Barretts esophagus (BE) Dysplasia in Barretts Esophagus Distal - - PowerPoint PPT Presentation

5/26/2018 1 Dysplasia in Barrett’s Esophagus and Inflammatory Bowel Disease: Diagnosis, Biomarkers, and Management

Won-Tak Choi, MD, PhD Assistant Professor Department of Pathology University of California San Francisco San Francisco, California

Barrett’s esophagus (BE)

BE is a complication of chronic GERD and a major risk factor for development of esophageal adenocarcinoma (EAC). The American College of Gastroenterology defines BE when there is extension of salmon-colored mucosa into the tubular esophagus extending ≥ 1 cm proximal to the GEJ with biopsy confirmation of intestinal metaplasia (IM).

Distal esophagus Proximal part of stomach

• Most EACs develop in the setting of BE through

chronic GERD-IM-dysplasia-carcinoma sequence.

• Historically, the grade of dysplasia on an initial

biopsy has determined endoscopic surveillance intervals.

• Risk of cancer progression per year:
• BE without dysplasia = 0.2-0.5%
• IND = 0.4-1.4%
• LGD = 0.4-13.4%
• HGD = 7-19%
• The variability in progression rates reflects

interobserver variability and presence of heterogeneous biological subsets within each morphologic dysplasia diagnosis category. Interobserver reproducibility was substantial at the ends of the spectrum (BE without dysplasia and HGD) but fair for LGD and slight for IND:

• BE = Moderate to substantial (k = 0.58)
• HGD = Substantial (k = 0.65)
• LGD = Fair (k = 0.32)
• IND = Slight (k = 0.15)

5/26/2018 2

Following histological review by two expert GI pathologists, 73-85% of LGD cases diagnosed in the community were downgraded to non- dysplastic BE or IND. For confirmed LGD, the risk of HGD/EAC was 9- 13% per patient-year, whereas patients downgraded to non-dysplastic BE or IND had a progression risk of 0.5-0.9%.

• Biopsies from 485 patients carrying a local

institutional diagnosis of HGD were reviewed by three GI pathologists.

• Only 51% were confirmed to have HGD.
• GI pathologists achieved a high-level agreement

(K = 0.77) for HGD.

Lam-Himlin D. et al. Mod Path. 2018. Mar;31(suppl. 2): 278.

• Among 1203 consecutive consultative cases

collected at 6 academic centers during 2016-17, upper GI cases accounted for 29%.

• Esophagus was the most common site (56%),

comprised mostly of BE-related consults (84%).

• Discordance with submitted diagnosis was 40%,

and downgrading of dysplasia was more common (67%).

Due to significant interobserver variability in the interpretation of dysplasia, review by two pathologists, at least one of whom has specialized expertise in GI pathology, is warranted for dysplasia of any grade.

5/26/2018 3

• The interpretation of staining results (such as p53

and AMACR) is highly variable and subjective.

• Non-neoplastic epithelium (~10-15%) frequently

shows positive staining (often strong and diffuse), especially when mucosa is inflamed or ulcerated.

• The diagnostic utility of genetic and chromosomal

abnormalities (including 9p loss of heterozygosity (LOH), 17p LOH, and mutations of p53 and cyclin- dependent kinase N2 (tumor suppressor genes)) is also limited, as these changes tend to occur early and frequently in BE even without dysplasia. Are Ancillary Studies Needed for Dysplasia Diagnosis or Risk Stratification in BE Patients? Challenges in interpreting p53 stain: (A, B) Two nondysplastic examples of BE showing scattered intensely staining nuclei. (C) Diffuse and intense staining.

Srivastava A. et al. AJSP. 2017. 41(5): e8-e21 Panarelli NC. et al. AJSP. 2016. 40(8): e83-e93

This biopsy sample obtained from an area adjacent to an ulcer shows mild nuclear enlargement with mucin depletion and infiltration by neutrophils that, at most, may be considered IND. However, the cells show strong, diffuse p53 staining.

HGD with strong and diffuse p53 staining (overexpression) correlating with TP53 mutations HGD with complete absence of p53 staining (null pattern, 20% of the time) also correlating with TP53 mutations

Srivastava A. et al AJSP. 2017. 41(5): e8-e21

5/26/2018 4

Srivastava A. et al. AJSP. 2017. 41(5): e8-e21

Scoring p53 IHC is subject to interobserver variability

46% 53% 64% 57% 3% 10% 27% 14%

A diagnosis of dysplasia remains a morphologic diagnosis. Existing data are insufficient to recommend ancillary stains as a prognostic marker.

5/26/2018 5

“DNA content abnormalities (using DNA flow cytometry)…defined patients at increased risk for progression to cancer…at the present time, no biomarkers or panels of biomarkers are ready for clinical practice.”

• Not widely available to pathologists and

clinicians, because it typically needs fresh tissue which requires special handling procedures and does not allow direct histology-flow cytometry correlation.

• Requires considerable technical and professional

expertise.

Outdated assumptions regarding DNA flow cytometry

All BE cases without dysplasia showed normal DNA content (no aneuploidy)

Diploid (2N) Tetraploid (4N) < 6%

The majority of HGD cases (95%) showed aneuploid population(s) (red and blue) that are visually distinguishable from the normal diploid population (green)

5/26/2018 6

A subset of LGD cases (21%) showed a distinct aneuploid population (red)

14% within 1 year 31% within 5 years HR = 7 85% within 1 year (p < 0.001) 100% within 5 years (p < 0.001)

A subset of IND cases (10%) showed a distinct aneuploid population (red)

100% within 2 years (p < 0.001) 6% within 13 years

HR = 20

5/26/2018 7

For LGD, continued surveillance every 12 months is an acceptable approach.

(in 3-6 months)

Endoscopically visible nodularity Endoscopic mucosal resection Shafa S. et al. AJG. 2017. 112(10): 1487-1490

After endoscopic eradication therapy, the baseline degree of dysplasia determines appropriate endoscopic surveillance intervals

Inflammatory Bowel Disease

• Dysplasia is the best marker of colorectal

cancer (CRC) risk in IBD patients, and surveillance colonoscopy is recommended to detect dysplasia or early CRC.

• Most dysplasia (~90%) is visible with newer

endoscopic technologies.

• Targeted biopsies detect similar proportions of

dysplasia as random biopsies.

• Such a paradigm shift may have important

implications for the surveillance and management of dysplasia.

• Endoscopically resectable polypoid dysplasia

(even for HGD) has a good prognosis (thus no colectomy required).

• Some other recent guidelines have suggested

colectomy for nonpolypoid dysplastic lesions (including flat and ‘invisible’ lesions), because they are usually not endoscopically resectable.

5/26/2018 8

(including “invisible”) (including “invisible”)

GI pathologist should review all biopsy specimens diagnosed initially as IND, LGD, or HGD. Progression rate: Flat HGD = 25-32% Flat LGD = 0-53% Synchronous CRC: Flat HGD = up to 67% Flat LGD = up to 27%

• Negative = fair to good (k = 0.4-0.5)
• HGD = fair to good (k = 0.4-0.5)
• LGD = poor (k = 0.2-0.3)
• IND = poor (k = 0.01-0.2)
• Overall = poor to fair (k = 0.3-0.4)
• The interpretation of staining results (such as

p53) is highly variable and subjective (depending on the location, intensity, extent, and/or number of positive cells; often using different cutoff values for positive immunoreactivity).

• Strong and diffuse staining (up to 10%) can be

seen in non-neoplastic mucosa, especially in areas of active inflammation (up to 100%).

Are Ancillary Stains Needed for Dysplasia Diagnosis or Risk Stratification in IBD Patients?

Strong nuclear p53 staining Weak nuclear p53 staining Strong cytoplasmic staining

• Correlated p53 expression of 44 IND cases with outcomes.
• p53 expression was determined as a percentage of epithelial cells

within a HPF showing strong nuclear, weak nuclear, or strong cytoplasmic staining.

• Composite p53 score = sum of (i) the percentage of epithelial cells

with weak nuclear staining, (ii) three times the percentage of epithelial cells with strong nuclear staining, and (iii) the percentage of epithelial cells with strong cytoplasmic staining.

5/26/2018 9

(Basal 1/3) (Middle 1/3) (Top 1/3)

Strong intensity p53 staining suggests dysplasia, and restriction of p53 staining to the basal 2/3 of the crypt appears to exclude HGD.

(9%) (100%)

Strong p53 staining?

(43%) (57%) (92%)

“Aneuploidy has been the most thoroughly investigated...However, evaluation of aneuploidy by flow cytometry requires considerable technical and professional expertise, which limits its use in routine practice.”

X

5/26/2018 10

Almost all flat HGD cases (93%) showed a distinct aneuploid population (red) A significant number of flat LGD cases (41%) showed a distinct aneuploid population (red)

57% within 1 year (p < 0.001) 100% within 12 years (p = 0.001) 5% within 1 year 33% within 12 years

HR = 5

The majority of IND cases (64%) showed normal DNA content

Wen KW and Choi WT. Under Review.

5/26/2018 11

A subset of IND cases (37%) showed a distinct aneuploid population (red)

5% within 7 years

HR = 13

12% within 1 year (p = 0.136) 34% within 3 years (p = 0.029) 59% within 7 years (p = < 0.001)

• A diagnosis of dysplasia in BE and IBD

remains a morphologic diagnosis at this time.

• Ancillary stains (such as p53) are not

currently recommended as a diagnostic

• r prognostic marker.
• DNA flow cytometry is a promising tool

that can provide supportive evidence to a morphologic impression or suspicion of HGD and risk stratify LGD or IND.

Conclusions

5/26/2018 12

Thank you! won-tak.choi@ucsf.edu