Cover Story

Promising Technologies Arise for Diagnosis of Barrett’s Esophagus

Recent and prospective advancements in diagnosing cancer in Barrett’s esophagus may allow for earlier and better detection. But as with most newer technologies, some of these alternatives — ranging from computer analysis to surveillance endoscopy to saliva — have their critics as well.

The standard diagnostic for detecting the disease is a high-quality endoscope, preferably high definition, which uses white light, according to Prateek Sharma, MD, a professor of medicine at the University of Kansas in Kansas City. The technique entails a careful examination of the distal esophagus “to look for a columnar-lined esophagus,” he told Healio Gastroenterology. “Once that is detected using endoscopy, you then take biopsies from the suspicious segment.”

Sharma said endoscopy is “highly accurate” for diagnosing suspected cases of Barrett’s esophagus (BE) or intestinal metaplasia. “However, for dysplasia, the accuracy falls significantly because the dysplasia within the Barrett’s segment is quite patchy and focal,” he said. “The dysplasia can occupy very small surface areas within the BE that may not be visible and may not be collected as part of the random biopsy.”

Other downsides of the biopsy are that it can also take up to 8 minutes to perform, depending on the segment length, and incurs additional costs due to instrument and pathology. “Each centimeter-level biopsy is placed in a separate container,” Sharma said.

Prateek Sharma

Emerging Technology

Wide area transepithelial sampling (WATS, CDx Diagnostics) addresses some of the shortcomings of the biopsy protocol, Sharma said. First, it uses a brush that allows movement over a larger surface area of the BE for more encompassing specimen collection.

“Although in the past we have used brushes, they were more superficial and looked at exfoliative cytology,” Sharma said. “In contrast, the brush we use for the WATS technique is more abrasive, so it collects not only the shed cells, but actually goes deeper to collect glandular tissue as well.”

Sharma was one of the lead investigators of a large multi-center study of WATS that was presented in May at Digestive Disease Week. “We found a 14% increase in the detection of patients with high-grade dysplasia/cancer when WATS was used in conjunction with random four-quadrant biopsies,” he said.

WATS also removes some of the pathologist’s objectivity because all the collected cells are fed into a computer, “which presents the 200 most abnormal cells to the pathologist, who then makes the diagnosis,” Sharma said. “For making a diagnosis of dysplasia, the addition of WATS to the biopsy protocol increases the detection rate of dysplasia.”

Improving surveillance in patients with BE is also possible with electronic chromoendoscopy, which is built into the scope. “This allows you to change the wavelength of light and to magnify into the Barrett’s mucosa, thereby detecting subtle lesions for targeted biopsy,” he said.

For diagnosing BE-associated neoplasia, “we are at a time frame where random biopsies are a thing of the past,” Sharma said. “We are looking at larger sampling, wider sampling, better diagnosis and targeted biopsy.”

Computer Analysis

Erik J. Schoon, MD, PhD, a gastroenterologist and head of endoscopy at Catharina Hospital in Eindhoven, the Netherlands, led a study published in Endoscopy, which assessed diagnosing BE by computer analysis.

Erik J. Schoon

“The results of the paper show that as ‘proof of principle’ and starting off with one of the most difficult subjects in endoscopy, BE, that the computer algorithm developed could almost recognize dysplasia and early cancer from still endoscopic images as accurately as four of the best European experts could,” Schoon told Healio Gastroenterology. “The computer program was also able to delineate the abnormal areas.”

For the study, 100 images from 44 patients with BE were analyzed via a computer algorithm, which employed specific texture, color filters and machine learning to detect early neoplastic lesions. The system identified early neoplastic lesions on a per-image analysis with a sensitivity and specificity of 0.83. At the patient level, the sensitivity and specificity were 0.86 and 0.87, respectively.

Figures 1 and 2. Barrett’s cancer and computer automated detection (CAD). Expert’s delineation on the left. Automatic delineation on the right.

Images: Schoon EJ

“A trade-off between the two performance metrics could be made by varying the percentage of training samples that showed neoplastic tissue,” Schoon said. “However, further research is required to improve the accuracy of the system and prepare it for real-time operation, before it can be applied in clinical practice.”

Schoon said it will take several more years before the program is suitable for clinical practice. “The aim is to make a real-time computer analysis and to validate the system,” he said. The Dutch Cancer Foundation and a National Technical Foundation from the Government are funding part of the further research.

“Computer analysis is appropriate for several image-based procedures in different fields of health care,” Schoon said.

Surveillance Endoscopy

Surveillance endoscopy is promising for patients with BE that is associated with significantly better EAC outcomes, including cancer-related mortality, compared with other non-surveillance endoscopy.

“Given that BE is the only known precursor/predisposing condition for esophageal adenocarcinoma (EAC), and that endoscopic surveillance of BE has the potential of uncovering precancer lesions (dysplasia) which can be effectively cured by endoscopic resection or radiofrequency ablation, higher detection of BE has the potential of reducing the incidence and mortality of esophageal adenocarcinoma,” Hashem B. El-Serag, MD, MPH, professor of medicine and chief of gastroenterology and hepatology at the Michael E. DeBakey Veterans Affairs Medical Center and Baylor College of Medicine in Houston, told Healio Gastroenterology.

Hashem B. El-Serag

“We look for evidence of neoplasia (dysplasia or cancer),” El-Serag said. “The frequency of these endoscopies is every 3 to 5 years among patients with non-dysplasia.”

The recommended careful examination consists of white light endoscopy and some magnification/high-resolution endoscopy (eg, narrow band imaging), coupled with taking biopsies from any suspicious areas (four quadrant samples every 1 cm to 2 cm).

El-Serag and colleagues looked at 29,536 patients with BE in a study published in Gut, 424 patients developed EAC during a mean follow-up period of 5 years. A total of 209 (49.3%) patients with EAC were assigned to a BE surveillance program and were diagnosed as a result of surveillance endoscopy. “These patients were more likely to be diagnosed at an early stage (stage 0 or 1: 74.7% vs. 56.2%; P < .001), survived longer (median 3.2 vs. 2.3 years; P < .001), have lower cancer-related mortality (34% vs. 54%, P < .0001) and were more likely to receive esophagectomy (51.2% vs. 42.3%) than the 215 patients diagnosed by non-BE surveillance endoscopy, 17.2% of whom were BE surveillance failure,” El-Serag said.

BE surveillance endoscopy was also associated with a decreased risk for cancer-related death (HR = 0.47, 95% CI, 0.35-0.64), “which was largely explained by the early stage of EAC at the time of diagnosis,” El-Serag said. Similarly, the adjusted mortality for patients with cancer in a prior surveillance program for overall death was 0.63 (0.47-0.84) compared with patients with cancer not enrolled in a surveillance program.

El-Serag said that most cases of EAC are diagnosed in patients who present with symptoms: difficulty swallowing, chest pain, weight loss, anemia. “In these patients, the BE can be related to the cancer,” he said. “Our study highlighted the value of knowing that among those with recognized (eg, diagnosed) BE before EAC diagnosis, endoscopic surveillance is of benefit.”

A separate study in Gut “reported that the mere knowledge of BE diagnosis before the cancer was a predictive factor of better cancer outcomes,” El-Serag said.

EAC has a “grave prognosis,” with a 5-year survival rate of only 16%, according to David G. Graham, MBBS, BSc, MRCP, a clinical research fellow pursuing his PhD in early detection of EAC at University College London in the United Kingdom. “BE is the only known precursor lesion to EAC. Although the majority of EAC arises in BE, the majority of individuals diagnosed with EAC are unaware of having BE.”

David G. Graham

Graham said studies have demonstrated that existing BE surveillance programs are ineffective. Large sums of resources are devoted to these programs, “even though the vast majority of individuals will not progress,” he said.

To meaningfully impact the prognosis of EAC, “we need to detect those with BE, and more importantly, those with high-risk, dysplastic BE and early EAC who will benefit from the effective, minimally invasive therapies, such as endoscopic resection and ablation,” Graham said. “However, these lesions are often asymptomatic or with symptoms indistinguishable from benign disease. Diagnostic symptoms and signs often only develop once cancer is well established and incurable. Thus, we need to find means to effectively screen an asymptomatic population.”

Graham noted that endoscopic surveillance of the population is invasive, impractical, expensive and likely to be ineffective. “Recent publications of endoscopic image computer analysis software and the WATS3D biopsy analysis, although promising, still involve endoscopy, not applicable to population screening,” he said.

A 2014 article in PLoS One attempted to identify at-risk individuals for BE through analysis of demographic, symptom and risk factor data alone, “but was impractical for usage, due to a low specificity (22%),” Graham said.

Another alternative to endoscopy is the Cytosponge (Medtronic GI Solutions), which is a capsule on a string that when swallowed becomes an abrasive sponge and then is pulled out of the mouth. Promising data were published last year in PLoS Med on this minimally invasive means to identify BE and EAC, but Graham expressed his doubts as to its use.

“We have been closely involved in these studies. Around 50% of patients approached to swallow the sponge on a second occasion declined to do so,” Graham said. “This augurs badly for a long-term surveillance strategy.”

Artificial Intelligence

To overcome some of the short comings of various diagnostic techniques, Graham and his colleagues have developed a tool to identify those with EAC and BE via artificial intelligence (AI) analysis of salivary transcriptomic data, alongside patient demographics and symptom and risk factor data obtained through a questionnaire. “Using AI to analyze ‘big data’ provides the ability to aggregate and mine vast sums of data and develop new insights into disease,” Graham said.

At DDW 2016, Graham presented results of an 80-patient study using analysis of saliva, a cross-section of data from a patient questionnaire and AI analysis. Patients were divided into four groups: 20 normal, 20 non-dysplastic BE (NDBE), 20 high-grade dysplasia (HGD) and 20 EAC. Recruited patients provided a 1-mL saliva sample (collected following fasting) and completed a specifically designed questionnaire (205 data points encompassing demographic, symptom and risk factor data linked to BE and EAC).

Figure 3. Examples of some of the significantly aberrantly expressed genes.

Images: Graham DG

Figure 4. The diagnoses of patients from low to high expression of CDKN2a.

“RNA was extracted from saliva using a QIAzol lysis reagent chloroform process that was quality controlled by the Nanodrop and Bioanalyzer,” Graham said. “Each sample also underwent a TURBO DNase step prior to complementary DNA (cDNA) creation.”

Targeted expression analysis was performed using quantitative real-time PCR. Target-specific primers with known links to BE and EAC were identified through a literature search and reanalysis of a public domain expression dataset. “Overall, 13 genes (AMY1, AMY2, CDKN2a, SMAD7, TLR6, TP53, BRAF, EGFR, KIT9, KRAS, NRAS, PIK3CA and PTEN) were selected with different exons within the genes targeted, creating 22 targets in total,” Graham said.

Expression data was then obtained and relative expression was calculated by comparing the cycle threshold of each amplicon against 18S rRNA (housekeeping gene). A relative expression/cycle threshold was also calculated to compare the four categories of patients. “Finally, a systematic Mann-Whitney U test was carried out between all the groups,” Graham said.

In summary, 11 of the 22 targets were found to be significantly aberrantly expressed at different stages of progression toward EAC. The eight genes with significant expression findings were: AMY2, CDKN2a, SMAD7, TLR6, EGFR, KIT9, PTEN and TP53.

“The questionnaire and expression data were analyzed using novel AI techniques to identify the key attributes for identifying those with EAC and those at risk through BE and HGD,” Graham said. For the tool, six of the selected attributes were aberrantly expressed genes, including AMY2, CDKN2a, SMAD7 and TP53, whereas five attributes were from the questionnaire data, including the waist–hip ratio, proton pump inhibitor use, previous cancer and smoking.

“Using combinations of these attributes, one could identify those with or at risk of EAC with 93% sensitivity and 73% specificity,” Graham said.

Looking toward the future, Graham believes his group’s work will benefit from pending whole transcriptome analysis, performed on selected samples, which will identify further genes of importance requiring validation through targeted expression analysis. “Additionally, we have performed epigenetic analysis on patient samples, with preliminary data suggesting that this significantly increases the tool’s accuracy and may be able to eliminate the need for the patient questionnaire,” he said.

The advances in molecular biology, particularly the ability to effectively analyze DNA and RNA “of poorer quantity and quality than ever before, allows the approach we have taken — and the promising field of ‘salivary diagnostics’ and other forms of ‘liquid biopsies’ — to potentially provide a noninvasive means to not only detect almost any disease, but also to longitudinally monitor an individual, including identification of potential relapses or recurrences,” Graham said. “The potential impact this will have on health care is immeasurable.”

Schoon said that genetic analysis could be a predictor of acquiring dysplasia in BE, “so select the high-risk patients. But BE patients still have to be endoscopically screened for dysplasia.” – by Bob Kronemyer

Disclosures: El-Serag, Graham and Schoon report no relevant financial disclosures. Sharma reports grant support from Olympus Inc., CDx Diagnostics and Cook Medical.

Recent and prospective advancements in diagnosing cancer in Barrett’s esophagus may allow for earlier and better detection. But as with most newer technologies, some of these alternatives — ranging from computer analysis to surveillance endoscopy to saliva — have their critics as well.

The standard diagnostic for detecting the disease is a high-quality endoscope, preferably high definition, which uses white light, according to Prateek Sharma, MD, a professor of medicine at the University of Kansas in Kansas City. The technique entails a careful examination of the distal esophagus “to look for a columnar-lined esophagus,” he told Healio Gastroenterology. “Once that is detected using endoscopy, you then take biopsies from the suspicious segment.”

Sharma said endoscopy is “highly accurate” for diagnosing suspected cases of Barrett’s esophagus (BE) or intestinal metaplasia. “However, for dysplasia, the accuracy falls significantly because the dysplasia within the Barrett’s segment is quite patchy and focal,” he said. “The dysplasia can occupy very small surface areas within the BE that may not be visible and may not be collected as part of the random biopsy.”

Other downsides of the biopsy are that it can also take up to 8 minutes to perform, depending on the segment length, and incurs additional costs due to instrument and pathology. “Each centimeter-level biopsy is placed in a separate container,” Sharma said.

Prateek Sharma

Emerging Technology

Wide area transepithelial sampling (WATS, CDx Diagnostics) addresses some of the shortcomings of the biopsy protocol, Sharma said. First, it uses a brush that allows movement over a larger surface area of the BE for more encompassing specimen collection.

“Although in the past we have used brushes, they were more superficial and looked at exfoliative cytology,” Sharma said. “In contrast, the brush we use for the WATS technique is more abrasive, so it collects not only the shed cells, but actually goes deeper to collect glandular tissue as well.”

Sharma was one of the lead investigators of a large multi-center study of WATS that was presented in May at Digestive Disease Week. “We found a 14% increase in the detection of patients with high-grade dysplasia/cancer when WATS was used in conjunction with random four-quadrant biopsies,” he said.

WATS also removes some of the pathologist’s objectivity because all the collected cells are fed into a computer, “which presents the 200 most abnormal cells to the pathologist, who then makes the diagnosis,” Sharma said. “For making a diagnosis of dysplasia, the addition of WATS to the biopsy protocol increases the detection rate of dysplasia.”

Improving surveillance in patients with BE is also possible with electronic chromoendoscopy, which is built into the scope. “This allows you to change the wavelength of light and to magnify into the Barrett’s mucosa, thereby detecting subtle lesions for targeted biopsy,” he said.

For diagnosing BE-associated neoplasia, “we are at a time frame where random biopsies are a thing of the past,” Sharma said. “We are looking at larger sampling, wider sampling, better diagnosis and targeted biopsy.”

Computer Analysis

Erik J. Schoon, MD, PhD, a gastroenterologist and head of endoscopy at Catharina Hospital in Eindhoven, the Netherlands, led a study published in Endoscopy, which assessed diagnosing BE by computer analysis.

Erik J. Schoon

“The results of the paper show that as ‘proof of principle’ and starting off with one of the most difficult subjects in endoscopy, BE, that the computer algorithm developed could almost recognize dysplasia and early cancer from still endoscopic images as accurately as four of the best European experts could,” Schoon told Healio Gastroenterology. “The computer program was also able to delineate the abnormal areas.”

For the study, 100 images from 44 patients with BE were analyzed via a computer algorithm, which employed specific texture, color filters and machine learning to detect early neoplastic lesions. The system identified early neoplastic lesions on a per-image analysis with a sensitivity and specificity of 0.83. At the patient level, the sensitivity and specificity were 0.86 and 0.87, respectively.

Figures 1 and 2. Barrett’s cancer and computer automated detection (CAD). Expert’s delineation on the left. Automatic delineation on the right.

Images: Schoon EJ

“A trade-off between the two performance metrics could be made by varying the percentage of training samples that showed neoplastic tissue,” Schoon said. “However, further research is required to improve the accuracy of the system and prepare it for real-time operation, before it can be applied in clinical practice.”

PAGE BREAK

Schoon said it will take several more years before the program is suitable for clinical practice. “The aim is to make a real-time computer analysis and to validate the system,” he said. The Dutch Cancer Foundation and a National Technical Foundation from the Government are funding part of the further research.

“Computer analysis is appropriate for several image-based procedures in different fields of health care,” Schoon said.

Surveillance Endoscopy

Surveillance endoscopy is promising for patients with BE that is associated with significantly better EAC outcomes, including cancer-related mortality, compared with other non-surveillance endoscopy.

“Given that BE is the only known precursor/predisposing condition for esophageal adenocarcinoma (EAC), and that endoscopic surveillance of BE has the potential of uncovering precancer lesions (dysplasia) which can be effectively cured by endoscopic resection or radiofrequency ablation, higher detection of BE has the potential of reducing the incidence and mortality of esophageal adenocarcinoma,” Hashem B. El-Serag, MD, MPH, professor of medicine and chief of gastroenterology and hepatology at the Michael E. DeBakey Veterans Affairs Medical Center and Baylor College of Medicine in Houston, told Healio Gastroenterology.

Hashem B. El-Serag

“We look for evidence of neoplasia (dysplasia or cancer),” El-Serag said. “The frequency of these endoscopies is every 3 to 5 years among patients with non-dysplasia.”

The recommended careful examination consists of white light endoscopy and some magnification/high-resolution endoscopy (eg, narrow band imaging), coupled with taking biopsies from any suspicious areas (four quadrant samples every 1 cm to 2 cm).

El-Serag and colleagues looked at 29,536 patients with BE in a study published in Gut, 424 patients developed EAC during a mean follow-up period of 5 years. A total of 209 (49.3%) patients with EAC were assigned to a BE surveillance program and were diagnosed as a result of surveillance endoscopy. “These patients were more likely to be diagnosed at an early stage (stage 0 or 1: 74.7% vs. 56.2%; P < .001), survived longer (median 3.2 vs. 2.3 years; P < .001), have lower cancer-related mortality (34% vs. 54%, P < .0001) and were more likely to receive esophagectomy (51.2% vs. 42.3%) than the 215 patients diagnosed by non-BE surveillance endoscopy, 17.2% of whom were BE surveillance failure,” El-Serag said.

BE surveillance endoscopy was also associated with a decreased risk for cancer-related death (HR = 0.47, 95% CI, 0.35-0.64), “which was largely explained by the early stage of EAC at the time of diagnosis,” El-Serag said. Similarly, the adjusted mortality for patients with cancer in a prior surveillance program for overall death was 0.63 (0.47-0.84) compared with patients with cancer not enrolled in a surveillance program.

El-Serag said that most cases of EAC are diagnosed in patients who present with symptoms: difficulty swallowing, chest pain, weight loss, anemia. “In these patients, the BE can be related to the cancer,” he said. “Our study highlighted the value of knowing that among those with recognized (eg, diagnosed) BE before EAC diagnosis, endoscopic surveillance is of benefit.”

A separate study in Gut “reported that the mere knowledge of BE diagnosis before the cancer was a predictive factor of better cancer outcomes,” El-Serag said.

EAC has a “grave prognosis,” with a 5-year survival rate of only 16%, according to David G. Graham, MBBS, BSc, MRCP, a clinical research fellow pursuing his PhD in early detection of EAC at University College London in the United Kingdom. “BE is the only known precursor lesion to EAC. Although the majority of EAC arises in BE, the majority of individuals diagnosed with EAC are unaware of having BE.”

PAGE BREAK
David G. Graham

Graham said studies have demonstrated that existing BE surveillance programs are ineffective. Large sums of resources are devoted to these programs, “even though the vast majority of individuals will not progress,” he said.

To meaningfully impact the prognosis of EAC, “we need to detect those with BE, and more importantly, those with high-risk, dysplastic BE and early EAC who will benefit from the effective, minimally invasive therapies, such as endoscopic resection and ablation,” Graham said. “However, these lesions are often asymptomatic or with symptoms indistinguishable from benign disease. Diagnostic symptoms and signs often only develop once cancer is well established and incurable. Thus, we need to find means to effectively screen an asymptomatic population.”

Graham noted that endoscopic surveillance of the population is invasive, impractical, expensive and likely to be ineffective. “Recent publications of endoscopic image computer analysis software and the WATS3D biopsy analysis, although promising, still involve endoscopy, not applicable to population screening,” he said.

A 2014 article in PLoS One attempted to identify at-risk individuals for BE through analysis of demographic, symptom and risk factor data alone, “but was impractical for usage, due to a low specificity (22%),” Graham said.

Another alternative to endoscopy is the Cytosponge (Medtronic GI Solutions), which is a capsule on a string that when swallowed becomes an abrasive sponge and then is pulled out of the mouth. Promising data were published last year in PLoS Med on this minimally invasive means to identify BE and EAC, but Graham expressed his doubts as to its use.

“We have been closely involved in these studies. Around 50% of patients approached to swallow the sponge on a second occasion declined to do so,” Graham said. “This augurs badly for a long-term surveillance strategy.”

Artificial Intelligence

To overcome some of the short comings of various diagnostic techniques, Graham and his colleagues have developed a tool to identify those with EAC and BE via artificial intelligence (AI) analysis of salivary transcriptomic data, alongside patient demographics and symptom and risk factor data obtained through a questionnaire. “Using AI to analyze ‘big data’ provides the ability to aggregate and mine vast sums of data and develop new insights into disease,” Graham said.

At DDW 2016, Graham presented results of an 80-patient study using analysis of saliva, a cross-section of data from a patient questionnaire and AI analysis. Patients were divided into four groups: 20 normal, 20 non-dysplastic BE (NDBE), 20 high-grade dysplasia (HGD) and 20 EAC. Recruited patients provided a 1-mL saliva sample (collected following fasting) and completed a specifically designed questionnaire (205 data points encompassing demographic, symptom and risk factor data linked to BE and EAC).

Figure 3. Examples of some of the significantly aberrantly expressed genes.

Images: Graham DG

Figure 4. The diagnoses of patients from low to high expression of CDKN2a.

“RNA was extracted from saliva using a QIAzol lysis reagent chloroform process that was quality controlled by the Nanodrop and Bioanalyzer,” Graham said. “Each sample also underwent a TURBO DNase step prior to complementary DNA (cDNA) creation.”

Targeted expression analysis was performed using quantitative real-time PCR. Target-specific primers with known links to BE and EAC were identified through a literature search and reanalysis of a public domain expression dataset. “Overall, 13 genes (AMY1, AMY2, CDKN2a, SMAD7, TLR6, TP53, BRAF, EGFR, KIT9, KRAS, NRAS, PIK3CA and PTEN) were selected with different exons within the genes targeted, creating 22 targets in total,” Graham said.

Expression data was then obtained and relative expression was calculated by comparing the cycle threshold of each amplicon against 18S rRNA (housekeeping gene). A relative expression/cycle threshold was also calculated to compare the four categories of patients. “Finally, a systematic Mann-Whitney U test was carried out between all the groups,” Graham said.

PAGE BREAK

In summary, 11 of the 22 targets were found to be significantly aberrantly expressed at different stages of progression toward EAC. The eight genes with significant expression findings were: AMY2, CDKN2a, SMAD7, TLR6, EGFR, KIT9, PTEN and TP53.

“The questionnaire and expression data were analyzed using novel AI techniques to identify the key attributes for identifying those with EAC and those at risk through BE and HGD,” Graham said. For the tool, six of the selected attributes were aberrantly expressed genes, including AMY2, CDKN2a, SMAD7 and TP53, whereas five attributes were from the questionnaire data, including the waist–hip ratio, proton pump inhibitor use, previous cancer and smoking.

“Using combinations of these attributes, one could identify those with or at risk of EAC with 93% sensitivity and 73% specificity,” Graham said.

Looking toward the future, Graham believes his group’s work will benefit from pending whole transcriptome analysis, performed on selected samples, which will identify further genes of importance requiring validation through targeted expression analysis. “Additionally, we have performed epigenetic analysis on patient samples, with preliminary data suggesting that this significantly increases the tool’s accuracy and may be able to eliminate the need for the patient questionnaire,” he said.

The advances in molecular biology, particularly the ability to effectively analyze DNA and RNA “of poorer quantity and quality than ever before, allows the approach we have taken — and the promising field of ‘salivary diagnostics’ and other forms of ‘liquid biopsies’ — to potentially provide a noninvasive means to not only detect almost any disease, but also to longitudinally monitor an individual, including identification of potential relapses or recurrences,” Graham said. “The potential impact this will have on health care is immeasurable.”

Schoon said that genetic analysis could be a predictor of acquiring dysplasia in BE, “so select the high-risk patients. But BE patients still have to be endoscopically screened for dysplasia.” – by Bob Kronemyer

Disclosures: El-Serag, Graham and Schoon report no relevant financial disclosures. Sharma reports grant support from Olympus Inc., CDx Diagnostics and Cook Medical.