Int J Colorectal Dis DOI 10.1007/s00384-015-2172-3
ORIGINAL ARTICLE
Water-enema multidetector computed tomography for planning surgery A. Venara & C. Ridereau-Zins & L. Toque & E. Cesbron & S. Michalak & E. Lermite & C. Aube & A. Hamy
Accepted: 15 February 2015 # Springer-Verlag Berlin Heidelberg 2015
Abstract Purpose Water-enema multidetector computed tomography (WE-MDCT) is a technique for the localization and preoperative T- and N-stage assessments of colon cancer. It may be a useful tool for planning surgery. The primary aim of this study was to evaluate the diagnostic accuracy of WE-MDCT for Tstaging and its ability to locate tumors for laparoscopy planning. The secondary aim was to assess reading reproducibility and diagnostic accuracy for the preoperative determination of N-stage. Methods We performed a study to evaluate preoperative WEMDCT for surgical planning in patients with symptomatic colon adenocarcinomas who underwent surgery between June 2010 and January 2014. A radiologist and a surgeon read the WE-MDCTs separately. Results were compared with colonoscopy and the surgical specimen. Results Seventy-one patients (42 men (59.1 %); mean age 73.1 years (range 45 to 95)) were included. Seventy-six tumors were assessed. The intraclass correlation coefficient
(ICC) for location as determined by surgery and that determined by WE-MDCT was 1, and the ICC for location between colonoscopy and WE-MDCT was 0.85 (95 % CI 0.75–0.91). For T-stage determination, sensitivity was 96 and 94 % and specificity 83 and 88 % for readers 1 and 2, respectively. The T-stage assessment allowed for the programing of surgical access and showed good sensitivity and specificity for the assessment of invasion in adjacent organs. Conclusion WE-MDCT is relatively easy to perform, and its results can be read effectively by radiologists and surgeons. WE-MDCT indicated the location of tumors perfectly and permitted a good determination of their T-stage. The technique is thus pertinent for the planning of laparoscopic surgery for colon cancer.
Keywords Neoplasia . Colon adenocarcinoma . Diagnostic . Staging
Work accepted for oral presentation at The American College of Surgeons 2014 Clinical Congress A. Venara : L. Toque : E. Lermite : C. Aube : A. Hamy L’UNAM, University of Angers, Angers, France A. Venara (*) : L. Toque : E. Lermite : A. Hamy Department of Visceral Surgery, CHU Angers, 4 rue Larrey, 49933 Angers Cedex 9, France e-mail: [email protected] C. Ridereau-Zins : C. Aube Department of Medical Imaging, CHU Angers, Angers, France E. Cesbron Department of Hepato-Gastroenterology, CHU Angers, Angers, France S. Michalak Department of Histopathology, CHU Angers, Angers, France
Background Although colonoscopy is still considered the gold standard investigation technique for malignant colon pathologies [1], radiological assessment is now possible as well, in particular by computed tomography colonography (CT-C) [1–7] or water-enema multidetector computed tomography (WEMDCT) [3, 8–12]. Studies have shown that CT-C has superior patient acceptability compared to colonoscopy [13]. As CT-C uses gas to distend the colon and WE-MDCT water, intraluminal contrast differs between the two techniques. We have been using WE-MDCT in our hospital since the 2000s. The examination is simple and requires neither colon preparation nor a dedicated console for posttreatment [3].
Int J Colorectal Dis
Assessing TNM stage before surgery is necessary to improve the management of colon adenocarcinoma. The pilot phase of the continuing FOxTROT trial demonstrated acceptable toxicity and perioperative morbidity for preoperative chemotherapy in locally advanced (T3 and T4 stage) colon cancer [14]. Another work in this same setting reported that perioperative chemotherapy was safe and caused no increase in early and medium-term complications [15]. Preoperative knowledge of the T3 or T4 status is thus useful to determine the pertinence of neoadjuvant chemotherapy. Computed tomography has become the reference examination to distinguish these stages, with sensitivity and specificity estimated between 78–92 and 71–84 %, respectively [14]. WE-MDCT has been shown to perfectly locate the tumor [12]. Laparoscopic surgery has many advantages but reduces the possibility of palpation to locate the tumor site [16]. Colonoscopy erroneously locates tumors in 4 to 34 % [17, 18] of interventions, essentially in the left or the sigmoid colon [18]. Surgeons need a usable tool to safely locate the tumor during an intervention if laparoscopy fails to do so. Our primary aim for the present work was to evaluate the reliability of WE-MDCT to plan surgery. For this, we assessed the performance of WE-MDCT in locating colon tumors and identifying their T-stage. Our secondary aims were to assess the reproducibility of the reading interpretation between the radiologist and the surgeon and to evaluate the reliability of the preoperative determination of tumor T- and N-stage by WE-MDCT.
Materials and methods Consecutive patients undergoing surgery for symptomatic colon adenocarcinoma, staged by preoperative WE-MDCT between 1 June 2010 and 1 January 2014, were included in this retrospective monocentric study, which was approved by the ethical committee of the University Hospital of Angers. All patients operated for a right colectomy, transverse colectomy, or left colectomy, with or without stomata, were included in the study. Patients who ultimately did not undergo surgery after the WE-MDCT staging and those without preoperative WEMDCT were excluded. Patients with rectal cancer were also excluded as it is now admitted that the gold-standard exploration for rectal adenocarcinoma is MRI [19]. Technical failure was defined as the absence of complete filling of the colon. An examination was considered nonanalyzable if the tumor was not visible or if the location or T-stage could not be identified. Technical failure was not a criterion of exclusion. Data from the medical files, including demographic information (surgical history, body mass index (BMI), reported symptoms), results of the colonoscopy, and results of the
histological examination, were retrospectively collected and entered anonymously in a computer database. Data for WE-MDCT were prospectively collected. The readers were a senior abdominal radiologist with more than 10 years of experience in reading WE-MDCT (reader 1) and a surgeon with no experience in reading WE-MDCT (reader 2). Both were blinded to the initial radiological interpretation, surgical findings, and final pathological results. The same, commonly used, WE-MDCT protocol as described in previous studies [3, 8, 12] was used for all examinations. No bowel cleaning preparation was performed. Patients were explored on a 16-row multidetector CT (Philips, Netherlands) with collimation at 1.5, pitch of 0.9, 120 kV, and mAs adapted to the morphology of each patient. For each patient, the colon was filled with water over 2 min. The acquisition was performed after injection of 2 mL/kg of iodine contrast media at a concentration of 350 g/L with a flow rate of 3.5 mL/s using an automated power injector. The water was evacuated at the end of the acquisition [12]. All examinations were read on a PACS workstation (PACS Synapse, Fujifilm Medical System, Stamford, US) using native images and multiplanar reconstruction (MPR). Tumor analysis criteria were as follows: location on the colon, growth pattern, external tumor margin, peritumoral fat, and peritumoral invasion [12] (Table 1). Criteria for radiological N-stage determination were size and density after injection of the lymph node [12]. Size greater than 5 mm and/or density higher than 100 HU were considered to indicate N positivity. The data of all tumors were reported; if a single patient had two concomitant tumors, we numbered two tumors in the database. Radiological readings allowed for the staging of the tumors analogously with the histological UICC 2002 TNM classification [20]. Tumors were classified into three groups according to T-stage: group 1 (adenoma, T1/T2), group 2 (T3), or group 3 (T4); then two groups according to N-stage: group A (N0) or group B (N+). The analysis performed in histopathology as recommended in the histological UICC 2002 TNM staging [20] was used as the reference for T-stage and N-stage in this study. The locations determined by radiology, endoscopy, and surgery were compared, with the latter serving as reference. Table 1
Criteria for the determination of radiological T-stage
Stage
Adenoma/T1/T2 T3 T4
External tumor margin Pericolic fat Peritumoral aspect
Well defined Clean
Convex and smooth or irregular and bowl-shaped aspect Invaded (fat densification) Spread to the retroperitoneal fascia or the adjacent organ
Int J Colorectal Dis
Statistical analyses were performed using SPSS version 15 Software (SPSS, Chicago, IL). Interobserver correlation and the correlation between histological and radiological stagings were assessed using the intraclass correlation index. Statistical significance was inferred as a confidence level of 5 % (p
ORIGINAL ARTICLE
Water-enema multidetector computed tomography for planning surgery A. Venara & C. Ridereau-Zins & L. Toque & E. Cesbron & S. Michalak & E. Lermite & C. Aube & A. Hamy
Accepted: 15 February 2015 # Springer-Verlag Berlin Heidelberg 2015
Abstract Purpose Water-enema multidetector computed tomography (WE-MDCT) is a technique for the localization and preoperative T- and N-stage assessments of colon cancer. It may be a useful tool for planning surgery. The primary aim of this study was to evaluate the diagnostic accuracy of WE-MDCT for Tstaging and its ability to locate tumors for laparoscopy planning. The secondary aim was to assess reading reproducibility and diagnostic accuracy for the preoperative determination of N-stage. Methods We performed a study to evaluate preoperative WEMDCT for surgical planning in patients with symptomatic colon adenocarcinomas who underwent surgery between June 2010 and January 2014. A radiologist and a surgeon read the WE-MDCTs separately. Results were compared with colonoscopy and the surgical specimen. Results Seventy-one patients (42 men (59.1 %); mean age 73.1 years (range 45 to 95)) were included. Seventy-six tumors were assessed. The intraclass correlation coefficient
(ICC) for location as determined by surgery and that determined by WE-MDCT was 1, and the ICC for location between colonoscopy and WE-MDCT was 0.85 (95 % CI 0.75–0.91). For T-stage determination, sensitivity was 96 and 94 % and specificity 83 and 88 % for readers 1 and 2, respectively. The T-stage assessment allowed for the programing of surgical access and showed good sensitivity and specificity for the assessment of invasion in adjacent organs. Conclusion WE-MDCT is relatively easy to perform, and its results can be read effectively by radiologists and surgeons. WE-MDCT indicated the location of tumors perfectly and permitted a good determination of their T-stage. The technique is thus pertinent for the planning of laparoscopic surgery for colon cancer.
Keywords Neoplasia . Colon adenocarcinoma . Diagnostic . Staging
Work accepted for oral presentation at The American College of Surgeons 2014 Clinical Congress A. Venara : L. Toque : E. Lermite : C. Aube : A. Hamy L’UNAM, University of Angers, Angers, France A. Venara (*) : L. Toque : E. Lermite : A. Hamy Department of Visceral Surgery, CHU Angers, 4 rue Larrey, 49933 Angers Cedex 9, France e-mail: [email protected] C. Ridereau-Zins : C. Aube Department of Medical Imaging, CHU Angers, Angers, France E. Cesbron Department of Hepato-Gastroenterology, CHU Angers, Angers, France S. Michalak Department of Histopathology, CHU Angers, Angers, France
Background Although colonoscopy is still considered the gold standard investigation technique for malignant colon pathologies [1], radiological assessment is now possible as well, in particular by computed tomography colonography (CT-C) [1–7] or water-enema multidetector computed tomography (WEMDCT) [3, 8–12]. Studies have shown that CT-C has superior patient acceptability compared to colonoscopy [13]. As CT-C uses gas to distend the colon and WE-MDCT water, intraluminal contrast differs between the two techniques. We have been using WE-MDCT in our hospital since the 2000s. The examination is simple and requires neither colon preparation nor a dedicated console for posttreatment [3].
Int J Colorectal Dis
Assessing TNM stage before surgery is necessary to improve the management of colon adenocarcinoma. The pilot phase of the continuing FOxTROT trial demonstrated acceptable toxicity and perioperative morbidity for preoperative chemotherapy in locally advanced (T3 and T4 stage) colon cancer [14]. Another work in this same setting reported that perioperative chemotherapy was safe and caused no increase in early and medium-term complications [15]. Preoperative knowledge of the T3 or T4 status is thus useful to determine the pertinence of neoadjuvant chemotherapy. Computed tomography has become the reference examination to distinguish these stages, with sensitivity and specificity estimated between 78–92 and 71–84 %, respectively [14]. WE-MDCT has been shown to perfectly locate the tumor [12]. Laparoscopic surgery has many advantages but reduces the possibility of palpation to locate the tumor site [16]. Colonoscopy erroneously locates tumors in 4 to 34 % [17, 18] of interventions, essentially in the left or the sigmoid colon [18]. Surgeons need a usable tool to safely locate the tumor during an intervention if laparoscopy fails to do so. Our primary aim for the present work was to evaluate the reliability of WE-MDCT to plan surgery. For this, we assessed the performance of WE-MDCT in locating colon tumors and identifying their T-stage. Our secondary aims were to assess the reproducibility of the reading interpretation between the radiologist and the surgeon and to evaluate the reliability of the preoperative determination of tumor T- and N-stage by WE-MDCT.
Materials and methods Consecutive patients undergoing surgery for symptomatic colon adenocarcinoma, staged by preoperative WE-MDCT between 1 June 2010 and 1 January 2014, were included in this retrospective monocentric study, which was approved by the ethical committee of the University Hospital of Angers. All patients operated for a right colectomy, transverse colectomy, or left colectomy, with or without stomata, were included in the study. Patients who ultimately did not undergo surgery after the WE-MDCT staging and those without preoperative WEMDCT were excluded. Patients with rectal cancer were also excluded as it is now admitted that the gold-standard exploration for rectal adenocarcinoma is MRI [19]. Technical failure was defined as the absence of complete filling of the colon. An examination was considered nonanalyzable if the tumor was not visible or if the location or T-stage could not be identified. Technical failure was not a criterion of exclusion. Data from the medical files, including demographic information (surgical history, body mass index (BMI), reported symptoms), results of the colonoscopy, and results of the
histological examination, were retrospectively collected and entered anonymously in a computer database. Data for WE-MDCT were prospectively collected. The readers were a senior abdominal radiologist with more than 10 years of experience in reading WE-MDCT (reader 1) and a surgeon with no experience in reading WE-MDCT (reader 2). Both were blinded to the initial radiological interpretation, surgical findings, and final pathological results. The same, commonly used, WE-MDCT protocol as described in previous studies [3, 8, 12] was used for all examinations. No bowel cleaning preparation was performed. Patients were explored on a 16-row multidetector CT (Philips, Netherlands) with collimation at 1.5, pitch of 0.9, 120 kV, and mAs adapted to the morphology of each patient. For each patient, the colon was filled with water over 2 min. The acquisition was performed after injection of 2 mL/kg of iodine contrast media at a concentration of 350 g/L with a flow rate of 3.5 mL/s using an automated power injector. The water was evacuated at the end of the acquisition [12]. All examinations were read on a PACS workstation (PACS Synapse, Fujifilm Medical System, Stamford, US) using native images and multiplanar reconstruction (MPR). Tumor analysis criteria were as follows: location on the colon, growth pattern, external tumor margin, peritumoral fat, and peritumoral invasion [12] (Table 1). Criteria for radiological N-stage determination were size and density after injection of the lymph node [12]. Size greater than 5 mm and/or density higher than 100 HU were considered to indicate N positivity. The data of all tumors were reported; if a single patient had two concomitant tumors, we numbered two tumors in the database. Radiological readings allowed for the staging of the tumors analogously with the histological UICC 2002 TNM classification [20]. Tumors were classified into three groups according to T-stage: group 1 (adenoma, T1/T2), group 2 (T3), or group 3 (T4); then two groups according to N-stage: group A (N0) or group B (N+). The analysis performed in histopathology as recommended in the histological UICC 2002 TNM staging [20] was used as the reference for T-stage and N-stage in this study. The locations determined by radiology, endoscopy, and surgery were compared, with the latter serving as reference. Table 1
Criteria for the determination of radiological T-stage
Stage
Adenoma/T1/T2 T3 T4
External tumor margin Pericolic fat Peritumoral aspect
Well defined Clean
Convex and smooth or irregular and bowl-shaped aspect Invaded (fat densification) Spread to the retroperitoneal fascia or the adjacent organ
Int J Colorectal Dis
Statistical analyses were performed using SPSS version 15 Software (SPSS, Chicago, IL). Interobserver correlation and the correlation between histological and radiological stagings were assessed using the intraclass correlation index. Statistical significance was inferred as a confidence level of 5 % (p
