Dig Dis Sci DOI 10.1007/s10620-014-3249-5
ORIGINAL ARTICLE
Endoscopic Resection as a First Therapy for Gastric Epithelial Atypia: Is It Reasonable? Chung Hoon Yu • Seong Woo Jeon • Sung Kook Kim Hyun Seok Lee • Jun Heo • Yong Hwan Kwon • Gyu Young Kim • Sun Zoo Kim • Han Ik Bae
•
Received: 27 March 2014 / Accepted: 4 June 2014 Ó Springer Science+Business Media New York 2014
Abstract Background and Aims Gastric atypical cell (GAC), an indefinite pathologic finding, often requires repeated biopsy or other diagnostic treatments, such as endoscopic mucosal resection (EMR), endoscopic submucosal dissection (ESD), or operation (OP). The aim of this study was to analyze the initial endoscopic and histologic findings of GAC and to discuss the necessity of EMR/ESD at establishing a correct diagnosis. Methods This retrospective study enrolled 96 patients proven as GAC on index forceps biopsy. ESD (17/96, 17.7 %), EMR (5/96, 5.2 %), OP (20/96, 20.8 %), and other treatment or follow-up (54/96, 56.3 %) were performed. We analyzed the initial endoscopic and histologic characteristics of GAC lesions, predictive of neoplasm. Results After diagnostic modalities, the final pathologic diagnoses were cancer (36/96, 37.6 %), dysplasia (9/96, 9.4 %), and non-neoplasm (51/96, 53.0 %). In univariate analysis, age [odds ratio (OR) 1.04, 95 % confidence interval (CI) 1.01–1.07], lesion size of 10 mm or greater (OR 3.94, 95 % CI 1.61–9.61), lesion with depressed type (OR 2.50, 95 % CI 1.09–5.72), and presence of H. pylori C. H. Yu S. W. Jeon S. K. Kim H. S. Lee J. Heo Y. H. Kwon G. Y. Kim Department of Internal Medicine, Kyungpook National University School of Medicine, Daegu, Republic of Korea S. W. Jeon (&) Gastric Cancer Center, Kyungpook National University Medical Center, 807 Hogukno, Buk-gu, Daegu 702-210, Republic of Korea e-mail: [email protected] S. Z. Kim H. I. Bae Department of Pathology, Kyungpook National University School of Medicine, Daegu, Republic of Korea
(OR 2.83, 95 % CI 1.11–7.25) were risk factors for neoplasm. In multivariate analysis, lesion size of 10 mm or greater (OR 3.63, 95 % CI 1.23–10.66), lesion with depressed type (OR 2.86, 95 % CI 1.11–7.38) were independent risk factors for cancer. Conclusion Considering the neoplastic risk of GAC, which could be missed on biopsy, more comprehensive tissue sampling via EMR/ESD might be necessary to establish a definite diagnosis. Keywords Gastric neoplasms Atypical epithelial cells Endoscopic surgical procedure Indefinite pathology
Introduction Gastric atypical cell (GAC) is classified as ‘‘indefinite for neoplasm/dysplasia’’ (Category 2) in the revised Vienna classification of gastrointestinal epithelial neoplasia [1]. This diagnosis can be used in a limited number of settings, such as when there is genuine difficulty in discerning reactive epithelial changes from dysplasia, when worrisome atypia is noted but is confined to a very small area, or when the tissue is limited in amount or quality. Due to the biological potential of gastric dysplasia, it is considered a precursor of gastric cancer, the most prevalent cancer in Eastern Asia, including Korea [2, 3]. Follow-up is absolutely necessary for GAC, and re-biopsy is commonly recommended. However, some studies have suggested that forceps biopsy is insufficient to confirm the real nature of the lesion, because it is not entirely representative of the gastric lesion. In addition, despite having an endoscopic appearance, like that of a malignant lesion, pathologic finding of forceps biopsy can be indefinite [4, 5]. Eventually, further diagnostic modalities are necessary for GAC in
123
Dig Dis Sci
order to confirm diagnoses and for treatment of premalignant or malignant lesions. Before the advent of endoscopic treatment modalities, patients underwent open abdominal surgery in cases where stomach cancer was strongly suspected, including early gastric cancer (EGC). Currently, however, endoscopic mucosal resection (EMR) or endoscopic submucosal dissection (ESD) is performed more commonly for the en bloc resection of a premalignant or malignant lesions, where possible. In addition, these procedures of en bloc resection can be useful diagnostic tools for cases with indeterminate pathology as atypical change due to more accurate pathologic information than forceps biopsy [5–7]. However, there is no sufficient evidence to support EMR or ESD as a first therapeutic modality in these situations. The aim of this study was to discuss the utility of advanced diagnostic modalities including EMR/ESD and to analyze the initial endoscopic and histologic risk factors predictive of neoplasm for cases of GAC.
Methods Patients This retrospective study included 110 patients with GAC that was histologically proven and based on initial endoscopic forceps biopsy between January 1999 and December 2012 at Kyungpook National University Hospital in Korea. Detailed information on the baseline characteristics of the patients, the endoscopic characteristics of the lesions, the pathologic diagnoses, and the clinical outcomes were obtained from the patients’ medical records. GAC on endoscopic forceps biopsy was defined as atypical change interpreted as questionable dysplasia. Cases of GAC with reactive or regenerative gastric epithelium [number (n = 10)] were excluded because reactivity is resolved when infection or inflammation shows improvement [8]. In addition, those lost to follow-up after the first forceps biopsy (n = 4) were excluded. This study was approved by our institutional review board. Endoscopic and Histopathologic Evaluations Diagnostic endoscopy was performed using a conventional gastroscope (Olympus, Tokyo, Japan). The initial endoscopic diagnosis was supervised by one of three expert endoscopists (SK Kim, HS Lee, and SW Jeon) of our center, and the size of lesion was measured with an open biopsy forceps (each arm of bite: 3 mm) as the reference device during endoscopic procedure. Endoscopic films and
123
reports were reviewed by a single expert endoscopist (SW Jeon) for size (mm), location (fundus, body, angle, and antrum), and type (elevated, flat or depressed). Two gastrointestinal pathology specialists (SZ Kim and HI Bae) were involved in the histologic diagnoses. In cases of discrepancy between the two pathologists, final diagnoses were made after discussion. Histologic diagnoses were based on the World Health Organization (WHO) classification of pathology and genetics of tumors of the digestive system [9]. In particular, atypical change interpreted as questionable dysplasia (Grade 2) was the basis of the definition of the Korean pathologic grading system for gastric epithelial proliferative disease [8]. This group included lesions that were difficult to discern as reactive or regenerative epithelium from dysplasia in terms of cellular and structural atypia (Fig. 1a, b). Gastric dysplasia was characterized by a mild architectural complexity, with glands lined by crowded, elongated cells with hyperchromatic and basilar pseudostratified nuclei. In contrast to dysplasia, regenerative or reactive atypia, in which the nuclei may be slightly larger than in normal epithelium, was characterized by the same degree of architectural atypia, with glands respecting their relative spaces without significant crowding or pseudostratification. Carcinoma showed a higher nuclear grade and either a prominent back-to-back or a syncytial growth pattern, abortive microglands, or small clusters of epithelial cells within the lamina propria between glands (Fig. 1c, d). The combined findings of dysplasia and carcinoma were classified as carcinoma. In this study, ‘‘neoplasm’’ was composed of cancer and dysplasia (including adenoma). An EGC (EGC) is defined as invasive gastric cancer involving the submucosa, irrespective of lymph node or distant metastasis (T1, any N, M) and advanced gastric cancer (AGC) as gastric cancer that invades into muscularis propria (T2 and above, any N, M) [9]. This study evaluated the correlation between the clinical, endoscopic diagnosis, and the final histologic diagnosis. A concordant diagnosis meant the initial endoscopic and final histologic diagnoses were the same. A concordant diagnosis of ‘‘neoplasm’’ was defined if the final histologic diagnosis showed within the range of neoplasm finding regardless of an upgraded or downgraded diagnosis, when the endoscopic diagnosis was ‘‘dysplasia/adenoma or cancer.’’ H. pylori infection was diagnosed by either histology or rapid urease test (RUT, CLO test, GASTREX, Warsaw, Poland). The specimens for RUT were obtained from both the antrum and the distal body of the stomach during EGD. Cresyl violet staining was used for the pathologic diagnosis of H. pylori infection. H. pylori infection was diagnosed when the result of one of the two methods described above was positive.
Dig Dis Sci
Fig. 1 Histologic findings of the initial endoscopic forceps biopsy (a, b) and the endoscopic submucosal dissection (c, d) from a patient with a depressed lesion. a Histologic section shows a proliferative glandular lesion, composed of atypical glands with mild architectural distortion and slightly inflamed intervening stroma (hematoxylin and eosin (H & E) stain, 9100). b The proliferative glands are lined by epithelium with large hyperchromatic nuclei, a high nucleus to
cytoplasm ratio, and loss of cellular polarity. There is some inflammation in the glandular epithelium and the surrounding stroma; however, it is insufficient to account for the epithelial abnormalities. The nuclei, while hyperchromatic and polymorphic, remain in a predominantly basal position, and maturation toward the surface is observed (H & E stain, 9400). c, d Slides show well-differentiated adenocarcinoma (H & E, 9100 and 9400)
Data and Statistical Analyses
biopsy, or operation as part of their diagnostic/therapeutic management. Then, some patients underwent additional procedures, if needed, for definite diagnoses or treatments. After complete work-up, the final pathologic diagnoses were cancer (n = 36, 37.6 %), dysplasia (n = 9, 9.4 %), and non-neoplasm (n = 51, 53.0 %). Note that three cases were persistently diagnosed as GAC on re-biopsy. All of these patients (n = 3) underwent a second re-biopsy consistent with regenerative atypia (n = 2) and gastritis (n = 1). Only on the third re-biopsy, done because of characteristics suspicious for malignancy, was a diagnosis of adenocarcinoma made in all cases.
Descriptive data are presented as the number of patients (%) or as mean ± standard deviation (SD). Continuous data were compared using independent t tests, and categorical variables were tested using the chi-square test. The results of univariate and multivariate logistic regression analyses were presented as odds ratio [OR; 95 % confidence intervals (CI)] for risk factors of neoplasm. Positive predictive value (PPV), negative predictive value (NPV), sensitivity, and specificity were evaluated for significant and independent risk factors in multivariate analysis. The criterion for statistical significance was p \ 0.05. Analyses were performed using SPSS (version 18.0; SPSS, Chicago, IL, USA).
Result Final Diagnoses of Patients Figure 2 shows a flow diagram of the study. Initially, the 96 enrolled patients underwent EMR/ESD (Fig. 3), re-
Endoscopic and Histologic Risk Factors for Gastric Neoplasm The number of patients with neoplasm was 45 (46.9 %). Patients in the neoplasm group (mean age 63 ± 11.1 years) were significantly older than those in the non-neoplasm group (56 ± 15.5 years). In endoscopic findings, the mean size of the neoplastic and non-neoplastic lesions was 18.87 ± 8.4 mm (range 5–42) and
123
Dig Dis Sci Fig. 2 Flow diagram of 96 GAC lesions diagnosed by forceps biopsy. n number, ESD endoscopic submucosal dissection, EMR endoscopic mucosal ressection, Bx biopsy, OP operation, EGC early gastric cancer, AGC advanced gastric cancer, F/U follow-up, Other other management, Small cell small cell carcinoma, Adenoca adenocarcinoma, Non-Tx nontreatment
Fig. 3 A case of the endoscopic diagnosis as ‘‘rule out early gastric cancer,’’ confirmed as adenocarcinoma after endoscopic submucosal dissection (ESD) in gastric lesion with atypical change interpreted as questionable dysplasia on forceps biopsy specimens. Endoscopic
123
imaging viewed (a) a 1.5-cm hyperemic mucosal lesion with depression on the posterior wall side of the low body, b marking dots for ESD, c the lesion completely dissected after procedure, and d the ESD specimen with a central lesion
Dig Dis Sci Table 1 Endoscopic and histologic risk factors of gastric neoplasm Variable [n (%)] Age [year (±SD)]
Non-neoplasm (n = 51)
Neoplasm (n = 45)
56 (15.56)
63 (11.10)
Male
30 (58.8)
31 (68.9)
Female
21 (41.2)
14 (31.1)
Sex
p value 0.020 0.308
Endoscopic findings Mean Size [mm(±SD)]
10.84 (7.2)
18.87 (8.4)
Size of long diameter
\0.001 0.030
\10 mm
27 (52.9)
10 (22.2)
C10 mm
24 (47.1)
35 (77.8)
1 (2.0)
2 (4.4)
Body
20 (39.2)
17 (37.8)
Angle
4 (7.8)
3 (6.7)
26 (51.0)
23 (51.1)
Elevated
19 (37.3)
13 (28.9)
0.387
Flat
15 (29.4)
8 (17.8)
0.186
17 (33.3)
24 (53.3)
0.030
Location Fundus
Antrum
0.918
Lesion type
Depressed Histologic findings Intestinal metaplasia
0.399
Presence
5 (9.8)
7 (15.5)
Absence
46 (90.2)
38 (84.5)
H. pylori
0.030
Presence
9 (17.6)
17 (37.8)
Absence
42 (82.4)
28 (62.2)
3.7 (1.4)
3.9 (1.8)
Number of biopsy fragments (±SD)a
0.546
n number, SD standard deviation a
Number of biopsy fragments per lesion in one endoscopic session
10.84 ± 7.2 mm (range 3–36). The neoplasm group (77.8 %) had larger lesions of 10 mm or greater compared with the non-neoplasm group (47.1 %). Depressed type was more frequent in the neoplasm group (53.3 %) than in the non-neoplasm group (33.3 %). In histologic findings, presence of H. pylori was more frequent in the neoplasm group (37.8 %) than in the non-neoplasm group (17.6 %) (Table 1). In univariate analysis, age (p = 0.020, OR 1.04, 95 % CI 1.01–1.07), lesion size of 10 mm or greater (p = 0.030, OR 3.94, 95 % CI 1.61–9.61), lesion with depressed type (p = 0.030, OR 2.50, 95 % CI 1.09–5.72), and presence of H. pylori (p = 0.030, OR 2.83, 95 % CI 1.11–7.25) were risk factors for neoplasm. However, in multivariate analysis, none of them was an independent risk factor for neoplasm (Table 2).
Endoscopic and Histologic Risk Factors of Gastric Cancer The number of patients with carcinoma was 36 (37.5 %). Patients in the carcinoma group (63 ± 11.5 years) were older than those in the non-carcinoma group (57 ± 14.9 years); however, this difference was not significant. In endoscopic findings, the mean size of the carcinoma and non-carcinoma lesions was 19.56 ± 8.3 mm (range 5–42) and 11.63 ± 7.6 mm (range 3–36). The carcinoma group (83.3 %) had larger lesions of 10 mm or greater compared with the non-carcinoma group (48.3 %). Depressed type was more frequent in the malignancy group (66.6 %) than in the non-carcinoma group (31.7 %). In histologic findings, no significant difference related to presence of intestinal metaplasia or H. pylori was observed between the two groups (Table 3). In univariate analysis, lesion size of 10 mm or greater (p = 0.001, OR 5.35, 95 % CI 1.94–14.71) and lesion with depressed type (p = 0.001, OR 4.32, 95 % CI 1.79–10.42) were risk factors for cancer related to GAC. In multivariate analysis, they were independent risk factors for cancer (Table 4). Positive Predictive Value, Negative Predictive Value, Sensitivity, and Specificity On the basis of significant and independent risk factors related with multivariate analysis mentioned above, we determined that the predictive capability of lesion size of 10 mm or greater for gastric cancer was: PPV 77.8 %, NPV 52.9 %, sensitivity 59.3 %, and specificity 72.9 %. The predictive capability of lesion with depressed type for gastric cancer was: PPV 58.5 %, NPV 61.8 %, sensitivity 53.3 %, and specificity 66.7 %; and the predictive capability of both factors for gastric cancer was: PPV 52.9 %, NPV 70.9 %, sensitivity 50.0 %, and specificity 73.3 %. The Correlation Between the Clinical, Endoscopic, and the Final Histologic Diagnoses Endoscopically, fifty cases (52.1 %) were diagnosed with gastritis, ten cases (10.4 %) with adenoma and thirty-six cases (37.5 %) with cancer. The concordance rate of total diagnoses and that of ‘‘neoplasm’’ was 75.0 % (n = 72, of 96) and 78.2 % (n = 36, of 46) (Table 5).
Discussion Gastric atypical cell (GAC) is an indefinite lesion for dysplasia; diagnosis must be confirmed by additional procedures. In our study, after additional diagnostic
123
Dig Dis Sci Table 2 Risk factors of gastric neoplasm in univariate and multivariate analyses
Variable
Univariate analysis
Age
Multivariate analysis
p value
OR
95 % CI
p value
OR
95 % CI
0.020
1.04
1.01–1.07
0.060
1.035
0.99–1.07
0.030
3.94
1.61–9.61
0.096
0.030
2.50
1.09–5.72
0.189
0.030
2.83
1.11–7.25
0.081
Endoscopic findings Lesion size C10 mm Lesion type Depressed Histologic findings OR odds ratio, CI confidence interval
H. pylori
Table 3 Endoscopic and histologic risk factors of gastric cancer Variable [n (%)]
Age [year (±SD)] Sex
Nonmalignancy (n = 60)
Malignancy (n = 36)
57 (14.99)
63 (11.56)
Male
36 (60.0)
25 (69.4)
Female
24 (40.0)
11 (30.6)
p value
0.063 0.353
Endoscopic findings Mean Size, mm (±SD)
11.63 (7.6)
19.56 (8.3)
Size of long diameter
\0.001 0.001
\10 mm
31 (51.7)
6 (16.7)
C10 mm
29 (48.3)
30 (83.3)
1 (1.6)
2 (5.6)
Body
22 (36.7)
15 (41.7)
Angle
4 (6.7)
3 (8.3)
33 (55.0)
26 (44.4)
Elevated Flat
20 (33.3) 21 (35.0)
6 (16.7) 6 (16.7)
0.081 0.186
Depressed
19 (31.7)
24 (66.6)
0.001
Presence
6 (10.0)
6 (16.7)
Absence
54 (90.0)
30 (83.3)
16 (26.7)
10 (27.8)
Location Fundus
Antrum
0.640
Lesion type
Histologic findings Intestinal metaplasia
0.344
H. pylori Presence Absence Number of biopsy fragments (±SD)a
0.906 44 (73.3)
26 (72.2)
3.6 (1.4)
4.1 (2.0)
0.163
n number, SD standard deviation a
Number of biopsy fragments per lesion in one endoscopic session
modalities, including OP (n = 20, 20.8 %) or EMR/ESD (n = 23, 23.9 %), the final pathologic diagnoses of GAC were cancer (n = 36, 37.6 %), dysplasia (n = 9, 9.4 %), and non-neoplasm (n = 51, 53 %). In univariate analysis,
123
endoscopic and histologic risk factors for gastric neoplasm were age, lesion size of 10 mm or greater, lesion with depressed type, and presence of H. pylori; and in multivariate analysis, lesion size of 10 mm or greater and lesion with depressed type were independent risk factors for gastric cancer. Before the advent of endoscopic treatment modalities, patients underwent open abdominal surgery for suspected gastric cancer. Indeed, in our study, nine patients underwent surgery due to strongly suspected gastric cancer until 2005. Their histologic reports were described as ‘‘GAC with more suspicious of adenocarcinoma’’ based on the initial biopsy; however, one of them was finally diagnosed as dysplasia with high-grade dysplasia after the operation. Currently, EMR/ESD are performed more commonly for en bloc resection of a premalignant or malignant lesion in possible cases of endoscopic procedures according to the extended endoscopic guidelines in Korea and Japan [10]. ESD contributes therapeutically to long-term outcomes and quality of life for patients with EGC [11, 12]. Considering the clinical merits of EMR/ESD, these procedures may be useful diagnostic tools for cases of GAC after endoscopic forceps biopsy. Although repeat forceps biopsy has been recommended by the guidelines for GAC [8], a discrepancy may exist between forceps biopsy samples and dissected specimens because forceps biopsy do not represent the entire lesion [13]. Some studies have reported discrepancy rates after EMR ranging from 18.7 to 44.5 %, including final upgraded and downgraded pathology [4, 14]. Results of our study showed that 45 cases of GAC were subsequently upgraded to dysplasia or cancer after more aggressive diagnostic sampling, and even three cases repeatedly misdiagnosed as GAC by re-biopsy were eventually found to be carcinomas. This cautions that repeat biopsy for GAC can miss the actual true diagnosis. Therefore, EMR/ESD procedures are needed for definite diagnoses and treatments of GAC lesions. On the other hand, endoscopic resection could be associated with more potential complication such as bleeding or perforation and
Dig Dis Sci Table 4 Risk factors of cancer in univariate and multivariate analyses
Variable
Univariate analysis
Age
Multivariate analysis
p value
OR
95 % CI
p value
OR
95 % CI
0.063
1.03
0.99–1.07
0.150
0.001
5.35
1.94–14.71
0.019
3.63
1.23–10.66
0.001
4.32
1.79–10.42
0.030
2.86
1.11–7.38
Endoscopic findings Lesion size C10 mm Lesion type OR odds ratio, CI confidence interval
Depressed
Table 5 The correlation between the endoscopic and the histologic diagnosis Histologic diagnosis (n)
Endoscopic diagnosis (n) Gastritis
Adenoma
Cancer
Gastritis
41
4
Adenoma
2
4
3
9
Cancer
7
2
27
36
50
10
36
96
Total
6
Total 51
n number
increase medical expenses compared with those of repeat biopsy. However, these complications could be managed safely and successfully by non-surgical methods [15]. And, we noted, in Eastern Asia including Korea, that medical costs were relatively inexpensive, but the prevalence rate of gastric cancer was higher than those of Western as stated above. Clinicians should consider the risk factors for neoplasm related to the host, endoscopic, or histologic characteristics. In several studies, age, grade of coexisting atrophic gastritis, depressed type, and large size were identified as risk factors for malignant transformation [16–19]. Our study also showed that, in univariate analysis, age (OR 1.04, 95 % CI 1.01–1.07), lesion size of 10 mm or greater (OR 3.94, 95 % CI 1.61–9.61), and lesion with depressed type (OR 2.50, 95 % CI 1.09–5.72) were risk factors for neoplasm. Coexisting atrophic gastritis was not included in analysis because most of the patients in our study had this factor. The World Health Organization’s International Agency for Research on Cancer classified H. pylori as a Group 1 or definite carcinogen [20]. Some studies have supported the concept of intestinal metaplasia and H. pylori infection as precursors for intestinal type gastric cancer [21–23]. In our study, presence of H. pylori (OR 2.83, 95 % CI 1.11–7.25) was a risk factor for neoplasm in univariate analysis. However, intestinal metaplasia showed no significant influence in any analysis. More specifically, we investigated risk factors for cancer. According to our results, in multivariate analysis, lesion size of 10 mm or greater (OR 3.63, 95 % CI 1.23–10.66) and lesion with
depressed type (OR 2.86, 95 % CI 1.11–7.38) were independent risk factors for cancer. The specificity related to cancer predictive value of lesion with depressed type, lesion size of 10 mm or greater, and both factors was 66.7, 72.9, and 73.3 %, respectively. These results demonstrate that the presence of these risk factors alone is insufficient to predict cancer. In our study, however, the rate of concordant diagnoses of ‘‘neoplasm’’ between the initial endoscopic and the final histologic diagnoses was 78.2 %. A recent study also showed high correlation between the clinic, endoscopic, and the final histologic diagnoses of GAC lesions, related with adenoma or cancer [24]. So, we assumed that there were more additional informations for endoscopic diagnoses of GAC lesions, although our study did not investigate. Other study reported that red coloration and friability in terms of endoscopic findings of GAC lesions were significant in the neoplastic group [25]. So, we suggest that an initial endoscopic diagnosis of GAC lesions can be more accurate and sharper, when endoscopists consider not only lesions of depression and size but also lesions of coloration, friability, and surface changes related with neoplasm. Also, in recent, clinicians can consider to perform magnifying narrowband imaging (M-NBI) for GAC lesion. Several studies demonstrated that M-NBI is more accurate and useful than conventional white-light imaging for differentiating between EGC of small, flat, and depressed lesions and gastritis by investigation of microvascular and mucosal patterns [26, 27]. Indeed, there were seven cases of gastritis-like cancer and six cases of cancerlike gastritis by conventional imaging in our study. So, using M-NBI would contribute to minimizing unnecessary biopsies or endoscopic resection of non-cancerous lesions taken in order to rule out gastritis-like cancer. Nevertheless, taking a histologic specimen cannot be omitted, because the histologic decision is definitely needed for diagnosis and treatment of a lesion suspicious of neoplasm. In clinical practice, four to six forceps biopsy specimens are taken from a lesion suspicious of malignancy. In our study, sixty-nine cases (71.8 %) including lesions suspicious of neoplasm reached the final diagnosis by approximately four pieces of re-biopsy. And, a prospective trial
123
Dig Dis Sci
emphasized that at least seven specimens of biopsy should be obtained for increasing the yield to greater than 98 % of malignancy diagnosis [28]. In another study, accurate acquisition of biopsy specimens comprising at least three to four pieces from the lesion of mucosal change is important; however, there is no international agreement regarding how many specimens are sufficient [29]. And, if the lesion is resectable by ESD or EMR, many of biopsies increase complication of endoscopic resection due to fibrosis and ulcers following the biopsies. Considering that mentioned above, clinicians should perform careful conventional endoscopy to investigate mucosal characteristics: depression, size, coloration, friability, and surface changes. Also, if needed and possible, M-NBI could be performed for additional investigation of microvascular and mucosal patterns. And then, clinicians must discuss with pathologists whether a GAC lesion is more suspicious of dysplasia/cancer in histologic finding or not. Finally, if a GAC lesion is resectable by ESD/EMR according to the extended endoscopic guidelines [10] and more malignant potential in totally endoscopic and histologic findings, an endoscopic resection could be more reasonable and comprehensive than repeated biopsies. Our study has several limitations. First, because this study was conducted retrospectively, there is a possibility of selection bias. However, in order to reduce this bias, we enrolled mostly GAC patients in our study and all pathologic slides were reviewed by two independent pathologists. Second, there were no standardized criteria regarding the next step that should be taken by individual clinicians across enrolled patients who underwent EMR/ESD, OP, or follow-up. However, our study indeed demonstrated the significance of the GAC lesion in clinical practice, considering the difficulty of conducting a prospective randomized controlled trial. Third, in cases of GAC, pathologists should consider performing immunohistochemical stains for p53 and Ki67. These stains are helpful for diagnoses of gastric adenoma or adenocarcinoma [30, 31]; However, our study did not investigate because they are neither definitive tools nor diagnostic factors of pathologic grading classification [1, 8]. Also, a recent study showed there was no significant difference in p53 or Ki67 expression between intestinal metaplasia with atypia and gastric carcinoma [32]. Finally, most patients diagnosed as gastritis on first re-biopsy were followed for a period of less than 2 years or lost to follow-up in our center. Therefore, the final diagnosis of these patients could be inadequate. However, we assumed that most of them received proper follow-up, because the Korea National Cancer Screening Program recommends and conducts biennial gastric cancer screening using upper gastrointestinal series or endoscopy in order to reduce the high rate of gastric cancer mortality in Korea [33]. Ultimately, conduct
123
of a further large-scale prospective study will be necessary in order to overcome these limitations. In conclusion, considering the neoplastic change of GAC, EMR/ESD procedures may be more useful for definite diagnoses and en bloc treatments, especially in cases accompanied by significant and independent risk factors, such as lesion size of 10 mm or greater and lesion with depressed type. Conflict of interest
None.
References 1. Dixon MF. Gastrointestinal epithelial neoplasia: Vienna revisited. Gut. 2002;51:130–131. 2. Bray F, Ren JS, Masuyer E, Ferlay J. Global estimates of cancer prevalence for 27 sites in the adult population in 2008. Int J Cancer. 2013;132:1133–1145. 3. Misdraji J, Lauwers GY. Gastric epithelial dysplasia. Semin Diagn Pathol. 2002;19:20–30. 4. Kim YJ, Park JC, Kim JH, et al. Histologic diagnosis based on forceps biopsy is not adequate for determining endoscopic treatment of gastric adenomatous lesions. Endoscopy. 2010;42:620–626. 5. Hull MJ, Mino-Kenudson M, Nishioka NS, et al. Endoscopic mucosal resection: an improved diagnostic procedure for early gastroesophageal epithelial neoplasms. Am J Surg Pathol. 2006;30:114–118. 6. Gotoda T. Endoscopic resection of early gastric cancer: the Japanese perspective. Curr Opin Gastroenterol. 2006;22:561–569. 7. Gotoda T, Yamamoto H, Soetikno RM. Endoscopic submucosal dissection of early gastric cancer. J Gastroenterol. 2006;41:929–942. 8. Kim HG, Jin SY, Jang JJ, et al. Grading system for gastric epithelial proliferative disease. Korean J Pathol. 1997;31:389–400. 9. Hamilton SRAL. World Health Organization classification of tumours: pathology and genetics of tumours of the digestive system. Lyon: IARC Press; 2000. 10. Japanese Gastric Cancer A. Japanese gastric cancer treatment guidelines 2010 (ver. 3). Gastric Cancer. 2011;14:113–123. 11. Lee H, Yun WK, Min BH, et al. A feasibility study on the expanded indication for endoscopic submucosal dissection of early gastric cancer. Surg Endosc. 2011;25:1985–1993. 12. Imagawa A, Okada H, Kawahara Y, et al. Endoscopic submucosal dissection for early gastric cancer: results and degrees of technical difficulty as well as success. Endoscopy. 2006;38:987–990. 13. Yoon WJ, Lee DH, Jung YJ, et al. Histologic characteristics of gastric polyps in Korea: emphasis on discrepancy between endoscopic forceps biopsy and endoscopic mucosal resection specimen. World J Gastroenterol. 2006;12:4029–4032. 14. Lee CK, Chung IK, Lee SH, et al. Is endoscopic forceps biopsy enough for a definitive diagnosis of gastric epithelial neoplasia? J Gastroenterol Hepatol. 2010;25:1507–1513. 15. Jeon SW, Jung MK, Kim SK, et al. Clinical outcomes for perforations during endoscopic submucosal dissection in patients with gastric lesions. Surg Endosc. 2010;24:911–916. 16. Park DI, Rhee PL, Kim JE, et al. Risk factors suggesting malignant transformation of gastric adenoma: univariate and multivariate analysis. Endoscopy. 2001;33:501–506.
Dig Dis Sci 17. Nakamura K, Sakaguchi H, Enjoji M. Depressed adenoma of the stomach. Cancer. 1988;62:2197–2202. 18. Jung MK, Jeon SW, Park SY, et al. Endoscopic characteristics of gastric adenomas suggesting carcinomatous transformation. Surg Endosc. 2008;22:2705–2711. 19. Rugge M, Leandro G, Farinati F, et al. Gastric epithelial dysplasia. How clinicopathologic background relates to management. Cancer. 1995;76:376–382. 20. Schistosomes, liver flukes and Helicobacter pylori. IARC Working Group on the Evaluation of Carcinogenic Risks to Humans. Lyon, 7–14 June 1994. IARC monographs on the evaluation of carcinogenic risks to humans/World Health Organization, International Agency for Research on Cancer. 1994; 61:1–241. 21. Shimoyama T, Fukuda S, Tanaka M, Nakaji S, Munakata A. Evaluation of the applicability of the gastric carcinoma risk index for intestinal type cancer in Japanese patients infected with Helicobacter pylori. Virchows Arch. 2000;436:585–587. 22. Parsonnet J, Vandersteen D, Goates J, Sibley RK, Pritikin J, Chang Y. Helicobacter pylori infection in intestinal- and diffusetype gastric adenocarcinomas. J Natl Cancer Inst. 1991;83:640–643. 23. Eslick GD, Lim LL, Byles JE, Xia HH, Talley NJ. Association of Helicobacter pylori infection with gastric carcinoma: a metaanalysis. Am J Gastroenterol. 1999;94:2373–2379. 24. Kim SI, Han HS, Kim JH, et al. What is the next step for gastric atypical epithelium on histological findings of endoscopic forceps biopsy? Dig Liver. 2013;45:573–577.
25. Lee H, Kim H, Shin SK, et al. The diagnostic role of endoscopic submucosal dissection for gastric lesions with indefinite pathology. Scand J Gastroenterol. 2012;47:1101–1107. 26. Yao K, Iwashita A, Tanabe H, et al. Novel zoom endoscopy technique for diagnosis of small flat gastric cancer: a prospective, blind study. Clin Gastroenterol Hepatol. 2007;5:869–878. 27. Ezoe Y, Muto M, Uedo N, et al. Magnifying narrowband imaging is more accurate than conventional white-light imaging in diagnosis of gastric mucosal cancer. Gastroenterology. 2011;141:2017e3–2025e3. 28. Graham DY, Schwartz JT, Cain GD, Gyorkey F. Prospective evaluation of biopsy number in the diagnosis of esophageal and gastric carcinoma. Gastroenterology. 1982;82:228–231. 29. Haber MM. Gastric biopsies: increasing the yield. Clin Gastroenterol Hepatol. 2007;5:160–165. 30. Shiao YH, Rugge M, Correa P, Lehmann HP, Scheer WD. p53 alteration in gastric precancerous lesions. Am J Pathol. 1994;144:511–517. 31. Zheng H, Tsuneyama K, Cheng C, et al. An immunohistochemical study of P53 and Ki-67 in gastrointestinal adenoma and adenocarcinoma using tissue microarray. Anticancer Res. 2006;26:2353–2360. 32. Zheng Y, Wang L, Zhang JP, Yang JY, Zhao ZM, Zhang XY. Expression of p53, c-erbB-2 and Ki67 in intestinal metaplasia and gastric carcinoma. World J Gastroenterol. 2010;16:339–344. 33. Yoo KY. Cancer control activities in the Republic of Korea. Jpn J Clin Oncol. 2008;38:327–333.
123
ORIGINAL ARTICLE
Endoscopic Resection as a First Therapy for Gastric Epithelial Atypia: Is It Reasonable? Chung Hoon Yu • Seong Woo Jeon • Sung Kook Kim Hyun Seok Lee • Jun Heo • Yong Hwan Kwon • Gyu Young Kim • Sun Zoo Kim • Han Ik Bae
•
Received: 27 March 2014 / Accepted: 4 June 2014 Ó Springer Science+Business Media New York 2014
Abstract Background and Aims Gastric atypical cell (GAC), an indefinite pathologic finding, often requires repeated biopsy or other diagnostic treatments, such as endoscopic mucosal resection (EMR), endoscopic submucosal dissection (ESD), or operation (OP). The aim of this study was to analyze the initial endoscopic and histologic findings of GAC and to discuss the necessity of EMR/ESD at establishing a correct diagnosis. Methods This retrospective study enrolled 96 patients proven as GAC on index forceps biopsy. ESD (17/96, 17.7 %), EMR (5/96, 5.2 %), OP (20/96, 20.8 %), and other treatment or follow-up (54/96, 56.3 %) were performed. We analyzed the initial endoscopic and histologic characteristics of GAC lesions, predictive of neoplasm. Results After diagnostic modalities, the final pathologic diagnoses were cancer (36/96, 37.6 %), dysplasia (9/96, 9.4 %), and non-neoplasm (51/96, 53.0 %). In univariate analysis, age [odds ratio (OR) 1.04, 95 % confidence interval (CI) 1.01–1.07], lesion size of 10 mm or greater (OR 3.94, 95 % CI 1.61–9.61), lesion with depressed type (OR 2.50, 95 % CI 1.09–5.72), and presence of H. pylori C. H. Yu S. W. Jeon S. K. Kim H. S. Lee J. Heo Y. H. Kwon G. Y. Kim Department of Internal Medicine, Kyungpook National University School of Medicine, Daegu, Republic of Korea S. W. Jeon (&) Gastric Cancer Center, Kyungpook National University Medical Center, 807 Hogukno, Buk-gu, Daegu 702-210, Republic of Korea e-mail: [email protected] S. Z. Kim H. I. Bae Department of Pathology, Kyungpook National University School of Medicine, Daegu, Republic of Korea
(OR 2.83, 95 % CI 1.11–7.25) were risk factors for neoplasm. In multivariate analysis, lesion size of 10 mm or greater (OR 3.63, 95 % CI 1.23–10.66), lesion with depressed type (OR 2.86, 95 % CI 1.11–7.38) were independent risk factors for cancer. Conclusion Considering the neoplastic risk of GAC, which could be missed on biopsy, more comprehensive tissue sampling via EMR/ESD might be necessary to establish a definite diagnosis. Keywords Gastric neoplasms Atypical epithelial cells Endoscopic surgical procedure Indefinite pathology
Introduction Gastric atypical cell (GAC) is classified as ‘‘indefinite for neoplasm/dysplasia’’ (Category 2) in the revised Vienna classification of gastrointestinal epithelial neoplasia [1]. This diagnosis can be used in a limited number of settings, such as when there is genuine difficulty in discerning reactive epithelial changes from dysplasia, when worrisome atypia is noted but is confined to a very small area, or when the tissue is limited in amount or quality. Due to the biological potential of gastric dysplasia, it is considered a precursor of gastric cancer, the most prevalent cancer in Eastern Asia, including Korea [2, 3]. Follow-up is absolutely necessary for GAC, and re-biopsy is commonly recommended. However, some studies have suggested that forceps biopsy is insufficient to confirm the real nature of the lesion, because it is not entirely representative of the gastric lesion. In addition, despite having an endoscopic appearance, like that of a malignant lesion, pathologic finding of forceps biopsy can be indefinite [4, 5]. Eventually, further diagnostic modalities are necessary for GAC in
123
Dig Dis Sci
order to confirm diagnoses and for treatment of premalignant or malignant lesions. Before the advent of endoscopic treatment modalities, patients underwent open abdominal surgery in cases where stomach cancer was strongly suspected, including early gastric cancer (EGC). Currently, however, endoscopic mucosal resection (EMR) or endoscopic submucosal dissection (ESD) is performed more commonly for the en bloc resection of a premalignant or malignant lesions, where possible. In addition, these procedures of en bloc resection can be useful diagnostic tools for cases with indeterminate pathology as atypical change due to more accurate pathologic information than forceps biopsy [5–7]. However, there is no sufficient evidence to support EMR or ESD as a first therapeutic modality in these situations. The aim of this study was to discuss the utility of advanced diagnostic modalities including EMR/ESD and to analyze the initial endoscopic and histologic risk factors predictive of neoplasm for cases of GAC.
Methods Patients This retrospective study included 110 patients with GAC that was histologically proven and based on initial endoscopic forceps biopsy between January 1999 and December 2012 at Kyungpook National University Hospital in Korea. Detailed information on the baseline characteristics of the patients, the endoscopic characteristics of the lesions, the pathologic diagnoses, and the clinical outcomes were obtained from the patients’ medical records. GAC on endoscopic forceps biopsy was defined as atypical change interpreted as questionable dysplasia. Cases of GAC with reactive or regenerative gastric epithelium [number (n = 10)] were excluded because reactivity is resolved when infection or inflammation shows improvement [8]. In addition, those lost to follow-up after the first forceps biopsy (n = 4) were excluded. This study was approved by our institutional review board. Endoscopic and Histopathologic Evaluations Diagnostic endoscopy was performed using a conventional gastroscope (Olympus, Tokyo, Japan). The initial endoscopic diagnosis was supervised by one of three expert endoscopists (SK Kim, HS Lee, and SW Jeon) of our center, and the size of lesion was measured with an open biopsy forceps (each arm of bite: 3 mm) as the reference device during endoscopic procedure. Endoscopic films and
123
reports were reviewed by a single expert endoscopist (SW Jeon) for size (mm), location (fundus, body, angle, and antrum), and type (elevated, flat or depressed). Two gastrointestinal pathology specialists (SZ Kim and HI Bae) were involved in the histologic diagnoses. In cases of discrepancy between the two pathologists, final diagnoses were made after discussion. Histologic diagnoses were based on the World Health Organization (WHO) classification of pathology and genetics of tumors of the digestive system [9]. In particular, atypical change interpreted as questionable dysplasia (Grade 2) was the basis of the definition of the Korean pathologic grading system for gastric epithelial proliferative disease [8]. This group included lesions that were difficult to discern as reactive or regenerative epithelium from dysplasia in terms of cellular and structural atypia (Fig. 1a, b). Gastric dysplasia was characterized by a mild architectural complexity, with glands lined by crowded, elongated cells with hyperchromatic and basilar pseudostratified nuclei. In contrast to dysplasia, regenerative or reactive atypia, in which the nuclei may be slightly larger than in normal epithelium, was characterized by the same degree of architectural atypia, with glands respecting their relative spaces without significant crowding or pseudostratification. Carcinoma showed a higher nuclear grade and either a prominent back-to-back or a syncytial growth pattern, abortive microglands, or small clusters of epithelial cells within the lamina propria between glands (Fig. 1c, d). The combined findings of dysplasia and carcinoma were classified as carcinoma. In this study, ‘‘neoplasm’’ was composed of cancer and dysplasia (including adenoma). An EGC (EGC) is defined as invasive gastric cancer involving the submucosa, irrespective of lymph node or distant metastasis (T1, any N, M) and advanced gastric cancer (AGC) as gastric cancer that invades into muscularis propria (T2 and above, any N, M) [9]. This study evaluated the correlation between the clinical, endoscopic diagnosis, and the final histologic diagnosis. A concordant diagnosis meant the initial endoscopic and final histologic diagnoses were the same. A concordant diagnosis of ‘‘neoplasm’’ was defined if the final histologic diagnosis showed within the range of neoplasm finding regardless of an upgraded or downgraded diagnosis, when the endoscopic diagnosis was ‘‘dysplasia/adenoma or cancer.’’ H. pylori infection was diagnosed by either histology or rapid urease test (RUT, CLO test, GASTREX, Warsaw, Poland). The specimens for RUT were obtained from both the antrum and the distal body of the stomach during EGD. Cresyl violet staining was used for the pathologic diagnosis of H. pylori infection. H. pylori infection was diagnosed when the result of one of the two methods described above was positive.
Dig Dis Sci
Fig. 1 Histologic findings of the initial endoscopic forceps biopsy (a, b) and the endoscopic submucosal dissection (c, d) from a patient with a depressed lesion. a Histologic section shows a proliferative glandular lesion, composed of atypical glands with mild architectural distortion and slightly inflamed intervening stroma (hematoxylin and eosin (H & E) stain, 9100). b The proliferative glands are lined by epithelium with large hyperchromatic nuclei, a high nucleus to
cytoplasm ratio, and loss of cellular polarity. There is some inflammation in the glandular epithelium and the surrounding stroma; however, it is insufficient to account for the epithelial abnormalities. The nuclei, while hyperchromatic and polymorphic, remain in a predominantly basal position, and maturation toward the surface is observed (H & E stain, 9400). c, d Slides show well-differentiated adenocarcinoma (H & E, 9100 and 9400)
Data and Statistical Analyses
biopsy, or operation as part of their diagnostic/therapeutic management. Then, some patients underwent additional procedures, if needed, for definite diagnoses or treatments. After complete work-up, the final pathologic diagnoses were cancer (n = 36, 37.6 %), dysplasia (n = 9, 9.4 %), and non-neoplasm (n = 51, 53.0 %). Note that three cases were persistently diagnosed as GAC on re-biopsy. All of these patients (n = 3) underwent a second re-biopsy consistent with regenerative atypia (n = 2) and gastritis (n = 1). Only on the third re-biopsy, done because of characteristics suspicious for malignancy, was a diagnosis of adenocarcinoma made in all cases.
Descriptive data are presented as the number of patients (%) or as mean ± standard deviation (SD). Continuous data were compared using independent t tests, and categorical variables were tested using the chi-square test. The results of univariate and multivariate logistic regression analyses were presented as odds ratio [OR; 95 % confidence intervals (CI)] for risk factors of neoplasm. Positive predictive value (PPV), negative predictive value (NPV), sensitivity, and specificity were evaluated for significant and independent risk factors in multivariate analysis. The criterion for statistical significance was p \ 0.05. Analyses were performed using SPSS (version 18.0; SPSS, Chicago, IL, USA).
Result Final Diagnoses of Patients Figure 2 shows a flow diagram of the study. Initially, the 96 enrolled patients underwent EMR/ESD (Fig. 3), re-
Endoscopic and Histologic Risk Factors for Gastric Neoplasm The number of patients with neoplasm was 45 (46.9 %). Patients in the neoplasm group (mean age 63 ± 11.1 years) were significantly older than those in the non-neoplasm group (56 ± 15.5 years). In endoscopic findings, the mean size of the neoplastic and non-neoplastic lesions was 18.87 ± 8.4 mm (range 5–42) and
123
Dig Dis Sci Fig. 2 Flow diagram of 96 GAC lesions diagnosed by forceps biopsy. n number, ESD endoscopic submucosal dissection, EMR endoscopic mucosal ressection, Bx biopsy, OP operation, EGC early gastric cancer, AGC advanced gastric cancer, F/U follow-up, Other other management, Small cell small cell carcinoma, Adenoca adenocarcinoma, Non-Tx nontreatment
Fig. 3 A case of the endoscopic diagnosis as ‘‘rule out early gastric cancer,’’ confirmed as adenocarcinoma after endoscopic submucosal dissection (ESD) in gastric lesion with atypical change interpreted as questionable dysplasia on forceps biopsy specimens. Endoscopic
123
imaging viewed (a) a 1.5-cm hyperemic mucosal lesion with depression on the posterior wall side of the low body, b marking dots for ESD, c the lesion completely dissected after procedure, and d the ESD specimen with a central lesion
Dig Dis Sci Table 1 Endoscopic and histologic risk factors of gastric neoplasm Variable [n (%)] Age [year (±SD)]
Non-neoplasm (n = 51)
Neoplasm (n = 45)
56 (15.56)
63 (11.10)
Male
30 (58.8)
31 (68.9)
Female
21 (41.2)
14 (31.1)
Sex
p value 0.020 0.308
Endoscopic findings Mean Size [mm(±SD)]
10.84 (7.2)
18.87 (8.4)
Size of long diameter
\0.001 0.030
\10 mm
27 (52.9)
10 (22.2)
C10 mm
24 (47.1)
35 (77.8)
1 (2.0)
2 (4.4)
Body
20 (39.2)
17 (37.8)
Angle
4 (7.8)
3 (6.7)
26 (51.0)
23 (51.1)
Elevated
19 (37.3)
13 (28.9)
0.387
Flat
15 (29.4)
8 (17.8)
0.186
17 (33.3)
24 (53.3)
0.030
Location Fundus
Antrum
0.918
Lesion type
Depressed Histologic findings Intestinal metaplasia
0.399
Presence
5 (9.8)
7 (15.5)
Absence
46 (90.2)
38 (84.5)
H. pylori
0.030
Presence
9 (17.6)
17 (37.8)
Absence
42 (82.4)
28 (62.2)
3.7 (1.4)
3.9 (1.8)
Number of biopsy fragments (±SD)a
0.546
n number, SD standard deviation a
Number of biopsy fragments per lesion in one endoscopic session
10.84 ± 7.2 mm (range 3–36). The neoplasm group (77.8 %) had larger lesions of 10 mm or greater compared with the non-neoplasm group (47.1 %). Depressed type was more frequent in the neoplasm group (53.3 %) than in the non-neoplasm group (33.3 %). In histologic findings, presence of H. pylori was more frequent in the neoplasm group (37.8 %) than in the non-neoplasm group (17.6 %) (Table 1). In univariate analysis, age (p = 0.020, OR 1.04, 95 % CI 1.01–1.07), lesion size of 10 mm or greater (p = 0.030, OR 3.94, 95 % CI 1.61–9.61), lesion with depressed type (p = 0.030, OR 2.50, 95 % CI 1.09–5.72), and presence of H. pylori (p = 0.030, OR 2.83, 95 % CI 1.11–7.25) were risk factors for neoplasm. However, in multivariate analysis, none of them was an independent risk factor for neoplasm (Table 2).
Endoscopic and Histologic Risk Factors of Gastric Cancer The number of patients with carcinoma was 36 (37.5 %). Patients in the carcinoma group (63 ± 11.5 years) were older than those in the non-carcinoma group (57 ± 14.9 years); however, this difference was not significant. In endoscopic findings, the mean size of the carcinoma and non-carcinoma lesions was 19.56 ± 8.3 mm (range 5–42) and 11.63 ± 7.6 mm (range 3–36). The carcinoma group (83.3 %) had larger lesions of 10 mm or greater compared with the non-carcinoma group (48.3 %). Depressed type was more frequent in the malignancy group (66.6 %) than in the non-carcinoma group (31.7 %). In histologic findings, no significant difference related to presence of intestinal metaplasia or H. pylori was observed between the two groups (Table 3). In univariate analysis, lesion size of 10 mm or greater (p = 0.001, OR 5.35, 95 % CI 1.94–14.71) and lesion with depressed type (p = 0.001, OR 4.32, 95 % CI 1.79–10.42) were risk factors for cancer related to GAC. In multivariate analysis, they were independent risk factors for cancer (Table 4). Positive Predictive Value, Negative Predictive Value, Sensitivity, and Specificity On the basis of significant and independent risk factors related with multivariate analysis mentioned above, we determined that the predictive capability of lesion size of 10 mm or greater for gastric cancer was: PPV 77.8 %, NPV 52.9 %, sensitivity 59.3 %, and specificity 72.9 %. The predictive capability of lesion with depressed type for gastric cancer was: PPV 58.5 %, NPV 61.8 %, sensitivity 53.3 %, and specificity 66.7 %; and the predictive capability of both factors for gastric cancer was: PPV 52.9 %, NPV 70.9 %, sensitivity 50.0 %, and specificity 73.3 %. The Correlation Between the Clinical, Endoscopic, and the Final Histologic Diagnoses Endoscopically, fifty cases (52.1 %) were diagnosed with gastritis, ten cases (10.4 %) with adenoma and thirty-six cases (37.5 %) with cancer. The concordance rate of total diagnoses and that of ‘‘neoplasm’’ was 75.0 % (n = 72, of 96) and 78.2 % (n = 36, of 46) (Table 5).
Discussion Gastric atypical cell (GAC) is an indefinite lesion for dysplasia; diagnosis must be confirmed by additional procedures. In our study, after additional diagnostic
123
Dig Dis Sci Table 2 Risk factors of gastric neoplasm in univariate and multivariate analyses
Variable
Univariate analysis
Age
Multivariate analysis
p value
OR
95 % CI
p value
OR
95 % CI
0.020
1.04
1.01–1.07
0.060
1.035
0.99–1.07
0.030
3.94
1.61–9.61
0.096
0.030
2.50
1.09–5.72
0.189
0.030
2.83
1.11–7.25
0.081
Endoscopic findings Lesion size C10 mm Lesion type Depressed Histologic findings OR odds ratio, CI confidence interval
H. pylori
Table 3 Endoscopic and histologic risk factors of gastric cancer Variable [n (%)]
Age [year (±SD)] Sex
Nonmalignancy (n = 60)
Malignancy (n = 36)
57 (14.99)
63 (11.56)
Male
36 (60.0)
25 (69.4)
Female
24 (40.0)
11 (30.6)
p value
0.063 0.353
Endoscopic findings Mean Size, mm (±SD)
11.63 (7.6)
19.56 (8.3)
Size of long diameter
\0.001 0.001
\10 mm
31 (51.7)
6 (16.7)
C10 mm
29 (48.3)
30 (83.3)
1 (1.6)
2 (5.6)
Body
22 (36.7)
15 (41.7)
Angle
4 (6.7)
3 (8.3)
33 (55.0)
26 (44.4)
Elevated Flat
20 (33.3) 21 (35.0)
6 (16.7) 6 (16.7)
0.081 0.186
Depressed
19 (31.7)
24 (66.6)
0.001
Presence
6 (10.0)
6 (16.7)
Absence
54 (90.0)
30 (83.3)
16 (26.7)
10 (27.8)
Location Fundus
Antrum
0.640
Lesion type
Histologic findings Intestinal metaplasia
0.344
H. pylori Presence Absence Number of biopsy fragments (±SD)a
0.906 44 (73.3)
26 (72.2)
3.6 (1.4)
4.1 (2.0)
0.163
n number, SD standard deviation a
Number of biopsy fragments per lesion in one endoscopic session
modalities, including OP (n = 20, 20.8 %) or EMR/ESD (n = 23, 23.9 %), the final pathologic diagnoses of GAC were cancer (n = 36, 37.6 %), dysplasia (n = 9, 9.4 %), and non-neoplasm (n = 51, 53 %). In univariate analysis,
123
endoscopic and histologic risk factors for gastric neoplasm were age, lesion size of 10 mm or greater, lesion with depressed type, and presence of H. pylori; and in multivariate analysis, lesion size of 10 mm or greater and lesion with depressed type were independent risk factors for gastric cancer. Before the advent of endoscopic treatment modalities, patients underwent open abdominal surgery for suspected gastric cancer. Indeed, in our study, nine patients underwent surgery due to strongly suspected gastric cancer until 2005. Their histologic reports were described as ‘‘GAC with more suspicious of adenocarcinoma’’ based on the initial biopsy; however, one of them was finally diagnosed as dysplasia with high-grade dysplasia after the operation. Currently, EMR/ESD are performed more commonly for en bloc resection of a premalignant or malignant lesion in possible cases of endoscopic procedures according to the extended endoscopic guidelines in Korea and Japan [10]. ESD contributes therapeutically to long-term outcomes and quality of life for patients with EGC [11, 12]. Considering the clinical merits of EMR/ESD, these procedures may be useful diagnostic tools for cases of GAC after endoscopic forceps biopsy. Although repeat forceps biopsy has been recommended by the guidelines for GAC [8], a discrepancy may exist between forceps biopsy samples and dissected specimens because forceps biopsy do not represent the entire lesion [13]. Some studies have reported discrepancy rates after EMR ranging from 18.7 to 44.5 %, including final upgraded and downgraded pathology [4, 14]. Results of our study showed that 45 cases of GAC were subsequently upgraded to dysplasia or cancer after more aggressive diagnostic sampling, and even three cases repeatedly misdiagnosed as GAC by re-biopsy were eventually found to be carcinomas. This cautions that repeat biopsy for GAC can miss the actual true diagnosis. Therefore, EMR/ESD procedures are needed for definite diagnoses and treatments of GAC lesions. On the other hand, endoscopic resection could be associated with more potential complication such as bleeding or perforation and
Dig Dis Sci Table 4 Risk factors of cancer in univariate and multivariate analyses
Variable
Univariate analysis
Age
Multivariate analysis
p value
OR
95 % CI
p value
OR
95 % CI
0.063
1.03
0.99–1.07
0.150
0.001
5.35
1.94–14.71
0.019
3.63
1.23–10.66
0.001
4.32
1.79–10.42
0.030
2.86
1.11–7.38
Endoscopic findings Lesion size C10 mm Lesion type OR odds ratio, CI confidence interval
Depressed
Table 5 The correlation between the endoscopic and the histologic diagnosis Histologic diagnosis (n)
Endoscopic diagnosis (n) Gastritis
Adenoma
Cancer
Gastritis
41
4
Adenoma
2
4
3
9
Cancer
7
2
27
36
50
10
36
96
Total
6
Total 51
n number
increase medical expenses compared with those of repeat biopsy. However, these complications could be managed safely and successfully by non-surgical methods [15]. And, we noted, in Eastern Asia including Korea, that medical costs were relatively inexpensive, but the prevalence rate of gastric cancer was higher than those of Western as stated above. Clinicians should consider the risk factors for neoplasm related to the host, endoscopic, or histologic characteristics. In several studies, age, grade of coexisting atrophic gastritis, depressed type, and large size were identified as risk factors for malignant transformation [16–19]. Our study also showed that, in univariate analysis, age (OR 1.04, 95 % CI 1.01–1.07), lesion size of 10 mm or greater (OR 3.94, 95 % CI 1.61–9.61), and lesion with depressed type (OR 2.50, 95 % CI 1.09–5.72) were risk factors for neoplasm. Coexisting atrophic gastritis was not included in analysis because most of the patients in our study had this factor. The World Health Organization’s International Agency for Research on Cancer classified H. pylori as a Group 1 or definite carcinogen [20]. Some studies have supported the concept of intestinal metaplasia and H. pylori infection as precursors for intestinal type gastric cancer [21–23]. In our study, presence of H. pylori (OR 2.83, 95 % CI 1.11–7.25) was a risk factor for neoplasm in univariate analysis. However, intestinal metaplasia showed no significant influence in any analysis. More specifically, we investigated risk factors for cancer. According to our results, in multivariate analysis, lesion size of 10 mm or greater (OR 3.63, 95 % CI 1.23–10.66) and lesion with
depressed type (OR 2.86, 95 % CI 1.11–7.38) were independent risk factors for cancer. The specificity related to cancer predictive value of lesion with depressed type, lesion size of 10 mm or greater, and both factors was 66.7, 72.9, and 73.3 %, respectively. These results demonstrate that the presence of these risk factors alone is insufficient to predict cancer. In our study, however, the rate of concordant diagnoses of ‘‘neoplasm’’ between the initial endoscopic and the final histologic diagnoses was 78.2 %. A recent study also showed high correlation between the clinic, endoscopic, and the final histologic diagnoses of GAC lesions, related with adenoma or cancer [24]. So, we assumed that there were more additional informations for endoscopic diagnoses of GAC lesions, although our study did not investigate. Other study reported that red coloration and friability in terms of endoscopic findings of GAC lesions were significant in the neoplastic group [25]. So, we suggest that an initial endoscopic diagnosis of GAC lesions can be more accurate and sharper, when endoscopists consider not only lesions of depression and size but also lesions of coloration, friability, and surface changes related with neoplasm. Also, in recent, clinicians can consider to perform magnifying narrowband imaging (M-NBI) for GAC lesion. Several studies demonstrated that M-NBI is more accurate and useful than conventional white-light imaging for differentiating between EGC of small, flat, and depressed lesions and gastritis by investigation of microvascular and mucosal patterns [26, 27]. Indeed, there were seven cases of gastritis-like cancer and six cases of cancerlike gastritis by conventional imaging in our study. So, using M-NBI would contribute to minimizing unnecessary biopsies or endoscopic resection of non-cancerous lesions taken in order to rule out gastritis-like cancer. Nevertheless, taking a histologic specimen cannot be omitted, because the histologic decision is definitely needed for diagnosis and treatment of a lesion suspicious of neoplasm. In clinical practice, four to six forceps biopsy specimens are taken from a lesion suspicious of malignancy. In our study, sixty-nine cases (71.8 %) including lesions suspicious of neoplasm reached the final diagnosis by approximately four pieces of re-biopsy. And, a prospective trial
123
Dig Dis Sci
emphasized that at least seven specimens of biopsy should be obtained for increasing the yield to greater than 98 % of malignancy diagnosis [28]. In another study, accurate acquisition of biopsy specimens comprising at least three to four pieces from the lesion of mucosal change is important; however, there is no international agreement regarding how many specimens are sufficient [29]. And, if the lesion is resectable by ESD or EMR, many of biopsies increase complication of endoscopic resection due to fibrosis and ulcers following the biopsies. Considering that mentioned above, clinicians should perform careful conventional endoscopy to investigate mucosal characteristics: depression, size, coloration, friability, and surface changes. Also, if needed and possible, M-NBI could be performed for additional investigation of microvascular and mucosal patterns. And then, clinicians must discuss with pathologists whether a GAC lesion is more suspicious of dysplasia/cancer in histologic finding or not. Finally, if a GAC lesion is resectable by ESD/EMR according to the extended endoscopic guidelines [10] and more malignant potential in totally endoscopic and histologic findings, an endoscopic resection could be more reasonable and comprehensive than repeated biopsies. Our study has several limitations. First, because this study was conducted retrospectively, there is a possibility of selection bias. However, in order to reduce this bias, we enrolled mostly GAC patients in our study and all pathologic slides were reviewed by two independent pathologists. Second, there were no standardized criteria regarding the next step that should be taken by individual clinicians across enrolled patients who underwent EMR/ESD, OP, or follow-up. However, our study indeed demonstrated the significance of the GAC lesion in clinical practice, considering the difficulty of conducting a prospective randomized controlled trial. Third, in cases of GAC, pathologists should consider performing immunohistochemical stains for p53 and Ki67. These stains are helpful for diagnoses of gastric adenoma or adenocarcinoma [30, 31]; However, our study did not investigate because they are neither definitive tools nor diagnostic factors of pathologic grading classification [1, 8]. Also, a recent study showed there was no significant difference in p53 or Ki67 expression between intestinal metaplasia with atypia and gastric carcinoma [32]. Finally, most patients diagnosed as gastritis on first re-biopsy were followed for a period of less than 2 years or lost to follow-up in our center. Therefore, the final diagnosis of these patients could be inadequate. However, we assumed that most of them received proper follow-up, because the Korea National Cancer Screening Program recommends and conducts biennial gastric cancer screening using upper gastrointestinal series or endoscopy in order to reduce the high rate of gastric cancer mortality in Korea [33]. Ultimately, conduct
123
of a further large-scale prospective study will be necessary in order to overcome these limitations. In conclusion, considering the neoplastic change of GAC, EMR/ESD procedures may be more useful for definite diagnoses and en bloc treatments, especially in cases accompanied by significant and independent risk factors, such as lesion size of 10 mm or greater and lesion with depressed type. Conflict of interest
None.
References 1. Dixon MF. Gastrointestinal epithelial neoplasia: Vienna revisited. Gut. 2002;51:130–131. 2. Bray F, Ren JS, Masuyer E, Ferlay J. Global estimates of cancer prevalence for 27 sites in the adult population in 2008. Int J Cancer. 2013;132:1133–1145. 3. Misdraji J, Lauwers GY. Gastric epithelial dysplasia. Semin Diagn Pathol. 2002;19:20–30. 4. Kim YJ, Park JC, Kim JH, et al. Histologic diagnosis based on forceps biopsy is not adequate for determining endoscopic treatment of gastric adenomatous lesions. Endoscopy. 2010;42:620–626. 5. Hull MJ, Mino-Kenudson M, Nishioka NS, et al. Endoscopic mucosal resection: an improved diagnostic procedure for early gastroesophageal epithelial neoplasms. Am J Surg Pathol. 2006;30:114–118. 6. Gotoda T. Endoscopic resection of early gastric cancer: the Japanese perspective. Curr Opin Gastroenterol. 2006;22:561–569. 7. Gotoda T, Yamamoto H, Soetikno RM. Endoscopic submucosal dissection of early gastric cancer. J Gastroenterol. 2006;41:929–942. 8. Kim HG, Jin SY, Jang JJ, et al. Grading system for gastric epithelial proliferative disease. Korean J Pathol. 1997;31:389–400. 9. Hamilton SRAL. World Health Organization classification of tumours: pathology and genetics of tumours of the digestive system. Lyon: IARC Press; 2000. 10. Japanese Gastric Cancer A. Japanese gastric cancer treatment guidelines 2010 (ver. 3). Gastric Cancer. 2011;14:113–123. 11. Lee H, Yun WK, Min BH, et al. A feasibility study on the expanded indication for endoscopic submucosal dissection of early gastric cancer. Surg Endosc. 2011;25:1985–1993. 12. Imagawa A, Okada H, Kawahara Y, et al. Endoscopic submucosal dissection for early gastric cancer: results and degrees of technical difficulty as well as success. Endoscopy. 2006;38:987–990. 13. Yoon WJ, Lee DH, Jung YJ, et al. Histologic characteristics of gastric polyps in Korea: emphasis on discrepancy between endoscopic forceps biopsy and endoscopic mucosal resection specimen. World J Gastroenterol. 2006;12:4029–4032. 14. Lee CK, Chung IK, Lee SH, et al. Is endoscopic forceps biopsy enough for a definitive diagnosis of gastric epithelial neoplasia? J Gastroenterol Hepatol. 2010;25:1507–1513. 15. Jeon SW, Jung MK, Kim SK, et al. Clinical outcomes for perforations during endoscopic submucosal dissection in patients with gastric lesions. Surg Endosc. 2010;24:911–916. 16. Park DI, Rhee PL, Kim JE, et al. Risk factors suggesting malignant transformation of gastric adenoma: univariate and multivariate analysis. Endoscopy. 2001;33:501–506.
Dig Dis Sci 17. Nakamura K, Sakaguchi H, Enjoji M. Depressed adenoma of the stomach. Cancer. 1988;62:2197–2202. 18. Jung MK, Jeon SW, Park SY, et al. Endoscopic characteristics of gastric adenomas suggesting carcinomatous transformation. Surg Endosc. 2008;22:2705–2711. 19. Rugge M, Leandro G, Farinati F, et al. Gastric epithelial dysplasia. How clinicopathologic background relates to management. Cancer. 1995;76:376–382. 20. Schistosomes, liver flukes and Helicobacter pylori. IARC Working Group on the Evaluation of Carcinogenic Risks to Humans. Lyon, 7–14 June 1994. IARC monographs on the evaluation of carcinogenic risks to humans/World Health Organization, International Agency for Research on Cancer. 1994; 61:1–241. 21. Shimoyama T, Fukuda S, Tanaka M, Nakaji S, Munakata A. Evaluation of the applicability of the gastric carcinoma risk index for intestinal type cancer in Japanese patients infected with Helicobacter pylori. Virchows Arch. 2000;436:585–587. 22. Parsonnet J, Vandersteen D, Goates J, Sibley RK, Pritikin J, Chang Y. Helicobacter pylori infection in intestinal- and diffusetype gastric adenocarcinomas. J Natl Cancer Inst. 1991;83:640–643. 23. Eslick GD, Lim LL, Byles JE, Xia HH, Talley NJ. Association of Helicobacter pylori infection with gastric carcinoma: a metaanalysis. Am J Gastroenterol. 1999;94:2373–2379. 24. Kim SI, Han HS, Kim JH, et al. What is the next step for gastric atypical epithelium on histological findings of endoscopic forceps biopsy? Dig Liver. 2013;45:573–577.
25. Lee H, Kim H, Shin SK, et al. The diagnostic role of endoscopic submucosal dissection for gastric lesions with indefinite pathology. Scand J Gastroenterol. 2012;47:1101–1107. 26. Yao K, Iwashita A, Tanabe H, et al. Novel zoom endoscopy technique for diagnosis of small flat gastric cancer: a prospective, blind study. Clin Gastroenterol Hepatol. 2007;5:869–878. 27. Ezoe Y, Muto M, Uedo N, et al. Magnifying narrowband imaging is more accurate than conventional white-light imaging in diagnosis of gastric mucosal cancer. Gastroenterology. 2011;141:2017e3–2025e3. 28. Graham DY, Schwartz JT, Cain GD, Gyorkey F. Prospective evaluation of biopsy number in the diagnosis of esophageal and gastric carcinoma. Gastroenterology. 1982;82:228–231. 29. Haber MM. Gastric biopsies: increasing the yield. Clin Gastroenterol Hepatol. 2007;5:160–165. 30. Shiao YH, Rugge M, Correa P, Lehmann HP, Scheer WD. p53 alteration in gastric precancerous lesions. Am J Pathol. 1994;144:511–517. 31. Zheng H, Tsuneyama K, Cheng C, et al. An immunohistochemical study of P53 and Ki-67 in gastrointestinal adenoma and adenocarcinoma using tissue microarray. Anticancer Res. 2006;26:2353–2360. 32. Zheng Y, Wang L, Zhang JP, Yang JY, Zhao ZM, Zhang XY. Expression of p53, c-erbB-2 and Ki67 in intestinal metaplasia and gastric carcinoma. World J Gastroenterol. 2010;16:339–344. 33. Yoo KY. Cancer control activities in the Republic of Korea. Jpn J Clin Oncol. 2008;38:327–333.
123
