Cell polarity (the ‘four lines’) distinguishes gastric dysplasia from epithelial changes in reactive gastropathy

Abstract

Aims:

Gastric dysplasia is a risk factor for synchronous and subsequent gastric carcinoma. Distinguishing gastric dysplasia from reactive changes is subject to interobserver disagreement and is a frequent reason for expert consultation. We previously used assessment of surface cell polarity (the ‘four lines’) as a key feature to decrease equivocal diagnoses in Barrett oesophagus. In the current study, we examined for the presence or absence of the four lines in gastric dysplasia and reactive gastropathy.

Materials and methods:

The study includes all (n = 91) in-house biopsies with at least gastric dysplasia from the surgical pathology archives of two academic institutions during a 5-year period from 2008 to 2012. A reactive gastropathy group (n = 60) was created for comparison.

Results:

The dysplasia/neoplasia group was comprised of 14 biopsies of gastric foveolar-type dysplasia, 59 of intestinal-type dysplasia, 14 with dysplasia in fundic gland polyps, three pyloric gland adenomas and one oxyntic gland adenoma. Loss of surface cell polarity was seen in all 88 dysplasia cases with evaluable surface epithelium. All 57 reactive gastropathy cases with evaluable surface epithelium showed intact surface cell polarity except in focal areas directly adjacent to erosions in 17 cases, where the thin wisp of residual surface mucin could not be appreciated on haematoxylin and eosin.

Conclusion:

Surface cell polarity (the four lines) was lost in all gastric dysplasia biopsies with evaluable surface epithelium and maintained in all biopsies of reactive gastropathy. Caution should be taken in using this feature adjacent to erosions in reactive gastropathy.

Introduction

Worldwide, there were more than 1 million incident gastric cancer cases and nearly 800 000 deaths in 2018, making it the fifth most common cancer site and the third most deadly.¹ Gastric cancer has a higher incidence in eastern Asia, eastern Europe and South America and relatively lower incidence in Africa, northern Europe and North America.¹

Correctly identifying precursor lesions and developing screening, monitoring and treatment protocols has the potential to greatly impact the incidence, morbidity and mortality associated with gastric cancer. Although it is hypothesised that gastric cancer commonly follows a sequence from intestinal metaplasia to dysplasia to carcinoma,^2,3 guidelines for surveillance of patients with gastric intestinal metaplasia in low-risk populations are not well established.⁴ Dysplasia has been associated with an increased risk for synchronous and subsequent gastric carcinoma.^5–11 Gastric dysplasia can be categorised based on grade and into intestinal and gastric foveolar subtypes.^12–14 Additionally, dysplasia can be seen in distinct types of polyps such as hyperplastic polyps (HP), fundic gland polyps (FGP), pyloric gland adenomas (PGA) and oxyntic gland adenomas (OGA).^15–19 European colleagues have embraced interpreting gastritis-associated polyp-forming dysplasia as ‘lesions’ rather than adenomas in management guidelines.²⁰

Diagnosing gastric dysplasia is challenging with significant interobserver disagreement.²¹ Distinguishing dysplasia from reactive/chemical gastropathy can be particularly problematic, as the epithelial changes associated with reactive/chemical gastropathy frequently show moderate nuclear atypia and depleted surface foveolar mucin. The distinction between reactive changes and dysplasia is also somewhat challenging, with significant interobserver variability in the setting of Barrett’s oesophagus.^22,23 We previously showed that using the maintenance of surface cell polarity (the ‘four lines’) as a feature of non-dysplastic epithelium has utility in assessing for dysplasia and minimising equivocal diagnoses in Barrett’s oesophagus.²⁴ We defined the four lines as: the gastric-type mucin vacuole (line 1); the base of the mucin vacuole (line 2); the cytoplasm (line 3); and the nuclei (line 4). The current study consisted of a comprehensive review of all gastric dysplasia cases at two tertiary hospitals to determine if surface cell polarity is also lost in gastric dysplasia.

Materials and methods

STUDY POPULATION

The institutional review board (IRB) approved study included all in-house biopsies with at least gastric dysplasia (n = 91) from the surgical pathology archives of both Cedars–Sinai Medical Center (n = 75) and Johns Hopkins Hospital (n = 16) collected during a 5 year-span from 2008 to 2012. A comparison group consisted of 60 patients with gastric biopsies showing chemical/reactive gastropathy collected in 2008.

STUDY VARIABLES

For each biopsy set, we collected basic demographic variables, including age at the time of collection of the index biopsy, sex and race/ethnicity. Each case of dysplasia was categorised based on the biopsy with the highest grade of dysplasia. Dysplasia was categorised as: negative for dysplasia (NFD), low-grade dysplasia (LGD), high-grade dysplasia (HGD) or at least intramucosal carcinoma (≥IMC). We characterised the cases based on the type of dysplasia: gastric foveolar-type (n = 14); intestinal-type (n = 59); dysplasia in an FGP (n = 14); dysplasia in a PGA (n = 3); and dysplasia in an OGA (n = 1). We collected information on the location of the biopsy, whether the dysplasia arose in flat mucosa or formed a visible lesion, and whether there was a history of Helicobacter pylori or autoimmune gastritis. Each case from both the dysplastic and chemical/reactive groups was evaluated by two gastrointestinal pathologists (K.M.W. and E.A.M.) to confirm the category of dysplasia and to assess for the presence or absence of the four lines. One case was recategorised from LGD to reactive gastropathy.

STATISTICAL ANALYSIS

Statistical testing was performed using the r statistical programming language (R Foundation, Vienna, Austria); t tests were used to compare mean values between two groups and χ² tests were used to compare differences in categorical variables. Tests of proportion were used to test for differences in proportions.

Results

DEMOGRAPHIC AND CLINICAL VARIABLES OF PATIENTS WITH DYSPLASIA AND REACTIVE GASTROPATHY

Patients in the dysplasia group were slightly older (mean age = 66.9 years) than those in the reactive gastropathy group (mean age = 62.6 years; P = 0.11; Table 1). The dysplasia patients were more likely to be Asian (19% versus 3%) and Hispanic (7% versus 0%) rather than Caucasian (55% versus 75%) compared with the reactive gastropathy group (P < 0.01). Dysplasia was encountered in both the antrum (47%) and gastric body (43%), while reactive gastropathy was typically antral (90%; P < 0.01). Dysplasia often manifested as a polyp or mass endoscopically (74%), while this finding was rare in reactive gastropathy (2%; P < 0.01).

Table 1.

Demographic variables and biopsy characteristics

	Dysplasia (n = 91)	Reactive gastropathy (n = 60)	P-value
Age, mean (range)	66.9 (17–99)	62.6 (15–86)	0.11
Female, n (%)	52 (57%)	37 (62%)	0.70
Race/ethnicity, n (%)
Asian	17 (19%)	2 (3%)	<0.01
African American	15 (16%)	9 (15%)
Hispanic	6 (7%)	0 (0%)
Caucasian	50 (55%)	45 (75%)
Unknown	3 (3%)	4 (7%)
Location of biopsy
Antrum	43* (47%)	54 (90%)†	<0.01
Body/fundus	39* (43%)	7 (12%)†
Cardia/proximal	5 (5%)	0 (0%)
Unknown	5 (5%)	2 (3%)
Polypoid, n (%)	67 (74%)	1 (2%)	<0.01
Helicobacter pylori
Negative	76 (84%)	60 (100%)	<0.01
Current	11 (12%)	0 (0%)
Treated	4 (4%)	0 (0%)
AMAG, n (%)	13 (14%)	1 (2%)	0.019
Loss of surface cytological architecture (‘four lines’)	88/88 (100%)‡	0/57 (0%)§	<0.01

AMAG, Autoimmune metaplastic atrophic gastritis.

^*One patient had two separate dysplastic lesions in the antrum and body/fundus.

^†Three patients had biopsy fragments with reactive changes in both the body and antrum.

^‡Three cases did not have surface dysplasia for evaluation.

^§Three cases did not have evaluable surface; 17 cases (27%) had focal loss at an area of erosion.

ASSESSMENT OF SURFACE CYTOLOGIC ARCHITECTURE: THE FOUR LINES

We found that the four lines (Figure 1A,B) were no longer intact in all 88 cases of dysplasia with an evaluable surface component (Table 1), irrespective of the type of dysplasia (Figure 1C–F). Gastric foveolar-type dysplasia typically showed jumbled nuclei with apical mucin, but no consistent cytoplasm space between the nuclei and mucin. High-grade lesions showed surface loss of nuclear polarity and glandular crowding with cribriform architecture. Intestinal-type dysplasia showed loss of the apical mucin cap and an abrupt transition from the non-dysplastic epithelium. In low-grade lesions, the nuclei were elongated and largely confined to the base of the cells. In high-grade lesions, the nuclei were enlarged and rounded with loss of nuclear polarity. The underlying glands were crowded and showed cribriform architecture. Like gastric-foveolar dysplasia, dysplasia in FGPs was characterised by a disorderly line of nuclei without a consistent cytoplasmic space between the nuclei and apical mucin. PGAs showed two lines with apical cytoplasm at the surface overlying a line of relatively bland nuclei. High-grade lesions showed architectural complexity and loss of nuclear polarity. The single OGA did not extend to the surface.

Figure 1.

A,B, Normal gastric mucosa showing intact surface cell polarity. The four lines are: line 1, apical mucin cap; line 2, base of the mucin cap; line 3; cytoplasm; line 4, nucleus. C, Gastric foveolar-type dysplasia shows no constant cytoplasmic space between line of apical mucin and jumbled nuclei. D, Intestinal-type dysplasia shows loss of four lines with lack of apical mucin cap. E, Dysplasia in a fundic gland polyp (FGP) shows similar loss of constant cytoplasmic space between the line of apical mucin cap and jumbled nuclei. F, Pyloric gland adenoma (PGA) shows modified two lines with apical cytoplasm over relatively monotonous nuclei at the cell base. G, Reactive gastropathy shows depleted but intact apical mucin cap, base of the mucin cap, cytoplasm, and nuclei. H, Erosion in reactive gastropathy shows that the four lines are focally imperceptible adjacent to the erosion. I, Periodic acid–Schiff Alcian Blue (PAS-AB) special stain highlights thin wisp of apical mucin demonstrating that the four lines are actually intact in this area.

Cases with marked reactive gastropathy showed decreased apical mucin and mild to moderate nuclear atypia, but the surface cell polarity (the four lines) was intact in all 57 cases with evaluable surface (Figure 1G–I). In 17 cases, the apical mucin cap was not identifiable by haematoxylin and eosin (H&E) in the area immediately adjacent to the fibrinous exudate of an erosion, but the four lines were present in the remainder of the biopsy. Periodic acid–Schiff Alcian Blue (PAS-AB) special stains, which were available in a small subset of biopsies, were helpful in this circumstance, highlighting a thin wisp of residual apical mucin demonstrating the presence of the lines (Figure 1I).

Discussion

We found that the loss of surface cell polarity (the four lines) also separates dysplasia from reactive changes in the stomach; however, its application is not as straightforward as it is in the columnar-lined epithelium of the oesophagus. Particular attention to the background stomach is needed. In gastric foveolar-type dysplasia, there is no longer a consistent line of cytoplasm separating the apical mucin from the disorderly line of nuclei. Dysplasia in a fundic gland polyp shows essentially the same pattern. Intestinal-type dysplasia shows loss of the apical mucin cap so that there is a line of cytoplasm and a line of elongated nuclei. PGAs show a line of cytoplasm and a line of round nuclei with little nuclear atypia. We only had a single case of OGA, which lacked surface involvement. In one series 30% (three of 10) OGAs involved the surface.¹⁹ Determining whether the four lines is applicable to OGAs with surface involvement will require further study. Cases with HGD showed further loss of the four lines, as loss of nuclear polarity was a common feature along with cribriform architecture. Additionally, an abrupt transition from non-dysplastic to dysplastic epithelium was a common feature in the dysplastic cases. As we previously noted in the oesophagus,²⁴ the four lines evaluating surface cell polarity cannot be used when the dysplasia does not involve the surface. Isolated ‘gastric pit/gland dysplasia’ has been described adjacent to gastric adenocarcinomas and may represent a precursor lesion.^25,26 If additional levels do not show surface involvement the pathologist must rely upon other features of dysplasia, such as nuclear atypia, compared to the non-dysplastic epithelium, glandular crowding and architecture and increased mitoses and apoptosis, while adjusting for any inflammation that may account for these changes.

Correctly identifying reactive gastropathy is important in reducing unnecessary clinical follow-up and procedures that may follow an overdiagnosis of dysplasia.^27,28 While noting that the four lines is helpful in confirming a diagnosis of reactive/chemical gastropathy, one factor which may complicate assessment is the presence of erosions. In 17 cases, we identified focal areas directly adjacent to the fibrinous exudate of an erosion where there was apparent loss of surface mucin. However, PAS-AB special stains showed that there was a thin wisp of apical mucin, indicating that the four lines were intact in these areas, but the surface mucin was simply imperceptible on H&E. For this reason, we recommend that pathologists should avoid using the four lines to diagnose focal dysplasia adjacent to an erosion in the setting of reactive gastropathy.

We found that surface cell polarity (the four lines) is a useful feature in distinguishing gastric dysplasia from reactive changes when applied correctly. The four lines are not applicable to dysplasia that does not reach the surface, and should be used with caution adjacent to erosions in the setting of reactive gastropathy.