SS 301b - Imaging of the colon and pelvic floor
Purpose: Unit costs of screening CT colonography (CTC) can be useful for cost-effectiveness analyses and for health care decision-making. We evaluated the unit costs of CTC as a primary screening test for colorectal cancer in the setting of a randomised trial in Italy.
Methods and Materials: Data were collected within the randomised SAVE trial. Subjects were invited to screening CTC by mail and requested to have a pre-examination consultation. CTCs were performed with 64- and 128-slice CT scanners after reduced or full bowel preparation. Activity-based costing was used to determine unit costs per-process, per-participant to screening CTC and per-subject with advanced neoplasia.
Results: Among 5242 subjects invited to undergo screening CTC, 1312 had pre-examination consultation and 1286 ultimately underwent CTC. Among 129 subjects with a positive CTC, 126 underwent assessment colonoscopy and 67 were ultimately diagnosed with advanced neoplasia (i.e. cancer or advanced adenoma). Cost per-participant of the entire screening CTC pathway was € 196.80. Average cost per-participant for the screening invitation process was € 17.04, while it was € 9.45 for the pre-examination consultation process. Average costs per-participant of the CTC execution and reading process was € 146.08 and of the diagnostic assessment colonoscopy process was € 24.23. Average cost per-subject with advanced neoplasia was € 3777.30.
Conclusion: Our data suggest that the more relevant cost of screening CTC, amenable of intervention, is related to CTC execution and reading process.
Retrospective analysis of patient position errors and impact of MSCT dose in virtual colonscopy: single centre experience
Purpose: Virtual-colonoscopy (VC) is a well known and valid technique for evaluation of colon wall and detection of new lesions. The main drawback for its execution is patient exposition to ionizing radiations. Indeed, guidelines recommend standardized low dose acquisition procedures aimed to minimize radiation exposition. We retrospectively examined the impact in term of radiation exposure of patient position and scan length mislead.
Methods and Materials: A two-year retrospective analysis of multi-slice-computed-tomography (MSCT) acquisition data of 106 patients that underwent a VC at our institution. Exams were performed in a MSCT 128 slice with low-dose protocol. The parameters that characterise the examinations and could induce an increase of absorbed dose were: patient position with respect to the gantry isocenter and scan length that could exceed the area of interest.
Results: Mean value of patient position resulted below isocenter for 48±25mm and 29±27mm in prone and supine positions, respectively. On the basis of previous literature experience the dose increase in our patient population, due to the incorrect positioning, was estimated to be about 30% (~1.5mSv) and 20% (1mSv) for prone and supine positions, respectively.Moreover, we found that the scan exceeded the correct position below the anal orifice for an average length of 36±25mm e 48±28mm in prone and supine positions, respectively, exposing sensitive anatomic organs.
Conclusion: Errors in acquisition parameters, eluding the optimized protocols, should be avoided given the increase in the dose.Nowadays acquisition technique is one crucial focus on daily practice, giving the major spread of ionizing radiations in routine exams.
Purpose: Computed tomography colonography (CTC) is an established tool in the screening for colon cancer. Furthermore CTC can be used in the staging of colon cancer, which is imperative with the possibility of neo-adjuvant treatment in the nearby future (FOXTROT trial). Therefore the aim of this study is to determine the accuracy of CTC in the detection of important prognostic factors being; tumour invasion beyond the bowel wall and nodal involvement of colorectal carcinomas.
Methods and Materials: A literature search was performed in Ovid, Embase and Pubmed to identify studies reporting on the accuracy of CTC for local staging of colon cancer with histology as reference standard. Data were extracted by two observers in consensus. A hierarchical summary ROC (HSROC) model was used to construct a summary ROC curve and to calculate summary estimates of sensitivity, specificity and diagnostic odds ratios (DOR).
Results: Fourteen studies fulfilled the inclusion criteria for detection of tumour invasion beyond the bowel wall (combined total of 1257 lesions). The pooled sensitivity, specificity and DOR were 90% (95% CI: 88-92%); 87% (95% CI: 83-90%) and 19,6 (95% CI: 10,2-41,5) respectively. For nodal involvement nine studies were included (combined total of 525 patients) the estimates were 84% (95% CI: 80-88%); 66% (95% CI: 59-72%); 5,8 (95% CI: 2,5-9,4) respectively.
Conclusion: CTC is an accurate tool in the staging of colon cancer with a good sensitivity for the detection of T3/T4 tumours and an improved specificity than standard CT. However, the accuracy in detecting nodal involvement is mediocre.
Purpose: To retrospectively identify the most significant morphologic findings at CT colonography (CTC) for differentiating between sigmoid cancer and chronic diverticular disease.
Methods and Materials: Two experienced radiologists reviewed the CTC examinations of 54 Patients, 33 with proven chronic diverticular disease and 21 with proven sigmoid cancer, and identified a subgroup of 26 Patients with a mass lesion (15 Patients with chronic diverticular disease and 11 with carcinoma). The readers looked at specific morphological criteria established from a literature review for CT findings of chronic diverticular disease and colon cancer. Sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV) and interval of confidence were calculated.
Results: The most suggestive morphologic findings for carcinoma are: absence of diverticula in the affected segment (NPV 82.6%, PPV 93.5%); straightened growth pattern (NPV 83.3%, PPV 83.3%); presence of shoulder phenomenon (NPV 93.1%, PPV 76%); completely distorted mucosal folds pattern (NPV 96.2, PPV 71.4). In the “mass lesion subgroup”, the distortion of mucosal folds (NPV 87.5%, PPV 55.6%) and the straightened growth pattern (NPV 75%, PPV 70%) have significantly lower diagnostic performance. In this latter group, the two strongest morphological features suggestive of carcinoma are the absence of diverticula in the affected segment (NPV 93.3%, PPV 90.9%) and the presence of shoulder phenomenon (NPV 91.7%, PPV 71.4%).
Conclusion: The most useful morphological sign to differentiate chronic diverticulitis from colorectal cancer is the presence of diverticula in the affected segment.
CT colonography: effect of electronic cleansing of tagged fecal residuals on the size of submerged colorectal polyps in screening patients
Purpose: To assess whether there is a difference in the size of colorectal polyps before and after electronic cleansing (EC) of tagged residuals in screening CT colonography (CTC) studies.
Methods and Materials: A database of 894 colonoscopy-validated CTC patient-datasets of a low prevalence cohort was retrospectively reviewed to identify patients with polyps ≥6mm that were entirely submerged in tagged residuals in at least 1 of 2 scanning positions. 2D measurements of the largest diameter of each identified polyp were performed independently by ten radiologists before and after EC in colon, bone, and soft tissue windows in a fully randomised order. Differences of size and of polyp count before and after EC were calculated for size-categories ≥6mm, ≥10mm. Statistical testing involved 95% CI, Intraclass Correlation and mixed model ANOVA.
Results: 37 patients with 48 submerged polyps (34 adenomas, 14 non-adenomas) ≥6mm were identified by reference-standard. Mean polyp size before EC was 9.8mm in colon, 9.9mm in bone and 8.2mm in soft tissue windows. After EC the mean polyp size decreased significantly to 9.4mm in colon, 9.1mm in bone, and 7.1mm in soft tissue windows. Compared to the unsubtracted colon window, EC, performed in bone and soft tissue windows, leads to a shift of 10 (20.8%) and 25 (52.1%) polyps ≥6mm and 1 (6.7%), and 5 (33.3%) polyps ≥10mm into the next smaller size category.
Conclusion: EC significantly reduces the size of polyps submerged in tagged residuals. Polyp measurements should be performed in unsubtracted image data in a colon window setting.
Diagnostic performance and influence on computer-assisted-diagnosis (CAD) of ultra-low dose CT colonography (ULD-CTC) with model-based iterative reconstruction (MBIR)
Purpose: To compare perfomance of standard-dose (SD) and ultra low-dose (ULD) in CT-Colonography (CTC), using hybrid-iterative reconstruction (HIR) and model-based iterative reconstruction (MBIR), and to asses the effect of different reconstruction algorithms on computed-aided diagnosis (CAD).
Methods and Materials: Fifty patients (31 males, 19 females, mean age 69,7) underwent CTC with SD (120kV, 50mAs 4,6±1,3mSv) and ULD (120kV, 10mAs, 0,95±0,2mSv); images were reconstructed with HIR and MBIR; CAD system was applied complementary to all reconstructions. Two independent radiologists review obtained images separately randomized to avoid recall-bias, and recorded colonic findings (polyps/cancer), extracolonic findings and CAD detections. Polyps were classified by size (<5mm, =5-10mm, ≥10mm), shape (flat, sessile, peduncolated) and location (rectum, sigmoid, descending, transverse, ascending, caecum). Extracolonic findings ≥E4 (C-RADS) were recorded. CAD results were reported and significant findings were selected.
Results: Nine polyps were found in 8 patients, six <5mm, two between 5-10mm and one ≥10mm; 8 were peduncolated and 1 flat, 2 were in caecum, 3 ascending, 2 transverse, 1 descending, 1 sigmoid tract; colonic-findings records were super-imposable for both radiologist between SD and ULD reconstructions; 2 extracolonic findings ≥E4 were recorded in all reconstructions. MBIR-CAD results compared to HIR-CAD showed a mean 13,56% increment at SD and 32,92% at ULD, with no change in terms of significant results. Colonic findings were confirmed by optical-colonscopy (OC).
Conclusion: ULD-MBIR CTC showed promising results in terms of diagnostic performance and inter-reader reproducibility for intra-colonic findings. CAD detection was influenced by MBIR, with no decrease in diagnostic value.
Protocol optimisation of MR colonography for polyp detection using pig colonic phantom: influence of magnetic field strength, colonic distension technique, and MRI sequence
Purpose: To compare the diagnostic performance and image quality of MR Colonography (MRC) using pig colon phantoms and to evaluate the influence of magnetic field strength (1.5-T or 3.0-T), colonic distension technique (bright- or dark-lumen), and MRI sequences.
Methods and Materials: Six pig colon segments (60-92 cm) with 56 artificial colon polyps (0.4-1.6 cm) were placed in plastic container containing soybean oil. The colon was distended using room air for dark-lumen MRC and with tap water or a gadolinium-chelate based enema fluid for bright-lumen MRC. Each colon phantom was scanned on both 1.5-T and 3.0-T scanners using the following sequences: 2D fast-imaging with steady-state precession, T2-weighted 2D single-shot fast-spin-echo (SSFSE), and/or T1-weighted 3D gradient-echo (GRE) sequences. Two radiologists evaluated the presence of polyps and analysed the image quality.
Results: For polyp detection sensitivity and image quality, MRC obtained at 1.5-T was better than that obtained at 3.0-T, and a bright-lumen technique was superior to a dark-lumen technique. Bright-lumen MRC at 1.5-T was most sensitive for polyp detection (p < 0.001) and gave the highest image quality (p < 0.05) regardless of polyp size and shape (flat or sessile). SSFSE and 3D GRE sequences at bright-lumen MRC at 1.5-T had highest sensitivity for polyp detection (83.9% and 83.0%, respectively) and highest image quality.
Conclusion: The most effective sequences of MRC for polyp detection were SSFSE- or 3D GRE-based bright-lumen MRC obtained with a 1.5-T scanner. These sequences had the highest polyp detection rate and the best image quality.
Purpose: To investigate the role of magnetic resonance imaging (MRI) as a potential tumour marker in predicting the aggressiveness of colon cancer tumours with the apparent diffusion coefficient (ADC), and its potential in discriminating between low and high-risk colon cancer patients.
Methods and Materials: Thirty patients (21M, 9F) were included retrospectively. All patients received a 1.5T MRI of the colon including T2 and DWI sequences. ADC maps were automatically constructed for all diffusion weighted magnetic images with b factors of 0 and 1000 s/mm(2). ADC measurements of each colon tumour were performed by two readers in concensus using regions of interest with IntelliSpace Discovery research platform (Philips Healthcare, The Netherlands). High-risk colon cancer patients were defined if one or more of the following parameters were present with histology as the reference standard: direct invasion of tumour into surrounding structures/organs (T4), nodal involvement (N+) and/or distant metastasis (M+). The student t-test was used to assess the differences between the ADC means of low and high-risk colon cancer tumours.
Results: The mean ADC of the high risk patients (n=16) was 1129.7 and for low-risk patients (n=14) 1231.8 with a significant difference between the mean ADC intensity of high versus low risk tumours (p=0.002). The optimal cut off value was 1178.8 with an area under the curve (AUC) of 0.83 and a sensitivity and specificity of 81% and 86% respectively.
Conclusion: ADC has potential to discriminate between low and high-risk colon cancer patients with lower ADC values significantly associated with high-risk colon tumours.
CT differentiation of poorly-differentiated colorectal neuroendocrine tumours from well-differentiated neuroendocrine tumours and colorectal adenocarcinoma
Purpose: To find differential CT features of colorectal poorly-differentiated neuroendocrine tumours (PD-NETs) from well-differentiated NETs (WD-NETs) and adenocarcinomas (ADCs).
Methods and Materials: CT features of 25 pathologically-proven colorectal WD-NETs, 36 PD-NETs and 36 ADCs were retrospectively reviewed. The following CT items were analysed: size, longitudinal location, morphology, homogeneity, the presence of intact overlying mucosa, necrosis, calcification, homogeneity, degree of enhancement on each CT phase, the presence of enlarged lymph node (LN), and metastasis. Significant CT variables were determined using the chi-square test, Fisher’s exact test, and Student t-test. Receiver operating characteristic analysis was used to determine the optimal cut-off value of tumour and LN size.
Results: Large size, rectum location, ulceroinfiltrative morphology, absence of intact overlying mucosa, heterogeneous CT attenuation with necrosis, presence of ≥3 enlarged LNs, and metastasis were found to be significant variables to differentiate PD-NETs from WD-NETs (P<0.05). The optimal cut-off value for tumour size in differentiating PD-NETs from WD-NETs was 1.5cm (AUC=0.958, sensitivity=100%, specificity= 84%, P<0.0001). High attenuation on arterial phase, persistent high enhancement pattern, presence of ≥6 enlarged LNs, large LN size, and wash-in/wash-out enhancement pattern of liver metastasis were significant variables to differentiate PD-NETs from ADCs (P<0.05).
Conclusion: Compared to WD-NETs, colorectal PD-NETs usually appear as large, ulceroinfiltrative, heterogeneous rectal mass without intact overlying mucosa and accompany with enlarged LNs and metastasis. High attenuation on arterial phase, presence of enlarged LNs with larger size and greater number, and wash-in/wash-out enhancement pattern of liver metastasis can be useful CT discriminators of colorectal PD-NETs from ADCs.
Purpose: To assess the value of MR-defecography (MRD) parameters used in diagnosis of obstructed defecation.
Methods and Materials: Twenty-two consecutive patients (16 women, 6 men; mean age 51±19.4) with obstructed defecation and twenty healthy volunteers (11 women, 9 men; mean age 33.4±11.5) underwent MRD in a closed-configuration 3.0T MRI-system in supine position. MRD included midsagittal T2-weighted images at rest and during defecation after filling the rectum with 250ml water-based gel. Two independent and blinded radiologists measured pelvic floor descent in reference to the pubococcygeal line (PCL) and HMO-lines during defecation, anorectal angle (ARA), qualitatively assessed grade of evacuation (GE), paradoxical contraction (PC), and missing sphincter relaxation (MSR).
Results: Interreader agreement for PCL-, HMO-, and ARA-measurements was good to excellent (Intraclass-correlation-coefficient [ICC], 0.60-0.99), for GE excellent (κ-value, 0.76-0.80), and for PC and MSR fair to moderate (κ-value, 0.24-0.47). Posterior compartment descent, H-, M-line and ARA were significantly greater in patients compared to volunteers (p<0.05). 30-50% of volunteers had mild anterior and middle compartment descent and hiatal enlargement. 50% of volunteers had moderate rectal descent. Impaired evacuation was seen in 9/20 (45%) volunteers and 9/22 (41%) patients. Also PC and MSR were equally found in volunteers (40%, 40%) and patients (41%, 32%).
Conclusion: Quantitative, but not qualitative MRD-parameters show both excellent interreader agreement and ability to differentiate patients with obstructed defecation from healthy volunteers. However, pelvic floor descent values considered abnormal according to PCL and HMO-system are found in a substantial number of healthy volunteers and therefore need to be redefined.
Dynamic MRI of the pelvic floor in different body positions: success rate of MR defecography in supine vs left lateral body position
Purpose: To assess the success rate of MR-defecography (MRD) performed in supine versus lateral body position and assess differences in pelvic floor measurements.
Methods and Materials: 22 consecutive patients (16 women, 6 men; mean age 51±19.4) with obstructed defecation underwent MRD in a closed-configuration 3.0T MRI-system in supine and lateral position. MRD included midsagittal T2-weighted images at rest and during defecation after filling the rectum with 250ml water-based gel. Two independent radiologists measured pelvic floor descent in reference to the pubococcygeal line (PCL) and assessed grade of evacuation (GE) in both body positions. Image quality (IQ) was rated on a 5-point-scale (5=excellent, 1=poor).
Results: Grades of middle and posterior compartment descent were similar in supine and lateral position (p>0.05). Anterior compartment descent was significantly higher in lateral position, but still normal to small in the majority of patients (4.6±23.1 cm vs. 8.9±25.2 cm, p<0.042). When attempting to defecate in supine position 6/22 (27%) patients showed no evacuation, while in lateral position only 3/22 (14%) were not able to evacuate. Image quality was equal at rest (4.4±0.5 and 4.7±0.6, p>0.05) and slightly better in supine compared to the lateral position during defecation (4.5±0.4 vs. 3.9±0.9, p<0.017).
Conclusion: In lateral position more patients were able to evacuate with similar grades of pelvic floor descent compared to supine position. Image quality was slightly degraded during defecation in lateral position. MRD in lateral position is a valuable alternative for patients unable to defecate in supine position.