Spanish 
Chinese 
Russian 
German 
French 
Japanese 
Portuguese 
Hindi Research Article, J Liver Dis Transplant Vol: 3 Issue: 1
Diagnostic Imaging of Patients with Pancreato-Biliary Diseases: Comparison between Ultrasound, Computed Tomography and Magnetic Resonance
| Simone Maurea*, Antonio Corvino, Pier Paolo Mainenti, Carmine Mollica, Massimo Imbriaco, Luigi Camera, Marcello Mancini, Fabio Corvino and Marco Salvatore | |
| Dipartimento di Scienze Biomediche Avanzate, Università degli Studi di Napoli Federico II (UNINA), Istituto di Biostrutture e Bioimmagini (IBB), Consiglio Nazionale delle Ricerche (CNR); Fondazione SDN (IRCCS); Napoli, Italy | |
| Corresponding author : Simone Maurea, MD, PhD via Ernesto Murolo n. 5, 80123 - Napoli, Italy Tel: (39) (81) – 746 3560 (2039); Fax: (39) (81) - 5457081 E-mail: maurea@unina.it |
|
| Received: March 24, 2014 Accepted: July 25, 2014 Published: July 31, 2014 | |
| Citation: Maurea S, Corvino A, Mainenti PP, Mollica C, Imbriaco M (2014) Diagnostic Imaging of Patients with Pancreato-Biliary Diseases: Comparison between Ultrasound, Computed Tomography and Magnetic Resonance. J Liver: Dis Transplant 3:1. doi:10.4172/2325-9612.1000121 |
Abstract
Diagnostic Imaging of Patients with Pancreato-Biliary Diseases: Comparison between Ultrasound, Computed Tomography and Magnetic Resonance
The aim of this study was to directly compare the results of MR cholangio-pancreatography (MRCP) with those of ultrasound (US) and multi-slice computed tomography (MSCT) in patients with pancreatico-biliary diseases. A total of 110 patients (62 M, 48 F), ranging in age from 22 to 89 years, was studied before surgery (n=99) or after cholecystectomy (n=11) for lithiasis. MRCP was performed in all patients while US was acquired in 55 patients and MSCT was performed in 76 patients. Histology (n=34), biopsy (n=38), endoscopic retrograde cholangio-pancreatography (ERCP) (n=28) and/or clinical-imaging follow-up (n=10) data were considered as standard of reference. Patient population was divided in three groups; Group 1 (n=55) consisted of a comparison between MRCP and US in biliary tract diseases; Group 2 (n=37) consisted of a comparison between MRCP and MSCT in biliary tract diseases; Group 3 (n=40) consisted of a comparison between MRCP and MSCT in pancreatic masses. A regional imaging qualitative evaluation of biliary and pancreatic duct system (gallbladder and cystic duct, intra- and extra-hepatic ducts, main pancreatic duct) was performed.
Keywords: Pancreatico-biliary disease; MR cholangiopancreatography; Bile duct stones
Keywords |
|
| Pancreatico-biliary disease; MR cholangiopancreatography; Bile duct stones | |
Introduction |
|
| Diagnostic imaging of pancreatico-biliary disease includes several techniques which are selected on the basis of clinical symptoms as well as technique invasiveness; in particular, ultrasonography (US) and computed tomography (CT), the most commonly noninvasive modalities used in this setting, appear to have high specificity but lower sensitivity especially in patients with common bile duct stones that require alternative techniques such as endoscopic retrograde cholangiopancreatography (ERCP) [1–4]. However, despite its high sensitivity, specificity and diagnostic accuracy and its inherent therapeutic possibilities, ERCP is invasive and thus it has limited use for diagnostic purposes alone [4–7]. This limitation has led to the use of magnetic resonance cholangiopancreatography (MRCP) that allows non-invasive accurate evaluation of pancreaticobiliary ducts. Similarly, diagnostic evaluation of pancreatic masses relies on several imaging modalities, such as US, CT, MRCP as well as CT combined with positron emission tomography (PET-CT) [8-10]. Therefore, the availability of multiple imaging modalities, each with different levels of invasiveness and accuracy, calls for protocols or guidelines to ensure that patients with pancreaticobiliary disease are correctly managed. | |
| Comparative studies regarding the imaging modalities used for diagnosing pancreaticobiliary disease indicate a fundamental role for MRCP, which has high sensitivity, specificity and accuracy [3,4,11-14]. In particular, in patients with cholelithiasis, diagnostic accuracy is dependent on the site, size and structure of the bile stones; in this setting, US has a good level of accuracy in the study of the extrahepatic common hepatic duct and of the intrahepatic bile ducts when dilated but performs less well in the assessment of the middle-distal portion of the extrahepatic ducts [13,15-17]. Multislice (MS) CT technology allows reconstruction of three-dimensional volumetric images with significant advantages in the evaluation of bile duct anatomy; however, given that CT cannot always visualise the cause of the obstruction because of the frequent isodensity of stones relative to bile, detection is based on the same indirect criteria as used at US, such as dilatation of the bile ducts above the stenosis; for this reason, it has been suggested that CT with biliary contrast material might help identify filling defects caused by the presence of radiolucent stones within the opacified bile duct lumen, especially in patients with contraindications to MRCP [18]. On the other hand, the choice between CT and MR imaging for diagnostic evaluation of pancreatic masses is controversial since the the majority of studies in the literature suggests that CT and MR imaging have comparable diagnostic accuracy [19-24], whereas other studies report conflicting results [25-28]. | |
| In this prospective study we report oue experience regarding the comparison of US, MSCT and MRCP in evaluating pancreaticobiliary diseases; for this purpose, we performed a comparative regional analysis of imaging results in the main anatomic sites of the pancreaticobiliary system as follow gallbladder, intrahepatic bile ducts, extrahepatic bile ducts, main pancreatic duct; furthermore, we directly compare the results of MSCT and MR imaging in tumor detection and resectability of malignant lesions in a group of patients with pancreatic masses. | |
Materials and Methods |
|
| Population | |
| A total of 110 patients (62 M, 48 F), ranging in age from 22 to 89 years, with pancreatico-biliary diseases was prospectively studied before surgery (n=99) or after cholecystectomy (n=11) for lithiasis. MRCP was performed in all patients while US was acquired in 55 patients and MSCT was performed in 77 patients. Histology (n=34), biopsy (n=38), endoscopic retrograde cholangio-pancreatography (ERCP) (n=28) and/or clinical-imaging follow-ussp (n=10) data were considered as standard of reference. Patient population was divided in three groups; Group 1 (n=55) consisted of a comparison between MRCP and US in biliary tract diseases; Group 2 (n=37) consisted of a comparison between MRCP and MSCT in biliary tract diseases; Group 3 (n=40) consisted of a comparison between MRCP and MSCT in pancreatic masses. | |
| Ultrasonography | |
| Patients were imaged with an ATL 5500 HDI unit after a 6-h fast and in various positions (supine, left lateral, upright). Examinations were carried out with a 3.5–4 MHz probe initially placed subcostally during deep breath-holding or between the 10th and 11th rib, without breath-holding. In all US studies, colour Doppler imaging was used to differentiate portal branches from dilated bile ducts. Transverse, longitudinal and oblique scans of the upper abdomen were performed, as reported in the literature [29,30]. | |
| Computed tomography | |
| CT examinations were performed with a four-detector row MSCT scanner (MX 8000-Marconi) with a rotation time of 0.5 s allowing acquisition of eight sections per second. All patients were instructed to drink 750 ml of water 10–15 min before the examination to improve definition of relationships between the pancreatic head and the second portion of the duodenum. Venous access for administration of contrast material was achieved with an 18- to 20-gauge needle cannula placed in an antecubital vein of the arm. An initial unenhanced baseline scan of the pancreatic region was obtained in the craniocaudal direction from the diaphragm down to the iliac crests, with the following parameters: collimation 4×1 mm, reconstruction interval 3 mm, pitch 0.875, kV 120, mAs 260 and field of view 25 cm. After contrast-medium administration, the pancreatic region was imaged in the cranio-caudal direction using the following parameters: collimation 4×1 mm, reconstruction interval 1.25 mm, pitch 0.875, kV 120, mAs 260 and field of view 25 cm. Acquisition of scans in the arterial phase was synchronised with the passage of contrast through the arterial tree using real-time bolus tracking to calculate scan delay. Scans were subsequently acquired in the pancreatic (40–45 s) and portal (70–80 s) phases. The contrast medium used was nonionic iodinated iopromide (Ultravist, Schering) with an iodine concentration of 370 mgI/ml administered at a rate of 3 ml/s via an automatic power injector. | |
| MR Cholangiography | |
| MR study was performed with a 1.5-Tesla scanner (Philips, Gyroscan Intera). After receiving 900 ml of oral superparamagnetic contrast material consisting of a suspension of silicon-coated iron oxide crystals (Ferumoxil, Lumirem, Guebert), all patients were studied with dedicated T1- and T2-weighted breath-hold (Bh) sequences using a phasedarray synergy body coil. Precontrast T1-weighted fast-field echo (T1 FFE-Bh) without fat saturation sequences were acquired in the axial plane using the following parameters: TR/TE 214/46 ms, flip angle 80°, matrix 192.512 and slice thickness 5 mm; T2-weighted single-shot turbo spin echo (SSTSE) (with and without fat saturation) were acquired with TR/TE 417/80 ms, flip angle 80°, matrix 192×512 and slice thickness 5 mm. The examination was completed by the acquisition of a magnetic resonance cholangiopancreatography (MRCPRad-Bh) with a 40-mmthick single-slab half-Fourier sequence, with acquisition matrix 256, reconstruction matrix 256×512, field of view 25 cm, effective echo time (TEeff) 1,050 ms and repetition time (TR) 2,600 ms. Scan time was approximately 2.7 s per Bh for nine radial coronal oblique scans. The T1 FFE-Bh sequences were complemented with contrastenhanced acquisitions using a multiphase dynamic technique with fat saturation and intravenous administration of 0.1 mmol/kg body weight of paramagnetic contrast medium (gadopentetic acid, Magnevist, Schering). The multiphase dynamic technique comprised an arterial phase synchronised with the passage of the paramagnetic contrast bolus using real-time bolus tracking to calculate scan delay, followed by a pancreatic (40–45 s) and portal (70–80 s) phase. | |
| Data analysis | |
| Images were reviewed separately by two radiologists for each technique who evaluated US images as well as read CT and MR images, respectively. The readers evaluated the images independently and were blinded to patients’ clinical findings; in case of discordant imaging interpretation, a third reader for each technique was invited to provide an additional evaluation. | |
| In groups 1 and 2, a regional analysis of the concordance or discordance of the results obtained with the two imaging techniques was carried out by comparing the main regions of the pancreaticobiliary system: gallbladder and cystic duct, intrahepatic bile ducts, extrahepatic bile ducts (divided into common hepatic duct and common bile duct) and main pancreatic duct. The results of MRCP were compared with those of US in group 1 and with those of CT in group 2. The assessment criteria have been previously reported [31-33] and include: gallbladder (degree of distension, dimorphism, heterogeneity of content and/or presence of filling defects, wall thickness and wall irregularity) cystic duct (calibre, normal tortuosity of course and filling defects), intrahepatic bile ducts (increase in calibre >2 mm, heterogeneity of content and/or presence of filling defects), common hepatic duct (increase in calibre >5 mm, heterogeneity of content and/or presence of filling defects), common bile duct (increase in calibre >7 mm, heterogeneity of content and/or presence of filling defects and regular course), and main pancreatic duct (according to proposed criteria) [32-34]. In the interpretation of results, postoperative histological (n=27), bioptic (n=5), ERCP (n=28) findings and/or those of the clinical–imaging (3–6 months) follow-up (n=10) were considered standards of reference. Based on these data, patients were classified as being affected by benign lesions (n=47) (29 cases of lithiasis, 13 of inflammation, two of primary biliary cirrhosis, two cysts of the main pancreatic duct and one iatrogenic lesion) or malignant lesions (n=12) (four cholangiocarcinomas and eight pancreatic tumours); the remaining 11 patients had no imaging evidence of biliary and/or pancreatic abnormalities and were thus considered free of disease. Statistical significance of results of the regional analysis in the two groups was evaluated using the McNemar test, with significance set at p<0.05 [35]. In particular, we tested the hypothesis that there is significant discordance between results observed in the two patient groups, as confirmed by the reference standards. | |
| In Group 3, each reader used a score from 1 to 5 for identifying and characterising the pancreatic masses (1=definitely benign, 2=probably benign, 3=indeterminate, 4=probably malignant, 5=definitely malignant). The readers also evaluated the resectability of the malignant lesions (n=32), according to previously reported criteria [36], using a score from 1 to 3 (1=resectable, 2=indeterminate, 3=unresectable). Malignant masses were considered unresectable in all cases of direct or indirect evidence of peripancreatic vascular involvement. Vessels evaluated were the coeliac trunk, mesenteric vessels (vein and artery) and splenic-mesenteric-portal confluence. Direct signs of vascular invasion were the degree of circumferential involvement (>180°) or a “tear-drop” appearance of involved vessels. Indirect signs were dilated pancreaticoduodenal arcades and/or gastrocolic venous trunk. Involvement of the splenic vessels (artery and vein) was considered to be an absolute contraindication for surgical resection. Other unresectability criteria were the invasion of adjacent tissues or organs, the presence of distant metastasis and the presence of peritoneal carcinosis. Results of the CT and MR imaging studies were compared with biopsy (n=33) and/or histology (n=7), which were considered the reference standards against which the diagnostic accuracy of CT and MR imaging was measured. In particular, histological reports were available for all resectable malignancies, whereas biopsy reports were available for unresectable malignancies or patients with chronic pancreatitis. Fine-needle biopsies were obtained under CT guidance. For each of the two modalities, we calculated sensitivity, specificity, diagnostic accuracy and positive (PPV) and negative (NPV) predictive values for both identifying and evaluating resectability of the pancreatic masses. To assess statistical significance of any differences between CT and MR imaging, we used the McNemar test, with significance set at p<0.05 [35]. In particular, we tested whether there was a statistically significant difference between results provided by the two modalities, as confirmed by the reference standards. | |
Results |
|
| Group 1 | |
| In this evaluation, 212 biliary regions were analysed: gallbladder and cystic duct in 47 cases, intrahepatic bile ducts in 55, extrahepatic ducts in 55 and main pancreatic duct in 55. As eight patients were undergoing follow-up after cholecystectomy, the gallbladder region was not evaluated. The analysis showed concordant results in most cases and in most anatomical regions of the pancreaticobiliary system. In the evaluation of the gallbladder region, there was concordance between MRCP and US in 45 cases (96%), which included negative findings (n=17), lithiasis (n=20) and inflammation (n=8), as confirmed by postoperative histology (n=24) or followup (n=4). Similarly, in the evaluation of the intrahepatic bile ducts, there was concordance between MRCP and US in 52 cases (95%), which included negative findings (n=39), lithiasis (n=7), primary biliary cirrhosis (n=2) or inflammation (n=4), as confirmed by biopsy (n=2), ERCP (n=5) or follow-up (n=6). In the evaluation of the main pancreatic duct, there was concordance between MRCP and US in 52 cases (95%) and, in particular, in cystic lesions of the main pancreatic duct (n=2) and pancreatic tumour (n=4), as confirmed by ERCP (n=2) or postoperative histology (n=4); MRCP and US were also concordant in yielding negative findings in the remaining 46 cases. In the study of the extrahepatic bile ducts, there was concordance between MRCP and US in 46 cases (84%), which included negative findings (n=34), lithiasis (n=7) and inflammation (n=5), as demonstrated by ERCP. In nine cases (16%) the results were discordant; in particular, MRCP showed filling defects related to middle-proximal stones (n=3) and distal stones (n=6) in the extrahepatic bile ducts, as confirmed by ERCP; these lesions went undetected at US (Figure 1). Based on statistical analysis, the 16% discordance between MRCP and US in the evaluation of the extrahepatic bile ducts was statistically significant (p=0.003, McNemar test). | |
 |
Figure 1: Choledocholithiasis: discordant findings between ultrasound and MRCP for the main biliary tract. A: Ultrasound: the choledocho is dilated (maximum diameter 11mm) but no signs suggestive for stones are detected. B: CPRM confirms the diagnosis of choledocho ectasia and shows two round “filling defects” in the distal tract of the choledocho suggestive for stones. |
| Group 2 | |
| In this evaluation, 143 biliary regions were analysed: gallbladder and cystic duct in 32 cases, intrahepatic bile ducts in 37, extrahepatic bile ducts in 37 and main pancreatic duct in 37. Five patients were studied during follow-up after cholecystectomy, so the gallbladder region was not assessed. The analysis of the two techniques yielded concordant results in most cases in the gallbladder region and the main pancreatic duct. In the evaluation of the gallbladder region, there was concordance between MRCP and CT in 29 cases (91%), which included lithiasis (n=9), as confirmed by postoperative data, and negative findings (n=20). Similarly, in the evaluation of the main pancreatic duct, there was concordance between MRCP and CT in most cases (92%), both with (n=11) and without (n=23) disease. Among the cases with positive findings on both methods, ERCP confirmed the presence of cystic lesions of the main pancreatic duct (n=2), whereas postoperative histology confirmed the presence of a tumour of the head of the pancreas (n=8) or cholangiocarcinoma of the terminal common bile duct (n=1). In contrast, in the study of the intra-hepatic and extra-hepatic ducts, the two methods yielded a rate of discordance of 19% (n=7) and 16% (n=6), with positive findings on MRCP compared with CT in all cases. Positive findings on MRCP were the presence of simple ductal ectasia of the intrahepatic bile ducts (n=7) (Figure 2), common bile duct (n=1), common hepatic duct (n=3) and proximal common bile duct (n=1) due to bile stones, or a lack of signal at the level of the proximal common bile duct (n=1) due to a malignant stenosing lesion (cholangiocarcinoma), as confirmed by ERCP. Based on statistical analysis with the McNemar test, the 19% discordance rate in evaluating the intrahepatic bile ducts was statistically significant (p=0.008), as was the 16% discordance rate in evaluating the extrahepatic bile ducts (p=0.01). | |
|
Figure 2: Intra-hepatic biliary lithiasis: discordance of TC and CPRM findings. A: TC shows intra-hepatic biliary ducts ectasia, but no hyperdence focal lesions are detected. B: MRCP shows signs of intra-hepatic biliary lithiasis represented by diffuse duct ectasia and by the presence of focal “filling defects” (arrow) near the hilum; furthermore, a stone of the gallbladder is detected in the infundibolo. |
| Group 3 | |
| Results of the CT and MR imaging studies in patients with ductal adenocarcinoma or cystic tumours of the pancreas, along with biopsy (n=25) and/or histology (n=7) results, consisted of 31 cases of pancreatic malignancies, of which 24 were ductal adenocarcinomas, six were mucinous cystadenocarcinomas and one was an intraductal mucinous papillary tumour. In one case, the lesion was benign and consisted of a serous cystadenoma. Lesions involved the pancreatic head (n=19), body (n=7) and tail (n=6) and were from 1 to 8 cm in size [2.6 ± 1.2 standard deviation (SD)]. In the remaining eight patients, the pancreatic masses corresponded to tissue degeneration due to chronic pancreatitis, which was diffuse in the majority of patients (7/8 cases) and focal in one case only. With regard to identification and localisation of pancreatic masses, the overall results of CT and MR imaging demonstrated a comparable diagnostic accuracy of CT and MR imaging (98%). Only one patient with focal chronic pancreatitis was negative for neoplastic lesions on both CT and MR imaging. Figures 3 and 4 show examples of concordant findings of CT and MR imaging in a patient with a cystic lesion of the pancreatic tail. The results of CT and MR evaluation in terms of tumor resectability consisted of 25 patients with unresectable tumours, of which 13 had neoplastic invasion of both mesenteric vessels, associated with hepatic metastases in one case, nine had involvement of the superior mesenteric vein (Figure 5), again associated with hepatic metastases in one case, and one patient, with a tumour of the pancreatic body, showed invasion of the celiac trunk. Finally, in two patients, unresectability was due to the presence of hepatic metastases. Thus, the diagnostic accuracy of CT and MR imaging in the judgement of tumour resectability was similar, with values of 94% and 90%, respectively (p=not significant). There were two cases in which both CT and MR imaging provided a false negative result, and one case in which MR imaging only provided a false negative result. The minimal difference in diagnostic accuracy between the two modalities did not reach statistical significance. | |
|
Figure 3: Cystic tumor mass of the pancreas tail: concordance of unehanced TC and RM findings. A: TC shows a round lesion with low density in the pancreatic tail with median calcification. B: MR T2-weighted image confirms the presence of the cystic mass showing signal hyperintensity. |
|
Figure 4: Cystic tumor mass of the pancreas tail: concordance of enhanced TC and RM findings. A: TC shows no significant contrast enhancement in the cystic mass which showed high density only in the capsula. B: MR T1-weighted image confirms CT finding. |
|
Figure 5: A: Enhanced TC in portal phase shows a round cystic mass in the head of the pancreas with “tear-drop” appearance of superior mesenteric vein (SMV) that was considered a sign of vascular involvement. B: MR T2-weighted image confirms the presence of the cystic mass in pancreatic head with similar (“tear-drop”) appearance of SMV which shows low signal intensity. C: Enhanced MR T1-weighted image shows similar CT characteristics of the tumor mass. |
Discussion |
|
| The availability of several imaging techniques, such as US, CT, MRI and ERCP, for evaluating patients with pancreatico-biliary diseases requires selecting the most accurate and less invasive modality for each clinical problem. This is also a prerequisite for defining the most appropriate diagnostic protocol according to the clinical presentation of each patient [36]. In this study, we report the results of a direct comparison of MRCP, US and CT in three different groups of patients with pancreatico-biliary diseases. The direct comparison of MRCP, US and CT was conducted by analysing the various regions of the pancreaticobiliary system namely, gallbladder, intrahepatic bile ducts, extrahepatic bile ducts, main pancreatic duct and pancreas parenchyma with the aim of establishing the most appropriate and reliable diagnostic technique for each region based on the type of disease. Evaluation of results was divided into three groups: in Group 1 MRCP was compared with US, whereas in Groups 2 and 3 MRCP was compared with CT. | |
| In group 1, the comparison between MRCP and US demonstrated that both imaging methods are of value in evaluating the gallbladder region, intrahepatic bile ducts and main pancreatic duct in both benign and malignant diseases, as well as in cases without organic lesions of the pancreatobiliary system. By contrast, results of the evaluation of the extrahepatic bile ducts showed a statistically significant superiority of MRCP, especially in detecting stones in the middle-distal tract of the extrahepatic bile ducts. This result is reasonable, given that the middle or distal tract of the extrahepatic bile ducts is not well visualised on US owing to interference from intestinal gas or in obese subjects. The results observed in our study confirm previous reports of direct comparisons of MRCP and US in choledocholithiasis. Varghese et al. [11] reported high levels of sensitivity (91%) and diagnostic accuracy (97%) for MRCP compared with US (38% and 89%), and Ferrari et al. [13] reported a significantly higher rate of false negative results with US compared with MRCP in detecting stones in the distal common bile duct. To overcome these limitations of conventional US in the study of the distal common bile duct, endoscopic US and new three-dimensional US techniques have been used that provide a more detailed analysis of the biliary tree [14,29,37]. Endoscopic US shows a higher diagnostic sensitivity (100%) than MRCP (88%) [14]. Hence, MRCP is no doubt the technique of choice for evaluating the middle-distal common bile duct, a frequent site of gallstones. On the other hand, before conventional or laparoscopic cholecystectomy for gallstones, it is crucial to ensure that there are no migrating stones in the common bile duct, as these can cause colic and jaundice in the postoperative period. Additionally, all cases of stasis jaundice require a careful study of the extrahepatic bile ducts. In such cases, MRCP allows the degree of ductal dilatation, the location of the stone – visualised as a biliary filling defect, the degree of stenosis caused by the lesion and any residual bile flow to be determined. In the diagnostic evaluation of patients with biliary malignancies, although our study showed no statistically significant difference in the analysis of the various biliary regions, MRCP with conventional T1- and T2-weighted sequences and intravenous paramagnetic contrast material has a higher contrast resolution than US and allows a detailed study of biliary tree anatomy (site, degree, extent and cause of the stenosis, vascular and/or lymph node involvement) [2]. | |
| In group 2, direct comparison between MRCP and CT demonstrated the diagnostic value of both techniques in evaluating the gallbladder and main pancreatic duct, both in benign and malignant diseases and in the case of no organic lesions affecting the pancreaticobiliary ducts. Conversely, results of the evaluation of the extrahepatic bile ducts demonstrated a statistically significant superiority of MRCP over CT, particularly for detecting bile stones of the intra- or extrahepatic bile ducts. In one case only did the higher contrast resolution of MRCP allow correct identification of a malignant stenosis due to cholangiocarcinoma at the confluence of the main bile ducts into the common hepatic duct (Klatskin type IV tumour). Our results coincide with those reported by Zandrino et al. [12], who showed a higher diagnostic sensitivity of MRCP (100%) compared with multislice CT (82%) in choledocholithiasis. In their study, CT missed two cases of common bile duct stones as well as an obstruction of the common hepatic duct. Indeed, despite the improved spatial and temporal resolution brought about by volumetric multislice CT and multiplanar reconstructions, it is well known that CT cannot detect stones that do not have calcium density. In contrast, a study by Kondo et al. [14] reported similar sensitivity (88%) for MRCP and CT in detecting choledocholithiasis, with both techniques yielding the same number of false negatives due to stones smaller than 5 mm. Despite having similar accuracy to MRCP, CT cholangiography with biliary contrast agents should be considered only in patients who cannot undergo MRCP or in the case of diagnostic doubts on US and/or MRI [30] owing to the risk of hypersensitivity reactions. With regard to malignant lesions, the high contrast resolution of MRCP with the different T1- and T2-weighted sequences enabled correct identification of a Klatskin type IV tumour and accurate definition of its hilar and perihilar extension. The role of MRCP in patients with cholangiocarcinoma has previously been well documented by others [38-40]. | |
| In Group 3, malignancies accounted for 78% of cases in our series and included 24 ductal adenocarcinomas, six cystadenocarcinomas and one intraductal mucinous papillary neoplasm; overall, CT and MR imaging had similar diagnostic accuracy both for the correct identification and characterization of primary lesions and to establish potential surgical resectability in case of malignancies. Considering that the most common pancreatic malignancy is ductal adenocarcinoma, a fatal cancer that is potentially curable by surgical resection, an early diagnosis and accurate evaluation of resectability are fundamental. Despite availability of numerous diagnostic modalities, the majority of lesions are generally detected at an advanced stage, as the patient is initially asymptomatic and does not undergo diagnostic investigations. On the other hand, symptomatic patients often cannot be treated because the diagnostic investigations show locoregional vascular invasion or distant metastases. Ultrasonography represents the first-line investigation, as it is widely available, noninvasive and allows complete evaluation of the upper abdomen, which is useful for locoregional staging of neoplastic masses; however, it is operator dependent and has limitations in obese patients or those with a large amount of bowel gas [16,29]. Consequently, in patients with pancreatic masses, ultrasonography must necessarily be integrated by a second-line investigation such as CT or MR imaging [8- 10]. The overall concordance we observed between CT and MR imaging, as demonstrated by the lack of statistical significance of the small differences identified, is similar to that reported by previous studies. Park et al. [19] reported that Dynamic 3D-GRE MRI with MRCP shows superior tumor conspicuity and a similar diagnostic performance compared with MDCT in the evaluating the resecability of pancreatic cancer. Arslan et al. [20] found a similar performance of CT and MR imaging in evaluating vascular invasion. Concordant results between CT and MR imaging were reported by Fukukura et al. [21], Sahani et al. [23] and Yamada et al. [24] in patients with intraductal papillary mucinous neoplasms, and similar results were also reported for other types of pancreatic neoplasms, such as islet cell tumours of the pancreas [22]. In contrast, a number of studies in the literature have reported conflicting results. For example, Irie et al. [25] suggested that MR imaging rather than CT should be used as the examination of choice in patients with small (<2 cm) pancreatic adenocarcinomas. Similarly, Andersson et al. [26] found MR imaging to be more accurate than CT in differentiating benign and malignant lesions in patients with periampullary tumours. Erturk et al. [27] demonstrated the superiority of multislice CT with multiplanar reconstructions compared with MR imaging for both detecting and assessing locoregional extension of disease. Song et al. [28] showed that MR imaging has better diagnostic performance than CT for differentiating intraductal papillary mucinous neoplasms from other cystic lesions of the pancreas. | |
| In Group 3, in the evaluation of pancreatic tumor resectability, both CT and MR imaging produced false negative results in two malignant lesions of the pancreatic head (a ductal adenocarcinoma and a mucinous cystadenocarcinoma), as the involved vessels had preserved flow signal. In another patient, a false negative result was produced by MR imaging alone. Finally, in one patient with focal chronic pancreatitis, both CT and MR imaging were false positive for malignancy. However, it is well known that pancreatic enlargement and contour disruption typical of focal chronic pancreatitis can mimic malignant disease [41]. Despite these erroneous results of both CT and MR imaging, statistical analysis did not reveal statistically significant differences between modalities in identifying or evaluating respectability of malignant masses, so we may hypothesise that either technique can be used in these patients, depending on availability. However, because our study was limited by a small study sample affected by neoplastic lesions of different histological type, our preliminary results need to be validated by larger studies conducted on more homogeneous patient populations. | |
Conclusion |
|
| In conclusion, results of our study suggest that compared with US and CT, MRCP provides greater diagnostic accuracy in evaluating the intrahepatic and extrahepatic bile ducts because it identifies a larger number of strictures secondary to gallstones and their impact on biliary dynamics (pre-stenotic dilatation, post-stenotic bile flow), on the basis of our experience and the results reported in the literature, we believe that MR imaging of the pancreas performed with a state-of-art-equipment and a dedicated study protocol represents a valuable alternative to CT in the diagnostic assessment of patients with pancreatic masses; in particular, MRCP provides both accurate identification and characterization of lesions and an appropriate evaluation of resectability in case of malignant masses; the noninvasiveness, lack of ionising radiation, panoramic capabilities and high spatial and contrast resolution of MRCP in depicting the pancreatobiliary system are significant technical advantages that justify its routine clinical use as a supplement to US, with CT being restricted to patients with contraindications to MRCP or to settings where MRI is not available. | |
References |
|
|
|
