Key words
diverticular disease - diverticulitis - colon - contrast - CT
Introduction
Diverticulosis, i. e. the presence of asymptomatic false diverticula in the gastrointestinal
tract, has a high prevalence of about 30 % in patients 50–70 years and up to 66 %
in patients older than 85 years [1]. Approximately 15–20 % of patients develop clinical symptoms, defined as diverticular
disease (DD). 75 % of symptomatic patients have pain, 1–2 % require hospitalization
and about 0.5 % require surgery [2]. Furthermore, up to 30 % of symptomatic patients present with complicated DD (i. e.,
abscess or perforation). Therefore, diverticulitis accounts for the most frequent
surgically treated disease after cancer in modern western countries [3]
[4].
Different classifications for the staging of DD have been in use, e. g. the Hinchey
classification [5] and the Hansen and Stock classification [6]. More recently, a new radiological classification of diverticular disease (CDD)
has been proposed [7]. Based on this new algorithm, patients may be stratified into the following categories:
outpatient care, in-house therapy, and surgical treatment. In this context, the accurate
diagnosis and staging of DD is crucial for appropriate therapy and is a prerequisite
to avoid insufficient or excessive treatment. Several national and international guidelines
attest that cross-sectional imaging in terms of CT and ultrasound has very high diagnostic
accuracy for the diagnosis of DD, and CT has been advocated as indispensable, especially
if ultrasound is inconclusive and to rule out other differential diagnoses [8]
[9]
[10]. However, the CT technique with respect to the need for and type of contrast is
often not specified in these guidelines. Still, many publications and studies advocate
the use of oral and/or rectal contrast for the appropriate diagnosis of DD [11]
[12]
[13]. On the other hand, especially oral or rectal contrast administration may result
in an unnecessary delay of diagnostic imaging [14].
To sum up, despite a high prevalence of DD in the general population, there is no
consensus regarding the necessity of contrast administration for evaluation. This
may also be due to a lack of high-quality comparative studies and a lack of detailed
imaging recommendations regarding contrast administration in national and international
guidelines. The purpose of our study was therefore to assess the added value of rectal
contrast administration in patients with DD. Specifically, we wanted to test the null
hypothesis, that CT scans without rectal contrast are not inferior to CT scans with
rectal contrast with regards to diagnostic accuracy in patients referred for diagnosis
of diverticular disease.
Materials and Methods
Patients
This retrospective study was approved by the local Institutional Review Board with
a waiver of patient consent granted. The Radiology Information System (RIS) of Hannover
Medical School (blinded for review) was searched for patients referred to CT because
of clinically suspected DD between June 2005 and October 2015. The inclusion criteria
were as follows: age ≥ 18 years, diagnostic CT scan, surgical or clinical follow-up
of at least 4 weeks. The exclusion criteria were non-diagnostic CT scan and no available
follow-up.
CT technique
Abdominal CT scans were acquired on a 64-slice (VCT; n = 398) or 16-slice (LightSpeed;
n = 62) scanner (both GE HealthCare; both slice thickness 1.25 mm) with patients in
supine position. Intravenous contrast (88 ml Imeron 350, Bracco Imaging; flow rate
3.5 ml/sec) was administered and image acquisition started 15 seconds after bolus
detection in the spleen (threshold 250 HU) [15]. In patients at risk (e. g., known contrast allergy or impaired kidney function)
i. v. contrast injection was omitted. Rectal contrast was applied at the discretion
of the radiologist in charge. In respective patients, up to 1000 ml of a 1:10 diluted
water-soluble contrast agent (Gastrografin, Bayer Healthcare) were administered via
a gravity drip into the colon using a blocked balloon catheter [15]. Based on differences regarding contrast administration, this study was divided
into two parts, as demonstrated in [Fig. 1]. Patients with i. v. contrast were included in the main study, while patients who
did not receive i. v. contrast were included in the supplementary study.
Fig. 1 Flowchart demonstrating inclusion of patients referred for CT evaluation of suspected
diverticular disease. Based on whether i. v. contrast was used or not, patients were
assigned either to the main or to the supplementary study. *includes 7 low-dose scans.
Abb. 1 Übersicht der Patienten, welche einer CT-Untersuchung bei klinischem V. a. Divertikulitis
erhielten. Basierend auf einer erfolgten oder fehlenden intravenösen KM-Gabe wurden
die Patienten entweder in die Haupt- oder Zusatzstudie eingeschlossen. *beinhaltet
7 Low-dose-Untersuchungen.
CT evaluation
Image evaluation was performed on a commercially available workstation (Visage 7.1,
Pro Medicus Inc). One board-certified radiologist (12 years of experience with abdominal
CT) and one 4th year radiology resident in consensus reviewed all images with regard to the presence
of DD or an alternative diagnosis. Transverse images as well as multiplanar reformations
were used for evaluation. If diverticular disease was present, findings were classified
using the newly proposed CDD algorithm [7], which is clearer than the Hansen-Stock or Hinchey classification with respect to
radiological diagnosis [7]. Type 0: diverticulosis; type I: acute uncomplicated diverticulitis (Ia: no reaction
of surrounding tissue; Ib: phlegmonous reaction of surrounding tissue); type 2: acute
complicated diverticulitis (IIa: microabscess; IIb: macroabscess; IIc: free perforation),
type III: chronic recurrent diverticulitis (IIIa: symptomatic uncomplicated DD, SUDD;
IIIb: recurrent diverticulitis without complications; IIIc: recurrent diverticulitis
with complications) and type 4: diverticular bleeding.
Patient follow-up
Electronic medical records were reviewed by the authors and patients were followed
up for at least 4 weeks, especially with regard to the final clinical diagnosis (confirmation
of DD or alternative diagnosis). In patients undergoing surgery, the radiological
findings and staging of DD were correlated with intraoperative and histopathological
staging, respectively. In conservatively treated patients clinical follow-up was the
reference standard.
Statistical analysis
Statistical analysis was performed using GraphPad Prism 7 (GraphPad Software Inc.).
To test for potential differences regarding age, sex, CRP value, and white blood cell
between patients between the groups without and with rectal contrast in both study
parts, the Mann-Whitney test and Fisher’s exact test were used (after exclusion of
a Gaussian and equal distribution using the D’Agostino and Pearson normality test).
A post-hoc power analysis was performed using G*Power 3.1 (Heinrich-Heine University,
Düsseldorf). Assuming an α-error of 0.05, an effective size f of 0.25, and a sample
size of 410 patients for the main study (CT with i. v. contrast), the calculated power
is 0.999. With a sample size of 50 patients, the calculated power of the supplementary
study (without. i. v. contrast) is 0.410. The accuracy, sensitivity, specificity,
and positive (PPV) and negative predictive value (NPV) for the CT diagnosis of DD
were calculated in each group. In patients undergoing surgery, CT staging of DD was
correlated with intraoperative and histopathological findings using a weighted Cohen-k
statistic. k-values were interpreted as follows: a value less than 0.20 indicated
poor agreement; a value between 0.21 and 0.40 fair agreement; a value between 0.41
and 0.60 moderate agreement; a value between 0.61 and 0.80 substantial agreement;
a value between 0.81 and 1.00 almost perfect agreement [16]. For all measurements, p < 0.05 indicated a significant difference.
Results
A total of 460 patients (253 males, 207 females; mean age 62 years, range 18–92 years)
with suspected DD of the colon referred for CT were included in this study. 224 patients
were diagnosed with DD, based on clinical follow-up or surgical findings as the reference
standard. This resulted in a prevalence of DD of 49.8 %. Examples of different patients
diagnosed with DD using varying CT acquisition protocols are presented in [Fig. 2].
Fig. 2 Presentation of diverticular disease in four different patients, using variable contrast
administration protocols. A CDD stage Ib (i. v. contrast only) with reaction of surrounding fat tissue (open
arrow). B CDD stage Ib (i. v. and rectal contrast) with reaction of surrounding fat tissue
(open arrow). C CDD stage IIb (neither i. v. nor rectal contrast) with depiction of a macroabscess
(*). D CDD stage IIc (rectal contrast only) with depiction of extraluminal contrast (open
arrow) and large amounts of free intraperitoneal gas (*).
Abb. 2 Beispiele verschiedener Patienten mit der Diagnose einer Divertikulitis im CT unter
Anwendung unterschiedlicher Akquisitionsprotokolle. A CDD-Stadium Ib (nur i. v.-KM) mit Verdichtungen des perikolischen Fettgewebes (offener
Pfeil). B CDD-Stadium IIb (i. v.- und rektales KM) mit Verdichtungen des perikolischen Fettgewebes
(offener Pfeil). C CDD-Stadium IIb (weder i. v.- noch rektales KM) mit Nachweis eines Makroabszesses
(*). D CDD-Stadium IIc (nur rektales KM) mit Detektion von extraluminalem KM (offener Pfeil)
und deutlichen Mengen freier Luft (*) intraperitoneal.
CT with i. v. contrast (main study)
410 patients (226 males, 184 females; mean age 61 years, range 18–92 years) underwent
CT with i. v. contrast injection. In 328 patients CT scans were acquired after i. v.
contrast injection only (group M1), and in 82 patients additional rectal contrast
was administered (group M2). Patient characteristics are presented in detail in [Table 1]. With regard to baseline characteristics (age, sex, CRP value and leukocytes), there
were no significant differences between patients who received rectal contrast and
those who did not.
Table 1
Characteristics of patients undergoing CT for suspected diverticular disease.
Tab. 1 Charakteristiken der Patienten, welche aufgrund des klinischen Verdachts einer Divertikelkrankheit
eine CT erhielten.
|
CT scans with i. v. contrast injection (main study)
|
|
group M1
(i. v. contrast)
|
group M2
(i. v. + rectal contrast)
|
statistical comparison
|
|
number [n]
|
328
|
82
|
n. a.
|
|
male/female [n]
|
180/148
|
48/36
|
p = 0.81 (Fisher’s exact test)
|
|
age [years]
|
61 (48–73)
|
60 (49–71)
|
p = 0.84 (MWU test)
|
|
CRP [mg/l]
(normal range ≤ 5)
|
104 (25–156)
|
101 (30–152)
|
p = 0.95 (MWU test)
|
|
leukocytes [1000/l]
(normal range 3600–10 500)
|
11.7 (7.8–14.9)
|
11.7 (8.4–14.8)
|
p = 0.97 (MWU test)
|
|
prevalence of DD[#] [n]
|
144 (43.9 %)
|
49 (59.8 %)
|
n. a.
|
|
CT scans without i. v. contrast injection (supplementary study)
|
|
group S1
(no contrast)
|
group S2
(rectal contrast)
|
statistical comparison
|
|
number [n]
|
31
|
19
|
n. a.
|
|
male/female [n]
|
15/16
|
12/7
|
p = 0.39 (Fisher’s exact test)
|
|
age [years]
|
61 (51–72)
|
62 (51–75)
|
p = 0.85 (MWU test)
|
|
CRP [mg/l]
(normal range ≤ 5)
|
113 (34–220)
|
120 (30–223)
|
p = 0.7 (MWU test)
|
|
leukocytes [1000/l]
(normal range 3600–10 500)
|
13.9 (9.7–16.7)
|
10.6 (6.4–13.6)
|
p = 0.05 (MWU test)
|
|
prevalence of DD[#] [n]
|
17 (54.8 %)
|
14 (73.7 %)
|
n. a.
|
Values are presented as total number (n) or mean with interquartile range in parentheses,
if not stated otherwise. n. a. = not applicable; DD = diverticular disease; CRP = C-reactive
protein.
Die Angaben entsprechen absoluten Zahlen (n) oder dem Mittelwert mit zugehörigem Interquartilbereich
in Klammern, sofern nicht anders angegeben. N. a. = nichtzutreffend, DD = Divertikelkrankheit;
CRP = C-reaktives Protein.
# as confirmed either by follow-up or surgery.
durch Verlauf oder OP bestätigt.
Diagnosis of DD
193 patients were diagnosed with DD (prevalence 47 %), confirmed either by surgery
(n = 57; 29.5 %) or clinical follow-up (n = 136; 70.5 %). In both groups, DD was most
often localized in the sigmoid colon (81.4 %), followed by the left hemicolon (8.3 %),
right hemicolon (6.2 %), and multifocal (3.1 %) and transverse colon (1 %). The accuracy,
sensitivity, specificity, PPV and NPV for studies with i. v. contrast only (group
M1) and with rectal contrast (group M2) were as follows: 95 %/93 %, 92 %/92 %, 97 %/94 %,
96 %/96 % and 94 %/89 %, respectively. The cumulated CT accuracy of the surgically
relevant CDD stage 2 (a–c combined) was 88 % (group M1: 84.8 %; group M2: 94.1 %).
Looking only at patients with confirmed diagnosis of DD, the rate of true positives,
true negatives, false positives, and false negatives for groups M1 and M2 were 91.7 %/91.8 %,
97.3 %/93.9 %, 2.7 %/6.1 % and 8.3 %/8.2 %, respectively ([Table 2]).
Table 2
Performance of CT for the diagnosis of DD of the colon, with clinical follow-up or
surgical findings as the reference standard.
Tab. 2 Abschneiden der CT im Hinblick auf die Diagnose Divertikelkrankheit des Dickdarms,
basierend auf dem klinischen Verlauf oder dem operativen Befund als Referenzstandard.
|
All i. v. contrast CT
|
i. v. only
|
i. v. + rectal
|
all non- contrast CT
|
non-contrast CT
|
rectal contrast only
|
|
|
(group M1)
|
(group M2)
|
|
(group S1)
|
(group S2)
|
|
accuracy
|
94.4 %
|
94.8 %
|
92.7 %
|
90 %
|
93.5 %
|
84.2 %
|
|
sensitivity
|
91.7 %
|
91.7 %
|
91.8 %
|
90.3 %
|
94.1 %
|
85.7 %
|
|
specificity
|
96.8 %
|
97.3 %
|
93.9 %
|
90.9 %
|
92.9 %
|
80 %
|
|
PPV
|
96.2 %
|
96.4 %
|
95.7 %
|
93.3 %
|
94.1 %
|
92.3 %
|
|
NPV
|
93 %
|
93.7 %
|
88.6 %
|
87 %
|
92.9 %
|
66.7 %
|
|
true positives[*]
|
91.7 %
|
91.7 %
|
91.8 %
|
90.3 %
|
94.1 %
|
85.7 %
|
|
true negatives[*]
|
96.8 %
|
97.3 %
|
93.9 %
|
89.5 %
|
92.9 %
|
80 %
|
|
false positives[*]
|
3.2 %
|
2.7 %
|
6.1 %
|
10.5 %
|
7.1 %
|
20 %
|
|
false negatives[*]
|
8.3 %
|
8.3 %
|
8.2 %
|
9.7 %
|
14.3 %
|
5.9 %
|
PPV = positive predictive value; NPV = negative predictive value.
PPV = positiv prädiktiver Wert; NPV = negativ prädiktiver Wert.
* calculations refer to patients with confirmed diagnosis of DD only.
Die Berechnungen beziehen sich nur auf Patienten mit bestätigter Diagnose einer Divertikelkrankheit.
Staging of DD
Out of 106 patients who underwent surgery, 57 patients were diagnosed with DD. At
CT, most of these patients were classified as DD type IIb (n = 17) or IIc (n = 21)
according to the CDD algorithm. There was almost perfect agreement of staging between
CT and intraoperative/histopathological findings, as indicated by a k-value of 0.823
for all CT scans with i. v. contrast. At subgroup analysis, there was also almost
perfect agreement and substantial agreement for CT without rectal contrast (k = 0.861,
group M1) and CT with rectal contrast (k = 0.782, group M2), respectively ([Table 3]). False-positive CT diagnoses of DD included one patient with colon cancer ([Fig. 3]), one patient with appendicitis (both group M1) and one patient with recurrent bladder
carcinoma (group M2). One patient was diagnosed with free fluid and liver cysts on
CT (intraoperative DD type IIa; group M1), one patient with an enterocutaneous fistula
(intraoperative DD type IIIc, group M2) and another patient with a large bowel obstruction
(intraoperative DD type IIa; group M2).
Table 3
Correlation of i. v. contrast-enhanced CT findings and histopathology in patients
undergoing surgery, with specific emphasis on DD type in respective cases.
Tab. 3 Korrelation von Befunden in der CT (mit i. v.-KM) und Histopathologie bei Patienten,
welche einer operativen Therapie zugeführt wurden.
|
all i. v. contrast
(n = 106)
|
i. v. only (group M1)
(n = 76)
|
i. v. + rectal (group M2)
(n = 30)
|
|
CT
|
surgery
|
CT
|
surgery
|
CT
|
surgery
|
|
diagnosis of DD
|
57 (53.8 %)
|
57 (53.8 %)
|
39 (51.3 %)
|
38 (50 %)
|
18 (60 %)
|
19 (63.3 %)
|
|
other diagnosis
|
49 (46.2 %)
|
49 (46.2 %)
|
37 (48.7 %)
|
38 (50 %)
|
12 (40 %)
|
11 (36.7 %)
|
|
DD stage
|
|
|
|
|
|
|
|
|
0
|
0
|
0
|
0
|
0
|
0
|
|
|
0
|
0
|
0
|
0
|
0
|
0
|
|
|
6
|
3
|
6
|
3
|
0
|
0
|
|
|
10
|
13
|
6
|
9
|
4
|
4
|
|
|
17
|
22
|
11
|
15
|
6
|
7
|
|
|
21
|
15
|
14
|
9
|
7
|
6
|
|
cumulated accuracy stage 2
|
88 %
|
|
84.4 %
|
|
94.1 %
|
|
|
|
0
|
0
|
0
|
0
|
0
|
0
|
|
|
2
|
1
|
2
|
1
|
0
|
0
|
|
|
1
|
3
|
0
|
1
|
1
|
2
|
|
|
0
|
0
|
0
|
0
|
0
|
0
|
|
correlation CT and pathology
|
k = 0.832
|
k = 0.861
|
k = 0.782
|
Values are presented as total numbers. DD = diverticular disease; k = weighted kappa
value.
Die Werte entsprechen absoluten Zahlen. DD = Divertikelkrankheit; k = gewichteter
Kappa.
Fig. 3 75-year-old female patient presenting to the emergency department with left lower
quadrant pain (CRP 144 mg/l, leucocytes 21.200/l). CT with intravenous contrast in
the axial A and coronal B plane depicts diverticula (arrow in B) in the sigmoid colon, discrete pericolic soft tissue changes (open arrows in A and B) and a pericolic abscess with a gas-fluid level (* in A and B). At CT, patient was diagnosed as diverticular disease type IIb (macroabscess, contained
perforation) and referred for surgery. Histopathological diagnosis showed a moderately
differentiated adenocarcinoma with infiltration of the pericolic fat tissue and transmural
inflammation.
Abb. 3 75-jährige Patientin, die sich in der Notaufnahme mit Schmerzen im linken Unterbauch
vorstellte (CRP 144 mg/l, Leukozyten 21 200/l). In der CT mit i. v.-KM in der transversalen
A und koronaren B Ebene ist neben Sigmadivertikeln (Pfeil in B) und einer Verdichtung des perikolischen Fettgewebes (offene Pfeile in A und B) ein perikolischer Abszess mit Luft-Flüssigkeitsspiegel (*in A und B) erkennbar. Im CT wurde die Diagnose einer Divertikulitis Stadium IIb gestellt (Makroabszess,
gedeckte Perforation) und die Patienten einer chirurgischen Therapie zugeführt. Histopathologisch
ergab sich der Befund eines Adenokarzinoms mit Infiltration des perikolischen Fettgewebes
und einer transmuralen Entzündung.
CT without i. v. contrast (supplementary study)
50 patients (27 males, 23 females; mean age 62 years, range 35–85 years) underwent
non-contrast-enhanced CT examinations, including 31 patients with neither i. v. nor
rectal contrast (group S1) and 19 patients with rectal contrast only (group S2). With
regard to baseline characteristics (age, sex, CRP value, and leukocytes), there were
no significant differences between patients who received rectal contrast and those
who did not ([Table 1]).
Diagnosis of DD
31 patients were diagnosed with DD (prevalence 62 %), confirmed either by surgery
(n = 12; 38.7 %) or clinical follow-up (n = 19; 61.3 %). DD was most often localized
in the sigmoid colon (71 %). The accuracy, sensitivity, specificity, PPV, and NPV
for studies without (group S1) and with rectal contrast (group S2) were 94 %/84 %,
94 %/86 %, 93 %/80 %, 94 %/92 %, 93 %/67 %, respectively ([Table 2]).
Staging of DD
Out of 19 patients undergoing surgery, 11 patients were diagnosed with DD. For CT
scans with neither i. v. nor rectal contrast, there was substantial agreement with
intraoperative findings regarding the staging of DD (k = 0.748). The agreement of
rectal contrast-enhanced CT scans with surgery was similar (k = 0.759) ([Table 4], [Fig. 4]). The cumulated CT accuracy of the surgically relevant CDD stage 2 (a-c combined)
was 100 %. In each group there was one false-positive diagnosis of DD at CT. Specifically,
this included one patient staged as Ib DD (intraoperative found to have peritoneal
carcinomatosis) and one patient staged as IIb DD (intraoperative diagnosis of recurrent
urothelial carcinoma).
Table 4
Correlation of non-contrast-enhanced CT findings and histopathology in patients undergoing
surgery, with specific emphasis on DD type in respective cases.
Tab. 4 Korrelation von Befunden in der CT (ohne i. v.-KM) und Histopathologie bei Patienten,
welche einer operativen Therapie zugeführt wurden.
|
all non-contrast
(n = 19)
|
no rectal contrast (group S1)
(n = 10)
|
with rectal contrast (group S2)
(n = 9)
|
|
CT
|
surgery
|
CT
|
surgery
|
CT
|
surgery
|
|
diagnosis of DD
|
14 (73.7 %)
|
12 (63.2 %)
|
6 (60 %)
|
5 (50 %)
|
8 (88.9 %)
|
7 (77.8 %)
|
|
other diagnosis
|
5 (26.3 %)
|
7 (36.8 %)
|
4 (40 %)
|
5 (50 %)
|
1 (11.1 %)
|
2 (22.2 %)
|
|
DD stage
|
|
|
|
|
|
|
|
|
0
|
0
|
0
|
0
|
0
|
0
|
|
|
0
|
0
|
0
|
0
|
0
|
9
|
|
|
2
|
1
|
1
|
0
|
1
|
1
|
|
|
0
|
1
|
0
|
0
|
0
|
0
|
|
|
5
|
5
|
2
|
3
|
3
|
3
|
|
|
5
|
4
|
3
|
2
|
2
|
2
|
|
cumulated accuracy stage 2
|
100 %
|
|
100 %
|
|
100 %
|
|
|
|
0
|
0
|
0
|
0
|
0
|
0
|
|
|
1
|
1
|
0
|
0
|
1
|
1
|
|
|
1
|
0
|
0
|
0
|
1
|
0
|
|
|
0
|
0
|
0
|
0
|
0
|
0
|
|
correlation CT and pathology
|
k = 0.772
|
k = 0.748
|
k = 0.759
|
Fig. 4 50-year-old male patient with polycystic kidney disease. The patient presented with
an acute abdomen to the emergency room (CRP 198 mg/l, leucocytes 8000/l). CT without
intravenous but rectal contrast depicts gas-filled diverticula in the sigmoid colon
with pericolic soft tissue changes (arrows in A) and free gas (open arrows in B) in the upper abdomen, consistent with diverticular disease type IIc (free perforation).
Findings were confirmed at surgery and patient underwent sigmoid resection. Of note,
there is no extraluminal contrast.
Abb. 4 50-jähriger Patient mit polyzystischer Nierenerkrankung. Der Patient stellte sich
mit einem akuten Abdomen in der Notaufnahme vor (CRP 198 mg/l, Leukozyten 8000/l).
In der nativen CT mit rektalem KM sind luftgefüllte Divertikel im Sigma, perikolische
Verdichtungen des Fettgewebes (Pfeile in A) und freie Luft (offene Pfeile in B) im Oberbauch erkennbar, passend zu einer Divertikulitis Stadium IIc (freie Perforation).
Der Befund bestätigte sich intraoperativ und der Patient erhielt eine Sigmaresektion.
Zu beachten ist, dass kein KM-Extraintestinat nachweisbar ist.
Discussion
In the present study, we demonstrated that CT without rectal and oral contrast has
a very high sensitivity and specificity for the diagnosis of DD and enables reliable
staging of disease severity. Therefore, we accept the null hypothesis that CT without
rectal contrast is not inferior to scans with additional positive rectal contrast
regarding diagnostic accuracy in patients referred for suspected DD. In a subgroup
analysis, we were further able to demonstrate that this is true not only for scans
with intravenous contrast enhancement, but also for CT scans without i. v. contrast
injection.
The omission of oral and rectal contrast administration at CT has several advantages
especially concerning clinical workflow and patient comfort. Oral contrast administration
causes a diagnostic delay for the patient due to the time needed for ingestion of
the contrast agent and gastrointestinal transport. In a randomized controlled trial
by Garra et al., the median time to ingest 2 liters of an oral contrast agent was
more than 100 minutes, potentially compromising individual patient safety in critical
and time-sensitive conditions [17]. Similarly, rectal contrast administration requires time, is dependent on patient
compliance, and may aggravate patient discomfort (also as continence may be impaired,
making adequate intraluminal filling improbable). Furthermore, administration of a
positive contrast agent (rectally or orally) may mask the presence of abnormal bowel
wall enhancement (e. g. in ischemia) or intestinal bleeding [18], impeding evaluation, especially with regard to differential diagnosis in patients
presenting with acute abdominal pain. It has been postulated that colonic opacification
helps to distinguish intraluminal air from fluid collections, e. g. in patients with
(contained) perforated DD [12]. However, studies have shown that extraluminal contrast is only visible in a small
number of patients with surgically proven alimentary tract perforation [19]. Incomplete luminal distention has been described as a common technical pitfall
that can hamper discrimination of a true focal wall thickening from apparent wall
thickening due to incomplete distention [11]. In our study, we did not find this to be a diagnostic problem especially concerning
differentiation of acute complicated DD. On the other hand, enteric contrast administration
may be beneficial for the assessment of fistula, e. g. in patients with recurrent
DD [20], a circumstance that we did not specifically assess in our study.
To the best of our knowledge, no study has compared CT with and without rectal contrast
specifically for the clinical workup of DD. A few studies have compared the value
of CT with and without rectal/enteric contrast in patients with appendicitis [21] and in patients with acute non-traumatic pain [22], concluding that enteric contrast does not improve the radiological diagnosis in
most patients [23]. The diagnostic accuracy (94.8 %), sensitivity (91.7 %), and specificity (97.3 %)
of CT with i. v. contrast alone in our study are comparable with previous studies
in patients with DD who underwent CT with varying enteric contrast protocols. Cho
et al. reported a sensitivity of 93 % in a small group of patients (27 with DD) for
CT scans with intravenous and oral contrast [24]. Ambrosetti et al. performed CT with intravenous, oral, and rectal contrast in patients
with acute left colonic diverticulitis and found a sensitivity of 97 % [25]. Similarly, in a prospective study of 120 patients with suspected left-sided diverticulitis,
CT with intravenous and rectal contrast had an accuracy of 98 %, a sensitivity of
97 %, and a specificity of 98 % [13]. To the best of our knowledge, there are only two recently published studies that
evaluate the newly proposed CDD algorithm in correlation with intraoperative findings
[26]
[27]. While La Torre et al. found a high interobserver agreement of CT data between two
radiologists (k = 0.905), the concordance of imaging and surgical findings was rather
weak (k = 0.213 and 0.248) [26]. In comparison, correlation with surgical findings was much higher in our study.
A possible explanation for these distinctly differing results may be the lack of experience
with this newer classification. Furthermore, La Torre et al. found a low sensitivity
for the classification of stages 2c1 (purulent peritonitis) and 2c2 (fecal peritonitis),
a differentiation we did not perform in our study. The low efficacy for the prediction
of postoperative outcome also somehow contradicts a recently published prospective
study by Lauscher and colleagues, who found that the CDD algorithm enables stratification
of different types of diverticulitis in terms of course of treatment [27]. It has been noted that this new classification is clearer than the Hansen-Stock
or Hinchey classification with respect to radiological diagnosis and treatment-relevant
categorization [7]. Implementation of this new classification in the clinical routine was straightforward
in our experience. The chosen clinical follow-up of 4 weeks may have led to underestimation
of chronic recurrent DD. There was a slight tendency toward overstaging at CT in patients
with stage IIb (acute complicated DD with macroabscess) in our study when compared
with intraoperative findings. This trend toward overestimation in less pronounced
inflammatory forms of diverticulitis was also reported in a study by Ritz et al. in
204 patients. They used the Hinchey classification and CT with intravenous, oral,
and rectal contrast [28].
Our study does have some limitations, also due to the retrospective design. First,
we did not assess the value of oral contrast in patients with suspected DD, but rather
focused on rectal administration. Second, the diagnosis of DD was confirmed in most
patients by means of clinical follow-up. This is due to the fact that DD is typically
diagnosed clinically (clinical presentation in correlation with elevated CRP) and
that patients with uncomplicated DD usually do not require surgery. Third, the group
of patients without intravenous contrast administration (supplementary study) was
small, as reflected also by the power analysis for this study part. However, the encouraging
results in this cohort (including 7 patients with a non-contrast-enhanced low-dose
protocol) warrant further investigation regarding the necessity of i. v. contrast
and the accuracy of low-dose scans for patients with suspected DD, as has already
been suggested [29]. Lastly, we did not look in detail at the value of rectal contrast administration
for discriminating DD and colon cancer, which is the most important differential diagnosis.
In conclusion, the results of our study show that CT without rectal contrast administration
has a high diagnostic accuracy in patients with suspected DD. Staging of disease severity
using the recently proposed CDD algorithm correlates substantially with intraoperative
findings. Adding rectal contrast administration did not contribute to diagnosis and
staging. In this regard, it did not matter if scans were acquired with or without
intravenous contrast. For the benefit of time, clinical workflow, and patient comfort,
rectal contrast administration may thus be omitted in patients referred for suspected
diverticular disease of the colon.
-
CT has a high sensitivity and specificity for the diagnosis of DD.
-
CT staging using the CDD algorithm correlates very well with surgery.
-
Positive rectal contrast administration does not improve diagnosis and radiological
staging.
