Rethinking Surveillance of Barrett Esophagus
Hello. I'm Dr. David Johnson, Professor of Medicine and Chief of Gastroenterology at Eastern Virginia Medical School in Norfolk, Virginia.
How often should Barrett esophagus surveillance be performed? Why do we do it? Is it really cost effective? We have relied on national society guidelines, which for the most part have been predicated on databases that are small with fairly short-term follow-up, and which suggest that the annualized progression from nondysplastic Barrett esophagus to high-grade dysplasia is approximately 0.9% per year and 0.4%-0.5% per year for adenocarcinoma. The rationale behind surveillance of Barrett esophagus is to prevent these patients from progressing to cancer or to recognize progression to high-grade dysplasia or early cancer. With the advent of therapeutic endoscopy and ablative strategies, we want to keep these patients away from progression toward a more ominous cancer and potentially avoid surgery.
The surveillance intervals that are recommended by national society guidelines are every 3 years. To be cost effective, it has been suggested that for nondysplastic Barrett esophagus, the incidence of adenocarcinoma would have to be in the range of 1.9% per year and surveillance should be performed at 5-year intervals. We are a ways away from cost-effective models in what we actually justify in clinical practice and set in national guidelines.
The most recent study is from Northern Ireland, which was a database analysis of 9300 patients over 22 years. Patients who had an index biopsy per high-grade dysplasia, adenocarcinoma, or did not have follow-up for more than a year were excluded. The follow-up range was 1-22 years, and about 8400 patients were followed. The annualized cancer, or high-grade dysplasia progression, was 0.22% per year. That is a news flash, and if you think about what I said before, high-grade dysplasia progression from standard databases in the United States has been about 0.9%; cancer risk progression has been about 0.4%-0.5%; and the combined risk from this Northern Ireland database is 0.22%. Is that applicable to the United States?
Only 46% of the patients in this database had actual biopsies because in Europe they don't require biopsies to establish the diagnosis of Barrett esophagus. In the United States, however, we require evidence of specialized columnar metaplasia to truly make the diagnosis of Barrett esophagus. If you look further at that data and tease out the patients who have only specialized columnar metaplasia, the annualized incidence is 0.38%, a little closer to what we have seen in the United States for high-grade dysplasia or cancer, but still a lot lower than the 1.4% that we would normally sum for high-grade dysplasia and cancer. If you look at the specialized columnar metaplasia vs the absence of specialized columnar metaplasia in the patients who had biopsy, there was nearly a 5.5 whole difference (0.38% vs 0.07%). Therefore, it is clearly specialized columnar metaplasia that had the cancer risk or high-grade dysplasia risk associated with it.
If we look further, it becomes very apparent that when they did a Kaplan-Meier analysis of the progression of cancer over time, that it was a uniform progression. This is very important because the most cost-effective models -- all of them, in fact -- assume that for these applications there is a uniform extrapolation of cancer risk. This is the first database that has actually had the power to show that it is an annualized incidence and that it doesn't change. The cohort was divided into 1-6 years, 6-11 years, and > 11 years. The annualized incidence seemed consistent through each of those time analyses. It was also apparent that the risk was higher for patients of a certain age. The highest risk was in the range of 60-68 years of age. The annualized incidence was around 0.27%, and it dropped off on either side. For patients under 50 years of age, it fell down to 0.12% -- very low. Even at an older age, over 80, the number was around 0.17%. Again, it is somewhat of a bell-shaped curve, although the numbers in the higher range of age were limited by the total number of patients. Nonetheless, the average risk seemed to peak in the fifth or sixth decade. The other issue was risk in men vs women. There was a 2-fold incidence risk in men vs women.
This study answers a couple of significant questions. First, what is the relevant risk for cancer progression? I would tell you that at least in these Northern Ireland data, if we hone in on the specialized columnar metaplasia, the risk is quite low -- 0.38% is quite low over time. The risk seems to be extended, annualized, and steadied. Second, is this an extrapolatable risk given time, or does it just occur or snowball as time progresses? The answer is that it doesn't appear to. Is there a gender difference? Yes, there is, and this has been successfully demonstrated in many previous publications. Men are more subject to risk than women, and specialized columnar metaplasia seems to be the considerable risk. That certainly makes sense, and we justify the diagnosis in the United States. In countries where this is not justified, it may be appropriate to go back and look at the surveillance recommendation where the overall risk was measurably low, 0.7% per year in patients who had no specialized columnar metaplasia.
I think we can take a couple of things from this unique study. One is that the risk is quite low. It is annualized and it certainly is evident over time. It justifies even more that we should be extending the intervals for Barrett esophagus surveillance, especially in patients without dysplasia, particularly women, and those of young age. Maybe they shouldn't even have the diagnosis established early. I think we would do a safety check on referral for patients for ablative therapies for nondysplastic Barrett esophagus, and that is supportive of what current consensus of recommendations and experts say at present. For your next conversation about Barrett esophagus, hopefully you will be forearmed with this new information and it will serve you well in your discussion.
I am Dr. David Johnson. Thanks for listening.
Hello. I'm Dr. David Johnson, Professor of Medicine and Chief of Gastroenterology at Eastern Virginia Medical School in Norfolk, Virginia.
How often should Barrett esophagus surveillance be performed? Why do we do it? Is it really cost effective? We have relied on national society guidelines, which for the most part have been predicated on databases that are small with fairly short-term follow-up, and which suggest that the annualized progression from nondysplastic Barrett esophagus to high-grade dysplasia is approximately 0.9% per year and 0.4%-0.5% per year for adenocarcinoma. The rationale behind surveillance of Barrett esophagus is to prevent these patients from progressing to cancer or to recognize progression to high-grade dysplasia or early cancer. With the advent of therapeutic endoscopy and ablative strategies, we want to keep these patients away from progression toward a more ominous cancer and potentially avoid surgery.
The surveillance intervals that are recommended by national society guidelines are every 3 years. To be cost effective, it has been suggested that for nondysplastic Barrett esophagus, the incidence of adenocarcinoma would have to be in the range of 1.9% per year and surveillance should be performed at 5-year intervals. We are a ways away from cost-effective models in what we actually justify in clinical practice and set in national guidelines.
The most recent study is from Northern Ireland, which was a database analysis of 9300 patients over 22 years. Patients who had an index biopsy per high-grade dysplasia, adenocarcinoma, or did not have follow-up for more than a year were excluded. The follow-up range was 1-22 years, and about 8400 patients were followed. The annualized cancer, or high-grade dysplasia progression, was 0.22% per year. That is a news flash, and if you think about what I said before, high-grade dysplasia progression from standard databases in the United States has been about 0.9%; cancer risk progression has been about 0.4%-0.5%; and the combined risk from this Northern Ireland database is 0.22%. Is that applicable to the United States?
Only 46% of the patients in this database had actual biopsies because in Europe they don't require biopsies to establish the diagnosis of Barrett esophagus. In the United States, however, we require evidence of specialized columnar metaplasia to truly make the diagnosis of Barrett esophagus. If you look further at that data and tease out the patients who have only specialized columnar metaplasia, the annualized incidence is 0.38%, a little closer to what we have seen in the United States for high-grade dysplasia or cancer, but still a lot lower than the 1.4% that we would normally sum for high-grade dysplasia and cancer. If you look at the specialized columnar metaplasia vs the absence of specialized columnar metaplasia in the patients who had biopsy, there was nearly a 5.5 whole difference (0.38% vs 0.07%). Therefore, it is clearly specialized columnar metaplasia that had the cancer risk or high-grade dysplasia risk associated with it.
If we look further, it becomes very apparent that when they did a Kaplan-Meier analysis of the progression of cancer over time, that it was a uniform progression. This is very important because the most cost-effective models -- all of them, in fact -- assume that for these applications there is a uniform extrapolation of cancer risk. This is the first database that has actually had the power to show that it is an annualized incidence and that it doesn't change. The cohort was divided into 1-6 years, 6-11 years, and > 11 years. The annualized incidence seemed consistent through each of those time analyses. It was also apparent that the risk was higher for patients of a certain age. The highest risk was in the range of 60-68 years of age. The annualized incidence was around 0.27%, and it dropped off on either side. For patients under 50 years of age, it fell down to 0.12% -- very low. Even at an older age, over 80, the number was around 0.17%. Again, it is somewhat of a bell-shaped curve, although the numbers in the higher range of age were limited by the total number of patients. Nonetheless, the average risk seemed to peak in the fifth or sixth decade. The other issue was risk in men vs women. There was a 2-fold incidence risk in men vs women.
This study answers a couple of significant questions. First, what is the relevant risk for cancer progression? I would tell you that at least in these Northern Ireland data, if we hone in on the specialized columnar metaplasia, the risk is quite low -- 0.38% is quite low over time. The risk seems to be extended, annualized, and steadied. Second, is this an extrapolatable risk given time, or does it just occur or snowball as time progresses? The answer is that it doesn't appear to. Is there a gender difference? Yes, there is, and this has been successfully demonstrated in many previous publications. Men are more subject to risk than women, and specialized columnar metaplasia seems to be the considerable risk. That certainly makes sense, and we justify the diagnosis in the United States. In countries where this is not justified, it may be appropriate to go back and look at the surveillance recommendation where the overall risk was measurably low, 0.7% per year in patients who had no specialized columnar metaplasia.
I think we can take a couple of things from this unique study. One is that the risk is quite low. It is annualized and it certainly is evident over time. It justifies even more that we should be extending the intervals for Barrett esophagus surveillance, especially in patients without dysplasia, particularly women, and those of young age. Maybe they shouldn't even have the diagnosis established early. I think we would do a safety check on referral for patients for ablative therapies for nondysplastic Barrett esophagus, and that is supportive of what current consensus of recommendations and experts say at present. For your next conversation about Barrett esophagus, hopefully you will be forearmed with this new information and it will serve you well in your discussion.
I am Dr. David Johnson. Thanks for listening.