CT colonography followed by elective surgery in patients with acute diverticulitis: a radiological-pathological correlation study

A non-specific written informed consent for the use of personal anonymized data in retrospective studies was obtained from all patients. Inclusion criteria were: a documented episode of AD (with availability of records of Emergency Department discharge records), a CTC examination performed after that AD episode, and subsequent elective surgery for diverticular disease. Exclusion criteria included: patients affected by Crohn’s disease or ulcerative colitis, previous colonic surgery, or colorectal cancer.

The final study sample included a series of 59 patients (31 males and 28 females), with a mean age of 58 ± 13 years. All CTC examinations and surgery were performed in our Hospital [Blinded], from 2018 and 2019. Among the 59 total patients other than a CTC, 37/56 (63%) patients had also had a conventional abdominal CT at the time of AD episode at emergency admission; the remaining 22 (37%) had either a clinical visit alone (8/22) or a clinic visit in association with ultrasound or conventional radiography (14/22). For all patients, it was possible to assess the severity of disease at the time of the AD episode was assessed, using the classification system proposed by Ambrosetti as complicated or not complicated [15].

Abdominal CT was performed at the time of emergency department admission. All examinations were performed using a 16-row unit (BrightSpeed, General Electric Healthcare, Waukesha, WI, USA). After an unenhanced CT scan, a second scan was acquired 60–70 s after intravenous injection of 100 mL of non-ionic iodinated contrast material (Iomeron® 400, Bracco Imaging SpA, Milan, Italy) followed by 50 mL of saline flush, both at a flow rate of 3 mL/s, thus obtaining images during the portal venous phase. This portal venous phase is particularly useful in evaluating the severity of the event; in particular it enhances the presence of abscesses, facilitating the diagnosis of complicated acute diverticulitis.

The CT imaging protocol was as follows: gantry rotation time 0.8 s; slice thicknes, 1.25 mm; pitch 1.375; reconstruction interval 1.25 mm; tube voltage 120 kVp (140 in obese patients), modulated tube current from 80 to 440 mA.

CTC exams were performed a mean of about 2 months (55 ± 18 days) after AD episode. Standard bowel preparation consisted of one sachet of mild oral laxative (Movicol; Norgine Italia srl; Milan, Italy) after breakfast, lunch and dinner for the 3 days before CTC with water-soluble iodinated contrast material (diatrizoate) taken 3 h before CTC examination for fecal tagging. Colonic distention was achieved with automated carbon dioxide (CO₂) insufflation. The patient was placed in the left decubitus position for the initial 1.5 L of CO₂, followed by the supine and right lateral decubitus positions for up to 1 L each, followed by the supine position. All the follow-up CTC examinations were performed using a 64-row scanner (LightSpeed; General Electric Medical Systems, Milwaukee, Wisconsin, USA). The CTC acquisition protocol was as follows: gantry rotation time 0.5 s; slice thickness 1.25 mm; table speed 27.5 mm/s; pitch 1.375; reconstruction interval 1 mm; tube voltage 120 kVp (140 in obese patients); tube current 100 mA (300 in obese patients).

Standard imaging protocol included supine and prone scans; a third additional lateral decubitus scan was performed in 28/59 (47%) to achieve the best distension of the sigmoid colon. In patients not able to lay in the prone decubitus, an alternative lateral scan was performed.

A radiologist with 15 years of experience in CTC retrospectively re-analyzed all examinations, including conventional abdominal CT performed at the time of AD in the Emergency Department, and the subsequent follow-up CTC examination after recovery. CTC assessment was blinded to the conventional CT findings, as well as to the clinical and pathologic results.

On conventional CT examination, maximum colonic wall thickening and its longitudinal extent along the site of acute inflammation were measured.

For each patient, the severity of disease at follow-up CTC was calculated using the established diverticular disease severity score (DDSS), a four-class classification proposed by Flor et al. [13], based on maximum sigmoid colon wall thickness and minimum lumen diameter, measured from inner-to-inner wall. The reader first selected the series (prone, supine or right lateral decubitus) in which the sigmoid colon was best distended. All the measurements were taken, in millimetres, on 2D multiplanar reformatted images from the best distended series, always on the images perpendicular to the long axis of the colon segment, using standard soft tissue windowing. For the purpose of this study, we assumed DDSS 1 and 2 to represent mild-moderate diverticular disease and DDSS 3 and 4 to represent severe diverticular disease [13]. The same radiologist evaluated the presence of colonic polyps with a diameter ≥ 6 mm.

All patients underwent elective surgery a mean of 8 ± 4 weeks after CTC at our Institution. Elective surgery was performed in an inflammation-free time and therefore postponed until after antibiotic therapy. Preoperative CTC findings were available to the surgeons, who followed the World Gastroenterology Organization surgical guidelines for diverticular disease [16]. The resection always involved the sigmoid colon and was performed using either laparoscopic or open resection techniques according to surgeon preference. The colon segment affected by AD was resected, extending the resection according to the vascular pedicle, severity, and extension of diverticular disease.

Sigmoid colon surgical biopsy sampling was performed on areas of maximum inflammation and bowel wall thickness. Specimens were fixed in formalin; paraffin-embedded samples and 5-µm-thick slides prepared with haematoxylin–eosin-stain were microscopically examined. A pathologist with 15 years of experience in abdominal intestinal diseases analyzed all specimens. Wall thickness was measured on histological slides from mucosal to serosal surface, in the area of greatest thickness. The presence of inflammation and/or fibrosis was recorded. In particular, inflammation was evaluated and classified as acute, chronic, or mixed (both acute and chronic). For each variable (acute inflammation; chronic inflammation; fibrosis), a 4-point scale was used: 0 (absence); 1 (low); 2 (moderate); 3 (severe) [17, 18]. However, for statistical analysis, all these three pathologic outcomes were dichotomized.

Findings of conventional CT and CTC were correlated with the results from pathologic specimen analysis. Being the three pathologic outcomes (acute inflammation, chronic inflammation, and fibrosis) probably correlated among them, one was chosen as primary outcome according to the strength of the mutual association. Once identified the primary outcome, bivariate associations with potential predictors were first investigated using the Chi square test. Then, variables that were found to be associated with the primary outcome were later entered into a multivariate regression analysis. A predicting model was developed using R² and statistical significance for deciding which predictors to keep in the model. Exploratory analysis was performed for secondary outcomes.