Gastrointestinal Imaging • Original Research Elmi et al. Early Colonoscopy in Diverticulitis

Downloaded from www.ajronline.org by 52.2.249.46 on 09/13/15 from IP address 52.2.249.46. Copyright ARRS. For personal use only; all rights reserved

Gastrointestinal Imaging Original Research

Is Early Colonoscopy Beneficial in Patients With CT-Diagnosed Diverticulitis? Azadeh Elmi1 Sandeep S. Hedgire Vivek Pargaonkar Kai Cao Shaunagh McDermott Mukesh Harisinghani Elmi A, Hedgire SS, Pargaonkar V, Cao K, McDermott S, Harisinghani M

OBJECTIVE. The purpose of this study was to establish CT criteria as an indication for colonoscopy in patients with acute diverticulitis. MATERIALS AND METHODS. The study was composed of patients diagnosed with acute diverticulitis on contrast-enhanced CT (CECT) from January 2000 to December 2004. Patients without subsequent colonoscopy were excluded. Findings on CT were correlated with patient outcomes over follow-up. Sensitivity, specificity, and predictive values of the imaging parameters for prediction of colon cancer were calculated. Outcomes were also compared between men and women. RESULTS. Of 1034 patients with a CT diagnosis of diverticulitis, 402 (235 women and 167 men; mean age, 63.3 years) had available endoscopic colonic evaluation after CECT. The mean follow-up was 5.3 years. Seventy-eight patients had polyps (71% adenomatous) and nine patients (2.2%) were diagnosed with colon cancer (seven women, two men). The odds ratio for diagnosis of colon cancer was 23.35 in patients with mesenteric or retroperitoneal lymphnodes, 4.67 for abscess, and 24.43 in patients with obstruction and localized mass reported on CT. A significant correlation was found between the location of diverticulitis and cancer (p < 0.001). The diagnosis of cancer was made within 6 months from the date of CECT in eight patients. The odds of cancer were 2.5 times higher in women. CONCLUSION. On the basis of the significant correlation of complicated diverticulitis or presence of lymph nodes with new diagnoses of colon cancer, we recommend early colonoscopy in patients with wall thickness more than 6 mm, abscess, obstruction, or lymph nodes seen on CECT.

T

Keywords: cancer, colon, CT, diverticular disease, diverticulitis, endoscopy DOI:10.2214/AJR.12.9539 Received July 3, 2012; accepted after revision October 1, 2012. 1

All authors: Division of Abdominal Imaging and Interventional Radiology, Massachusetts General Hospital, Harvard Medical School, 55 Fruit St, White 270, Boston, MA 02114. Address correspondence to A. Elmi ([email protected]).

AJR 2013; 200:1269–1274 0361–803X/13/2006–1269 © American Roentgen Ray Society

he prevalence of diverticular disease and colorectal cancer has markedly increased in the past century [1]. It is estimated that 60% of humans over age 60 years develop colonic diverticula, which can be inflamed in approximately 10–25% of cases [2, 3]. Contrastenhanced CT (CECT) depicts diverticulitis in the early stage and has become the optimal method for evaluation of diverticular disease in the emergency setting [4], with reported accuracy of 80–100% using high-resolution CECT [5]. In addition to establishing the diagnosis, objective grading of the severity and quantification of the disease resolution and classification into complicated and uncomplicated disease by use of CECT have a direct impact on the management of patients [6]. Although, the immediate role of CECT in acute diverticulitis has now been largely defined, excluding the possibility of underlying colon cancer as a possible cause for diverticulitis can still be chal-

lenging. Anatomic distortion from numerous diverticular outpouchings along with the associated inflammatory response makes it difficult to identify the underlying colon cancer in the setting of acute diverticulitis. Several studies have evaluated morphologic criteria, such as length of involvement, presence of a mass, and pericolonic inflammation, in patients with equivocal CT findings; however, most patients required further investigation to establish a firm diagnosis [7–9]. There are also controversies regarding the use of colonoscopy to exclude colon cancer in these patients. Some experts recommend evaluation of the entire length of the colon after resolution of an episode of diverticulitis in all patients presenting with acute diverticulitis using colonoscopy, barium enema and sigmoidoscopy, or CT colonography to exclude other diagnoses [10, 11]. On the other hand, others conclude that subsequent evaluation of the colon may not be re-

AJR:200, June 2013 1269

Downloaded from www.ajronline.org by 52.2.249.46 on 09/13/15 from IP address 52.2.249.46. Copyright ARRS. For personal use only; all rights reserved

Elmi et al. quired because the yield of advanced colonic neoplasia has been shown to be equivalent to or less than that detected on screening asymptomatic average-risk individuals and they recommend routine screening [12]. In addition, there is concern about the potential risk of perforation or other exacerbation of the disease process during air insufflation and scope manipulation if colonoscopy is performed soon after an attack of acute diverticulitis. Given the frequency of the diagnosis of diverticulitis, its association with increased age and the overlapping features of diverticulitis with cancer on CT, there is a clear need for evidence-based recommendations regarding follow-up after an attack of acute diverticulitis. The aim of this study was to establish CT criteria as indications for colonoscopy in patients with CECT-documented acute diverticulitis after initial successful medical management. Materials and Methods Patient Selection The study was conducted at a single university-affiliated teaching hospital. Institutional review board approval was obtained for this retrospective study that was undertaken in compliance with HIPAA. From January 2000 to December 2004, we retrospectively reviewed the CECT reports containing the word “diverticulitis” (in history, findings, or impression) for patients older than 49 years. The reports were assessed and patients with acute diverticulitis were identified. Subsequently, the patients in whom evaluation of the colon using colonoscopy had been performed after diagnostic CECT during the follow-up period were included in this study. Among these patients, we excluded patients with history of colorectal cancer to minimize selection bias. All outcomes were examined until September 30, 2011.

Imaging Technique and Analysis CT studies were performed using a helical MDCT system using the standard protocol within the department. Scanning included the axial and coronal planes with a slice thickness of 5 mm. Each patient ingested oral contrast medium before the examination. IV administration of contrast medium was routinely given except for 12 patients because of poor renal function or previous allergic history. The CT images were retrospectively reviewed by one radiologist with more than 10 years of experience in abdominal imaging. Wall thickness of the involved segment was measured as the maximum thickness of the affected segment and reported on a two-level scale (≤ 6 mm and > 6 mm) [13]. Location of diverticulitis, presence of extraluminal air, abscess, fistula, localized mass, obstruction, lymph

1270

nodes, and enhancement pattern were evaluated. To determine location, the colon was divided into segments: cecum, ascending, transverse, descending, sigmoid, and rectum. Extraintestinal gas was defined as bubbles or pockets of extraintestinal air without a wall (not located inside an abscess). Fluid-containing collections, surrounded by a contrast-enhancing rim were defined as abscesses. Enhancement was graded subjectively as none, minimal, or obvious on the basis of comparison with the normal bowel wall segments. The visibility of mesenteric and retroperitoneal lymph nodes was assessed and the shortest diameter was measured.

Follow-Up The reports of colonoscopy during follow-up were reviewed. The finding of polyps or masses and the histologic findings from the endoscopic biopsy were recorded. The site of the lesions was retrieved from the endoscopy reports, and the location of the lesions reported on the colonoscopy reports were corresponded with the segmental location. For this purpose, the colon was divided into the following six segments: rectum, sigmoid, descending colon, transverse colon, ascending colon, and cecum. Patients were stratified according to the most advanced lesions identified on histopathology. The first group included patients with nonneoplastic conditions, such as lipoma, granulation, angiodysplasia, hyperplasic polyps, and hamartomas. The patients with adenomatous polyps and cancer were considered group 2. The outcomes were also compared between men and women.

Statistical Analysis Continuous variables were compared between groups by using the Student t test. The Fisher exact test was performed to analyze differences among the categoric data. A univariate statistical analysis was performed for CECT to determine the correlation with patient outcomes over the follow-up period. Sensitivity, specificity, and positive and negative predictive values of the imaging parameters for prediction of colon cancer were also calculated. The site of the mass on the colonoscopy report was correlated with the location of diverticulitis on CT by the involved segment. A match was accepted when the mass was located in the same colonic segment reported for diverticulitis on initial CT. Statistical significance was defined as a p value of less than 0.05.

Results A total of 4881 CT studies were retrieved using the defined search criteria. After reviewing the reports, 1034 patients with a CT diagnosis of acute diverticulitis were identified. The records showed that a cohort of 402 patients (235

women and 167 men) had available colonic endoscopic evaluation after CECT, and these patients were included in the study (Fig. 1). The mean age at CECT was 63.3 years (range, 50–94 years). Analysis of CECT images revealed that 82 (20.3%) patients had perforation, seven (1.7%) had fistula, 41 (10.2%) had abscess, and 27 (6.7%) had adenopathy. The mean follow-up was 5.3 years (1 month to 11 years). Figure 2 shows the timing of follow-up colonoscopy in our cohort of patients. During the first 6 months after the first attack of diverticulitis, 34.8% of the patients had colonoscopy. Colonic polyps were present and biopsied in 78 patients (19.4%), including 55 adenomatous polyps, 21 hyperplastic polyps, and two polypoid granulations. Overall, 345 patients had nonneoplasitc lesions (group 1) and 57 patients had neoplastic pathology and were categorized to group 2. The prevalence of different morphologic findings on initial CECT in each group is presented in Table 1. The rate of adenopathy, obstruction, and localized mass was significantly higher in patients with neoplastic lesions. Nine patients (2.2%) were diagnosed with colon cancer (seven women and two men; mean age, 68.11 years) during follow-up colonoscopy (Fig. 3). The sensitivity, specificity, and predictive values of individual CECT criteria for predicting colon cancer are summarized in Table 2. Presence of enhancement showed the highest sensitivity for detecting both neoplastic lesions and cancer. Localized mass and presence of obstruction showed the highest specificity for predicting both neoplastic lesions and cancer. Bowel thickness of more than 6 mm was observed in patients with neoplastic lesions at a significantly higher rate (p = 0.033). The odds ratio of a diagnosis of colon cancer on colonoscopy was 23.35 (95% CI, 5.56– 97.94) in patients with presence of mesenteric or retroperitoneal lymph node, 4.67 (1.12– 19.43) for abscess, and 24.43 (2.00–297.85) in patients with obstruction and localized mass reported on CECT. Diverticulitis and cancer were reported predominantly in the sigmoid colon (79.6% and 66.7%, respectively), and a significant correlation was found between the approximate location of diverticulitis on the original CECT and cancer reported on colonoscopy (matched colonic segment in eight cases, p < 0.001). The diagnosis of malignancy was made within 6 months from the date of CECT in eight patients (1–6 months), and only one patient had a diagnosis of colon cancer after 6 months during our follow-up. It should

AJR:200, June 2013

Early Colonoscopy in Diverticulitis Fig. 1—Flow chart shows patient summary.

Acute diverticulitis on CT report (n = 4881)

CT diagnosis of acute diverticulitis (n = 1034) No colonic evaluation (n = 632 ) Interval colonoscopy (n = 402)

be noted that in this patient malignant tumor was detected in the right colon whereas acute diverticulitis was reported in the descending colon. Most of the patients with cancer (seven of nine) had nodal diseases with advanced tumor (pT3 or pT4) at the time of diagnosis (Table 3). We also noted an unequal distribution of men and women. Women accounted for a higher number of cases in our diverticulitis cohort, and colon cancer was detected at higher rate in women as well (seven vs two, p = 0.041). The odds of a diagnosis of colon cancer in women increased more than 2.5fold in comparison with men. Discussion There are no generally accepted criteria to guide the need for colonoscopy in patients with acute diverticulitis. To the best of our knowledge, our study is the largest series of patients with CT evidence of diverticulitis who underwent colonoscopy over almost 5.5 years at a large teaching hospital. Of 402 patients identified in our series, as many as 14.2% of patients (57 patients with neoplastic lesions) had additional findings on early colonoscopy with the potential to change management strategy and outcomes. Nine patients were found to have adenocarcinoma of colon (2.2%). Noticeably, of these patients with new diagnoses of cancer, only one patient had imaging evidence suspicious for a mass according to the initial CT report.

Fig. 2—Graph shows timing of follow-up colonoscopy in our cohort of patients. Timing is represented as intervals from date of initial CT to initial colonoscopy in months.

The rate of colon cancer in our patients seems to be higher than the reported prevalence in the general population. A recent metaanalysis involving 10 studies of screening colonoscopy in asymptomatic people with a total of 68,324 participants showed that the yield of invasive carcinoma in a heterogeneous population of high- and average-risk asymptomatic individuals was 0.8% [14]. The rate of colon cancer in patients with diverticulitis is variable in different reports [12, 15]. A recent meta-analysis of the prevalence of colorectal cancer in patients with CT-diagnosed acute diverticulitis proposed that the rate of colon cancer in such patients (2.1%) compares to a calculated estimated prevalence among U.S. adults older than 55 years [16]. The authors concluded that there are limited data to support the recommendation to perform colonoscopy after a diagnosis of acute diverticulitis. The conclusion of this meta-analysis is limited by the sparse data about follow-up colonoscopy after CT diagnosis of acute diverticulitis in

Interval From CT to Colonoscopy (mo)

Downloaded from www.ajronline.org by 52.2.249.46 on 09/13/15 from IP address 52.2.249.46. Copyright ARRS. For personal use only; all rights reserved

No acute diverticulitis (n = 3847)

the present literature, which was mentioned by the authors. They also commented that these available studies did not specifically examine CT or clinical findings that would predict or exclude the possibility of colorectal cancer. Accordingly, despite numerous studies, the association between colon cancer and diverticulitis remains unclear. This discrepancy may be a result of the highly heterogeneous study designs and settings in the different series. The authors of a large nationwide case-controlled study in Sweden suggested that diverticular disease does not increase the risk of colon cancer in the long term, and a history of diverticular disease does not affect colon cancer mortality [17]. In parallel, Morini et al. [18] performed a comprehensive evaluation of the available data and concluded that the data are not yet strong enough to suggest more aggressive colon cancer screening in diverticular disease. To evaluate this issue we tried to correlate the approximate anatomic site of diverticulitis as documented on original CECT with the location of colon cancer reported on colonoscopy. Interestingly, all but one case had fair correlation of the site of the pathology with early diagnosis of cancer (within 6 months of CECT). However, in one case, the tumor was detected in the right colon whereas diverticulitis was reported on the left and this patient was the only case in our series with a remote diagnosis of colon cancer (after 5 years). Accordingly, we propose that the higher rate of cancer in our series might be attributed to the fact that cancer can mimic clinical and CT features of acute diverticulitis and is more likely to be detected due to close surveillance of diverticulitis patients.

> 60 48−60 36−48 24−36 12−24 6−12 0−6 0

40

80

120

160

No. of Patients

AJR:200, June 2013 1271

Elmi et al. TABLE 1: Comparison of Imaging Parameters in Nonneoplastic (n = 345) and Neoplastic (n = 57) Lesions Total Cohort Morphologic Criteria

Nonneoplastic

Neoplastic

No.

Percentage

pa

Downloaded from www.ajronline.org by 52.2.249.46 on 09/13/15 from IP address 52.2.249.46. Copyright ARRS. For personal use only; all rights reserved

Diverticulitis location Left

314

57

371

92.2

Right

24

7

31

7.7

≤ 6 mm

126

13

140

34.8

> 6 mm

218

44

262

65.2

Local perforation

74

8

82

20.3

0.097

Abscess

37

4

41

10.2

0.598

6

1

7

1.7

0.991

17

10

0.001

Wall thickness

Fistula Adenopathy

0.033

27

6.7

0.031

—

—

0.470

1

3

0.7

0.001

1

3

0.7

0.001

283

46

329

81.8

5

4

9

2.2

Size (mm)

8.8

7.5

Obstruction

2

Localized mass

2

Minimal Obvious

Degree of enhancementb 0.001

Note—Nonneoplastic includes lipoma, granulation, angiodysplasia, hyperplasic polyps, and hamartomas. Neoplastic includes adenomatous polyp and cancer. Dashes indicate not applicable. aComparing nonneoplastic lesions with neoplastic lesions. bSubjective evaluation was used for enhancement. Fig. 3—Acute diverticulitis. A and B, 61-year-old woman with acute diverticulitis. Contrast-enhanced axial CT image (A) shows focal sigmoid diverticulitis with perforation (arrow). Colonoscopy image (B) shows partially obstructing malignant-appearing tumor in sigmoid. Biopsy revealed invasive adenocarcinoma. C and D, 73-year-old woman with acute diverticulitis. Contrast-enhanced axial CT image (C) shows 12-cm segment of sigmoid colon with diffuse wall thickening with haziness of surrounding mesentery without evidence of perforation or abscess formation (arrow). On colonoscopy image (D) 10-mm polyp in sigmoid colon was identified. Biopsy revealed adenomatous polyp.

No genuine increased risk of cancer was experienced in patients with diverticulitis after 6 months of the acute attack in our cohort. We also noted that the incidence rate for cancer in women was higher compared with men, despite equal age distribution. Although, the overall incidence of colon cancer in the litera-

1272

ture is slightly higher in men, increased rates of diverticulitis and colon cancer diagnosed after an attack of diverticulitis in women have been previously reported by Lam et al. [19]. Additionally, a recent study showed increased odds of colorectal cancer in the women in a diverticulitis cohort [20]. We could not find any risk

factor for this discrepancy; however, the possibility of a higher rate of diverticulitis attacks in colon cancer in women cannot be excluded. In several studies, quantitative CT perfusion measurements as well as morphologic criteria have been evaluated to establish practical diagnostic criteria to differentiate colonic diverticulitis from colon cancer [13, 21]. Morphologic features, such as length of involved bowel, presence of inflamed diverticulum, and pericolic infiltration, for the diagnosis of colonic diverticulitis and excluding colon cancer have been reported. On the other hand, combinations of morphologic and quantitative CT perfusion criteria have been evaluated and shown to be strong discriminators between cancer and diverticulitis [13].

AJR:200, June 2013

1.39 (0.6–3.22) 100 (63/63) 89.0 (56/63)

TABLE 3: Clinical Features and Outcomes for Patients With Colorectal Cancers

Note—Except for odds ratio, data in parentheses are no./total. PPV = positive predictive value, NPV = negative predictive value.

24.43 (2–297)

3.65 (0.21–63.5)

3.05 (0.3–34.13) 98.0 (384/392) 86.0 (341/397)

14.8 (50/338)

33.3 (1/3) 99.4 (384/386) 99.4 (341/343)

16.3 (63/393) 100 (9/9) Enhancement

11.1 (1/9) 1.7 (1/57)

87.7 (50/57)

Obstruction

16.2 (56/344)

33.3 (1/3)

1.98 (0.48–8.1)

2.6 (9/339)

24.43 (2–297.85)

0.60 (0.27–1.34) 98.1 (313/319) 85.0 (270/318) 14.2 (8/56) Local perforation

33.3 (3/9)

78.4 (270/344)

80.0 (313/392)

10.0 (8/82)

3.6 (3/82)

23.35 (5.56–97.9) 2.49 (1.15.49)

3.05 (0.27–34.1)

99.2 (372/375)

98.0 (384/392)

87.4 (317/364)

86.0 (341/397) 33.3 (1/3)

22.2 (6/27) 27.3 (10/37)

33.3 (1/3) 99.5 (384/386)

94.6 (372/393)

1.7 (1/57)

11.1 (1/9)

66.6 (6/9) 17.5 (10/57) Adenopathy

99 (341/343)

97.7 (386/395)

Localized mass

95.1 (317/344)

4.67 (1.12–19.43)

2.71 (0.14–51.01)

0.46 (0.14–1.55)

1.01 (0.12–8.54)

98.0 (355/361) 85.0 (307/361)

86.0 (339/395) 0 (0/7)

7.3 (3/41) 7.5 (3/40)

14.2 (1/7) 98.0 (386/393)

90.3 (355/393) 90 (307/344)

1.7 (1/57)

98.2 (339/345) 0 (0/9)

5.2 (3/57) Abscess

Fistula

33.3 (3/9)

1.95 (1.02–3.7)

CRC Neoplastic

1.91 (0.4–9.24)

CRC

99.0 (138/140) 91.0 (126/139)

Neoplastic CRC

2.6 (7/262) 17.0 (44/262)

Neoplastic CRC

35.0 (138/393) 37 (126/344)

Neoplastic CRC Neoplastic

77.0 (44/57)

Morphologic Criteria

Wall thickness > 6 mm

78.0 (7/9)

Odds Ratio (95% CI) NPV (%) PPV (%) Specificity (%) Sensitivity (%)

TABLE 2: Sensitivity, Specificity, and Predictive Values of Individual CT Criteria for Predicting Neoplastic Lesions and Colorectal Cancer (CRC)

Downloaded from www.ajronline.org by 52.2.249.46 on 09/13/15 from IP address 52.2.249.46. Copyright ARRS. For personal use only; all rights reserved

Early Colonoscopy in Diverticulitis

Age (y)

Sex

Location of Diverticulitis

75

F

Sigmoid

68

M

74

Location of Cancer

Staging

Interval to Diagnosisa

Sigmoid

pT4N1M0

5 mo

Alive after colectomy and chemotherapy

Cecum

Cecum

pT3N1M0

2 mo

Alive after colectomy and chemotherapy

F

Sigmoid

Sigmoid

pT4N1M0

5 mo

Alive after colectomy and chemotherapy

66

F

Sigmoid

Sigmoid

pT3N1M1

2 mo

Death in 4 y due to advanced disease

71

F

Sigmoid

Sigmoid

pT3N1M0

5 mo

Alive after colectomy and chemotherapy

61

F

Sigmoid and Sigmoid descending colon

pT4N1M0

6 mo

Alive after colectomy and chemotherapy

73

F

Sigmoid and Sigmoid descending colon

pT3N0M0

4 mo

Alive after colectomy and chemotherapy

57

F

Sigmoid

Sigmoid

pT1N0M0

2 mo

Alive after resection

68

F

Descending colon

Ascending colon pT4N2M1

5y

Death in 2 y due to advanced disease

Follow-Up

aFrom the date of CT to definite diagnosis of malignancy.

Likewise, our results show that some characteristics observed on CT, such as presence of an abscess, obstruction, or wall thickness more than 6 mm, may be useful predictors of neoplastic lesions and specifically colon cancer. The presence of a mass and pericolonic nodes had high specificity, whereas the presence of enhancement showed sensitivity of 100% for predicting colon cancer. Shen et al. [21] reported that in the presence of colonic obstruction colon cancer is more likely than diverticulitis. Obstruction was reported on CT only in three cases in our series but had specificity of more than 99%. Accordingly, the low sensitivity is a major limiting factor for considering this finding as a consistent predictive imaging feature. Regarding adenopathy, it should be noted that in our study adenopathy was reported significantly more frequently in patients with neoplastic lesions; however, there was no significant difference between the sizes of the lymph nodes when comparing the two groups. We also evaluated the enhancement pattern of the involved part of the colon. Although the sensitivity of presence of enhancement was 100% for colon cancer, the specificity was 16.3%, limiting the use of enhancement parameters for this purpose.

Recently, Lau et al. [20] reported a substantial risk of cancer in patients after acute leftsided diverticulitis within 1 year of CT, in particular, when abscess, local perforation, and fistula are observed. They recommended routine colonoscopy in patients who have not had a recent colonic luminal evaluation. They also reported the overall prevalence of 2.1% for colorectal cancer among the cohort, which seems comparable to our findings (2.2%). Our findings were different from the study by Lau et al. in several aspects. First of all they only included left-sided diverticulitis and followed the patients for only 1 year. Additionally, we did not find any significant correlation between presence of perforation or fistula and colon cancer. However, abscess formation and obstruction were reported with higher odds of colon cancer diagnosis in our cohort of patients. It should be noted that in our series only patients older than 49 years were included because typically colon cancer is regarded as a disease of the elderly, with 90% of patients diagnosed beyond the age of 55 years [22, 23]. Considering the low incidence of colon cancer in those younger than 50 years, we excluded this group to make the study more consistent. Larger trials are necessary to address the role of accelerated

AJR:200, June 2013 1273

Downloaded from www.ajronline.org by 52.2.249.46 on 09/13/15 from IP address 52.2.249.46. Copyright ARRS. For personal use only; all rights reserved

Elmi et al. colonoscopy in younger patients presenting with acute diverticulitis. There are some limitations to our study that should be considered when interpreting these findings. The main drawback was the retrospective design of the study, which limited evaluation of colonic lesions in a timely manner. Also, it was not possible to explore the reason for colonoscopy in a significant number of cases. Importantly, because we only included patients with colonoscopy during our follow-up period in our institution, there is a selection bias because those patients who did not undergo colonoscopy during our follow-up period were excluded. It should be noted that we did not include polyp size in our study, and one can make the argument that the presence of advanced adenoma may have resulted in accelerated colonoscopy in some patients leading to the diagnosis of colon cancer. However, on the basis of our findings, advanced colon cancer was detected on the first colonoscopy after the diverticulitis attack in eight of nine cases, and the finding of advanced adenoma probably did not contribute to the higher rate of colon cancer in our series. We acknowledge that we did not include clinical information and medical conditions of the patients, and the small number of patients with cancer in our study limits evaluation of accuracy of the imaging criteria for predicting cancer in diverticulitis patients. Further studies in larger populations of patients with acute diverticulitis should be performed to confirm our findings. Conclusions We showed an increased risk of occult colon cancer in patients with acute diverticulitis, especially in women, during the first 6 months of an acute diverticulitis attack (2.2% vs 0.8% in screening studies). Accordingly, we recommend early colonoscopy after an attack of acute diverticulitis in patients with wall thickness greater than

1274

6 mm, presence of abscess, obstruction, or lymph node on initial CT. References 1. Stollman N, Raskin JB. Diverticular disease of the colon. Lancet 2004; 363:631–639 2. Parks TG. Natural history of diverticular disease of the colon. Clin Gastroenterol 1975; 4:53–69 3. Heise CP. Epidemiology and pathogenesis of diverticular disease. J Gastrointest Surg 2008; 12:1309–1311 4. Rao PM, Rhea JT, Novelline RA, et al. Helical CT with only colonic contrast material for diagnosing diverticulitis: prospective evaluation of 150 patients. AJR 1998; 170:1445–1449 5. Rotert H, Noldge G, Encke J, Richter GM, Dux M. The value of CT for the diagnosis of acute diverticulitis. Radiologe 2003; 43:51–58 6. Ambrosetti P. Acute diverticulitis of the left colon: value of the initial CT and timing of elective colectomy. J Gastrointest Surg 2008; 12:1318–1320 7. Balthazar EJ, Megibow A, Schinella RA, Gordon R. Limitations in the CT diagnosis of acute diverticulitis: comparison of CT, contrast enema, and pathologic findings in 16 patients. AJR 1990; 154:281–285 8. Chintapalli KN, Esola CC, Chopra S, Ghiatas AA, Dodd GD 3rd. Pericolic mesenteric lymph nodes: an aid in distinguishing diverticulitis from cancer of the colon. AJR 1997; 169:1253– 1255 9. Chintapalli KN, Chopra S, Ghiatas AA, Esola CC, Fields SF, Dodd GD 3rd. Diverticulitis versus colon cancer: differentiation with helical CT findings. Radiology 1999; 210:429–435 10. Rafferty J, Shellito P, Hyman NH, Buie WD. Practice parameters for sigmoid diverticulitis. Dis Colon Rectum 2006; 49:939–944 11. Ambrosetti P, Jenny A, Becker C, Terrier TF, Morel P. Acute left colonic diverticulitis: compared performance of computed tomography and water-soluble contrast enema—prospective evaluation of 420 patients. Dis Colon Rectum 2000; 43:1363–1367 12. Westwood DA, Eglinton TW, Frizelle FA. Routine colonoscopy following acute uncomplicated diverticulitis. Br J Surg 2011; 98:1630–1634

13. Goh V, Halligan S, Taylor SA, Burling D, Bassett P, Bartram CI. Differentiation between diverticulitis and colorectal cancer: quantitative CT perfusion measurements versus morphologic criteria— initial experience. Radiology 2007; 242:456–462 14. Niv Y, Hazazi R, Levi Z, Fraser G. Screening colonoscopy for colorectal cancer in asymptomatic people: a meta-analysis. Dig Dis Sci 2008; 53:3049–3054 15. Stefánsson T, Ekbom A, Sparèn P, Páhlman L. Association between sigmoid diverticulitis and left-sided colon cancer: a nested, population-based, case control study. Scand J Gastroenterol 2004; 39:743–747 16. Sai VF, Velayos F, Neuhaus J, Westphalen AC. Colonoscopy after CT diagnosis of diverticulitis to exclude colon cancer: a systematic literature review. Radiology 2012; 263:383–390 17. Granlund J, Svensson T, Granath F, et al. Diverticular disease and the risk of colon cancer: a population-based case-control study. Aliment Pharmacol Ther 2011; 34:675–681 18. Morini S, Zullo A, Hassan C, Tomao S, Campo SM. Diverticulosis and colorectal cancer: between lights and shadows. J Clin Gastroenterol 2008; 42:763–770 19. Lam TJ, Meurs-Szojda MM, Gundlach L, et al. There is no increased risk for colorectal cancer and adenomas in patients with diverticulitis: a retrospective longitudinal study. Colorectal Dis 2010; 12:1122–1126 20. Lau KC, Spilsbury K, Farooque Y, et al. Is colonoscopy still mandatory after a CT diagnosis of left-sided diverticulitis: can colorectal cancer be confidently excluded? Dis Colon Rectum 2011; 54:1265–1270 21. Shen SH, Chen JD, Tiu CM, et al. Differentiating colonic diverticulitis from colon cancer: the value of computed tomography in the emergency setting. J Chin Med Assoc 2005; 68:411–418 22. Atkin WS, Cuzick J, Northover JM, Whynes DK. Prevention of colorectal cancer by once-only sigmoidoscopy. Lancet 1993; 341:736–740 23. Meyer A, Merkel S, Bruckl W, et al. Cdc2 as prognostic marker in stage UICC II colon carcinomas. Eur J Cancer 2009; 45:1466–1473

AJR:200, June 2013