepresentation somehow?
Stefan Enroth 16:22
Well, there is an obvious skew in what we look at because it's women only. So naturally, it's only female patients. Ovarian cancer is also a disease that's usually occurring in women that are older than 70, so they're postmenopausal in most cases, or elderly. So that could also mean that there are different processes that are ongoing in the body as compared to younger women, for instance, and that could be a big factor, especially if you're looking for protein biomarkers in blood. There are huge differences in circulating proteins and other markers based on the hormone cycle, for instance, with the menstrual cycle. So one big issue is post and premenopausal women, and something to really consider. And it's easy to make the assumption that it might look the same and you just put everyone in the same pool, but it could be that most of the benign cases are maybe younger women, and the malign cases are sort of have a different distribution towards elderly women. And then you might be just looking at being older, biomarkers for being older, rather than biomarkers for disease.
Natalie von der Lehr 16:35
Yeah, the hormonal cycle, of course, must have a huge impact on gynecological cancers. So is that anything that you consider in your research?
Stefan Enroth 17:42
Since most of the women are postmenopausal, so in that case, it doesn't affect as much, at least not what we know of. But there are many, many other conditions that are affected by by the hormone cycle, especially with the estrogen levels and progesterone levels that have been in recent years, connected to many female specific conditions. And I think there is many more that we're not maybe aware of yet that could be connected to these things as well. And there's also an issue of using contraceptives, for instance, based on hormones. We're not really sure what that means. In a long time, sort of lifelong exposure to changes in your hormone sort of natural hormone cycle.
Natalie von der Lehr 18:24
Many things to consider there really.
Stefan Enroth 18:26
Yeah, many things to consider. And usually these are things that are maybe not recorded in the databases as well. It might seem obvious to ask women where in their menstrual cycle there are, for instance, if they know. So that could be recorded alongside when the samples are collected. But in many cases, it is not.
Natalie von der Lehr 18:45
In a way I was thinking, this kind of test that you are developing for screening of cancers, could one see this also as an empowerment to women to have a bit more control over their health and disease development, so to say?
Stefan Enroth 19:01
I have to think about it from my own perspective. If I could do similar testing for some disease that were applicable for me as a man as compared to a woman, I guess that would feel like a good thing, that I could do that myself. But I think it maybe has to do more with with a self sampling strategy, rather than going somewhere to have your body checked by somebody else. And the difference between having a time booked and say, come here at this time slot and we will look inside you and see if we see something, as compared to choosing when you can take the test yourself.
Natalie von der Lehr 19:43
So you're looking into self sampling in a greater sense, what does this mean and imply, and what kind of possibilities and also limitations are there to self sampling?
Stefan Enroth 19:59
I think there's great promise in self sampling for healthcare as a whole, both as a very cost efficient way of finding precursors to disease, and the earlier we can find or place a correct diagnosis, the better prognosis it is for the individual and for cost efficientness in the healthcare system. It's usually easier to cure someone if you find a condition early, but it's also, it might also be a big responsibility. So where should we store all of these samples, and is it ethical to maybe collect samples and do genome wide screening if we sample everyone's DNA, what kinds of things do we find, and should it be done on everyone, or should we limit it to the risk groups? And should we give the information to the insurance companies, for instance? And there are already, today, lots of private healthcare providers that can run many biomarker tests for you and tell you what your blood lipids are and your cholesterol levels, and maybe you can measure your testosterone levels, and you will maybe act upon all these things. And some of them might be good, but some of them might be something that you need to maybe go to the standard healthcare, and that will, of course, be a cost issue. So finding out who's going to be responsible for all the information, I think, would be a major problem with the sort of general just measuring everything that we can in a general population.
Natalie von der Lehr 21:24
I think there's a trend right now also among people who are very aware of their health, and who like maybe training, working out, to take a lot of glucose tests or do checkups of different, as you mentioned, health markers to sort of keep an eye on their health, so to say.
Stefan Enroth 21:45
There are quite a few studies ongoing, as well into wellness. So not sort of monitoring biomarkers as compared to disease, but maybe as a marker for how well you are, sort of trying to stay healthy. But that's that's a kind of a different question. Could we just monitor everything continuously and look for things that are departing, and should we treat that as an indication of of disease and do additional examinations, or is it just natural variation that could occur? If you're you're having a cold, maybe you didn't sleep so well, you ate a lot yesterday, things like this could actually influence your biomarker levels.
Natalie von der Lehr 22:23
If you could wish for something in your research or in your progress. What do you need there and what would be a dream scenario?
Stefan Enroth 22:33
For the research questions that we focus on, the thing that we would really need is samples collected before diagnosis, but that would mean sampling basically everyone once in a while, and just store the samples and then wait basically for the correct diagnosis, and then we could trace back samples that were collected beforehand. That would be the absolute best resource to have. I don't think we have any such specific samples collections in Sweden. There are samples collections that are collecting samples with regular intervals. There is one really good one in Umeå, for instance, in the north of Sweden that collects samples every 10 years, roughly, from everyone that are participating. But 10 years might be a little bit too long of a time span. Those samples - resources are really good, but it's expensive to collect these samples, so that's a major hurdle, really. There is one cohort in the UK that has collected samples longitudinally from millions of women where some of them have developed ovarian cancer, and those samples are collected before diagnosis, and we're actually discussing with them now, if we can access those samples and see if we can use some of our methods also in those samples collection, to see when, when we can sort of start seeing changes in these biomarkers before diagnosis, what kind of time span or time frame that the test would be applicable in relation to diagnosis.
Natalie von der Lehr 24:01
What would it take to start such a screening program in Sweden?
Stefan Enroth 24:06
Yeah, lots of money, I think. But that's also a major benefit from the self sampling strategies, because that's an easy sample to do, and we don't have to bring everybody into the clinic to deposit a sample and biobank the samples. But we can just, we could send home a kit containing everything you need for sampling, and then we can take the sample. Or the individual could take the sample whenever it suits them, and then we can just mail them back and put them on the shelf in a box. And of course, we could expand this to males, also in both men and women. And one thing that we've been investigating as well is to sample blood from a finger blood, capillary blood. So you send home a paper card, and you send home a little needle, and then you do a small puncture yourself, and you drop a couple of drops of blood on a paper card. And then there's the same procedure, basically. So you allow this to dry, and then you can send it through ordinary mail and then you can store it.
Natalie von der Lehr 25:15
You were a fellow at SCAS within the Natural Sciences Program in the spring of 2022. What was your experience of this multi- and interdisciplinary research environment?
Stefan Enroth 25:26
I think it was a very stimulating environment, both in terms of having the possibility to focus on my research questions without sort of having to put time on the other parts of being in academia, so Supervision and administration and all of these things, just focus on your research question was a very nice opportunity. But also to interact with the other fellows was very, very rewarding, and have a little peek inside other people's minds and how they address their research questions. They're very different from my own, and what types of methodologies they use and methods to reach their conclusions, which in many cases are strikingly different from natural sciences compared to humanities, for instance, where we always try for tons of observations based on statistics and things that we can measure objectively, while if you're studying a single manuscript, you can't really - you can't repeat or validate your findings in the next manuscript, because there is only one. So there's quite a different way of thinking on how you draw your conclusions and how you approach research. I think that was a very good experience.
Natalie von der Lehr 26:36
What kind of input did you get on your own project here?
Stefan Enroth 26:39
Maybe not so much specifically, but I guess we discussed things that are maybe general to any kind of, let's say, screening programs. We touched upon this, the fact that you might introduce worry, what would you do, sort of, in the meantime, is it better to know or not to know things like this?
Natalie von der Lehr 26:56
Is there anything from this experience here at SCAS that you had during this half year or so that you were here, that you brought back to your home department at Uppsala University?
Stefan Enroth 27:05
I think I did. So one obvious thing is the sort of wealth of pure factual knowledge, just fun facts, really, in topics that I've never even considered studying. But also, I think a broadened perspective on how you can do research and what it means to be in academia. A bigger mind maybe, and also maybe sort of realizing a little bit of what kind of a luxury it is, really, to work based on your own curiosity, try and find answers to questions, things you find interesting yourself and also be able to do some greater good in a long sense, that's rewarding, I think.
Natalie von der Lehr 27:54
I would like to thank you for coming here to the studio and talking to me and our listeners, of course,
Stefan Enroth 28:00
Thank you for inviting me.
Natalie von der Lehr 28:08
And thank you for listening to SCAS Talks, a podcast by the Swedish Collegium for Advanced Study. In this episode, I've talked to Stefan Enroth, Associate Professor of Computational Genomics at Uppsala University, and fellow within the Natural Sciences Program at SCAS during the spring of 2022. This was the second episode in our theme "Women in Medicine", and we have talked about Stefan Enroth's research on detecting biomarkers for screening and diagnosis for gynecological cancer. In the previous episode within this theme, we heard Anat Biegon, Professor of Radiology and Neurology at Stony Brook University, about the importance of recognizing sex differences in different diseases, as well as bias and education of health professionals in order to increase awareness about women's health. SCAS Talks features a broad variety of topics, which is a reflection of the multi- and interdisciplinary research environment at the collegium. We are sure that there is something of interest for everyone. Tune in, find your favorite topic or surprise yourself with something new. And as always, we are very happy if you can recommend SCAS Talks to your colleagues and friends. Subscribe to us and you won't miss any new content. SCAS Talks is available on podbean, Apple podcast, Spotify and most podcast apps. I would like to thank Stefan Enroth once again for talking to me, and thanks to you for listening. Bye for now.
Transcribed by https://otter.ai