- Email: [email protected]
Usefulness of hydrogel-CT for detecting and staging of rectosigmoid colon cancer
Accepted Manuscript Title: Usefulness of hydrogel-CT for detecting and staging of rectosigmoid colon cancer Author: Seonji Jeong Se Hyung Kim Ijin Joo Su Joa Ahn Joon Koo Han PII: DOI: Reference:
S0720-048X(16)30053-5 http://dx.doi.org/doi:10.1016/j.ejrad.2016.02.024 EURR 7411
To appear in:
European Journal of Radiology
Received date: Revised date: Accepted date:
22-12-2015 18-2-2016 23-2-2016
Please cite this article as: Jeong Seonji, Kim Se Hyung, Joo Ijin, Ahn Su Joa, Han Joon Koo.Usefulness of hydrogel-CT for detecting and staging of rectosigmoid colon cancer.European Journal of Radiology http://dx.doi.org/10.1016/j.ejrad.2016.02.024 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Usefulness of Hydrogel-CT for Detecting and Staging of Rectosigmoid Colon Cancer
Seonji Jeong1,2, Se Hyung Kim1,2, Ijin Joo1,2, Su Joa Ahn1,2, Joon Koo Han1,2,3
Department of Radiology, Seoul National University Hospital1 and Seoul National University College of Medicine2 Institute of Radiation Medicine, Seoul National University Medical Research Center3
This research was supported by Basic Science Research Program through the National Research Foundation of Korea [NRF] funded by the Ministry of Science, ICT & Future Planning [2013R1A1A3005937] and by the Seoul National University Hospital Research Fund No. 04-2015-620
Address correspondence to: Se Hyung Kim, M.D. Department of Radiology, Seoul National University Hospital and Seoul National University College of Medicine 101 Daehak-ro, Jongno-gu, Seoul, 110-744, Korea Tel: 82-2-2072-2057 ; Fax: 82-2-743-6385 E-mail: [email protected] 1
Key points
Rectum distension using a sonographic transmission gel is convenient and safe.
Hydrogel-CT is useful to detect primary lesion of rectosigmoid colon cancer.
More accurate preoperative T/N staging can be achieved with hydrogel-CT.
Abstract Purpose: To demonstrate the usefulness of hydrogel-CT for detecting and staging of rectosigmoid colon cancer. Materials and Methods: Fifty-four patients with rectosigmoid colon cancers underwent routine CT without (n=27) and with (n=27) rectum distension using a sonographic gel. Rectum distensibility and tumor visualization were evaluated. T and N stages on CT independently recorded by two radiologists were correlated with pathologic staging. Staging accuracies were compared using Fisher’s exact test. Diagnostic performances in differentiating
rectosigmoid colon cancer.
Keywords Rectum; Neoplasms; Multidetector computed tomography; Hydrogel; Neoplasm staging
Introduction The prognosis of rectosigmoid colon cancer is closely related to its stage at diagnosis and the choice of treatment involved [1, 2]. As appropriate preoperative treatment decisionmaking requires knowledge of the exact stage of a tumor, accurate radiologic T (depth of cancer invasion) and N staging (lymph node metastasis) present great clinical importance. At present, radiologic staging can be performed using endorectal sonography (ERUS), CT, and MRI. Although ERUS and MRI have shown promise in the field of rectosigmoid colon cancer staging, CT has been the most commonly used method for the staging of rectosigmoid colon cancer. However, its accuracy is still a matter of debate with some studies reporting high accuracy while others have reported less satisfactory results [3]. The major sources of staging errors reported thus far have been due to CT’s inability to separate the rectal wall layers and the inability to detect small rectosigmoid lesions especially when the rectosigmoid colon has collapsed [3]. Thus, to improve the diagnostic performance of CT in T staging of primary rectal cancers, water has been used to induce rectal distention [4-6]. However, although water is cheap and easy to handle, it creates an air-fluid level that may cause beam hardening artifacts at the interface, which in turn would result in a limited capacity to evaluate rectal cancers in the anterior wall [7]. In addition, water can easily move upward through the sigmoid colon resulting in inadequate distention of the rectum. 3
To the contrary, the sonography transmission gel is inert and semisolid, while still remaining inexpensive and easy to handle [8]. Indeed, its usefulness for the visualization of rectosigmoid colon cancer in hydrogel-MRI has been proven in a previous study [8]. However, to the best of our knowledge, there has been no study in which this agent was used to evaluate rectosigmoid colon cancer staging with CT. Furthermore, there have been no studies in which radiologists’ performance using hydrogel-CT for the staging of rectosigmoid colon cancer has been compared with that of conventional CT without rectum distention. Therefore, the purpose of our study is to demonstrate the usefulness of hydrogel-CT for the detection and staging of rectal cancer by comparing radiologists’ performance without and with rectum distention.
Materials and Methods Patients This retrospective study was approved by our institutional review board and the requirement for informed consent was waived as the sonography transmission gel has been used for rectal MRI for several years at our institute. Between July 2013 and April 2014, a total of 33 patients who underwent CT with rectum distention using a sonography transmission gel (Supersonic, Sung Heung Medical Co., Pucheon-si, Korea) were selected. Among them, we only included patients who fulfilled the following criteria: (a) patients pathologically proven to have rectosigmoid colon cancer either through endoscopic mucosal resection or surgery and (b) patients who did not undergo neoadjuvant concomitant chemoradiotherapy (CCRT). From the above criteria, we excluded six patients who had rectal mesenchymal tumors (n = 2), patients whose rectal lesions were confirmed only by biopsy (n = 2), and patients who underwent neoadjuvant CCRT (n = 2). Finally, 27 patients comprised our study population and 4
were designated as our rectum distended group. As control, we separately selected 250 patients with surgically proven rectosigmoid colon cancers who underwent conventional CT without rectum distention between January 2010 and February 2012. Control patients were matched at a 1:1 ratio in terms of age, sex, and T stage with patients who underwent hydrogel-CT with rectum distention. Thus, 27 patients were matched and designated as the control group. Because tumor relationship to anterior peritoneal reflection is important for tumor staging and for deciding treatment strategy, we tried to identify the anterior peritoneal reflection and its relationship to the tumor on axial and sagittal (if available) CT images. On axial image, the apex of the peritoneum attaches to the anterior rectal wall in a V-shaped configuration. In men, this is generally at a point just above the tip of the seminal vesicle while in women, the point of attachment is more variable. We categorized the tumor into three groups; tumors entirely above anterior peritoneal reflection, tumors saddling anterior peritoneal reflection, and tumors entirely below anterior peritoneal reflection. For distention group, 4, 7, and 16 tumors were located entirely above, saddling, and entirely below anterior peritoneal reflection, respectively. For control group, 5, 6, and 16 tumors were located entirely above, saddling, and entirely below anterior peritoneal reflection, respectively. Patients’ demographics including pathologic T and N staging in both the distended and control groups are presented in Table 1.
CT Techniques For rectum distention in 27 patients, approximately 80 - 100 ml of a sonography transmission gel was administered using an enema syringe with the patient placed on the CT table in the right lateral genupectoral position with their knees on their chest [8]. The administration was stopped immediately if the patient experienced intolerable pain. After the
5
administration, the patient was placed in the supine position. Anti-peristaltic agents were not administered in any patient. A variety of CT scanners were used in our study owing to its retrospective design. All patients underwent MDCT with 4, 8, 16, 32, 64, 128, or 320 detector-rows either at our institution (n = 50) or at outside hospitals (n = 4). The parameters used for these MDCTs were as follows: detector configuration, 0.63 - 1 mm; pitch, 0.89 - 1.35; rotation time, 0.5 - 0.75 sec; tube voltage, 120 kVp; tube current, 150 - 250 mAs; slice thickness, 2 - 5 mm; and reconstruction interval, 2 - 3 mm. Contrast-enhanced CT images were obtained after administration of an iodinated contrast agent (Ultravist 370, Bayer Schering Pharma, Berlin, Germany) at a dose of 1.5 ml/kg and at a rate of 3 - 5 ml/sec using an automatic power injector. In most cases (n = 45, 83.3%), dynamic enhanced images during arterial and portal phases were obtained. The remaining nine patients underwent single phase CT using the portal phase only.
Image Analysis For all CT images of the 54 patients, the degree of rectal distention and tumor visualization was assessed by two radiologists (
with 4 years of experience and
with
18 years of experience) in consensus using criteria reported in previous studies [7, 8]. Optimal rectal distention was assessed on a 4-point scale [8] (Fig. 1A). Grade 1 is defined as when the entire rectum (arrowheads) is collapsed; grade 2 when the rectum (arrowheads) is partially distended with more than one collapsed segment; grade 3 when the rectum (arrowheads) is suboptimally distended with one collapsed segment; grade 4 when the entire rectum (arrowheads) is fully distended without a collapsed segment. Tumor visualization was also assessed on a 4-point scale [7, 8] (Fig. 1B). Grade 1 is defined as when the tumor could not be 6
identified; grade 2 when the tumor (arrow) is identified but its intraluminal margin is not defined; grade 3 when the tumor (arrows) is identified and its intraluminal margin is partially defined; grade 4 when the whole tumor margin (arrow) is well defined.
Two different radiologists (
with 9 years of experience and
with 6
years of experience) independently reviewed all CT images for tumor (T) and lymph node (N) staging. All images were evaluated on a 2,000 × 2,000 PACS monitor (INFINITT, Seoul, Korea) with adjustment of the optimal window setting in each case. They were requested to stage the tumor. The depth of cancer invasion on CT was interpreted as follows: cT1, tumor shows enhancing wall thickening with an intact low-attenuation-strip outer wall; cT2, tumor shows enhancing wall thickening with a destroyed low-attenuation-strip without perirectal fat infiltration; cT3, tumor shows full thickness enhancement with perirectal fat infiltration; cT4, tumor shows direct invasion to adjacent structures [9, 10]. If tumors were invisible on CT, reviewers were asked to record it as “not evaluable.” In addition, reviewers independently scored CT images for ≥ cT3 on a 5-point confidence scale: 1, definitely < cT3, 2, probably < cT3; 3, possibly ≥ cT3; 4, probably ≥ cT3; 5, definitely ≥ cT3. For N staging, lymph node metastasis was considered to be present (≥ N1) if there were any lymph nodes ≥ 8 mm in shortaxis diameter, if a cluster of three or more peri-lesional nodes was identified regardless of their size, if the lymph nodes showed round configuration, or if the lymph nodes contained necrotic foci [9]. In addition, reviewers independently scored CT images for ≥ cN1 on a 5-point confidence scale: 1, definitely cN0, 2, probably cN0; 3, possibly ≥ cN1; 4, probably ≥ cN1; 5, definitely ≥ cN1.
7
Statistical Analysis Scores for rectum distention and tumor visualization were compared between the distended and control groups using the Student t test. Correlation between rectum distention and tumor visualization was analyzed using Pearson’s correlation test. For independent review, interobserver agreement between the two radiologists was also evaluated using weighted kappa (κ) statistics. We considered a κ value of ≥ 0.81 to represent almost perfect agreement and values of 0.61~0.80 and 0.41~0.60 to represent substantial and moderate agreement, respectively. Values of 0.21~0.40 and 0~0.20 were considered to represent slight and poor agreement, respectively [11]. Pathologic staging was correlated with CT staging by the two radiologists in terms of T and N staging using weighted κ statistics and Pearson’s correlation test. Accuracy of the two reviewers for T and N staging was calculated and compared between CT without and with rectum distention using Fisher’s exact test. Diagnostic performance of the two radiologists in terms of differentiating < T3 from ≥ T3 and N0 from ≥ N1 was evaluated using areas under the receiver operating characteristic curves (AZ), which were calculated using the nonparametric method. In addition, sensitivities and specificities in terms of differentiating < T3 from ≥ T3 and N0 from ≥N1 were calculated using only those patients allocated a rating of 3 or greater in the level of confidence by each radiologist. Sensitivities and specificities were compared using the Fisher’s exact test between the distended and control groups. All statistical analyses were performed using commercially available software, SPSS version 21.0 for Windows (SPSS Inc, Chicago, IL, USA) and free web-version Graphad Quickcalc software (www.graphpad.com/quickcalcs/). P values of < 0.05 were considered to indicate a statistical significance.
8
Results Successful administration of the sonography transmission gel was achieved in all 27 patients (27/27, 100%) of the rectum distended group. The mean score (mean ± standard deviation, 3.52 ± 0.75) for rectum distention in the distended group was significantly higher than that (1.44 ± 0.64) in the control group (P<0.0001). The number of patients showing full rectal distention (score 4) was 18 in the distended group while no patients showed full rectal distention in the control group. In addition, the overall tumor visualization scores in the distended group were nearly perfect and its mean value (mean ± standard deviation, 3.70 ± 0.67) was also significantly higher than that (2.04 ± 0.90) in the control group (P<0.0001). In the control group, the number of patients who has invisible rectosigmoid colon cancers on CT was eight while it was zero in the rectum distended group. The correlation coefficient between scores for rectum distention and tumor visualization was high (0.815, P<0.0001). At independent review, T staging was not evaluable in six and three patients, respectively, due to non-visualization of the tumor on CT by both reviewers. Among these nine patients, eight were in the control group and one was in the distended group. Those nine patients were excluded from further analysis of T staging. Interobserver agreement between the two radiologists for T and N staging was moderate (weighted κ value = 0.401 and 0.487, respectively). For subgroup analysis, weighted κ values were greater in the distended group (0.402 and 0.523) than in the control group (0.373 and 0.459) for both T and N staging. Agreements between pathologic staging and CT staging in both reviewers are presented in Table 2. Weighted κ values between pathologic and CT staging of both reviewers were 0.245 and 0.672 for T staging and 0.237 and 0.647 for N staging, respectively. Pearson’s correlation coefficient between pathologic and CT staging of both reviewers were 0.509 and 9
0.765 for T staging and 0.341 and 0.715 for N staging, respectively (P < 0.0001). On subgroup analysis, weighted κ values and Pearson’s correlation coefficients were greater in the distended group than in the control group for both T and N staging by both reviewers (Table 2). For T staging, CT staging accuracies (50% [13/26] for reviewer 1 and 85.2% [23/27] for reviewer 2) in the distended group was higher than those (45.5% [10/22] and 62.5% [15/24]) in the control group for both reviewers, albeit not statistically significant (P= 0.780 and 0.107). For N staging, CT staging accuracies (59.3% [16/27] for reviewer 1 and 92.6% [25/27] for reviewer 2) in the distended group was also greater than those (33.3% [9/27] for reviewer 1 and 59.3% [16/27] for reviewer 2) in the control group. This difference was statistically significant for reviewer 2 (P=0.0091). Table 3 summarized the diagnostic performance expressed as AZ, sensitivity, and specificity of the two reviewers for T and N staging. The AZ value for differentiating < T3 from ≥ T3 in the distended group was greater than those in the control group for both reviewers. Sensitivity and specificity for differentiating < T3 from ≥ T3 in the distended group was also greater than those in the control group for both reviewers, albeit not statistically significant (P>0.05). The AZ value for differentiating N0 from ≥ N1 in the distended group was greater than those in the control group for both reviewers. Sensitivity and specificity for differentiating N0 from ≥ N1 in the distended group was also greater than those in the control group. The difference in sensitivity was statistically significant for reviewer 2 (P=0.0411). Representative examples are presented in Figures 2 - 5.
10
Discussion In our study, we found that rectal administration of a sonography transmission gel for rectal CT was technically and clinically feasible, with all patients showing good tolerance without any complications. Compared with other rectal contrast agents such as a warm water solution or a 5% methylcellulose solution, the sonography transmission gel is similarly simple and easy to use but can be utilized in a smaller volume (≤ 100 ml). A previous article reported that water enema CT can be performed for rectal cancer staging with 500 ml of tap water but that it may cause patient discomfort leading to poor compliance owing to its large amount [5]. As for the 5% methylcellulose solution as a rectal contrast agent, although Goh et al. reported its usefulness, the use of that agent required the additional step of dissolving the compound granules prior to the procedure [12]. In contrast, in our study, effective rectal wall distention was achieved without any additional steps further complicating the process. Considering that the sonography transmission gel also has inert characteristics, provides high contrast on CT and effective distention of the rectosigmoid colon lumen owing to its own semisolid state with low cost (less than 50 cents per patient) and ease of manipulation, hydrogel-CT can be the best method to achieve rectal distention. As expected, scores for tumor visualization and rectum distention were significantly higher in patients with rectum distention than in those without distention. The correlation coefficient between scores for tumor visualization and rectum distention was also high, indicating that better distention of the rectosigmoid colon provides better detection of the tumor. The overall tumor visualization score (3.70) in the distended group of our study was slightly lower than the results (3.68 - 3.95) of a previous study in which MRI was used [8]. Considering the similar scores for rectum distention between our study and the previous study, the better score reported for tumor visualization in the previous study may have been due to the high soft 11
tissue contrast provided by MRI. In our study, the accuracy for T staging was better with hydrogel-CT than with conventional CT for both radiologists. As a result, the numbers of over- and under-staged cancers were lower with hydrogel-CT. Also radiologists’ performance for diagnosing ≥ T3 cancer was better in patients with rectum distention than without distention for both radiologists. Considering that an adequate distention of the bowel with good contrast between the tumor and bowel lumen is required to accurately assess the T stage of cancer with CT, our results seem to be quite natural and well coincides with those of previous studies. In previous studies, collapsed bowels were reported as sources of errors for incorrect T staging of colon cancers [3]. The preoperative correct diagnosis of advanced colorectal cancer ≥ cT3 (clinical T3) is of clinical importance as many current guidelines including that from the National Comprehensive Cancer Network (NCCN) recommend neoadjuvant CCRT for advanced rectal cancers ≥ cT3 and adjuvant chemotherapy for advanced colon cancers ≥ pT3 [13, 14]. In this regard, our study might be noteworthy. In terms of N staging, there was a greater effect of rectum distention on radiologists’ performance. The accuracy for N staging was better with hydrogel-CT than with conventional CT for both radiologists. It was an unexpected finding as we believed that the accuracy of N staging might depend on the criteria used for discriminating positive LNs, not on whether the rectum is distended or not. However, considering most involved mesorectal lymph nodes are found at or proximal to the level of the tumor, the exact knowledge of the primary tumor site and tumor staging obtained via rectum distention may have been responsible for the better N staging performance of both radiologists. Indeed, among the nine patients in whom primary tumor was not visible, eight were in the control group and only one was in the distention group. Considering that many guidelines including the NCCN guidelines recommend neoadjuvant 12
CCRT for cN1 in rectal cancer [14], the exact identification of metastatic LNs preoperatively may have a clinical impact. We also found that the agreement and correlation between pathologic staging and CT staging were better in the distended group than in the control group for both T and N staging in both reviewers. It may demonstrate the effect of rectum distention on radiologists’ diagnostic performance for T and N staging. Thus, considering its easy use, well availability, and the cheap cost of the sonographic gel, we believe that the use of hydrogel-CT should be the standard CT protocol for patients with rectosigmoid colon cancer. We acknowledge the following limitations of our study. First, as our study was retrospectively designed, various MDCT scanners were used, especially in the control group, resulting in differences in CT acquisition parameters. However, all acquisitions in our study used reconstruction interval of less than 3 mm which may be acceptable for imaging interpretation [3]. Second, the number of patients in our study population (n=54) is relatively small, making the generalization of our study results difficult. Therefore, future studies with a larger study population are warranted to validate the results of our study. Third, the difference in the diagnostic performances between the two reviewers was slightly large for both T and N staging, although the trend toward better performances in the rectum distended group than in the control group was similar between the two reviewers. Nonetheless, further studies recruiting more radiologists with different amounts of expertise may be needed to overcome this limitation. Fourth, CT accuracies for both T and N staging in our study were relatively low compared to those of previous studies [3]. However, it may have been due to the blinding of the results of endoscopy in our study, which would be different from that in real clinical practice. Finally, we did not analyze the presence of mesorectal fascial involvement on CT. Because involvement of circumferential resection margin (CRM) is very important for patients’ 13
prognosis and further treatment plan, description regarding CRM involvement might be critical. Therefore, further study dealing with this issue should be strongly warranted. In conclusion, hydrogel-CT using a sonography transmission gel is a convenient and valuable tool, providing better diagnostic performance for T and N staging of rectosigmoid colon cancer compared to conventional CT without rectal distention.
Conflict of interest All authors have no conflict of interest to declare.
References 1. Visser BC, Varma MG, Welton ML (2001) Local therapy for rectal cancer. Surg Oncol 10:61-69 2. Reynolds JV, Joyce WP, Dolan J, Sheahan K, Hyland JM (1996) Pathological evidence in support of total mesorectal excision in the management of rectal cancer. Br J Surg 83:11121115 3. Dighe S, Purkayastha S, Swift I, et al. (2010) Diagnostic precision of CT in local staging of colon cancers: a meta-analysis. Clin Radiol 65:708-719 4. Gazelle GS, Gaa J, Saini S, Shellito P (1995) Staging of colon carcinoma using water enema CT. J Comput Assist Tomogr 19:87-91 5. Palkó A, Gyulai C, Fedinecz N, Balogh A, Nagy F (2000) Water enema CT examination of rectum cancer by reduced amount of water. Rofo 172:901-904 6. Cademartiri F, Luccichenti G, Rossi A, Pavone P (2002) Spiral hydro-CT in the evaluation of colo-sigmoideal cancer. Radiol Med 104:295-306 14
7. Kim MJ, Lim JS, Oh YT, et al. (2004) Preoperative MRI of rectal cancer with and without rectal water filling: an intraindividual comparison. Am J Roentgenol 182:1469-1476 8. Kim SH, Lee JM, Lee MW, Kim GH, Han JK, Choi BI (2008) Sonography transmission gel as endorectal contrast agent for tumor visualization in rectal cancer. Am J Roentgenol 191:186189 9. Jin KN, Lee JM, Kim SH, et al. (2006) The diagnostic value of multiplanar reconstruction on MDCT colonography for the preoperative staging of colorectal cancer. Eur Radiol 16:22842291 10. Lee IJ, Lee JM, Kim SH, et al. (2010) Diagnostic performance of 64-channel multidetector CT in the evaluation of gastric cancer: differentiation of mucosal cancer (T1a) from submucosal involvement (T1b and T2). Radiology 255:805-814 11. Landis JR, Koch GG (1977) The measurement of observer agreement for categorical data. Biometrics 33:159-174 12. Goh JS, Goh JP, Wansaicheong GK (2002) Methylcellulose as a rectal contrast agent for MR imaging of rectal carcinoma. Am J Roentgenol 178:1145-1146 13. National Comprehensive Cancer Network (2015) Clinical practice guidelines in oncology (NCCN
guidelines):
colon
cancer
version
3.
Available
via
http://www.nccn.org/professionals/physician_gls/pdf/colon.pdf. Accessed August 30, 2015 14. National Comprehensive Cancer Network (2015) Clinical practice guidelines in oncology (NCCN
guidelines):
rectal
cancer
version
3.
Available
http://www.nccn.org/professionals/physician_gls/pdf/rectal.pdf. Accessed August 30, 2015
15
via
Legends for Figures
Figure 1. Semi-quantitative grading for the degree of rectum distention (A) and tumor visualization (B) on CT images. A, Grading for the degree of rectum distention on sagittal CT images. Grade 1 is defined as when the entire rectum (arrowheads) is collapsed; grade 2 when the rectum (arrowheads) is partially distended with more than one collapsed segment; grade 3 when the rectum (arrowheads) is suboptimally distended with one collapsed segment; grade 4 when the entire rectum (arrowheads) is fully distended without a collapsed segment. B, Grading for the degree of tumor visualization on axial CT images. Grade 1 is defined as when the tumor could not be identified; grade 2 when the tumor (arrow) is identified but its intraluminal margin is not defined; grade 3 when the tumor (arrows) is identified and its intraluminal margin is partially defined; grade 4 when the whole tumor margin (arrow) is well defined. 16
Figure 2. A 61-year old woman with distal sigmoid colon cancer in the distended group. A, B. On axial (A) and coronal (B) contrast-enhanced CT images, a 1.5 cm polypoid lesion (arrow) is well visualized at the distal sigmoid colon. Both reviewers recorded this lesion as cT1 stage with a confidence score of 1. Histopathology after lower anterior resection confirmed the stage as T1. c = CT staging. .
17
Figure 3. A 46-year old man with rectal cancer in the distended group. Axial contrast-enhanced CT image shows a focal full-thickness wall thickening (arrow) at the upper rectum (6~8 o’clock). However, perirectal fat seems to be clear. Therefore, both reviewers recorded this lesion as cT2 stage with a confidence score of 1. Histopathology after lower anterior resection confirmed the stage as T2. c = CT staging.
18
Figure 4. A 53-year old man with rectal cancer in the distended group. Axial contrast-enhanced CT image shows a focal full-thickness wall thickening (arrow) with perirectal fat infiltration at the mid-rectum (3~6 o’clock). Both reviewers recorded this lesion as cT3 stage with a confidence score of 5. Histopathology after ultra-low anterior resection confirmed the stage as T3. c = CT staging.
19
Figure 5. A 53-year old man with rectal cancer in the distended group. Serial sagittal contrastenhanced CT images demonstrate focal concentric wall thickening with perirectal infiltration (arrows) at the upper rectum. Note the multiple conglomerated enhancing lymph nodes (arrowheads) around the primary lesion. Both reviewers recorded this lesion as cT3 and cN2 stage with a confidence score of 5. Histopathology after low anterior resection confirmed a rectal adenocarcinoma of T3N2 stage.
20
Tables Table 1. Patients’ Demographics in the Distended and Control Groups Distended Group Control Group P Value Sex (M:F) 17:10 16:11 0.780 Mean Age (Age Range, Years) 61.6 (39~84) 60.1 (40~79) 0.624 Pathologic T pT1 7 6 0.926 Staging pT2 6 7 pT3 14 14 Pathologic N pN0 16 14 0.112 Staging pN1 5 11 pN2 6 2 Operation AR 5 0 0.085 LAR 17 25 ULAR 2 1 Hartmann’s Operation 1 0 Others 2* 1† * = endoscopic submucosal dissection, † = transanal excision. AR = anterior resection, LAR = lower anterior resection, ULAR = ultra-low anterior resection.
21
Table 2. Correlations between Pathologic and CT Staging Distended Group Reviewer 1
pT1 2 1 3 0
T staging pT2 pT3 0 0 2 0 4 9 0 5
pT4 0 0 0 0
N staging pN0 pN1 11 1 3 2 2 2
pN2 0 3 3
cT1 cN0 cT2 cN1 cT3 cN2 cT4 Weighted κ value 0.346 0.451 Correlation coefficient 0.637 (P < 0.0001) 0.558 (P=0.002) Reviewer 2 cT1 6 0 0 0 cN0 15 0 0 cT2 1 5 1 0 cN1 1 5 1 cT3 0 1 12 0 cN2 0 0 5 cT4 0 0 1 0 Weighted κ value 0.836 0.909 Correlation coefficient 0.898 (P < 0.0001) 0.944 (P < 0.0001) Control T staging N staging Group pT1 pT2 pT3 pT4 pN0 pN1 pN2 Reviewer 1 cT1 0 0 0 0 cN0 7 5 1 cT2 2 0 2 0 cN1 3 1 0 cT3 1 4 10 0 cN2 4 5 1 cT4 0 1 2 0 Weighted κ value 0.072 0.043 Correlation coefficient 0.280 (P=0.208) 0.107 (P=0.595) Reviewer 2 cT1 0 0 0 0 cN0 9 4 1 cT2 3 5 2 0 cN1 5 6 0 cT3 1 1 10 0 cN2 0 1 1 cT4 0 0 2 0 Weighted κ value 0.434 0.302 Correlation coefficient 0.552 (P = 0.005) 0.344 (P=0.079) *Correlation coefficients were calculated using Pearson’s correlation analysis. p = pathologic staging, c = CT staging
22
Table 3. Diagnostic Performances of the Two Reviewers for T and N Staging Staging Reviewers Differentiating Reviewer 1 < T3 and ≥ T3
Groups AZ Sensitivity Distended 0.827 100% (14/14) Control 0.817 92.9% (13/14) P Value 1.000 Reviewer 2 Distended 0.989 92.9% (13/14) Control 0.907 85.7% (12/14) P Value 1.000 Differentiating Reviewer 1 Distended 0.858 90.9% (10/11) N0 and ≥ N1 Control 0.544 61.5% (8/13) P Value 0.166 Reviewer 2 Distended 0.980 100% (11/11) Control 0.654 61.5% (8/13) P Value 0.041 AZ = area under the receiver operating characteristic curve
23
Specificity 33.3% (4/12) 12.5% (1/8) 0.603 92.3% (12/13) 80% (8/10) 0.560 56.2% (9/16) 42.9% (6/14) 0.715 87.5% (14/16) 64.3% (9/14) 0.204
S0720-048X(16)30053-5 http://dx.doi.org/doi:10.1016/j.ejrad.2016.02.024 EURR 7411
To appear in:
European Journal of Radiology
Received date: Revised date: Accepted date:
22-12-2015 18-2-2016 23-2-2016
Please cite this article as: Jeong Seonji, Kim Se Hyung, Joo Ijin, Ahn Su Joa, Han Joon Koo.Usefulness of hydrogel-CT for detecting and staging of rectosigmoid colon cancer.European Journal of Radiology http://dx.doi.org/10.1016/j.ejrad.2016.02.024 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Usefulness of Hydrogel-CT for Detecting and Staging of Rectosigmoid Colon Cancer
Seonji Jeong1,2, Se Hyung Kim1,2, Ijin Joo1,2, Su Joa Ahn1,2, Joon Koo Han1,2,3
Department of Radiology, Seoul National University Hospital1 and Seoul National University College of Medicine2 Institute of Radiation Medicine, Seoul National University Medical Research Center3
This research was supported by Basic Science Research Program through the National Research Foundation of Korea [NRF] funded by the Ministry of Science, ICT & Future Planning [2013R1A1A3005937] and by the Seoul National University Hospital Research Fund No. 04-2015-620
Address correspondence to: Se Hyung Kim, M.D. Department of Radiology, Seoul National University Hospital and Seoul National University College of Medicine 101 Daehak-ro, Jongno-gu, Seoul, 110-744, Korea Tel: 82-2-2072-2057 ; Fax: 82-2-743-6385 E-mail: [email protected] 1
Key points
Rectum distension using a sonographic transmission gel is convenient and safe.
Hydrogel-CT is useful to detect primary lesion of rectosigmoid colon cancer.
More accurate preoperative T/N staging can be achieved with hydrogel-CT.
Abstract Purpose: To demonstrate the usefulness of hydrogel-CT for detecting and staging of rectosigmoid colon cancer. Materials and Methods: Fifty-four patients with rectosigmoid colon cancers underwent routine CT without (n=27) and with (n=27) rectum distension using a sonographic gel. Rectum distensibility and tumor visualization were evaluated. T and N stages on CT independently recorded by two radiologists were correlated with pathologic staging. Staging accuracies were compared using Fisher’s exact test. Diagnostic performances in differentiating
rectosigmoid colon cancer.
Keywords Rectum; Neoplasms; Multidetector computed tomography; Hydrogel; Neoplasm staging
Introduction The prognosis of rectosigmoid colon cancer is closely related to its stage at diagnosis and the choice of treatment involved [1, 2]. As appropriate preoperative treatment decisionmaking requires knowledge of the exact stage of a tumor, accurate radiologic T (depth of cancer invasion) and N staging (lymph node metastasis) present great clinical importance. At present, radiologic staging can be performed using endorectal sonography (ERUS), CT, and MRI. Although ERUS and MRI have shown promise in the field of rectosigmoid colon cancer staging, CT has been the most commonly used method for the staging of rectosigmoid colon cancer. However, its accuracy is still a matter of debate with some studies reporting high accuracy while others have reported less satisfactory results [3]. The major sources of staging errors reported thus far have been due to CT’s inability to separate the rectal wall layers and the inability to detect small rectosigmoid lesions especially when the rectosigmoid colon has collapsed [3]. Thus, to improve the diagnostic performance of CT in T staging of primary rectal cancers, water has been used to induce rectal distention [4-6]. However, although water is cheap and easy to handle, it creates an air-fluid level that may cause beam hardening artifacts at the interface, which in turn would result in a limited capacity to evaluate rectal cancers in the anterior wall [7]. In addition, water can easily move upward through the sigmoid colon resulting in inadequate distention of the rectum. 3
To the contrary, the sonography transmission gel is inert and semisolid, while still remaining inexpensive and easy to handle [8]. Indeed, its usefulness for the visualization of rectosigmoid colon cancer in hydrogel-MRI has been proven in a previous study [8]. However, to the best of our knowledge, there has been no study in which this agent was used to evaluate rectosigmoid colon cancer staging with CT. Furthermore, there have been no studies in which radiologists’ performance using hydrogel-CT for the staging of rectosigmoid colon cancer has been compared with that of conventional CT without rectum distention. Therefore, the purpose of our study is to demonstrate the usefulness of hydrogel-CT for the detection and staging of rectal cancer by comparing radiologists’ performance without and with rectum distention.
Materials and Methods Patients This retrospective study was approved by our institutional review board and the requirement for informed consent was waived as the sonography transmission gel has been used for rectal MRI for several years at our institute. Between July 2013 and April 2014, a total of 33 patients who underwent CT with rectum distention using a sonography transmission gel (Supersonic, Sung Heung Medical Co., Pucheon-si, Korea) were selected. Among them, we only included patients who fulfilled the following criteria: (a) patients pathologically proven to have rectosigmoid colon cancer either through endoscopic mucosal resection or surgery and (b) patients who did not undergo neoadjuvant concomitant chemoradiotherapy (CCRT). From the above criteria, we excluded six patients who had rectal mesenchymal tumors (n = 2), patients whose rectal lesions were confirmed only by biopsy (n = 2), and patients who underwent neoadjuvant CCRT (n = 2). Finally, 27 patients comprised our study population and 4
were designated as our rectum distended group. As control, we separately selected 250 patients with surgically proven rectosigmoid colon cancers who underwent conventional CT without rectum distention between January 2010 and February 2012. Control patients were matched at a 1:1 ratio in terms of age, sex, and T stage with patients who underwent hydrogel-CT with rectum distention. Thus, 27 patients were matched and designated as the control group. Because tumor relationship to anterior peritoneal reflection is important for tumor staging and for deciding treatment strategy, we tried to identify the anterior peritoneal reflection and its relationship to the tumor on axial and sagittal (if available) CT images. On axial image, the apex of the peritoneum attaches to the anterior rectal wall in a V-shaped configuration. In men, this is generally at a point just above the tip of the seminal vesicle while in women, the point of attachment is more variable. We categorized the tumor into three groups; tumors entirely above anterior peritoneal reflection, tumors saddling anterior peritoneal reflection, and tumors entirely below anterior peritoneal reflection. For distention group, 4, 7, and 16 tumors were located entirely above, saddling, and entirely below anterior peritoneal reflection, respectively. For control group, 5, 6, and 16 tumors were located entirely above, saddling, and entirely below anterior peritoneal reflection, respectively. Patients’ demographics including pathologic T and N staging in both the distended and control groups are presented in Table 1.
CT Techniques For rectum distention in 27 patients, approximately 80 - 100 ml of a sonography transmission gel was administered using an enema syringe with the patient placed on the CT table in the right lateral genupectoral position with their knees on their chest [8]. The administration was stopped immediately if the patient experienced intolerable pain. After the
5
administration, the patient was placed in the supine position. Anti-peristaltic agents were not administered in any patient. A variety of CT scanners were used in our study owing to its retrospective design. All patients underwent MDCT with 4, 8, 16, 32, 64, 128, or 320 detector-rows either at our institution (n = 50) or at outside hospitals (n = 4). The parameters used for these MDCTs were as follows: detector configuration, 0.63 - 1 mm; pitch, 0.89 - 1.35; rotation time, 0.5 - 0.75 sec; tube voltage, 120 kVp; tube current, 150 - 250 mAs; slice thickness, 2 - 5 mm; and reconstruction interval, 2 - 3 mm. Contrast-enhanced CT images were obtained after administration of an iodinated contrast agent (Ultravist 370, Bayer Schering Pharma, Berlin, Germany) at a dose of 1.5 ml/kg and at a rate of 3 - 5 ml/sec using an automatic power injector. In most cases (n = 45, 83.3%), dynamic enhanced images during arterial and portal phases were obtained. The remaining nine patients underwent single phase CT using the portal phase only.
Image Analysis For all CT images of the 54 patients, the degree of rectal distention and tumor visualization was assessed by two radiologists (
with 4 years of experience and
with
18 years of experience) in consensus using criteria reported in previous studies [7, 8]. Optimal rectal distention was assessed on a 4-point scale [8] (Fig. 1A). Grade 1 is defined as when the entire rectum (arrowheads) is collapsed; grade 2 when the rectum (arrowheads) is partially distended with more than one collapsed segment; grade 3 when the rectum (arrowheads) is suboptimally distended with one collapsed segment; grade 4 when the entire rectum (arrowheads) is fully distended without a collapsed segment. Tumor visualization was also assessed on a 4-point scale [7, 8] (Fig. 1B). Grade 1 is defined as when the tumor could not be 6
identified; grade 2 when the tumor (arrow) is identified but its intraluminal margin is not defined; grade 3 when the tumor (arrows) is identified and its intraluminal margin is partially defined; grade 4 when the whole tumor margin (arrow) is well defined.
Two different radiologists (
with 9 years of experience and
with 6
years of experience) independently reviewed all CT images for tumor (T) and lymph node (N) staging. All images were evaluated on a 2,000 × 2,000 PACS monitor (INFINITT, Seoul, Korea) with adjustment of the optimal window setting in each case. They were requested to stage the tumor. The depth of cancer invasion on CT was interpreted as follows: cT1, tumor shows enhancing wall thickening with an intact low-attenuation-strip outer wall; cT2, tumor shows enhancing wall thickening with a destroyed low-attenuation-strip without perirectal fat infiltration; cT3, tumor shows full thickness enhancement with perirectal fat infiltration; cT4, tumor shows direct invasion to adjacent structures [9, 10]. If tumors were invisible on CT, reviewers were asked to record it as “not evaluable.” In addition, reviewers independently scored CT images for ≥ cT3 on a 5-point confidence scale: 1, definitely < cT3, 2, probably < cT3; 3, possibly ≥ cT3; 4, probably ≥ cT3; 5, definitely ≥ cT3. For N staging, lymph node metastasis was considered to be present (≥ N1) if there were any lymph nodes ≥ 8 mm in shortaxis diameter, if a cluster of three or more peri-lesional nodes was identified regardless of their size, if the lymph nodes showed round configuration, or if the lymph nodes contained necrotic foci [9]. In addition, reviewers independently scored CT images for ≥ cN1 on a 5-point confidence scale: 1, definitely cN0, 2, probably cN0; 3, possibly ≥ cN1; 4, probably ≥ cN1; 5, definitely ≥ cN1.
7
Statistical Analysis Scores for rectum distention and tumor visualization were compared between the distended and control groups using the Student t test. Correlation between rectum distention and tumor visualization was analyzed using Pearson’s correlation test. For independent review, interobserver agreement between the two radiologists was also evaluated using weighted kappa (κ) statistics. We considered a κ value of ≥ 0.81 to represent almost perfect agreement and values of 0.61~0.80 and 0.41~0.60 to represent substantial and moderate agreement, respectively. Values of 0.21~0.40 and 0~0.20 were considered to represent slight and poor agreement, respectively [11]. Pathologic staging was correlated with CT staging by the two radiologists in terms of T and N staging using weighted κ statistics and Pearson’s correlation test. Accuracy of the two reviewers for T and N staging was calculated and compared between CT without and with rectum distention using Fisher’s exact test. Diagnostic performance of the two radiologists in terms of differentiating < T3 from ≥ T3 and N0 from ≥ N1 was evaluated using areas under the receiver operating characteristic curves (AZ), which were calculated using the nonparametric method. In addition, sensitivities and specificities in terms of differentiating < T3 from ≥ T3 and N0 from ≥N1 were calculated using only those patients allocated a rating of 3 or greater in the level of confidence by each radiologist. Sensitivities and specificities were compared using the Fisher’s exact test between the distended and control groups. All statistical analyses were performed using commercially available software, SPSS version 21.0 for Windows (SPSS Inc, Chicago, IL, USA) and free web-version Graphad Quickcalc software (www.graphpad.com/quickcalcs/). P values of < 0.05 were considered to indicate a statistical significance.
8
Results Successful administration of the sonography transmission gel was achieved in all 27 patients (27/27, 100%) of the rectum distended group. The mean score (mean ± standard deviation, 3.52 ± 0.75) for rectum distention in the distended group was significantly higher than that (1.44 ± 0.64) in the control group (P<0.0001). The number of patients showing full rectal distention (score 4) was 18 in the distended group while no patients showed full rectal distention in the control group. In addition, the overall tumor visualization scores in the distended group were nearly perfect and its mean value (mean ± standard deviation, 3.70 ± 0.67) was also significantly higher than that (2.04 ± 0.90) in the control group (P<0.0001). In the control group, the number of patients who has invisible rectosigmoid colon cancers on CT was eight while it was zero in the rectum distended group. The correlation coefficient between scores for rectum distention and tumor visualization was high (0.815, P<0.0001). At independent review, T staging was not evaluable in six and three patients, respectively, due to non-visualization of the tumor on CT by both reviewers. Among these nine patients, eight were in the control group and one was in the distended group. Those nine patients were excluded from further analysis of T staging. Interobserver agreement between the two radiologists for T and N staging was moderate (weighted κ value = 0.401 and 0.487, respectively). For subgroup analysis, weighted κ values were greater in the distended group (0.402 and 0.523) than in the control group (0.373 and 0.459) for both T and N staging. Agreements between pathologic staging and CT staging in both reviewers are presented in Table 2. Weighted κ values between pathologic and CT staging of both reviewers were 0.245 and 0.672 for T staging and 0.237 and 0.647 for N staging, respectively. Pearson’s correlation coefficient between pathologic and CT staging of both reviewers were 0.509 and 9
0.765 for T staging and 0.341 and 0.715 for N staging, respectively (P < 0.0001). On subgroup analysis, weighted κ values and Pearson’s correlation coefficients were greater in the distended group than in the control group for both T and N staging by both reviewers (Table 2). For T staging, CT staging accuracies (50% [13/26] for reviewer 1 and 85.2% [23/27] for reviewer 2) in the distended group was higher than those (45.5% [10/22] and 62.5% [15/24]) in the control group for both reviewers, albeit not statistically significant (P= 0.780 and 0.107). For N staging, CT staging accuracies (59.3% [16/27] for reviewer 1 and 92.6% [25/27] for reviewer 2) in the distended group was also greater than those (33.3% [9/27] for reviewer 1 and 59.3% [16/27] for reviewer 2) in the control group. This difference was statistically significant for reviewer 2 (P=0.0091). Table 3 summarized the diagnostic performance expressed as AZ, sensitivity, and specificity of the two reviewers for T and N staging. The AZ value for differentiating < T3 from ≥ T3 in the distended group was greater than those in the control group for both reviewers. Sensitivity and specificity for differentiating < T3 from ≥ T3 in the distended group was also greater than those in the control group for both reviewers, albeit not statistically significant (P>0.05). The AZ value for differentiating N0 from ≥ N1 in the distended group was greater than those in the control group for both reviewers. Sensitivity and specificity for differentiating N0 from ≥ N1 in the distended group was also greater than those in the control group. The difference in sensitivity was statistically significant for reviewer 2 (P=0.0411). Representative examples are presented in Figures 2 - 5.
10
Discussion In our study, we found that rectal administration of a sonography transmission gel for rectal CT was technically and clinically feasible, with all patients showing good tolerance without any complications. Compared with other rectal contrast agents such as a warm water solution or a 5% methylcellulose solution, the sonography transmission gel is similarly simple and easy to use but can be utilized in a smaller volume (≤ 100 ml). A previous article reported that water enema CT can be performed for rectal cancer staging with 500 ml of tap water but that it may cause patient discomfort leading to poor compliance owing to its large amount [5]. As for the 5% methylcellulose solution as a rectal contrast agent, although Goh et al. reported its usefulness, the use of that agent required the additional step of dissolving the compound granules prior to the procedure [12]. In contrast, in our study, effective rectal wall distention was achieved without any additional steps further complicating the process. Considering that the sonography transmission gel also has inert characteristics, provides high contrast on CT and effective distention of the rectosigmoid colon lumen owing to its own semisolid state with low cost (less than 50 cents per patient) and ease of manipulation, hydrogel-CT can be the best method to achieve rectal distention. As expected, scores for tumor visualization and rectum distention were significantly higher in patients with rectum distention than in those without distention. The correlation coefficient between scores for tumor visualization and rectum distention was also high, indicating that better distention of the rectosigmoid colon provides better detection of the tumor. The overall tumor visualization score (3.70) in the distended group of our study was slightly lower than the results (3.68 - 3.95) of a previous study in which MRI was used [8]. Considering the similar scores for rectum distention between our study and the previous study, the better score reported for tumor visualization in the previous study may have been due to the high soft 11
tissue contrast provided by MRI. In our study, the accuracy for T staging was better with hydrogel-CT than with conventional CT for both radiologists. As a result, the numbers of over- and under-staged cancers were lower with hydrogel-CT. Also radiologists’ performance for diagnosing ≥ T3 cancer was better in patients with rectum distention than without distention for both radiologists. Considering that an adequate distention of the bowel with good contrast between the tumor and bowel lumen is required to accurately assess the T stage of cancer with CT, our results seem to be quite natural and well coincides with those of previous studies. In previous studies, collapsed bowels were reported as sources of errors for incorrect T staging of colon cancers [3]. The preoperative correct diagnosis of advanced colorectal cancer ≥ cT3 (clinical T3) is of clinical importance as many current guidelines including that from the National Comprehensive Cancer Network (NCCN) recommend neoadjuvant CCRT for advanced rectal cancers ≥ cT3 and adjuvant chemotherapy for advanced colon cancers ≥ pT3 [13, 14]. In this regard, our study might be noteworthy. In terms of N staging, there was a greater effect of rectum distention on radiologists’ performance. The accuracy for N staging was better with hydrogel-CT than with conventional CT for both radiologists. It was an unexpected finding as we believed that the accuracy of N staging might depend on the criteria used for discriminating positive LNs, not on whether the rectum is distended or not. However, considering most involved mesorectal lymph nodes are found at or proximal to the level of the tumor, the exact knowledge of the primary tumor site and tumor staging obtained via rectum distention may have been responsible for the better N staging performance of both radiologists. Indeed, among the nine patients in whom primary tumor was not visible, eight were in the control group and only one was in the distention group. Considering that many guidelines including the NCCN guidelines recommend neoadjuvant 12
CCRT for cN1 in rectal cancer [14], the exact identification of metastatic LNs preoperatively may have a clinical impact. We also found that the agreement and correlation between pathologic staging and CT staging were better in the distended group than in the control group for both T and N staging in both reviewers. It may demonstrate the effect of rectum distention on radiologists’ diagnostic performance for T and N staging. Thus, considering its easy use, well availability, and the cheap cost of the sonographic gel, we believe that the use of hydrogel-CT should be the standard CT protocol for patients with rectosigmoid colon cancer. We acknowledge the following limitations of our study. First, as our study was retrospectively designed, various MDCT scanners were used, especially in the control group, resulting in differences in CT acquisition parameters. However, all acquisitions in our study used reconstruction interval of less than 3 mm which may be acceptable for imaging interpretation [3]. Second, the number of patients in our study population (n=54) is relatively small, making the generalization of our study results difficult. Therefore, future studies with a larger study population are warranted to validate the results of our study. Third, the difference in the diagnostic performances between the two reviewers was slightly large for both T and N staging, although the trend toward better performances in the rectum distended group than in the control group was similar between the two reviewers. Nonetheless, further studies recruiting more radiologists with different amounts of expertise may be needed to overcome this limitation. Fourth, CT accuracies for both T and N staging in our study were relatively low compared to those of previous studies [3]. However, it may have been due to the blinding of the results of endoscopy in our study, which would be different from that in real clinical practice. Finally, we did not analyze the presence of mesorectal fascial involvement on CT. Because involvement of circumferential resection margin (CRM) is very important for patients’ 13
prognosis and further treatment plan, description regarding CRM involvement might be critical. Therefore, further study dealing with this issue should be strongly warranted. In conclusion, hydrogel-CT using a sonography transmission gel is a convenient and valuable tool, providing better diagnostic performance for T and N staging of rectosigmoid colon cancer compared to conventional CT without rectal distention.
Conflict of interest All authors have no conflict of interest to declare.
References 1. Visser BC, Varma MG, Welton ML (2001) Local therapy for rectal cancer. Surg Oncol 10:61-69 2. Reynolds JV, Joyce WP, Dolan J, Sheahan K, Hyland JM (1996) Pathological evidence in support of total mesorectal excision in the management of rectal cancer. Br J Surg 83:11121115 3. Dighe S, Purkayastha S, Swift I, et al. (2010) Diagnostic precision of CT in local staging of colon cancers: a meta-analysis. Clin Radiol 65:708-719 4. Gazelle GS, Gaa J, Saini S, Shellito P (1995) Staging of colon carcinoma using water enema CT. J Comput Assist Tomogr 19:87-91 5. Palkó A, Gyulai C, Fedinecz N, Balogh A, Nagy F (2000) Water enema CT examination of rectum cancer by reduced amount of water. Rofo 172:901-904 6. Cademartiri F, Luccichenti G, Rossi A, Pavone P (2002) Spiral hydro-CT in the evaluation of colo-sigmoideal cancer. Radiol Med 104:295-306 14
7. Kim MJ, Lim JS, Oh YT, et al. (2004) Preoperative MRI of rectal cancer with and without rectal water filling: an intraindividual comparison. Am J Roentgenol 182:1469-1476 8. Kim SH, Lee JM, Lee MW, Kim GH, Han JK, Choi BI (2008) Sonography transmission gel as endorectal contrast agent for tumor visualization in rectal cancer. Am J Roentgenol 191:186189 9. Jin KN, Lee JM, Kim SH, et al. (2006) The diagnostic value of multiplanar reconstruction on MDCT colonography for the preoperative staging of colorectal cancer. Eur Radiol 16:22842291 10. Lee IJ, Lee JM, Kim SH, et al. (2010) Diagnostic performance of 64-channel multidetector CT in the evaluation of gastric cancer: differentiation of mucosal cancer (T1a) from submucosal involvement (T1b and T2). Radiology 255:805-814 11. Landis JR, Koch GG (1977) The measurement of observer agreement for categorical data. Biometrics 33:159-174 12. Goh JS, Goh JP, Wansaicheong GK (2002) Methylcellulose as a rectal contrast agent for MR imaging of rectal carcinoma. Am J Roentgenol 178:1145-1146 13. National Comprehensive Cancer Network (2015) Clinical practice guidelines in oncology (NCCN
guidelines):
colon
cancer
version
3.
Available
via
http://www.nccn.org/professionals/physician_gls/pdf/colon.pdf. Accessed August 30, 2015 14. National Comprehensive Cancer Network (2015) Clinical practice guidelines in oncology (NCCN
guidelines):
rectal
cancer
version
3.
Available
http://www.nccn.org/professionals/physician_gls/pdf/rectal.pdf. Accessed August 30, 2015
15
via
Legends for Figures
Figure 1. Semi-quantitative grading for the degree of rectum distention (A) and tumor visualization (B) on CT images. A, Grading for the degree of rectum distention on sagittal CT images. Grade 1 is defined as when the entire rectum (arrowheads) is collapsed; grade 2 when the rectum (arrowheads) is partially distended with more than one collapsed segment; grade 3 when the rectum (arrowheads) is suboptimally distended with one collapsed segment; grade 4 when the entire rectum (arrowheads) is fully distended without a collapsed segment. B, Grading for the degree of tumor visualization on axial CT images. Grade 1 is defined as when the tumor could not be identified; grade 2 when the tumor (arrow) is identified but its intraluminal margin is not defined; grade 3 when the tumor (arrows) is identified and its intraluminal margin is partially defined; grade 4 when the whole tumor margin (arrow) is well defined. 16
Figure 2. A 61-year old woman with distal sigmoid colon cancer in the distended group. A, B. On axial (A) and coronal (B) contrast-enhanced CT images, a 1.5 cm polypoid lesion (arrow) is well visualized at the distal sigmoid colon. Both reviewers recorded this lesion as cT1 stage with a confidence score of 1. Histopathology after lower anterior resection confirmed the stage as T1. c = CT staging. .
17
Figure 3. A 46-year old man with rectal cancer in the distended group. Axial contrast-enhanced CT image shows a focal full-thickness wall thickening (arrow) at the upper rectum (6~8 o’clock). However, perirectal fat seems to be clear. Therefore, both reviewers recorded this lesion as cT2 stage with a confidence score of 1. Histopathology after lower anterior resection confirmed the stage as T2. c = CT staging.
18
Figure 4. A 53-year old man with rectal cancer in the distended group. Axial contrast-enhanced CT image shows a focal full-thickness wall thickening (arrow) with perirectal fat infiltration at the mid-rectum (3~6 o’clock). Both reviewers recorded this lesion as cT3 stage with a confidence score of 5. Histopathology after ultra-low anterior resection confirmed the stage as T3. c = CT staging.
19
Figure 5. A 53-year old man with rectal cancer in the distended group. Serial sagittal contrastenhanced CT images demonstrate focal concentric wall thickening with perirectal infiltration (arrows) at the upper rectum. Note the multiple conglomerated enhancing lymph nodes (arrowheads) around the primary lesion. Both reviewers recorded this lesion as cT3 and cN2 stage with a confidence score of 5. Histopathology after low anterior resection confirmed a rectal adenocarcinoma of T3N2 stage.
20
Tables Table 1. Patients’ Demographics in the Distended and Control Groups Distended Group Control Group P Value Sex (M:F) 17:10 16:11 0.780 Mean Age (Age Range, Years) 61.6 (39~84) 60.1 (40~79) 0.624 Pathologic T pT1 7 6 0.926 Staging pT2 6 7 pT3 14 14 Pathologic N pN0 16 14 0.112 Staging pN1 5 11 pN2 6 2 Operation AR 5 0 0.085 LAR 17 25 ULAR 2 1 Hartmann’s Operation 1 0 Others 2* 1† * = endoscopic submucosal dissection, † = transanal excision. AR = anterior resection, LAR = lower anterior resection, ULAR = ultra-low anterior resection.
21
Table 2. Correlations between Pathologic and CT Staging Distended Group Reviewer 1
pT1 2 1 3 0
T staging pT2 pT3 0 0 2 0 4 9 0 5
pT4 0 0 0 0
N staging pN0 pN1 11 1 3 2 2 2
pN2 0 3 3
cT1 cN0 cT2 cN1 cT3 cN2 cT4 Weighted κ value 0.346 0.451 Correlation coefficient 0.637 (P < 0.0001) 0.558 (P=0.002) Reviewer 2 cT1 6 0 0 0 cN0 15 0 0 cT2 1 5 1 0 cN1 1 5 1 cT3 0 1 12 0 cN2 0 0 5 cT4 0 0 1 0 Weighted κ value 0.836 0.909 Correlation coefficient 0.898 (P < 0.0001) 0.944 (P < 0.0001) Control T staging N staging Group pT1 pT2 pT3 pT4 pN0 pN1 pN2 Reviewer 1 cT1 0 0 0 0 cN0 7 5 1 cT2 2 0 2 0 cN1 3 1 0 cT3 1 4 10 0 cN2 4 5 1 cT4 0 1 2 0 Weighted κ value 0.072 0.043 Correlation coefficient 0.280 (P=0.208) 0.107 (P=0.595) Reviewer 2 cT1 0 0 0 0 cN0 9 4 1 cT2 3 5 2 0 cN1 5 6 0 cT3 1 1 10 0 cN2 0 1 1 cT4 0 0 2 0 Weighted κ value 0.434 0.302 Correlation coefficient 0.552 (P = 0.005) 0.344 (P=0.079) *Correlation coefficients were calculated using Pearson’s correlation analysis. p = pathologic staging, c = CT staging
22
Table 3. Diagnostic Performances of the Two Reviewers for T and N Staging Staging Reviewers Differentiating Reviewer 1 < T3 and ≥ T3
Groups AZ Sensitivity Distended 0.827 100% (14/14) Control 0.817 92.9% (13/14) P Value 1.000 Reviewer 2 Distended 0.989 92.9% (13/14) Control 0.907 85.7% (12/14) P Value 1.000 Differentiating Reviewer 1 Distended 0.858 90.9% (10/11) N0 and ≥ N1 Control 0.544 61.5% (8/13) P Value 0.166 Reviewer 2 Distended 0.980 100% (11/11) Control 0.654 61.5% (8/13) P Value 0.041 AZ = area under the receiver operating characteristic curve
23
Specificity 33.3% (4/12) 12.5% (1/8) 0.603 92.3% (12/13) 80% (8/10) 0.560 56.2% (9/16) 42.9% (6/14) 0.715 87.5% (14/16) 64.3% (9/14) 0.204