鶹ýӰ

AI and Breast Cancer Screening; Cancer After Treatment for Sickle Cell

— Also in TTHealthWatch: mail-in cervical cancer screening kits

MedpageToday

TTHealthWatch is a weekly podcast from Texas Tech. In it, Elizabeth Tracey, director of electronic media for Johns Hopkins Medicine in Baltimore, and Rick Lange, MD, president of the Texas Tech University Health Sciences Center in El Paso, look at the top medical stories of the week.

This week's topics include AI and breast cancer screening, vaccination and long COVID, mailed cervical cancer screening kits, and cancers after treatment for sickle cell.

Program notes:

0:42 Long COVID after vaccination

1:42 Full range of vaccines included

2:42 Two vaccines

3:00 Gene therapy for sickle cell and cancer

4:00 Related to chemotherapy?

5:00 Some way to screen cells

5:20

6:20 Get education or not, or queried on desire

7:20 Not surprised for overdue

8:20 AI to improve early detection of breast cancer

9:20 Achieve 0.2% more

10:20 Troubling for women with too much recall

11:18 End

Transcript:

Elizabeth: Can we improve cervical cancer screening with mailed HPV self-sampling kits?

Rick: Why do people with gene therapy for sickle cell disease develop malignancies?

Elizabeth: What's the impact of COVID vaccination on the development of long COVID?

Rick: And using artificial intelligence to improve early detection of breast cancer.

Elizabeth: That's what we're talking about this week on TT HealthWatch, your weekly look at the medical headlines from Texas Tech University Health Sciences Center in El Paso. I'm Elizabeth Tracey, a Baltimore-based medical journalist.

Rick: And I'm Rick Lange, president of Texas Tech University Health Sciences Center in El Paso, where I'm also dean of the Paul L. Foster School of Medicine.

Elizabeth: Rick, how about if we turn first to the BMJ? We have in the past always treated COVID material first, and then we've kind of abandoned that practice. I'm going to recall it, at least for today.

In this study, performed in Sweden, they looked at just shy of 600,000 individuals and looked at their COVID-19 vaccine status, and examined, "What was the impact of that on the development of so-called long COVID?", which they identify as "post-COVID-19 condition," or PCC. This, of course, points up the need for some international agreement on what we're going to call this condition, since it seems to be so prevalent.

What they did was look at all adults with COVID-19 first registered between December 2020 and February 2022 in the two largest regions in Sweden. They looked at individuals who had received at least one dose of a COVID-19 vaccine before infection. It was the full range of the COVID vaccines that were considered in this study. The vaccine effectiveness against post-COVID condition for one dose, two doses, and three or more doses was 21%, 59%, and 73% respectively. This is a pretty powerful reason to at least get one vaccine in view of the fact that so many people report post-COVID sequelae as pretty dire.

Rick: We've already identified that the vaccine is effective in reducing mortality. It reduces the need for ICU hospitalization. It reduces the need for overall hospitalization. Now, we're saying that it's also effective in reducing long COVID symptoms as well. That's pretty powerful if you ask me.

Elizabeth: I think so, too. I hope that there is not anyone out there who's not thinking that they want to get vaccinated again, and it's sure looking like this is going to become a seasonal thing.

Rick: Many people got one or two vaccines, and I'm hearing people getting COVID fatigue and not getting the third vaccine. But, again, I want to highlight, as you mentioned, getting two vaccines reduces the risk of long COVID by 59%, and three vaccines by 73%. Getting fully vaccinated is beneficial, not only reducing the severity of disease, but also the long-term symptoms as well.

Elizabeth: Let's turn now to Nature Medicine, something that is a little bit maybe less positive spin, what's called clonal selection of hematopoietic stem cells after gene therapy for sickle cell disease, just approved by the European Union for people with sickle cell.

Rick: We're going to see a lot more gene therapy treatments for various diseases become increasingly available to people. There are currently hundreds of clinical trials. The early success in gene therapy was actually dampened by reports of patients who actually developed leukemias related to how the genes were inserted.

Now, it's been reduced by what you call improved vector design. However, there are still reports of people that develop myeloid, or bone cancer malignancies, leukemias 3 to 6 years after their transplantation. The question is, why does that occur? Is it because whatever the person has -- they just have an increased rate of just mutations? Is there increased mutations as a result of the hematopoietic stem cells that they put into the individual for the gene therapy?

The third possibility is it's related to the chemotherapy the person gets. Fourth is -- and this is what the authors found out -- there is a selective pressure on the hematopoietic stem cells that actually contain a pre-existing driver mutation. These hematopoietic stem cells have their own drivers for developing malignancy and whatever the condition is, however they were manipulated, or however they were put in, or subsequently how the patient recovers afterwards, there is an increased driver for these particular hematopoietic stem cells, either modified or unmodified, to actually increase the risk of malignancy.

Elizabeth: This, of course, is pretty sobering because we have these very high hopes for these kinds of therapies, especially for folks with sickle cell because they have a whole lifetime ahead of them.

Rick: I want to stress that it happens fairly infrequently, under 5%. Before we can understand how best to avoid it, or to treat it, or to assess for it or prevent it, we need to understand the mechanism of it, and that's why this particular study ends up being incredibly important.

Elizabeth: It seems to me that there will be some way to screen these cells for this particular driver before re-infusing them into someone.

Rick: Or we have better surveillance. Or perhaps there is a therapy we can give prophylactically to prevent it. So there are a lot of different ways to address it.

Elizabeth: Let's turn to JAMA. This is looking at increasing cervical cancer screening by mailing kits to women who are overdue for that particular type of screening. I would just note that this mailing kits so that women in particular can perform their own kind of screening tests and then mail them back in so that they can be processed has been underway for quite some time, especially for common STIs [sexually transmitted infections]. This strategy has already gotten all sorts of the bones in place in order to make it more successful.

The authors say that mailing these kits to women who are actually compliant most of the time with screening has never really been evaluated here in the U.S. So they decided to evaluate that effectiveness. They recruited as a Kaiser Permanente population individuals who were 30 to 64 years of age, women.

They stratified them into several different groups. Basically, they could either get education or not, and then, "Do you want to get a mailed kit sent to you, or not?" Then they said, "All right, did this help? Did you complete screening within 6 months?"

Their randomization included 31,000+ people. They found that, in fact, it did increase cervical cancer screening by more than 14% in individuals who were due or overdue for cervical cancer screening. To me, this says, "This issue of direct mail? It's probably a good idea."

Rick: This cervical cancer screening is actually looking for human papilloma virus. If it's negative, then the women don't need to undergo a Pap smear. But if it's positive, then it's recommended that the women come in to the clinic and get a Pap smear.

Now, obviously, you can't do Pap smears at home, but you can detect HPV and that's exactly what this test does. I'm not surprised that for women who were overdue to have their screening done, that mailing them a kit encouraged them to mail it back.

I am a little surprised that women that were not overdue -- that is, who routinely come in -- thought, "Hey, this is easier. I think I'll do it." And it increased cancer screening 14% as well. I think this has a lot of merit.

Again, they didn't give the woman the option. They just mailed it automatically. Because if women opted in, they oftentimes did not. There was no benefit to that. But if they got it, then they used it and they send it back. I think this is good news.

It also makes it easier to screen women that are in rural areas or don't have access, and women that are working and may not want to take off from work to go to the doctor's office, so a lot of benefit to this particular type of screening.

Elizabeth: I absolutely agree and certainly we're seeing tons of that -- for colorectal cancer screening, for example, lots of things where people get mailed things and then they just send it back in, reduces the barriers.

Rick: Not only did it increase screening participation, but it also resulted in a decreased time to screen. They were more likely to do it because it was so convenient.

Elizabeth: Win-win. Finally, let's turn back to Nature Medicine: breast cancer and AI.

Rick: This is using AI to improve early detection of breast cancer. In the past, we've looked at computer-aided detection systems that were supposed to improve cancer detection, but the benefits in the real world really weren't there.

Now, what modern artificial intelligence (AI ) does is, it has deep learning -- it's a different technology -- and it actually has a higher potential quality for reading. More importantly, it's less likely to call out things that aren't positive.

Here is what these investigators did. They didn't interfere with the routine way that mammograms were read. First of all, they took an off-the-shelf, AI-assisted additional reading process. The usual way is they have two readers; both look at the mammogram. If they agree that there is something that's fishy, they flag that and they call it positive or they do additional studies. But if they both agree that there is nothing unusual about it, they don't recall the patient.

Now, here is where AI is inserted. It would look at those mammograms and say, "Whoa, wait a minute. There may be something you missed." The AI-assisted additional-reader process could achieve an additional 0.7 to 1.6 additional cancer detections per thousand cases, 0.2%. The additional recalls, less than 1%. Unnecessary recalls, less than 1%.

The positive predictive value went up and the majority of cancers detected by this AI-assisted were invasive and they were small, therefore those are the ones you want to treat. Those are the ones you want to catch early. Using AI-assisted reading can actually improve early detection of breast cancer with minimal to no unnecessary recalls.

Elizabeth: This is good stuff because we have talked about many times in the past that one of the issues with even tomosynthesis and other more sophisticated analysis and imaging techniques still results in a pretty significant number of recalls that turn out to not be anything. That's very troubling for women when that happens.

Rick: Yep. I mean, what you want to do is you want to find the right balance. You don't want to "over read" things where you're increasing the workload on the physicians and you're bringing women back, and you don't want to miss things either. The nice thing about this is that it's an off-the-shelf system; you can routinely use it.

Elizabeth: I'll just note, of course, that imaging is low-hanging fruit in terms of AI. We're going to be talking a lot more about AI and then about a lot more multifactorial kinds of things. I feel absolutely certain.

Rick: Yeah. But having a technique that can increase the rate of early cancer detection by 5% to 13%, and these are invasive and small ones, I call that a big win.

Elizabeth: On that note then, that's a look at this week's medical headlines from Texas Tech. I'm Elizabeth Tracey.

Rick: And I'm Rick Lange. Y'all listen up and make healthy choices.