Poor Agreement between Preoperative Biopsies and Pathological Resection Findings in IBD-Associated Dysplasia

Special Article - Inflammatory Bowel Disease

Austin J Gastroenterol. 2015;2(1): 1032.

Poor Agreement between Preoperative Biopsies and Pathological Resection Findings in IBD-Associated Dysplasia

Althumairi AA1*, Monn MF2, Derck JE3, Wick EC1, Efron JE1, Lazarev MG4, Montgomery EA5, Gearhart SL1

1Department of Surgery, Johns Hopkins Medical Institutions, Baltimore, MD, USA

2Bloomberg School of Public Health, Johns Hopkins University, Baltimore, MD, USA

3School of Medicine, University of Michigan, Ann Arbor, MI, USA

4Department of Gastroenterology, Johns Hopkins Medical Institutions, Baltimore, MD, USA

5Department of Pathology, Johns Hopkins Medical Institutions, Baltimore, MD, USA

*Corresponding author: Althumairi AA, Department of Surgery, Johns Hopkins Medical Institutions, Blalock 618 600 North Wolfe Street, Baltimore, MD 21287

Meeting Presentation: This paper has been presented as podium presentation at the annual meeting of the American Society of Colon and Rectal Surgeons, Hollywood, FL May 17 - 21, 2014

Received: December 20, 2014; Accepted: February 02, 2015; Published: February 04, 2015


Background: The decision to perform a proctocolectomy in patients with inflammatory bowel disease (IBD)-associateddysplasia is based on the degree of dysplasia on endoscopic biopsies. Unexpected pathological findings at the time of surgery can be troubling to patients. Therefore, we aimed to determine the extent of pathological agreement between endoscopic biopsies and surgical resection specimens in patients undergoing surgery at a tertiary referral center and to identify risk factors for poor agreement.

Methods: Aretrospective review of patients who underwent surgery for IBDassociated dysplasia was performed. Data including demographics, disease history, endoscopic surveillance, and procedure type were collected. Risk factors for poor agreement were assessed using regression analysis.

Results: 81 patients were identified; 60 (74%) male with a mean age of 54 years. Ulcerative colitis was seen in 70 (86%) while 11 (14%) had Crohn’s disease. In the colectomy specimens, newly diagnosed adenocarcinoma was identified in 16 (20%) and no dysplasia was seen in 16 (20%). Agreement between preoperative endoscopic biopsies and whole specimen pathology occurred in 33 (41%) patients (r=0.17, p=0.14). Highest agreement was flat low grade dysplasia, while lowest agreement was indefinite dysplasia and polypoid low grade dysplasia. The diagnosis of cancer was more common with a preoperative diagnosis of high grade dysplasia. A repeat endoscopic evaluation at our institution was associated with lower likelihood of the findings of no dysplasia on the final surgical specimen.

Conclusions: Agreement between preoperative biopsies and final pathology remains low, however, newer endoscopic techniques may provide better pathological correlation. This study highlights the necessity of preoperative counseling and joint decision making prior to surgery for dysplasia in IBD.

Keywords: Inflammatory Bowel Disease; Dysplasia; Colorectal Cancer; Endoscopy; Surveillance; Colectomy


IBD: Inflammatory Bowel Disease; HGD: high grade dysplasia; LGD: low grade dysplasia; UC:Ulcerative Colitis; CD: Crohn’s disease; CPT: Current Procedural Terminology; ICD-9:International Classifications of Disease 9th version; ALM: adenoma like mass; DALM: dysplasia associated lesion or mass; PSC: primary sclerosing cholangitis;ANOVA: Analysis of Variance; OR: Odds Ration


Inflammatory Bowel Disease (IBD) was first linked to increased risk of colorectal cancer in 1925. Since then, a recognized transition in IBD is the evolution of reparative changes in the setting of inflammation to dysplasia and then carcinoma [1]. The reported risk of detecting invasive carcinoma in patients with a preoperative diagnosis of high grade dysplasia (HGD) is 42 - 58% and is 16 - 27% in patients with pre-operative biopsies showing low grade dysplasia (LGD) [2,3]. In the presence of a dysplastic polypoid lesion, the reported risk has been even more variable, ranging from 19% - 42% [4,5]. Therefore, IBD patients are prescribed to strict endoscopic surveillance regimens to screen for dysplasia beginning 8 years after initial diagnosis [5]. Surveillance regimens include 4 quadrant biopsies collected every 10 cm throughout the colon [6,7,8]. Screening and surveillance colonoscopy have been shown to reduce the development of colorectal cancer in IBD [9,10]. Similar to non- IBD-associated colorectal cancer patients, the mortality rates in patients with IBD-associated neoplasia may be reduced by detection and early removal [9,11,12].

The decision to perform a colectomy on IBD patients for dysplastic findings is complex for both the patient and the health care provider. Recommendations for colectomy in IBD-associated dysplasia have been outlined by experts in this field [10]. Currently, the decision is based on endoscopic surveillance biopsies and often there is little opportunity for shared decision making due to a lack of patients’ understanding of their disease [13,14,15]. Furthermore, histological diagnosis of dysplasia in IBD can also be difficult with significant inter-observer variation among pathologists [16]. The goal of this study is to assess level of agreement between preoperative endoscopic biopsies and postoperative pathological findings and to identify risk factors for poor agreement. We hope to gain a further understanding of the expected pathological outcome following resection for IBD-associated dysplasia at a tertiary referral center, and to use this information in the preoperative counseling with the patients to enhance shared decision making.

Materials and Methods


We performed a retrospective study of patients with IBD who were referred to a tertiary center for surgical colectomy and underwent resection at Johns Hopkins Hospital from 2003 to 2013. The Johns Hopkins pathology databaseand physician’s billing codes were utilized to identify patients. Patients were included if they underwent laparoscopic or open restorative proctocolectomy, total proctocolectomy with end ileostomy, or subtotal colectomy. The following procedure codes were utilized: Current Procedural Terminology (CPT) 44150, 44155, 44157, 44158, 44210, 44211, 44212, and 45113. The following International Classification of Disease codes (ICD-9) for IBD were matched with CPT codes: 555.0-555.9, 556.0-556.6, 556.8, and 556.9. All patients who underwent evaluation or re-evaluation of their pathology at Johns Hopkins Hospital and were found to have IBD-associated dysplasia or a dysplasia associated lesion or mass (DALM) were eligible for inclusion in the study. Patients considered to have an adenoma like mass (ALM) were not included in this study. Patient variables and disease characteristics collected included age, gender, race, disease type, time with IBD, time with dysplasia, personal history of primary sclerosing cholangitis (PSC), and personal or family history of colorectal cancer. This study was approved by the Johns Hopkins institutional review board.

Endoscopic evaluation

Patients who underwent outside endoscopic evaluation which lacked a sufficient number of random biopsies with regards to assessment of the entire colon, were offered repeat surveillance at out institution. Surveillance endoscopy at Johns Hopkins is performed by a group of gastroenterologists specializing in IBD. Guidelines published by the American College of Gastroenterology are followed [17]. For the purpose of dysplasia surveillance, patients who have had a diagnosis of ulcerative colitis (UC) for at least 8 years, colonoscopy is performed with four quadrant biopsies taken every 10 cm. Repeat endoscopic surveillance is performed approximately every 1 to 3 years.

For the appearance of a mass or raised lesion on endoscopy performed at our institution, a biopsy of the mass and surrounding tissue, tattooing, and/or polypectomy was performed at the discretion of the endoscopist. The decision to label these lesions as polypoid was determined by an endoscopist and was dependent upon the ability of the polyp to be removed by polypectomy. All other LGD found on random biopsy or in non-polypoid lesions were labelled flat LGD. Furthermore, chromoendoscopy, narrow band imaging, and high definition endoscopy was performed at the discretion of endoscopist.

Pathologic evaluation

All preoperative biopsies performed elsewhere were re-evaluated at Johns Hopkins Medical Institution by experts in gastrointestinal and IBD pathology.Specimens obtained by referring institutions or at Johns Hopkins were initially read by one pathologist and the diagnosis was then confirmed by showing the case at the daily interdepartmental quality assurance conference. Preoperative biopsies were classified into one of the following 4 categories: indefinite for dysplasia, flat LGD, polypoid LGD, high grade dysplasia (HGD). For the purpose of this analysis, classification was based upon the highest degree of dysplasia found.

All operative specimens were reviewed by pathologists specializing in gastrointestinal and IBD pathology at Johns Hopkins Medical Institutions. Postoperative colectomy diagnoses were categorized into 6 groups: no dysplasia, indefinite for dysplasia, flat LGD, polypoid LGD, high grade dysplasia (HGD), or invasive adenocarcinoma. All operative specimens underwent standardized tissue sampling with sectioning every 10 cm. Additional histological sections were performed for any focal lesions, ulcers, or polyps identified on gross inspection of the entire specimen by pathologist at the time of resection. For the purpose of this study these additional sections were labeled as sections of areas of interest. All sections taken were examined at 5 micron intervals. Colectomy specimens were evaluated to see if the final pathology diagnosis agreed with the preoperative diagnosis.

Statistical methods

Patient characteristics including demographics and medical history were analyzed. Preoperative diagnosis on biopsy specimens were compared with postoperative whole specimen pathology. The percent agreement overall and within each biopsy or specimen category was examined separately. Agreement between preoperative and postoperative diagnoses was assessed using Lin’s concordance correlation coefficient to test correlation between the two specimens [18].

Using univariate analysis, we examined whether there was a relationship between preoperative factors and three specific outcomes: 1) no dysplasia, 2) invasive carcinoma on final pathology, and 3) biopsy agreement with final pathology. The variables that we examined were age, gender, diagnosis (UC vs. Crohn’s disease (CD)), time with IBD, time with dysplasia, history of PSC, family history of colorectal cancer, family history of IBD, endoscopy, biopsy pathology, polypectomy and numbers of total histological sections and sections of area of interest. Additional multivariate logistic regressions were performed examining the impact of the variables used in univariate analysis on the outcomes of 1) no dysplasia on final pathology and 2) cancer on final pathology. In the no dysplasia on final pathology model, history of PSC was excluded as it was collinear with the outcome.

Comparisons of the number of total histological sections and sections of interest with the disease type, final pathology, type of surgery and level of pathological agreement were performed using descriptive statistics, specifically the Mann-Whitney test and Analysis of Variance (ANOVA). For all analyses, p<0.05 was used as the accepted level of statistical significance.


Patient characteristics

A total number of 363 cases meeting the study criteria were performed between the years 2003 and 2013. Of these procedures, a total of 81 (22%) patients underwent resection for IBD-associated dysplasia and were included in this study. The clinical features of these patients are summarized in table 1. The majority of patients was male, had UC, and underwent a total proctocolectomy. Furthermore, 51 (63%) of patients had a repeat endoscopic evaluation at our institution and of these patients, 18 (35%) were evaluated with newer methods of endoscopy