European Journal of Radiology 84 (2015) 1701–1707

Contents lists available at ScienceDirect

European Journal of Radiology journal homepage: www.elsevier.com/locate/ejrad

CT colonography with rectal iodine tagging: Feasibility and comparison with oral tagging in a colorectal cancer screening population Emanuele Neri a,∗ , Annalisa Mantarro a , Lorenzo Faggioni a , Paola Scalise a , Pietro Bemi a , Francesca Pancrazi a , Giuseppe D’Ippolito b , Carlo Bartolozzi a a b

Diagnostic and Interventional Radiology – Department of Translational Research and New Technologies in Medicine and Surgery, University of Pisa, Italy Federal University of São Paulo – Sena Madureira 1500 – Vila Mariana, UNIFESP, São Paulo, SP, Brazil

a r t i c l e

i n f o

Article history: Received 24 March 2015 Received in revised form 11 May 2015 Accepted 18 May 2015 Keywords: CT colonography Colorectal cancer Colorectal polyps Fluid tagging Image quality

a b s t r a c t Purpose: To evaluate feasibility, diagnostic performance, patient acceptance, and overall examination time of CT colonography (CTC) performed through rectal administration of iodinated contrast material. Materials and methods: Six-hundred asymptomatic subjects (male:female = 270:330; mean 63 years) undergoing CTC for colorectal cancer screening on an individual basis were consecutively enrolled in the study. Out of them, 503 patients (group 1) underwent CTC with rectal tagging, of which 55 had a total of 77 colonic lesions. The remaining 97 patients (group 2) were randomly selected to receive CTC with oral tagging of which 15 had a total of 20 colonic lesions. CTC findings were compared with optical colonoscopy, and per-segment image quality was visually assessed using a semi-quantitative score (1 = poor, 2 = adequate, 3 = excellent). In 70/600 patients (11.7%), CTC was performed twice with both types of tagging over a 5-year follow-up cancer screening program. In this subgroup, patient acceptance was rated via phone interview two weeks after CTC using a semi-quantitative scale (1 = poor, 2 = fair, 3 = average, 4 = good, 5 = excellent). Results: Mean per-polyp sensitivity, specificity, positive and negative predictive values of CTC with rectal vs oral tagging were 96.1% (CI95% 85.4 ÷ 99.3%) vs 89.4% (CI95% 65.4 ÷ 98.1%), 95.3% (CI95% 90.7 ÷ 97.8%) vs 95.8% (CI95% 87.6 ÷ 98.9%), 86.0% (CI95% 73.6 ÷ 93.3) vs 85.0% (CI95% 61.1 ÷ 96.0%), and 98.8% (CI95% 95.3 ÷ 99.8%) vs 97.2% (CI95% 89.4 ÷ 99.5%), respectively (p > 0.05). Polyp detection rates were not statistically different between groups 1 and 2 (p > 0.05). Overall examination time was significantly shorter with rectal than with oral tagging (18.3 ± 3.5 vs 215.6 ± 10.3 minutes, respectively; p < 0.0001). Conclusions: Rectal iodine tagging can be an effective alternative to oral tagging for CTC with the advantages of greater patient acceptance and lower overall examination time. © 2015 Elsevier Ireland Ltd. All rights reserved.

1. Introduction CT colonography (CTC) has gained wide reliability in the clinical practice and nowadays is a standardized radiological procedure for the evaluation of colonic lesions that has superseded double contrast barium enema owing to its higher diagnostic accuracy, better patient tolerance, and lower radiation dose [1,2]. Patient’s compliance, the safety of the CTC procedure, and the diagnostic accuracy of CTC depend on multiple factors (e.g. bowel preparation

∗ Corresponding author at: Diagnostic and Interventional Radiology – Department of Translational Research and New Technologies in Medicine and Surgery, University of Pisa, Italy. E-mail address: [email protected] (E. Neri). http://dx.doi.org/10.1016/j.ejrad.2015.05.025 0720-048X/© 2015 Elsevier Ireland Ltd. All rights reserved.

and tagging regimen, colonic distention, data acquisition protocol and image interpretation), which are strongly dependent on an appropriate methodological approach and radiologist’s experience [3]. To ensure that a correct methodological approach is used, compliance to guidelines is mandatory. The most recent consensus statement of the European Society of Gastrointestinal and Abdominal Radiology (ESGAR) suggests that CTC should be carried out in accordance to specific standard levels of quality, including adequate bowel cleansing and oral tagging with barium and/or iodine contrast agents as essential key factors [4]. To improve bowel preparation, oral tagging of both residual fluid and stool with positive (either iodine or barium) contrast material is a crucial step that can be adopted in conjunction, or can follow, bowel cleansing [5–7].

1702

E. Neri et al. / European Journal of Radiology 84 (2015) 1701–1707

Several tagging regimens have been proposed, however no definite consensus has been reached so far about the most effective scheme in terms of contrast agent, dose, administration protocol, etc. [8,9]. Sodium and meglumine amidotrizoate is an ionic iodinated contrast agent, mostly used for oral tagging, that may encounter variable patient compliance due to its laxative effect and taste, which may be unpleasant to some patients. Its oral administration is safe, yet rare anaphylactoid reactions after oral administration have been reported [10]. Moreover, the quality of fluid tagging depends on the patient’s transit time through the small bowel, and overdilution of contrast medium may result in poor fluid tagging, potentially impairing colonic assessment. Small bowel transit time is also an issue when a same day tagging protocol is used, requiring the patient to be admitted to the CT suite at least 3–4 hours before CTC in order to start the oral assumption of contrast agent in a water solution. Although the majority of such limitations can be overcome if oral tagging is taken the day before, with such policy the outpatients are not under direct medical control during the assumption of iodinated contrast, which may cause (even if rare) allergic reactions [9,11]. In this setting, rectal administration of iodinated contrast material for fluid tagging immediately prior to colonic insufflation on the CT table, through the same rectal probe, may be an attractive alternative to same-day oral tagging in terms of patient safety and overall tolerance. With the aim to improve CTC patient’s acceptance and to overcome the limitations of the same-day oral tagging, we tested the introduction of iodinated contrast material through a rectal tube. The purpose of our study was to evaluate feasibility, homogeneity of fluid tagging, and diagnostic accuracy of CTC performed through rectal administration of a solution of iodinated contrast, and to assess image quality, patient’s acceptance, and overall examination time, compared with oral iodine tagging. 2. Methods and materials 2.1. Inclusion criteria Six-hundred asymptomatic subjects (male:female = 270:330, age 50–76 years, mean 63 years) undergoing CTC for colorectal cancer screening on an individual basis were prospectively enrolled in a consecutive manner. All subjects were referred to CTC by a gastroenterologist, general practitioner, internal medicine specialist or surgeon in case of refusal of colonoscopy or a previous incomplete examination. Out of them, in 503 patients (group 1) CTC was carried out with rectal tagging, while the remaining 97 subjects (group 2) were randomly selected to get CTC with oral tagging using a sameday preparation protocol. Patients undergoing CTC with rectal and oral tagging, respectively were randomized with an allocation ratio of 5:1. In this respect, we had estimated our sample size taking into account data from a large screening population study (ACRIN Trial), reporting a frequency of adenomatous polyp detection of 15.4% [12]. On this basis, we assumed that at least 500 patients should be recruited in the rectal tagging group to obtain a cohort of at least 70 patients with positive findings. In 70/600 patients (11.7%), CTC was performed twice over a 5year program for colorectal cancer screening using both types of tagging. In particular, the first CTC was conducted with oral tagging and the second one with rectal tagging. Optical colonoscopy was carried out in all subjects with positive CTC findings 10–20 days after CTC. 2.2. Exclusion criteria Exclusion criteria included history of familial adenomatous polyposis or hereditary non-polyposis cancer syndromes, known

colorectal cancer, a suspected diagnosis of inflammatory bowel disease, bowel obstruction, acute diverticulitis, and contraindications to the administration of iodinated contrast agents (including previous allergic reactions to iodine products or iodinated contrast material) [9,13]. The study was performed in accordance with the Declaration of Helsinki, and written informed consent was obtained from all patients after the purpose and protocol of the study had been fully explained. 2.3. Bowel preparation 2.3.1. Diet and colonic cleansing In the 3 days before CTC, patients were asked to avoid products containing fibers and to restrict their diet to meat, eggs, fish, and milk. In addition, patients were also asked to assume a light supper as vegetable or meat broth the day before the examination. Fasting was required for the day of the examination [14,15]. The day before CTC, all patients received 4 tablets of Bisacodyl 5 mg (Lovoldyl, Promefarm, Milan, Italy), followed about 3 hours later by 2 l of Macrogol 4000 solution (Lovol-Esse, Promefarm, Milan, Italy). 2.3.2. Fluid tagging and colonic insufflation Oral tagging – 50 mL of diatrizoate dimeglumine with an iodine concentration of 370 mg/mL (Gastrografin® , Bayer Schering, Germany) diluted into 500 mL of water were orally administered three hours before carbon dioxide (CO2 ) insufflation (same-day preparation). CO2 delivery was controlled by a dedicated electronic device (PROTOCO2 L® , General Electric, Italy). Rectal tagging – No oral iodinated contrast agent was administered. Immediately before CTC, all patients received an enema of 50 mL of diatrizoate dimeglumine (Gastrografin® ) diluted in 300 mL warm tap water through a dedicated ballooned rectal enema tube equipped with an antireflux valve to avoid reflux of fecal material. Subsequently, patients were instructed to turn themselves once on the CT table in a left-to-right direction at least 1 time, so as to ensure homogeneous intraluminal distribution and mixing of contrast medium with the residual fluid. Colonic distention was obtained by automatic CO2 insufflation (2–4 l) with the patients turned in the right decubitus. As with oral tagging, CO2 delivery was controlled by a dedicated electronic device (PROTOCO2 L® ). 2.4. CTC image acquisition protocol All data were acquired on a 64-row CT scanner (LightSpeed VCT, General Electric, Milwaukee, WI) in the supine and prone position with the following parameters: detector configuration 64 × 0.625 mm, tube voltage 120 kV, tube current 50–80 mA depending on patient size, rotation time 500ms, beam pitch 1.375:1, reconstructed slice thickness 1.25 mm, reconstruction increment 1 mm, and Soft reconstruction kernel. A spasmolytic agent (Hyoscine-N-butylbromide; Buscopan® , Boehringer Ingelheim, Germany) was administered only when colonic insufflation had been judged inadequate on scout view images. A single breathhold CT acquisition encompassing the entire colon was performed and repeated twice, with the patient first in the supine and then in the prone position. 2.5. Data analysis 2.5.1. Lesion detection Three abdominal radiologists with about 12–15 years of experience in gastrointestinal imaging and CTC interpreted all CTC studies blindly and independently in 2D mode using axial and multiplanar reconstructions (MPR), using 3D analysis (Virtual Endoscopy)

E. Neri et al. / European Journal of Radiology 84 (2015) 1701–1707

1703

Table 1 Diagnostic performance of CTC for detecting polypoid lesions in group 1 (rectal tagging): overall and size-wise analysis. PPV = positive predictive value, NPV = negative predictive value.

Sensitivity (CI95% ) Specificity (CI95% ) PPV (CI95% ) NPV (CI95% )

Overall (N = 77)

30 mm (N = 20)

0.961 (0.854 ÷ 0.993) 0.953 (0.907 ÷ 0.978) 0.86 (0.736 ÷ 0.933) 0.988 (0.953 ÷ 0.998)

1 (0.56 ÷ 1) 0.97 (0.929 ÷ 0.989) 0.583 (0.286 ÷ 0.835) 1 (0.971 ÷ 1)

0.946 (0.805 ÷ 0.991) 0.988 (0.953 ÷ 0.998) 0.946 (0.805 ÷ 0.991) 0.988 (0.953 ÷ 0.998)

1 (0.561 ÷ 1) 0.994 (0.962 ÷ 1) 0.875 (0.467 ÷ 0.993) 1 (0.971 ÷ 1)

1 (0.799 ÷ 1) 1 (0.954 ÷ 1) 1 (0.799 ÷ 1) 1 (0.954 ÷ 1)

for problem solving. Consensus reading was performed in case of disagreement among readers. All lesions were measured in 2D mode with axial and MPR on both the supine and prone acquisitions, and the maximum lesion diameter was considered for polyp size classification. CTC findings were compared with optical colonoscopy, used as reference standard, to calculate per-polyp sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) of CTC with rectal tagging [13,16]. Moreover, CTC findings obtained by means of rectal tagging were compared with those achieved by oral tagging.

and patient acceptance scores were compared between groups 1 and 2 using the two-tailed Mann–Whitney test, and inter-rater agreement for image quality was quantified by means of the Cohen k coefficient. Sensitivity, specificity, PPV and NPV were expressed as mean and 95% confidence interval (CI95% ), while intraluminal density values and all scores were expressed as mean ± standard deviation. For all tests, a p-value less than 0.05 indicated statistical significance. Statistical analysis was performed using software (GraphPad Prism v. 5, www.graphpad.com).

2.5.2. Image quality All CTC images were evaluated on a commercial workstation (Advantage Windows 4.6, General Electric) equipped with a dedicated plugin (Colon VCAR® ) enabling generation of synchronized axial, MPR, Volume Rendering, and Virtual Endoscopy images. Intraluminal enhancement in CT numbers (Hounsfield Units, HU) was measured in the various colonic segments (i.e. cecum-ascending colon, transverse colon, descending colon, and sigmoid-rectum) by placing circular regions of interest inside the contrast-filled lumen, taking care to avoid the normal bowel wall, polyps, or stool residues (Fig. 1). Per-segment image quality was visually assessed in blind by the three readers using a semi-quantitative score (1 = poor, 2 = adequate, 3 = excellent). Persegment image quality was defined as excellent if there were no residual liquid or faecal material, inhomogeneous tagging in the colonic segments, and poor colonic distension. Image quality was ranked as adequate if the following conditions were satisfied: polyps smaller than 6 mm could be ruled out, no or little fecal residues, sufficiently homogeneous distribution of iodine tagging, and acceptable colonic distention. CTC images not fulfilling the criteria for adequate quality were ranked as poor quality. The overall examination time was recorded, spanning from the beginning of patient evaluation by performing radiologists to the end of every CTC examination [17,18].

3.1. Polyp detection

2.5.3. Patient acceptance Patient acceptance was assessed through a phone interview two weeks after CTC in the subset of patients (70/600, 11.7%) who had undergone CTC twice with both rectal and oral tagging, and included the following: bowel preparation (including laxative preparation and faecal tagging), colonic insufflation, patient condition 24 h after CTC, and overall acceptance. Patient acceptance was rated using a semi-quantitative scale (1 = poor, 2 = fair, 3 = average, 4 = good, 5 = excellent). Overall examination time was evaluated as well in this subset of patient s [17–20]. 2.5.4. Statistical analysis Per-polyp sensitivity, specificity, PPV and NPV of CTC with oral and rectal tagging were calculated and compared by means of the Fisher exact test. Intraluminal density in the various colonic segments (in terms of mean value of density measured in the supine and prone datasets) was compared using the Kruskal–Wallis test with the Dunn test for post-hoc analysis. Overall examination time

3. Results

In group 1, 55 subjects had a total of 77 colonic lesions (12 sized less than 6 mm in their maximum diameter; 37 sized between 6 mm and 9 mm, 8 between 9 mm and 30 mm, and 20 larger than 30 mm). Sensitivity, specificity, PPV and NPV of CTC in group 1 classified by lesion size are tabulated in Table 1. Overall sensitivity, specificity, PPV, and NPV were 96.1% (CI95% 85.4–99.3%), 95.3% (CI95% 90.7–97.8%), 86.0% (CI95% 73.6–93.3%), and 98.8% (CI95% 95.3–99.8%), respectively. In group 2, 15 subjects showed an overall number of 20 colonic lesions (10 sized less than 6 mm in maximum diameter; 6 sized between 6 mm and 9 mm, 4 between 9 mm and 30 mm). Sensitivity, specificity, PPV and NPV of CTC in group 2 classified by lesion size are tabulated in Table 2. Overall sensitivity, specificity, PPV, and NPV were 89.4% (CI95% 65.4–98.1%), 95.8% (CI95% 87.6–98.9%), 85.0% (CI95% 61.1–96%), and 97.2% (CI95% 89.4–99.5%), respectively. Polyp detection rate was not significantly different between the two groups (p = 0.549). 3.2. Image quality In group 1, overall mean intraluminal density was 1058 ± 220.6HU (image quality 2.95 ± 0.08). Intraluminal density values in the various colonic segments were 944.4 ± 222.7 in the cecum-ascending colon (image quality 2.85 ± 0.30), 1008.9 ± 237.8 in the transverse colon (image quality 2.97 ± 0.13), 1182 ± 327.1 in the descending colon (image quality 3.0 ± 0.0), and 1096.5 ± 263.1 in the sigmoid-rectum (image quality 3.0 ± 0.0) (Table 3). In group 2, overall mean intraluminal density was 1084.3 ± 337.5HU (image quality 2.72 ± 0.40). Intraluminal density values in the various colonic segments were 1091.9 ± 357.9 in the cecum-ascending colon (image quality 2.87 ± 0.35), 1107.6 ± 354.4 in the transverse colon (image quality 2.87 ± 0.35), 1193.4 ± 431.4 in the descending colon (image quality 2.80 ± 0.41), and 944.1 ± 343.9 in the sigmoid-rectum (image quality 2.33 ± 0.82) (Table 4). Inter-rater agreement for image quality was very good (k = 0.83) (Fig. 2) [21]. 3.3. Patient acceptance and overall examination time Patient acceptance was comparable between group 1 and 2 (bowel preparation mean score 4.5 ± 0.8 vs 4.7 ± 0.5, p = 0.57; insuf-

1704

E. Neri et al. / European Journal of Radiology 84 (2015) 1701–1707

Fig. 1. Placement of regions of interest for measurement of intraluminal enhancement.

Table 2 Diagnostic performance of CTC for detecting polypoid lesions in group 2 (oral tagging): overall and size-wise analysis. PPV = positive predictive value, NPV = negative predictive value.

Sensitivity (CI95% ) Specificity (CI95% ) PPV (CI95% ) NPV (CI95% )

Overall (N = 20)

CT colonography with rectal iodine tagging: Feasibility and comparison with oral tagging in a colorectal cancer screening population.

To evaluate feasibility, diagnostic performance, patient acceptance, and overall examination time of CT colonography (CTC) performed through rectal ad...
2MB Sizes 0 Downloads 12 Views