Press "Enter" to skip to content

AI and Lung Ca Screening; Managing Mild Asthma: It’s PodMed Double T!

PodMed Double T is a weekly podcast from Texas Tech. In it, Elizabeth Tracey, director of electronic media for Johns Hopkins Medicine, 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. A transcript of the podcast is below the summary.

This week’s topics include best management of mild asthma, suicide data, bariatric surgery in adults and adolescents, and AI and lung cancer screening.

Program notes:

0:45 An AI model to improve lung cancer screening

1:47 Improve model’s predictive value

2:47 Had 42,000 CT scans

3:37 Management of mild asthma

4:38 Eosinophils in their sputum

5:37 Rise in suicides in 10-14 year old girls

6:40 Guns in 15-19 year olds

7:40 Social media relationship?

8:40 Use of bariatric surgery in adults and adolescents

9:40 Adolescents more likely to experience remission of diabetes

10:30 Shared decision making

11:15 End


Elizabeth Tracey: Can artificial intelligence be used to improve lung cancer screening?

Rick Lange, MD: How should we treat patients with mild asthma?

Elizabeth: A disturbing rise in suicides among 10- to 14-year-old girls.

Rick: And bariatric surgery in adolescents.

Elizabeth: That’s what we’re talking about this week on PodMed TT, your weekly look at the medical headlines from Texas Tech University Health Sciences Center in El Paso. I’m Elizabeth Tracey, a medical journalist at Johns Hopkins, and this will be posted on May 24th, 2019.

Rick: And I’m Rick Lange, President of the 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 Nature Medicine, something that I think is pretty provocative? All of us are hearing tons and tons about artificial intelligence, and in this case, they used a model to try to improve lung cancer screening in people who are at risk. Some of the background, of course — lung cancer is the most common cause of cancer death in the United States. And for folks who are at risk, it’s been shown that low-dose computed tomography, CT scan, has been shown to reduce mortality by 20% to 43% and is now part of the U.S. screening guidelines. However, there are an awful lot of false positives and false negatives with regard to that screening.

In this case, they did this model that performed really well, almost 95% of the things that should have been detected were. They called this an end-to-end approach where they attempted to look at, “All right, what happens in a radiologist’s workflow when they make this kind of a diagnosis?” They also used something that they called “deep convolutional neural networks.” And I’m not even going to go into that because I don’t know what that is, really, but it turns out that they think it’s important in improving this model’s predictive value.

So they used this in a national lung screening trial dataset with 42,000+ CT cases from just shy of 15,000 patients. Over the follow up, 578 of those folks developed biopsy-confirmed cancer, which I thought was really an interesting and a low number. They showed that the model achieved significantly better sensitivity and specificity than the average radiologist. The authors then conclude that we need to do this prospectively and retrospectively for a much bigger group of folks because it’s interesting that 14,000 patients and only 600 or so actually developed lung cancer during this follow up. They do say that they think that this approach, however, this deep-learning technique, artificial intelligence could also be employed with other 3-D imaging data, and that’s pretty powerful.

Rick: To accomplish something like this, the authors needed to have a large dataset. They had approximately 42,000 different CT scans and a number of individuals that had not suspected cancer, but biopsy-proven cancer. What happens is the more you have, the better it learns, the better it becomes. And as a result, there was a significant specificity boost of about 12% compared to well-trained, experienced radiologists.

Elizabeth: I think pretty exciting and some fruition relative to AI, which, as I said, I feel like every single day I hear at least one thing about it and its application in medicine.

Rick: And radiology is particularly suited to this because it’s image recognition, and when you have tens of thousands or hundreds of thousands of x-rays as input, it increases, again, the sensitivity and specificity of the test.

Elizabeth: Let’s turn to the New England Journal of Medicine, one of yours. Would you rather do the bariatric surgery or the asthma management?

Rick: Let’s do the asthma management first, and there were two studies that looked at the insight of how we should treat mild asthma. And this treatment of asthma has really been evolving. As you know, there are several different types of treatments one could get. Some are called short-acting beta agonists or SABAs, long-acting muscarinic agents or agonists, and steroids, as well. There’s a reluctance to use steroids because of their side effects.

Now if someone has persistent asthma, previous studies have suggested that although short-acting beta agonists can be beneficial in some individuals, if they have persistent asthma, they require the use of steroids. And initially, people were put on maintenance-daily steroids, but they’re not really very compliant with that. Subsequent studies showed that as-needed steroids were just as good. What about mild asthma because it’s thought that many of these individuals may respond to steroids, that is they have eosinophils in their sputum? Other individuals may not respond because they don’t have eosinophils in their sputum.

In these two studies, they took individuals that had mild asthma, and first of all, they asked, “Do they have eosinophils in their sputum or not?” indicating that they might respond to steroids. Interesting enough, three-fourths of the patients didn’t. When they looked at what agents were particularly useful in these individuals, they determined two things. One is that as-needed therapy was just as good as maintenance therapy, and secondly, many of these individuals don’t need steroids. They responded just as well to long-acting muscarinic agents or agonists, such as tiotropium, as they did to steroids.

Elizabeth: Well, that sounds like a really good outcome. Help me to understand this thing about eosinophils because, gosh, I thought that they were really important in this reaction.

Rick: They are in some individuals. They listed “this allergic response.” These individuals, by the way, if they do have eosinophils in their sputum, they do respond better to steroids than the other agents. However, in individuals that don’t have eosinophils, they respond just as well to the non-steroid treatments, i.e. the tiotropium, as they do to steroids.

Elizabeth: Let’s turn from here to JAMA Open, a very disturbing rise in suicides among — it turns out the population most at risk or the one in whom this is rising most quickly is girls age 10 to 14. This is from national data from the United States between 1975 and 2016. During this time period, 85,051 youth suicide deaths occurred in the United States, the vast majority — 80% of which — were among males, and almost 20%, of course, among females.

When they looked at this over time, they found that the ratio of male to female suicides was declining among those white youth who are aged 10 to 14 years, so the girls were catching up with the boys with regard to the suicide risk. The male-to-female risk for firearm use increased significantly in those 15 to 19 years. So guns, of course, still a really popular choice, if you will, relative to committing suicide. They did decline hanging and suffocation, and there was no difference in the poisoning selection over that time period.

In the editorial, the editorialists speculate that one reason for this sex disparity that may have girls choosing suicide more often now is that their social media use may be more likely to result in interpersonal stress, which is a common factor associated with suicide attempts in youth. Compared with boys, girls use social media more frequently and are also more likely to experience cyberbullying, so they suggest that this association between social media use and mental health outcomes also needs to be looked at a little bit more carefully.

Rick: The study was not designed, in fact, it didn’t even aim to understand what the underlying mechanisms were, but other studies, as you suggest, indicate that cyberbullying may be more of a problem with girls than boys. Girls spend more time on social media. There’s more cyberbullying for girls as opposed to boys. Other studies have shown that. It’s an interesting hypothesis and certainly needs additional studies.

Elizabeth: Yeah, what do you think about the choice of firearms? Not surprising in some ways.

Rick: We’ve talked before about the fact that just the availability of firearms in the home and the fact that they’re not locked up. For kids that make rash decisions, it’s unfortunately an easy and available way to commit suicide. Again, all of our listeners should be aware that if you do have firearms, really they should be locked up.

Elizabeth: And would you also note that parents probably need to be paying attention to the social media that their children are accessing?

Rick: They do. I’d say it’s not only parents, but all of us that interact with adolescents. Pediatricians need to ask about it. Teachers need to be aware of it as well. Counselors. And unfortunately, it’s a really prevalent issue and a really prevalent problem right now. As you suggested, it appears to affect girls more than boys.

Elizabeth: OK. On that hypothesis-generating note, let’s turn to your final one in the New England Journal of Medicine, the use of bariatric surgery among adolescents.

Rick: It was two parallel studies, both looking at gastric bypass, just one in adolescents and the other in adults. What these authors attempted to do was to say, “Listen, in teens with severe obesity, can bariatric surgery wait until adulthood?” They looked at a particular type of gastric bypass called a Roux-en-Y in a cohort of 161 adolescents and a cohort of almost 400 adults. And they looked at the percent weight change between adolescents and also in the remission of other diseases associated with it such as diabetes and/or hypertension.

What they found out was at the end of 5 years the percent weight change between adolescents and adults was the same. It ranged between 25% and 30%. However, when they looked at remission of diabetes, 86% of children experience [that] versus only 53% of adults. When they looked at remission of hypertension, it was 68% in children versus 41% in adults. Both adolescents and adults had marked weight loss, similar magnitude after 5 years, but adolescents were much more likely to experience remission of diabetes and hypertension than adults.

Elizabeth: I think another outcome that’s worth considering is the idea that, for adolescents, the whole social structure that’s surrounding them I think would say, “Hey, let’s go ahead and have the surgery,” because they’re so sensitive to the whole body image and acceptance by their peers.

Rick: So it’s interesting, Elizabeth. If you look at it long term, there are negative health outcomes associated with bariatric surgery both reported in adolescents and also in adults. These include the potential for self-harm, increased risk of alcohol or drug abuse in those that have had bariatric surgery. In addition, there’s mortality and it ranges between 1.5% and 2%. That was true for both adolescents and adults. So, armed with all of this information, I think that the weight-control specialists can speak to adolescents and to their parents, by the way, about the risk and benefits of the surgery so they can be involved in that decision-making process. And I say the parents need to be involved with that as well because adolescent patients may really not have the fully developed capacity for making these important decisions by themselves.

Elizabeth: Absolutely. I agree with that. OK, on that note, that’s a look at this week’s medical headlines from Texas Tech. I’m Elizabeth Tracey.

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