TY - JOUR AU - MD, David Binion, AB - Background Subjective physician assessment is the cornerstone of routine ulcerative colitis (UC) management. Endoscopic and histologic assessment of UC provides objective measures of inflammatory disease activity. The level of agreement between physician impression of UC activity and endoscopic disease activity has not been evaluated. The aim was to assess the level of agreement between physician's clinical impression of UC disease activity and endoscopic and histologic findings of inflammation. Methods Using the Medical Archival Retrieval System at the University of Pittsburgh Medical Center, we reviewed clinical information on all UC patients between 1995 and 2008 who had clinic visits recorded prior to colonoscopy. Clinical UC disease activity was defined by the physician's clinical impression and the endoscopic and histologic activity by colonoscopy with biopsy. The level of agreement between colonoscopy assessment of UC with histologic and clinical assessment was determined by sensitivity, specificity, positive and negative predictive values, and the kappa coefficient. Results There were 369 UC patients who had a clinic visit proximate to a colonoscopy. The mean age of patients was 46 ± 16 years (50% female). The performance of clinical impression in recognizing disease activity, as determined by endoscopy, was relatively poor: sensitivity = 56.0%, predictive value negative = 56.8%, kappa coefficient = 0.35. In contrast, the performance of histological evaluation in recognizing disease activity was markedly better: sensitivity = 93.5%, predictive value negative = 89.1%, kappa coefficient = 0.70. Conclusions The physician's clinical impression of UC activity shows poor agreement with endoscopy and histology, with over one-third of patients with chronic inflammation underrecognized by clinical impression. The consequences of underestimated UC activity by clinical assessment may include undertreatment of active disease and uncontrolled chronic inflammation. (Inflamm Bowel Dis 2011;) physician assessment, ulcerative colitis, endoscopic correlation Ulcerative colitis (UC) is a lifelong chronic inflammatory disorder of the large intestine characterized by variable disease activity. Patients may exhibit periods of remission, mild, moderate, and severe colitis. The consequence of long-term colitis are cumulative colonic damage and an increased risk of colorectal cancer, both of which may require surgery. Current treatment goals emphasize implementing induction treatment for active UC and use of medications to maintain remission and prevent flares of colitis. Clinical factors that guide therapeutic decision-making begin with patient symptoms and physician assessment. At present, there are no ideal, objective measures of inflammatory activity that are uniformly employed in the assessment of UC patients. Assessment of UC disease activity in clinical research trials has relied on a variety of subjective indices or scoring systems.1,–4 Endoscopically derived scores and assessment provide more objective definitions of UC activity.5,–7 Historically, treatment decisions for UC have been made by the physician's clinical assessment of the patient. This approach has relied heavily on the patient's ability to communicate symptoms indicative of active colitis. In contrast, endoscopic assessment with pinch biopsies obtained for histologic analysis may provide a more objective measure of inflammatory activity. Whether endoscopic assessment for mucosal inflammation and healing in response to medical treatment should be routinely performed in clinical practice has been debated.8,9 The concept of silent inflammatory activity in Crohn's disease whereby patients feel well but have endoscopically and histologically active disease has gained momentum.10 In this study we evaluated the ability of gastroenterologists to recognize endoscopic and histologic disease activity based on clinical assessment of the patient. Patients and Methods All UC patients ≥18 years of age seen at the University of Pittsburgh Medical Center (UPMC) Digestive Disorder Center between January 1, 1995 and June 30, 2008 were identified by ICD-9 CM (556.0–556.9) using the Medical Archival Retrieval System (MARS).11 This system contains a permanent archive of both the electronic ambulatory and hospital data at UPMC over the past 20 years. Clinical records including laboratory results, medications, radiology reports, transcribed notes, and other types of information are fed to the archive continuously. The MARS repository houses more than 215 million clinical reports and was originally developed in 1988 to integrate computer systems that support medical care at the departmental level. MARS is implemented in a UNIX-based, distributed parallel processing environment. MARS accepts all machine-readable data without requiring structure at the point of data entry. MARS is indexed on every word and number in the database with parallel use of a proximity operator, which makes it possible to recognize individual terms as well as multiword terms in structured or unstructured data. This capability provides the means to study conditions that were not considered or defined at the time the data was collected. To meet Health Insurance Portability and Accountability Act (HIPAA0 guidelines and ensure patient confidentiality, all data was deidentified (De-ID, University of Pittsburgh) using an honest broker system. This study met the criteria for exemption of informed consent by the University of Pittsburgh Institutional Review Board. Inclusion criteria for the study were patients with a diagnosis of UC confirmed by endoscopy and histopathology. All evaluable UC patients had a clinic visit within 6 months of a colonoscopy, i.e., a paired visit. The clinical notes, endoscopy reports, and pathology reports were obtained through a search by a certified honest broker. Three investigators separately recorded the physician's clinical impression of UC activity, the endoscopy findings, and histology. For patients with more than one clinic visit we selected the most recent clinic visit for analysis. The note from the clinic visit was evaluated to see if the physician deemed UC to be clinically active or in clinical remission. The following terms were searched and recorded to define the physician's impression of disease activity, “exacerbation of UC, active UC, mildly, moderately, severely active UC, quiescent UC, inactive UC, and UC in remission.” Notes that did not indicate a clear clinical impression were not included in the analysis. The corresponding colonoscopy report was reviewed for active UC. This was defined by an impression including any of the following terms: “erythema, edema, friability ulceration, inflammation, active disease, and/or active UC.” The corresponding pathology reports were also reviewed for active UC. These were considered to display active disease if the report contained the terms: “cryptitis, crypt abscesses, infiltration of plasma cells or neutrophils,” or the pathology impression was “consistent with active UC.” Multiple characteristics were recorded for each UC patient including age, gender, tobacco use, current UC medication, family history of inflammatory bowel disease (IBD), duration of UC, extent of disease (proctitis, left-sided colitis, pancolitis), and the presence of extraintestinal manifestations. Statistical Analysis To describe the study cohort we classified the UC patients based on the results of colonoscopy and clinical assessment into four mutually exclusive groups: 1) no evidence of disease (i.e., remission); 2) endoscopic and/or histologic evidence of disease in the absence of clinical evidence; 3) clinical evidence of disease in the absence of endoscopic and histologic evidence; and 4) endoscopic and/or histologic evidence of disease concomitant with clinical evidence. For these groups and the full study cohort, continuous variables are presented as mean, standard deviation (SD); categorical variables are presented as percentages. The primary goal of the analysis was to assess the level of agreement between the finding from endoscopy (i.e., gold standard) and both histologic and clinical assessment. To this end, five measures of concordance were calculated, as defined in Table 1. These included sensitivity, specificity, predictive value positive, predictive value negative, and the kappa coefficient. Results were calculated for the full cohort and within subgroups (e.g., age, gender, etc.). Interaction terms within logistic regression analysis models were used to evaluate whether the performance of histology and clinical impression in predicting endoscopically defined UC activity varied by patient characteristics (age, gender, duration of UC). For interpretation of level of agreement, kappa values were defined as: poor (less than 0.20); fair (0.20–0.40); moderate (0.40–0.60); good (0.60–0.80); and very good (0.80–1.00).12 Table 1 Measures of Concordance Between Endoscopy (Gold Standard) and Concurrent Histologic/Clinical Evaluation     View Large Table 1 Measures of Concordance Between Endoscopy (Gold Standard) and Concurrent Histologic/Clinical Evaluation     View Large Results Descriptive Characteristics Between 1995 and 2008 there were a total of 634 UC patients identified. Of these, 369 patients had a clinic visit in proximity to a colonoscopy and were included in the study; 265 patients were excluded due to lack of recorded clinical impression (n = 38) or a clinic visit within 6 months of colonoscopy (n = 227). The median time between the paired clinic visit and the colonoscopy was 2.6 months (range 0.5–5.2 months). The indications for colonoscopy were assessment of disease activity (n = 236; 64%) and/or surveillance for dysplasia/cancer (n = 133; 36%). The median time from clinic visit to colonoscopy was shorter, 1.1 months (range 0.5–1.9 months), for patients with active symptoms who had a subsequent colonoscopy for assessment of disease activity compared to those without symptoms undergoing colonoscopy for surveillance, 3.4 months (range 0.9–5.2 months). There were 144 patients (39%) who also had a clinical impression recorded on the day of colonoscopy; 99 of these patients had active symptoms and 45 had no symptoms at the preceding clinic visit. In this subset of patients there was no difference between the recorded clinical impression at the time of the preceding clinic visit and colonoscopy. As seen in Table 2, the mean age of patients was 46 ± 16 years with 18% being under the age of 30 and 22% being age 60 or older. The study cohort was equally split by gender, with extent of UC being distributed as: left-sided (21%); pancolitis (71%); and proctitis (8%). Table 2 Characteristics of Study Cohort     View Large Table 2 Characteristics of Study Cohort     View Large The duration of UC was unknown in 58% of all patients. Among those with available data, 62% had UC for less than 10 years and 38% had disease for more than 10 years. Approximately 8% of all patients had a family history of UC, most often in first-degree rather than second-degree relatives (data not shown). We analyzed treatment modailities in the UC study cohort. All patients were receiving at least one medication for UC at the time of the clinic visit/colonoscopy. When we analyzed the breakdown of medications used for the treatment of UC, we found that the patients were often using different combinations of agents. Approximately 1 in 5 patients (21%) were taking corticosteroids, about half (53%) were taking a 5-aminosalicylate (5-ASA) agent, 23% were taking sulfasalazine, and 24% were taking a 6-mercaptopurine or azathioprine agent in various combinations. Given that the study duration was 1995 to 2008 and the Food and Drug Administration approval of infliximab for UC was late in 2005, infliximab use was limited in our study population and not included in the analysis. We next analyzed the patterns of disease activity which were identified in the clinical record. There were 93 patients (25% of the total UC cohort) without evidence of active UC. These individuals were older, on average, (mean age = 50 years, range 20–80 years) than the remaining patients with evidence of active UC (mean age = 45 years, range 18–86 years) (P = 0.001). The group of patients who had evidence of clinical activity but no confirmation endoscopically was small (n = 35; 9.5% of the total UC cohort). This group of patients with clinical indication of active UC in the absence of endoscopic and histologic evidence was overrepresented by females (69%) and those having duration of UC for more than 10 years. Corticosteroid use was highest (41.2%) in the group of 114 patients with concurrent endoscopic/histologic/clinical evidence of active UC and lowest in those (6.5%) classified as being in remission. Level of Agreement Between Measures of Disease Activity Among the 369 patients, there was agreement between endoscopy and clinical impression in 207 patients (56.1%). This included 93 patients with no evidence of active disease and 114 patients with evidence of active UC by both measures (Fig. 1). Conversely, in terms of lack of agreement, 127 patients had endoscopic evidence of active inflammatory disease but had been deemed to be in clinical remission. In addition, there were 35 patients who had no endoscopic evidence of active UC, but had a clinical impression of active disease. Figure 1 View largeDownload slide Classification of the 369 study patients according to endoscopy versus clinical impression of UC activity. Figure 1 View largeDownload slide Classification of the 369 study patients according to endoscopy versus clinical impression of UC activity. In the full study cohort the degree of agreement with endoscopic evaluation was much higher for histologic evaluation (kappa = 0.70) than for clinical assessment (kappa = 0.35) (Table 3). This can be interpreted as only “fair” agreement between clinical assessment and endoscopy compared to “good” agreement between histology and endoscopy. In examining the relatively poor performance of clinical assessment, the sensitivity of only 56% means that nearly half of all cases of active UC, as determined endoscopically, were not identified on the basis of clinical evaluation. Also noteworthy is that nearly 1 in 5 patients (19%) classified as having clinical indication of disease showed no concurrent evidence from endoscopic evaluation (i.e., false-positive). Table 3 Measures of Concordance for Clinical and Histologic Assessment with Endoscopic Evaluation     View Large Table 3 Measures of Concordance for Clinical and Histologic Assessment with Endoscopic Evaluation     View Large A limitation of this analysis was that there was a wide range of time between the clinic visit and colonoscopy in the full study cohort; as short as 0.5 months to as long as 5.2 months. Since it is conceivable that the clinical impression could change over time, we analyzed the subset of patients in whom the clinical assessment was within 1 month of endoscopy. In this analysis a kappa of 0.76 was observed between endoscopic evaluation and histologic evaluation compared to a kappa of 0.25 between endoscopic evaluation and clinical assessment. These results were consistent with the full study cohort. Subgroup Analyses Analyses by subgroups showed several interesting patterns. First, the utility of clinical assessment showed an inverted “U” shape relationship with endoscopy with only “fair” agreement (kappa estimates from 0.21–0.28) in patients under the age of 40 and older than 60 compared to “moderate” agreement (kappa estimates of 0.49 and 0.52) among patients age 40 to 59 (Table 3; Fig. 2). In contrast, the extent of agreement between histology and endoscopy increased consistently as patient age increased (P-value for interaction = 0.05). Agreement between histology and endoscopy was only “moderate” in patients age 18 to 29 (kappa = 0.49) compared to “very good” in patients 60 and older (kappa = 0.84). Figure 2 View largeDownload slide Kappa estimates of the extent of agreement between endoscopy and clinical assessment (left side) and histologic assessment (right side) overall and by age. Figure 2 View largeDownload slide Kappa estimates of the extent of agreement between endoscopy and clinical assessment (left side) and histologic assessment (right side) overall and by age. The agreement between clinical assessment and endoscopy was “fair” (kappa = 0.27) in females compared to “moderate” (kappa = 0.43) in males (P-value for interaction = 0.02). In contrast, nearly identical “good” levels of agreement between histology and endoscopy were observed for females and males (kappa = 0.70 and 0.71, respectively). Finally, there was an indication that agreement between clinical assessment and endoscopy was better (P-value for interaction = 0.12), albeit still suboptimal, among patients with duration of UC less than 10 years (kappa = 0.48) compared to more than 10 years (kappa = 0.26). Discussion The present study found agreement between physician clinical impression of UC activity and endoscopic mucosal findings in only 56% of our patient cohort. The clinical impression underestimated inflammatory activity in approximately one-third of UC patients and overestimated disease activity in 10%. We believe that the long duration of the study and number of physicians involved provides an experience similar to that of practicing gastroenterologists. Our results imply that symptom-based assessment of UC is a poor predictor of mucosal inflammation. The importance of detecting persistent, subclinical inflammation in the long-term management of UC has recently been recognized. Analysis of UC cancer registries at St. Mark's Hospital, Mount Sinai Medical Center, and the University of Chicago have all suggested that endoscopic and histologic evidence of inflammation at the time of routine surveillance colonoscopies are independent risk factors for the development of dysplasia and cancer.13,–15 The fact that subclinical inflammation was found to be a culprit also implies that many UC patients may have “smoldering” colonic inflammation that is not detected unless a colonoscopy is performed. The importance of detecting subclinical inflammation in the progression to dysplasia/cancer is further emphasized, as this may be one of the few modifiable risk factors in the long-term care of IBD colitis. Thus, patients with long-standing, active inflammation may develop complications of disease and ultimately require surgery.16,17 The lack of agreement between clinical and endoscopic UC activity seemed to be most pronounced in females and patients in the younger and older age ranges. The kappa values for concordance in the 18–29 year old patients (k = 0.28), 30–39 (k = 0.21), and older than 60 (k = 0.25) were nearly half of those in the 40–49 and 50–59 year ages (k = 0.52 and 0.49, respectively). Similarly, the degree of agreement between the clinical and endoscopic UC activity was lower in females than in males (k = 0.27 versus 0.43). The reason for such disparity is not clear, but could suggest that gender and age may influence the reporting of UC symptoms. Although the detection of active inflammation in someone deemed to be in clinical remission is an important finding and the most common discrepancy found in our study, it is equally interesting that we identified a small subgroup of patients labeled as clinically active UC without any evidence of colonoscopic inflammation. These 35 patients (9.5% of the study population) were disproportionately female and had suffered from UC for longer than 10 years. The mechanisms leading to gastrointestinal symptoms in these individuals are not precisely defined. It is conceivable that these patients had altered neurosensory function or motility akin to irritable bowel syndrome, and, as a consequence, their bowel-related symptoms were not due to active inflammation. Importantly, symptom-based treatment decisions may lead to unnecessary treatment with corticosteroids or immunobiologics in this subgroup of patients. Currently, endoscopic evaluation of the colon is the best objective measure of UC activity. However, the cost and ability of patients to tolerate repeated colonoscopies pose a significant limitation on this approach for monitoring. Noninvasive modalities for assessment of UC disease activity are needed. Erythrocyte sedimentation rate (ESR) and C-reactive protein (CRP) are two serologic markers that can parallel inflammatory activity. CRP elevation has been shown in multiple studies to be associated with severe UC clinical disease activity.18,19 While CRP elevation is concerning for active inflammation in UC and signals more severe disease, it is not sufficiently sensitive for diagnosing patients with mild mucosal disease activity. CRP elevation is more sensitive at detecting more extensive disease with proximal colonic involvement.20,21 Fecal calprotectin and fecal lactoferrin have been more accurate in detecting endoscopic disease activity in UC compared to CRP.22,23 One study in particular reported that fecal lactoferrin had a high diagnostic accuracy (83%) for detecting active UC.22 Ultrasound, multidetector computed tomographic colonography, and magnetic resonance colonography have also shown promise as noninvasive studies of UC.24,–27 Until these tests have undergone further study and become more widely available, colonoscopy remains the procedure of choice for objectively defining UC activity. The limitations of this study are the retrospective nature of the study design, the lack of formal clinical, endoscopic, or histologic scoring, and the lack of a clinical impression at the time of colonoscopy. Without a recorded clinical assessment on the day of colonoscopy, it is conceivable that the patient's clinical activity would have changed between the clinic visit and the colonoscopy. The fact that patients with active symptoms had a colonoscopy within 2 months (median 1.1 months; range = 0.5–1.9 months) of the clinic visit makes it unlikely that the disease activity would significantly change by the time of the colonoscopy. Similarly, it is unlikely that the clinical impression of disease activity would significantly change for patients undergoing surveillance colonoscopy for dysplasia/cancer, despite the longer interval between the clinic visit and colonoscopy (median 3.4 months, range = 0.9–5.2 months). Ideally, the clinical impression would have been made on the day of colonoscopy, but given the retrospective nature and lack of reported clinical activity on the day of colonoscopy in most patients, this was not possible. Nonetheless, there were 39% of patients who had a recorded clinical impression at the time of colonoscopy, and there was no difference between the clinical impression at the clinic visit and colonoscopy. Additionally, we were able to analyze a subset of patients who had a recorded clinical impression within 1 month of colonoscopy and found no difference in kappa values for agreement between this subgroup and the overall study cohort. Therefore, we believe that the disease activity listed at the time of the clinic visit did not significantly change by the time the colonoscopy was performed. Despite the limitations, the study provides a “real world” assessment of UC based on physician clinical and endoscopic impression. The power of the UPMC electronic medical archival retrieval system is the ability to identify paired clinic visits with colonoscopy on 369 patients over the course of a 13 year time period. To the best of our knowledge, this is the largest such study to evaluate the level of agreement between clinical impression of UC activity and histology with endoscopy. In an attempt to limit bias in the unblinded nature of the retrospective study, we had separate investigators record clinical activity, endoscopic findings, and histopathology results. Ideally, our results would be tested in a large prospective study; however, to do such a study on 369 patients would be difficult. In conclusion, we found that the physician's clinical assessment of UC activity shows poor agreement with the colonoscopic assessment. Among patients with endoscopically defined objective quantification of mucosal inflammation in UC, only half of these patients were identified on the basis of symptom-based assessment. Whether this suboptimal agreement between symptom reporting and objective quantification of inflammatory activity remains stable over time is an important clinical question. Reliance on symptom assessment to guide treatment decisions means an over- or underestimation of disease activity in the majority of UC patients. Until better surrogate markers or noninvasive gastrointestinal imaging studies for disease activity have been validated and made readily available, colonoscopy remains the most accurate method to assess mucosal inflammation in UC patients. References 1 Truelove SC, Witts LJ. Cortisone in ulcerative colitis: preliminary report on a therapeutic trial. Br Med J.  1954; 2: 375– 378. CrossRef Search ADS PubMed  2 Seo M, Okada M, Tsuneyoshi Y, et al. An index of disease activity in patients with ulcerative colitis. Am J Gastroenterol.  1992; 87: 971– 976. PubMed  3 Walmsley RS, Ayres RC, Punder RE, et al. A simple clinical colitis activity index. Gut.  1998; 43: 29– 32. CrossRef Search ADS PubMed  4 Jowett SL, Seal CJ, Phillips E, et al. Defining relapse of ulcerative colitis using a symptom based activity index. Scand J Gastroenterol.  2003; 38: 164– 171. CrossRef Search ADS PubMed  5 Dick AP, Holt P, Dalton ER. Persistence of mucosal abnormality in ulcerative colitis. Gut.  1966; 7: 355– 360. CrossRef Search ADS PubMed  6 Schroeder KW, Tremaine WJ, Ilstrup DM. Coated oral 5-aminosalicylic acid therapy for mildly to moderately active ulcerative colitis. A randomized study. N Engl J Med.  1987; 24; 317: 1625– 1629. CrossRef Search ADS   7 Rachmilewitz D. Coated mesalazine (5-aminosalicylic acid) versus sulphaslazine in the treatment of active ulcerative colitis: a randomized trial. Br Med J.  1989; 298: 82– 86. CrossRef Search ADS   8 Higgins PD, Schwartz M, Mapali J, et al. Is endoscopy necessary for the measurement of disease activity in ulcerative colitis? Am J Gastroenterol.  2005; 100: 355– 361. CrossRef Search ADS PubMed  9 Higgins PD, Leug J, Schwartz M, et al. The quantitative validation of non-endoscopic disease activity indices in ulcerative colitis. Aliment Pharmacol Ther.  2007; 25: 333– 342. CrossRef Search ADS PubMed  10 Regueiro M, Schraut W, Baidoo L, et al. Infliximab prevents Crohn's disease recurrence after ileal resection. Gastroenterology.  2009; 136: 441– 450. CrossRef Search ADS PubMed  11 Yount RJ, Vries JK, Councill CD. The Medical Archival System: an information retrieval system based on distributed parallel processing. Inf Proc Manag.  1991; 27: 379– 389. CrossRef Search ADS   12 Altman DG. Practical Statistics for Medical Research. London: Chapman and Hall;  1991. 13 Rutter M, Saunders B, Wilkinson K, et al. Severity of inflammation is a risk factor for colorectal neoplasia in ulcerative colitis. Gastroenterology.  2004; 126: 451– 459. CrossRef Search ADS PubMed  14 Gupta RB, Harpaz N, Itzkowitz S, et al. Histologic inflammation is a risk factor for progression to colorectal neoplasia in ulcerative colitis: a cohort study. Gastroenterology.  2007; 133: 1099– 1105. CrossRef Search ADS PubMed  15 Rubin DT, Huo D, Rothe JA, et al. Increased inflammatory activity is an independent risk factor for dysplasia and colorectal cancer in ulcerative colitis: a case-control analysis with blinded prospective pathology review. Gastroenterology.  2006; 130: A2. CrossRef Search ADS   16 Froslie JF, Jansen J, Moum BA, et al. Mucosal healing in inflammatory bowel disease: results from a Norwegian population based cohort. Gastroenterology.  2007; 133: 412– 422. CrossRef Search ADS PubMed  17 Rutgeerts P, Vermiere S, Van Assche G. Mucosal healing in inflammatory bowel disease: impossible ideal or therapeutic target. Gut.  2007; 56: 453– 455. CrossRef Search ADS PubMed  18 Solem CA, Loftus EV, Tremaine WJ, et al. Correlation of C-reactive protein with clinical, endoscopic, histologic and radiologic activity in inflammatory bowel disease. Inflamm Bowel Dis.  2005; 11: 707– 712. CrossRef Search ADS PubMed  19 Rodgers AD, Cummins AG. CRP correlates with clinical scores in ulcerative colitis but not in Crohn's disease. Dig Dis Sci.  2007; 52: 2063– 2068. CrossRef Search ADS PubMed  20 Osada T, Ohkusa T, Okayasu I, et al. Correlations among total colonoscopic findings, clinical symptoms, and laboratory markers in ulcerative colitis. J Gastroenterol Hepatol.  2008;( suppl 2): S262– S267. 21 Lok KH, Ng CH, Hung HG, et al. Correlation of serum biomarkers with clinical severity and mucosal inflammation in Chinese ulcerative colitis patients. J Dig Dis.  2008; 9: 219– 224. CrossRef Search ADS PubMed  22 Langhorst J. Elsenbruch S, Koelzer J. et al. Non-invasive markers in the evaluation of intestinal inflammation in inflammatory bowel disease: performance of fecal lactoferrin, calprotectin, and PMN-elastase, CRP and clinical indices. Am J Gastroenterol.  2008; 103: 162– 169. CrossRef Search ADS PubMed  23 Xiang JY, Ouyang Q, Li GD, et al. Clinical value of fecal calprotectin in determining disease activity of ulcerative colitis. World J Gastroenterol.  2008; 14: 53– 57. CrossRef Search ADS PubMed  24 Maconi G, Ardizzone S, Parente F, et al. Ultrasonography in the evaluation of extension, activity and follow up of ulcerative colitis. Scand J Gastroenterol.  1999; 34: 1103– 1107. CrossRef Search ADS PubMed  25 Parente F, Molteni M, Marino B, et al. Bowel ultrasound and mucosal healing in ulcerative colitis. Dig Dis.  2009; 27: 285– 290. CrossRef Search ADS PubMed  26 Andersen K, Vogt C, Blondin D, et al. Multi detector CT colonography in inflammatory bowel disease: prospective analysis of CT findings to high resolution video colonoscopy. Eur J Radiol.  2006; 58: 140– 146. CrossRef Search ADS PubMed  27 Ajaj WM, Lauenstein TC, Pelster G, et al. Magnetic resonance colonography for the detection of inflammatory diseases of the large bowel: quantifying the inflammatory activity. Gut.  2005; 54: 257– 263. CrossRef Search ADS PubMed  Copyright © 2010 Crohn's & Colitis Foundation of America, Inc. TI - Physician assessment of ulcerative colitis activity correlates poorly with endoscopic disease activity JF - Inflammatory Bowel Diseases DO - 10.1002/ibd.21445 DA - 2011-04-01 UR - https://www.deepdyve.com/lp/oxford-university-press/physician-assessment-of-ulcerative-colitis-activity-correlates-poorly-3GIA8sL2wq SP - 1008 EP - 1014 VL - 17 IS - 4 DP - DeepDyve ER -