Surgical resection of the affected bowel is required in almost two-thirds of Crohn's disease (CD) patients.1 Postoperative recurrence after ileocolonic resection is a feature of CD.2–6 Follow-up of CD patients after surgery includes the assessment of both clinical and endoscopic recurrence, ileocolonoscopy representing the gold standard for this purpose. As ileocolonoscopy is quite an invasive procedure,2 alternative noninvasive techniques are needed in order to assess the postoperative recurrence of CD.7–17
Transabdominal ultrasonography (TUS) has been proposed for detecting small bowel lesions in patients with suspected or known CD, showing a sensitivity of 67%–84% and 81%–95%, respectively.18–25 The use of oral contrast significantly increases the sensitivity of ultrasonography for assessing small bowel lesions in CD patients with suspected or known diagnosis (>95%).24–26 In particular, small intestine contrast ultrasonography (SICUS) performed by an experienced sonographer may visualize not only established CD lesions (i.e., stenosis with possible prestenotic dilation), but also minor changes of the small bowel wall.26–28 In experienced hands, SICUS has been shown to detect intestinal lesions in patients with suspected small bowel diseases with a high (>95%) sensitivity and specificity, when compared with small bowel follow-through and enema.26 SICUS has been shown to be more accurate than TUS for assessing small bowel CD lesions, although the experience of the sonographer significantly affects the accuracy of both techniques, particularly of TUS.26
The sensitivity of TUS as assessed by the bowel wall thickness (BWT) in identifying the endoscopic CD recurrence after ileocolonic resection has been investigated in 2 studies, showing 82% sensitivity.29, 30 The use of a noninvasive technique, such as ultrasonography, in the follow-up of CD patients after surgery is advisable in order to reduce the radiation exposure and the use of the quite invasive ileocolonoscopy.
The possible role of SICUS in the assessment of CD recurrence is under investigation as, to our knowledge, only 2 studies evaluated this issue.31, 32 In a prospective longitudinal study, our findings supported the usefulness of SICUS for assessing CD recurrence after ileocolonic resection when using ileocolonoscopy as a gold standard.31 This finding was confirmed by a different group, reporting that a BWT >4 mm assessed by SICUS is the best cutoff for differentiating the severity of CD recurrence.32
Current management of CD patients after curative resection includes a more aggressive medical treatment in patients with early postoperative asymptomatic recurrence, for possible relapse prevention.33 Therefore, the development of noninvasive and repeatable techniques able to detect CD recurrence in the early postoperative period are needed for proper follow-up and treatment of resected patients.
On the basis of these observations, we aimed to investigate the possible usefulness of SICUS for assessing the postoperative recurrence of CD in patients under regular follow-up after ileocolonic resection. The specific aim was to evaluate the possible correlation between the severity of CD recurrence as assessed by SICUS and by ileocolonoscopy, considered the gold standard.
- Top of page
- MATERIALS AND METHODS
Clinical characteristics of the study population are reported in Table 1. As indicated, during the study period 72 patients (38 female; median age 44, range 16–73 years, 58 clinically inactive, median disease duration 114 months, range 12–492) with ileocolonic resection for CD had ileocolonoscopy and SICUS performed within 6 months. All 14 patients clinically active at the time of endoscopy underwent SICUS within a median of 1 month (range 1–90 days), and in both clinically active and inactive (n = 58) patients the CDAI was comparable at the time of the 2 investigations. Treatment at time of endoscopy and SICUS was similar, including anti-TNFα therapies in 3 patients (2 certolizumab, 1 infliximab; with concomitant azathioprine [AZA], and/or 5-aminosalicylate [5-ASA] in 2), steroids in 17 (prednisone 2, budesonide 15, with concomitant AZA, 1), AZA in 3, 5-ASA only in 47, while 2 patients used no drugs.
Table 1. Clinical Characteristics of the 72 CD Patients with Previous Ileocolonic Resection
|Disease location|| |
| Neoterminal ileum||37 (51%)|
| Neoterminal ileum and colon||31 (44%)|
| Colon only||1 (1%)|
| Jejunum and neoterminal ileum||3 (4%)|
|Disease behavior|| |
| Fibrostricturing||42 (58%)|
| Fistulating||30 (42%)|
|Smoking habits|| |
| Nonsmoker (%)||32 (44.5%)|
| Smoker (%)||21 (29%)|
| Exsmoker (%)||19 (26.5%)|
|Surgical resections|| |
| 1||54 (75%)|
| ≥ 2||18 (25%)|
|Indication for surgery|| |
| Abscesses||14 (19.5%)|
| Occlusion||52 (72 %)|
| Fistulae (2 E-E, 1 E-V, 1 E-C)||4 (5.5%)|
| Suspected appendicitis||2 (3%)|
|Time from the last resection (months)||18 (3-396)|
|CDAI (median, range)||85 (0-416)|
Recurrence was detected by ileocolonoscopy in 67/72 (93%) patients. The degree of recurrence was Grade 4 in 34/67 (51%) patients (associated with stenosis not passed by the endoscope in 22), Grade 3 in 9/67 (13%) patients, Grade 2 in 17/67 (25%) patients, and Grade 1 in 7/67 (11%) patients (Grade 0 in 5 patients). The median degree of endoscopic recurrence in the 67 patients was 3 (range 1–4). Among the 22 patients with endoscopic stenosis, at the time of investigation 6 were clinically active (4 with subostructive symptoms) and 16 inactive (2 undergoing ileocolonic resection within 3 months for subostructive symptoms). The endoscopic degree of recurrence was significantly correlated with the CDAI value (r = 0.27; P = 0.03).
The time from the last resection showed a marked variation between patients (Table 1). A comparable amount of patients were observed when subgrouped according to previous resection performed in the previous <3 or ≥3 years (n = 39 and n = 33, respectively). When the frequency of endoscopic recurrence was compared in these 2 subgroups, no significant difference was observed between patients with ileocolonic resection performed <3 versus ≥3 years before ileocolonoscopy (patients with recurrence: 37/39; 94.8% versus 30/33; 90.9%, respectively; P = NS).
Additional endoscopic findings were observed in 13 out of 72 (18%) patients, including: information regarding the colorectal anastomosis in the only 2 patients with both ileocolonic and colocolonic anastomoses (showing recurrence in 1), inflammatory micropolyps (<0.5 cm) (n = 4), angiodysplasia (n = 1), aphtoid or deep ulcers in the colon (n = 5), diverticulae in the sigmoid colon (n = 1).
SICUS was well tolerated in all patients and no side effects were observed during or after this procedure. Sonographic findings compatible with recurrence were detected in 62/72 (86%) patients. When considering the whole group of 72 patients, the median BWT was 5 mm (range 3.5–10 mm). SICUS detected strictures in 31/72 (43%) patients, associated with bowel dilation above stricture in 16/31 (51.6%), showing a median lumen of 28 mm (range 25–32). Different from the endoscopic score, no significant correlation was observed between the BWT and CDAI value at time of sonographic assessment (r = 0.2; P = 0.1). When subgrouping patients according to indication for surgical resection, the BWT did not significantly differ between patients with fistulating (n = 32) versus fibrostricturing (n = 40) disease (BWT: 5 mm, range 3–9.5 versus 4.5 mm, range 3–10; P = NS).
Additional sonographic findings included: enteroenteric fistulae in 3/72 (4%), lymph nodes enlargement in 1/72 (1.4%), mesenteric adipose tissue alteration 5/72 (7%), and jejunal involvement in 6/72 (8%) patients. No abdominal abscesses were detected.
SICUS provided additional findings in the subgroup of 22 patients showing at endoscopy a stenosis not allowing visualization of the ileum. In this subgroup of 22 patients, SICUS allowed the assessment of the neoterminal ileum above the anastomotic stenosis, showing the extent of the ileal lesions (median 13.75 cm, range 5–25), the presence of lumen narrowing at the anastomotic level in all 22 patients (median 4.5 mm, range 2–9), and ileal dilation above anastomotic stenosis in 9 out of 22 patients (41%) (median 20 mm, range 20–35).
Comparison Between Sonographic and Endoscopic Findings
Sonographic findings compatible with recurrence were detected in 62/72 (86%) patients showing endoscopic recurrence (Figs. 1A,B, 2A,B). When using ileocolonoscopy as a gold standard for assessing CD recurrence, using SICUS there were 4 FP, 62 TP, 1 TN, and 5 FN findings. SICUS therefore showed a 92.5% sensitivity, 20% specificity, 94% PPV, 16.6% NPV, and 87.5% accuracy in detecting CD recurrence. The endoscopic degree of recurrence was ≤2 in all 5 patients showing FN sonographic findings (Grade 2: n = 4; Grade 1: n = 1). The BWT was ≤4.5 mm in all 4 patients showing FP sonographic findings (3.5 mm: n = 1; 4 mm: n = 2; 4.5 mm: n = 1). Figure 3 shows, for each patient, the BWT and the endoscopic degree of recurrence, including FP and FN findings. As indicated, a significant correlation was observed between the perianastomotic BWT and the Rutgeerts' score (P < 0.0001; r = 0.67) (Fig. 3).
Figure 1. A,B: Perianastomotic area from a CD patient (P.M.) with ileocolonic anastomosis (side-to-side), as assessed by ileocolonoscopy (A) and SICUS (B) 6 months after surgery. A: Ileocolonoscopy showing CD recurrence at the anastomosis and a diffuse aphthous ileitis in neoterminal ileum (Rutgeerts' score: Grade 3). B: Perianastomotic area from the same patient, showing an increased BWT (5 mm; n.v. ≤3 mm) (arrows), with no stricture or loop dilation above lesions.
Download figure to PowerPoint
Figure 2. A,B: Perianastomotic area from a CD patient (D.L.A.) with ileocolonic anastomosis (side-to-side), as assessed by ileocolonoscopy (A) and SICUS (B) 12 months after surgery. A: Ileocolonoscopy showing recurrence involving the anastomosis and the neoterminal ileum, with more than 5 aphthae (Rutgeerts' score: Grade 2). B: Perianastomotic area from the same patient showing an increased BWT (6 mm; n.v. ≤3 mm) (arrows), with no stricture or ileal loop dilation.
Download figure to PowerPoint
Figure 3. The graph shows, for each of the 72 patients, the correlation between the BWT at the perianastomotic level as assessed by SICUS (mm) and the endoscopic degree of recurrence (Rutgeerts' score) (white squares = FP; white triangles = FN). As shown, a significant correlation was observed (r = 0.67; P = 0.0001). White circles represent patients with endoscopic score of Grade 4, associated with stenosis. Dotted line indicates the median BWT (n.v. = normal cutoff value of BWT, 3 mm).
Download figure to PowerPoint
A significant correlation was also observed between the perianastomotic BWT and the endoscopic degree of recurrence when subgrouping patients according to the time interval from surgery (≥3 years: P = 0.001, r = 0.63 in 33 patients; <3 years: P = 0.0001, r = 0.57; n = 39) (Fig. 4A,B).
Figure 4. A,B: The graphs show the correlation between the BWT in the perianastomotic area as assessed by SICUS (mm) and the endoscopic degree of recurrence (Rutgeerts' score; Grade 0–4) in patients with ileocolonic resection with a time interval <3 years (n = 39) (A) or ≥3 years (n = 33) (B) from enrolment. As shown, a significant correlation was observed in both subgroups.
Download figure to PowerPoint
As SICUS allows the detection of small bowel lesions, comparisons between sonographic and endoscopic findings were analyzed separately in patients showing established ileal lesions visualized by SICUS (Grade ≥3: diffuse aphthous ileitis, larger ulcers, nodules, and/or narrowing) or in patients with recurrence limited to the anastomosis with possible minor ileal lesions (Grade ≤2: aphthoid ulcers or lesions confined to the ileocolonic anastomosis).3 This analysis showed that the median BWT was significantly higher in patients with an endoscopic score of recurrence ≥3 (5.5 mm, range 4–10; n = 43) versus ≤2 (4 mm, range 3–7; n = 29; P = 0.0001). In patients with an endoscopic score ≥3 versus ≤2, a significantly higher median BWT, extent of the lesions, and prestenotic dilation were observed (median prestenotic dilation 20 mm, range 20–25 versus 20 mm, range 30–35, respectively; P = 0.001) (Table 2). Accordingly, the lumen diameter was significantly lower in patients with a Rutgeerts' score ≥3 versus ≤2 (Table 2).
Table 2. SICUS Findings in Patients with an Endoscopic Degree of Recurrence ≤2 vs. ≥3 (Rutgeerts' Score)
|SICUS Parameters||Rutgeerts' Score ≤2 (n = 29)||Rutgeerts' Score ≥3 (n = 43)||P-value|
|BWT (mm)||4 (3-7)||5.5 (4-10)||P = 0.0001|
|Lumen diameter (mm)||10 (5-15)||6.5 (2-15)||P = 0.0001|
|Lesion extent (cm)||5 (0-15)||10 (4-30)||P = 0.0001|
In order to evaluate and compare the clinical outcome in patients grouped according to endoscopic and sonographic findings, among the 58 quiescent CD patients at the time of examination, 11 relapsed within 1 year. When considering the Rutgeerts' score as a predictor, clinical relapse was observed in 1 out of the 25 (4%) patients with a score ≤2, and in 10 out of the 33 (30%) patients with a score ≥3. However, when considering the sonographic BWT as a predictor, clinical relapse was observed in 5 out of 40 (12.5%) patients with a BWT between 3–5 mm (4 mm in 2; 5 mm in 3 patients), and in 6 out of the 18 (33%) patients with a BWT ≥6 mm. Additional surgical resection within 1 year was required by only 3 out of the 58 (5%) inactive patients.
- Top of page
- MATERIALS AND METHODS
The growing use of radiologic techniques in CD management, together with the younger age at diagnosis of CD during the last years,35 is raising concern about the radiation exposure, indicating the need of radiation-sparing techniques in these patients. We therefore investigated the role of sonography using oral contrast for evaluating the postoperative recurrence of CD in patients with ileocolonic resection when using ileocolonoscopy as a gold standard. Although sonography is promoted in continental Europe,36 very few studies investigated the possible role of this technique for assessing the postoperative recurrence of CD, reporting a high sensitivity (81%–82%), and specificity (86%–100%).29, 30 The use of oral contrast, together with operator experience, significantly increases the accuracy of sonographic assessment of ileal lesions in CD,26-28 although only 2 studies investigated the usefulness of this technique in assessing CD recurrence.31, 32 As recent evidence supports the need of a more aggressive treatment in patients with early endoscopic recurrence, even if asymptomatic,33, 37 a noninvasive technique in the early postoperative period may identify patients at risk of a more aggressive course after surgery. These observations prompted us to investigate the possible usefulness of SICUS for assessing the postoperative recurrence of CD. In our series, SICUS showed a high sensitivity (92.5%), PPV (94%), and accuracy (87.5%) for detecting lesions compatible with CD recurrence as assessed by ileocolonoscopy. These results were observed by 2 independent investigators unaware of previous endoscopic or sonographic findings, but not blind in terms of both a previous diagnosis of CD and of an ileocolonic anastomosis for the disease. However, in our CD population SICUS showed a low specificity (20%) and NPV (16.6%), most likely in relation to the low number of TN findings (n = 5). When compared with previous findings, a lower sensitivity (82%) and a comparable PPV (93%) was observed, although associated with a higher specificity (94%) and NPV (84%).32 Discrepancies between the 2 studies are most likely related to the different number of tested patients (72 versus 40), particularly regarding the subgroup with no recurrence (5 versus 18) and to the different time intervals from surgery.32
In our CD population, 5 FN findings were detected by SICUS when using 3 mm as a normal cutoff value.26, 38 However, all these 5 FN findings were detected in patients showing an endoscopic degree of recurrence ≤2, and therefore included patients with no established ileal lesions detectable by SICUS (i.e., <5 aphthoid ulcers and/or anastomosis). We also detected 4 FP findings, in 4 patients with a BWT ranging from 3.5–4.5 mm, including values below the observed median BWT in our study population (5 mm). Discrepant findings arising from sonographic and endoscopic assessment of CD recurrence may also be related to the different view of the lesions provided by these 2 techniques. While sonography allows the visualization of the bowel wall, ileocolonoscopy shows the inner mucosal surface. Nevertheless, when pooling sonographic and endoscopic findings from the whole group of 72 CD patients with ileocolonic resection, we unexpectedly found a significant correlation between the BWT detected by SICUS and the Rutgeerts' score of recurrence (r = 0.67; P = 0.0001). This finding appeared not related to the time interval from surgery, as a significant correlation was also observed when subgrouping patients resected <3 versus ≥3 years before enrolment. Although this cutoff value was quite arbitrary, it was chosen in relation to a comparable amount of patients included in each subgroup. Different from the present findings, in our previous study31 correlation between severity of recurrence as assessed by endoscopy and SICUS at 1 year was at the limit of statistical significance (P = 0.05; r = 0.42). The observed discrepancy may be related to the different number of tested patients (72 versus 22) and/or to the different time interval between surgery and recurrence assessment.31 The time interval between the 2 techniques (within 6 months) may limit the relevance of this study. However, in all 72 patients both clinical activity (CDAI) and treatment strategies (steroids or biologics) were comparable at the time of SICUS and ileocolonoscopy. These observations suggest that the time interval between the 2 procedures may not significantly influence our findings.
In the present study the time interval from surgery included a wide range (3–396 months). The protocol was indeed designed as a cross-sectional study in order to assess the role of SICUS in evaluating CD recurrence in any patient with a previous ileocolonic resection performed at any time before assessment. This issue may assume relevance in clinical assessment of resected CD patients referred for the first time to an IBD service including an available expert sonographer.
Although the retrospective analysis and the small number of patients relapsing within 1 year limit the search for a role of endoscopy versus SICUS as predictors of clinical relapse, our findings further support that a higher Rutgeerts' score is associated with early relapse and also suggest that a higher BWT is observed in almost one-third of patients undergoing relapse.
The finding of a significant correlation between the CDAI and the endoscopic score of recurrence, but not between the CDAI and the BWT, may well be explained by the different view of the lesions provided by these 2 techniques. Different from our study, a significant correlation between the CDAI and BWT has been reported by Maconi et al,41 although this finding was observed in a study population including a higher number of CD patients not selected on the basis of a previous ileocolonic resection.39 The clinical relevance of the observed statistically significant difference between patients with an endoscopic score <2 versus >3 in terms of prestenotic dilation needs further investigation.
The blood flow in the anastomosis was not measured in our CD population, also in relation to previous studies showing that surgical resection may affect superior mesenteric artery blood flow.40, 41
When comparing sonography versus endoscopy in terms of additional findings, ileocolonoscopy was obviously more appropriate in the only 2 patients with not only ileocolonic, but also colocolonic anastomosis. No significant additional lesions were detected by endoscopy. Different from ileocolonoscopy, SICUS also detected extraluminal lesions, including enteroenteric fistulae (4%), lymph nodes enlargement (1.4%), mesenteric adipose tissue alterations (7%), and proximal small bowel involvement (8%). This relatively low frequency of additional findings may be related to the quite short time interval from surgery in most of the patients. In our study, the role of SICUS appeared more relevant in visualizing the neoterminal ileum in patients showing an anastomotic stenosis not passed by the endoscope. In all 22 patients showing a stenosis of the ileocolonic anastomosis, SICUS indeed provided additional information regarding the neoterminal ileum, including the presence and extent of lesions compatible with recurrence and possible dilation above stenosis. These findings suggest that in CD patients with suspected anastomotic stenosis, SICUS is more appropriate than colonoscopy, as the visualization of the neoterminal ileum provided by sonography may avoid the need of small bowel radiology after colonoscopy. In our study, a comparable amount of time was required to perform SICUS (median 40 minutes, range 35–90) and colonoscopy.
Taken together, results from the present study suggest that sonography using oral contrast is a noninvasive technique useful for assessing CD recurrence after ileocolonic resection, providing findings comparable to colonoscopy. Different from colonoscopy, SICUS also allows the visualization of the neoterminal ileum in patients with a stenosis of the anastomosis not allowing passage of the endoscope. However, FN findings may be observed by using SICUS in patients with minor lesions related to CD recurrence. Results from our cross-sectional study, however, suggest that SICUS performed by an experienced sonographer is a noninvasive technique useful for assessing CD recurrence in patients with ileocolonic resection, particularly in those patients at high risk for complications related to colonoscopy (i.e., older age or comorbidities), as also in young patients with a history of high radiation exposure.