Introduction
Mechanical small bowel obstruction (SBO) was one of the main complications requiring emergency surgery, accounting for 20~50% of emergency surgeries [
1]. Among them, about 60% of the incidence was caused by intestinal adhesion, especially after pelvic surgery, and most patients could be improved by medical care [
2]. Mechanical SBO requiring surgery includes reversible ischemic changes and transmural necrosis of the small bowel [
3]. Determining the occurrence of small bowel necrosis is an important indicator to judge the severity of intestinal obstruction and the indications of surgery [
4]. Early and timely surgical intervention could reduce patients’ mortality and improve their prognosis [
5]. Therefore, if intestinal necrosis could be accurately predicted, it would be beneficial to improve the clinical management of patients with mechanical SBO. However, at present, there is no a standardized and evidence-based criteria to judge small bowel necrosis [
6]. In clinical practice, when a patient is diagnosed with signs of peritonitis through physical examination, the condition is often severe, and even accompanied with septic shock. The sensitivity and specificity of laboratory examination for diagnosing mechanical SBO with small bowel necrosis are relatively low [
7].
Currently, computed tomography (CT) is the first choice for diagnosing intestinal obstruction [
8], and is recommended by several studies [
9‐
11]. However, most studies focused on enhanced CT [
12], which may underestimate the diagnostic value of unenhanced CT [
13]. Some studies had found that the absent or diminished enhanced of small bowel loop [
13‐
15] and increased bowel wall attenuation in an isolated loop could be regarded as typical signs of small bowel necrosis in unenhanced CT [
3,
16,
17]. However, these studies usually only paid attention to a few CT signs, and the sensitivity and specificity of the same signs in different literatures were quite different, and failed to establish a predictive model. In addition, some patients may not be suitable for enhanced CT because they are allergic to contrast media, or complicated with renal insufficiency and septic shock, etc. Therefore, the diagnostic value of enhanced CT for patients with mechanical SBO is better than that of unenhanced CT, but we still believe that the study of unenhanced CT also has important clinical value.
Therefore, our study aimed to find objective risk factors of mechanical SBO with small bowel necrosis on unenhanced CT imaging, and establish a predictive model.
Discussion
Imaging diagnosis is of great significance to the treatment of patients with mechanical SBO. Chuong et al. [
32] revealed that the combination of unenhanced and enhanced CT could improve the diagnostic efficiency of mechanical SBO. However, unenhanced CT is more economical, rapid and effective, and it can also avoid contraindications related to the enhanced CT. Our results also suggested that many features in unenhanced CT imaging had good diagnostic value for mechanical SBO with small intestinal necrosis, and the predictive model had achieved good AUC value.
Intestinal vascular circulation disorder could cause pathological changes of various tissues through various mechanisms (such as exudation, deterioration and proliferation), which could be reflected on CT imaging [
33].
The exudation in different tissues presented different characteristics on CT imaging. (1) The exudation in intestinal wall was usually presented as a thickened, double-layered ring sign or target sign. Sometimes it could also cause oozing blood in the intestinal wall, which could be easily identified as a mixed high-density shadow on unenhanced CT imaging [
18]. Unfortunately, the CT value of small-bowel wall failed to pass the ICC test, which should be related to the measurement of completely different areas of the small bowel wall by two radiologists. However, we still believe that the CT value of small-bowel wall would be a meaningful image indicator as long as we find a suitable measurement standard. (2) The exudation in mesentery was manifested as a cloudy water-like density which was usually distributed along the mesangial side of the involved intestinal wall [
26,
34]. Corresponding features could be observed on CT imaging, such as mesenteric ground glass sign, mesenteric vasodilation, diffuse mesenteric haziness, change of CT value of mesenteric vessels, and whirlpool sign. In our study, the diffuse mesenteric haziness, and the difference of CT values between mesenteric vessel and aorta were all independent risk factors of mechanical SBO with small bowel necrosis. The mesenteric vasodilatation and whirlpool sign were not related to mechanical SBO with small bowel necrosis, which was consistent with previous studies [
28,
35]. (3) The exudation in the intestinal lumen was presented as lumen effusion and expansion [
4,
36], which were common CT signs of mechanical SBO, but not specificity in mechanical SBO with small bowel necrosis. In our study, there was no significant difference between the two groups in the small bowel lumen dilation and maximum diameter of small bowel lumen dilation. In addition to difference of exudation sites, difference of exudation components also presented different characteristics on CT imaging. When the exudative components were mainly liquid or protein, the CT value of ascites was often less than 10 Hu. If the CT value of ascites was above 10 Hu, it generally indicated co-infection, hemorrhagic effusions or necrosis [
37]. In clinical practice, ascites could be easily obtained by paracentesis and had more objective results. Therefore, we did not measure the CT value of ascites. The results showed that the proportion of ascites in mechanical SBO patients with small bowel necrosis was higher, but there was not significant difference in multivariable logistic regression analysis.
In the process of metamorphism, according to the different stages of pathological changes, the intestinal wall shows successively thickening, weakening or disappearance of reinforcement, or thin paper-like shape [
38]. In our study, increased attenuation of small bowel wall and small bowel wall annular thickening showed significant differences between the two groups and univariable logistic regression analysis. However, small bowel wall annular thickening may only be related to mechanical SBO with intestinal necrosis on unenhanced CT images, because this result had not been found in the current research on enhanced CT images [
3,
25]. The increased attenuation of small bowel wall was also an independent predictor of mechanical SBO with small bowel necrosis. Previous literatures had shown that increased attenuation of small bowel wall was the most specific CT sign of intestinal ischemia [
3,
39,
40]. Meanwhile, mucosal cells were sensitive to ischemia and hypoxia. Mucosal barrier dysfunction could lead to submucosal gas accumulation, gas-liquid inversion, submucosal gas accumulation and portal venous gas accumulation [
41]. Small bowel fecal gas sign and pneumatosis intestinalis were caused by this reason, but they were not the risk factor of mechanical SBO with small bowel necrosis. Only one patient was identified with portal venous gas in our study. This is similar to the research of Lebert et al. [
27]. The portal venous gas is very rare, and it is not specific for small bowel necrosis.
The CT findings of intestinal hyperplasia stage are thickening of intestinal wall and lead-like ankyloses, which usually indicates severe or intestinal ischemia [
25]. U-/C-shaped small bowel loop signs were typical manifestations [
42]. Our results also showed that U-/C-shaped small bowel loop was an independent risk factor of mechanical SBO with small bowel necrosis. As we all know, on unenhanced CT imaging, the signs of closed loop obstruction may predict small bowel necrosis better than the U-/C-shaped small bowel loop signs [
43], but the signs of closed loop obstruction are usually accompanied by the U-/C-shaped small bowel loop in cross section, while the appearance of the U-/C-shaped small bowel loop does not represent closed loop obstruction [
44]. Therefore, we finally chose the U-/C-shaped small bowel loop as one of the predictive indicators. Moreover, inflammatory factors released by necrotic intestinal wall could cause obvious peritoneal inflammation, leading to significant ascites, peritoneal thickening and fascia-like thickening [
45]. In our study, there was a significant difference in peritoneal thickening between the two groups, but peritoneal thickening was not a risk factor in the logistic regression analysis.
Our study also found that mechanical SBO patients without history of abdominal surgery were more likely to be complicated with small bowel necrosis. This may be the deviation caused by our small case. Theoretically, the history of abdominal surgery is related to intestinal adhesion, and adhesion bands and internal hernias caused by intestinal adhesion are more likely to cause closed-loop intestinal obstruction, leading to small bowel necrosis [
3,
22,
23] Nevertheless, this also gives a warning to gastrointestinal surgeons that mechanical SBO patients who had no history of abdominal surgery should also paid attention to in clinical practice.
Darras et al. [
46] suggested that contrast-enhanced dual energy CT with virtual mono energetic image reconstruction at 70 keV maximizes the contrast to noise ratios of small bowel mural enhancement and could increase the diagnosis in assessing small bowel wall enhancement in patients with SBO. However, more scholars believed that dual energy CT may be more valuable in the diagnosis of acute mesenteric ischemia, rather than mechanical SBO [
47]. Moreover, dual energy CT is expensive and not a routine examination item. The major advantage of this study was to measure the difference of CT values between mesenteric vessel and aorta, which was confirmed as an independent risk factor of mechanical SBO with small bowel necrosis. Through the vitro experiment, Kirchhof et al. [
48] found that the CT value of normal human abdominal aorta blood is 35–50 Hu, and that of venous blood is 55 ± 5 Hu, and the higher the content of deoxyhemoglobin is, the higher the CT value. It was also believed that the changes of CT values on unenhanced CT could be used as evidence to suspect intravascular obstruction. Meanwhile, Morita et al. [
49] used the CT value changes in cerebrovascular density on unenhanced CT to diagnose cerebrovascular diseases. We also found that the difference of CT values between the experimental group and the control group was statistically significant, but there was a lot of overlap, which may need more clinical studies to confirm. At the same time, we established a predictive model for mechanical SBO with small bowel necrosis, and obtained satisfactory AUC and moderate calibration results.
The limitations of this study were as follows: Firstly, the small sample size and high proportion of surgical patients may lead to potential bias. Secondly, the predictive model was only validated internally instead of externally. Thirdly, this study did not distinguish simple and complicated mechanical SBO. According to the classification of mechanical SBO, it would be better to distinguish simple and complicated mechanical SBO. However, it is always difficult to accurately distinguish them in clinical practice [
50].
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.