Validity of MDCT cholangiography in differentiating benign and malignant biliary obstruction

MDCT cholangiography can be utilized to identify variant biliary anatomy to guide preoperative planning of biliary surgery, and determine the cause and level of biliary obstruction. Early tumor detection and staging of biliary cancer are key factors for a possible cure by surgical resection. Between December 2019 and October 2020, 69 patients with clinically suspected biliary obstruction were enrolled in the study, subjected to clinical assessment (full history taking and clinical examination) and imaging assessment by MDCT cholangiography. Our findings were correlated to standard reference examinations including operative/ERCP/biopsy and histopathology findings. The most affected age was between 60 and 70 years old, and males were more affected. The commonest clinical presentation was yellowish discoloration of the skin and sclera followed by biliary colic. Right hypochondrial tenderness was the main clinical finding on clinical examination. Our patients were categorized according to the etiology of biliary obstruction into 7 groups: malignant stricture 52.2% (36 patients), calcular 24.6% (17 patients), iatrogenic 5.8% (4 patients), portahepatis lesions 5.8% (4 patients), benign stricture 4.3% (3 patients), inflammatory 4.3% (3 patients), and congenital 1.4% (1 patient). The malignant group shows dominant mass, moderate biliary obstruction, and arterial and venous enhancement. The overall sensitivity was 94% for malignancy. MDCT cholangiography is non-invasive, fast, and highly sensitive and specific in the diagnosis of different causes and levels of biliary obstruction and is useful in the characterization of the lesion in cases of malignant obstruction and differentiating it from benign stricture. It can be used as an effective alternative to ERCP or PTC.


Background
Obstructive jaundice is a common surgical problem that happens when there is a blockage to the passages of conjugated bilirubin from liver cells to the intestine. This can lead to life-threatening complications such as ascending cholangitis, malabsorption, and hepatorenal syndrome [1].
Causes of biliary obstruction can be divided into intrahepatic or extrahepatic causes and can be divided also according to its pathology into benign or malignant causes. Diseases of the biliary tract affect a large portion of the worldwide population, and the majority of cases are caused by cholelithiasis [2][3][4][5][6].
Ultrasound (US) is the initial imaging technique as a non-invasive cost-effective modality for evaluating the biliary obstruction but of lower sensitivity and specificity with limited value in the evaluation of the peripheral intrahepatic ductal lesions [7,8].
Magnetic resonance cholangiopancreatography (MRCP) is considered the most reliable non-invasive technique that produces high-contrast and high-resolution images of the biliary tree and allows the evaluation of the solid organs. However, contraindications including patients with cardiac pacemakers, cerebral aneurysm clips, or claustrophobia added to that high cost and not readily available [9][10][11].
Direct cholangiography (percutaneous cholangiography (PTC) or endoscopic retrograde cholangiopancreatography (ERCP)) is considered the gold standard technique. They enable direct visualization of the biliary tree and at the same time propose the therapeutic intervention. The drawbacks include invasiveness and even life-threatening complications [12,13].
Recently, in the last decade, multidetector computed tomography (MDCT) has led to the acquisition of true isotropic voxels that can be post-processed to yield images in any plane [14]. The combined use of multiplaner reformatting (MPR) and minimum intensity projection (MinIP) techniques significantly improves the visualization of the biliary ducts and their site of confluence compared with those obtained by axial CT. Moreover, MinIP technique enables us to depict the small biliary duct and the pancreatic duct more clearly [15][16][17].
MDCT cholangiography has a vital role as a postprocessing technique that can be utilized to identify variant biliary anatomy to guide preoperative planning of biliary-related surgery [18] and determine the level and the cause of biliary obstruction; early tumor detection and staging of the biliary cancer are the key factors for a possible cure by surgical resection [19,20].
MDCT cholangiography, as a non-invasive modality, committed to the biliary tract denotes a good substitute to MRCP. CT systems are particularly useful when magnetic resonance imaging (MRI) is contraindicated or unapproachable or when the value of MRCP images is suboptimal. The benefit of MDCT due to thinner (1 mm) sections, quick scanning, and the parallel escalation in the abilities of the workstations has permitted the development from section-based to volume-based systems [21][22][23][24]. MDCT cholangiography with VR images has some advantages over formerly mentioned imaging tools. The pre-contrast series is useful in those with an elevated bilirubin level or abnormal liver or renal functions. By using CPR/MPR, all intra-hepatic ducts can be viewed, a feature that is not possible at ERCP or PTC. Contrast series data set is used as well to reproduce angiography images to delineate tumoral vascular invasion [25].
The aim of this work was to evaluate the validity of MDCT cholangiography in differentiating benign and malignant biliary obstruction.

Study design and population
This was a prospective diagnostic study in concordance with the STARD guidelines conducted during the period from December 2019 to October 2020 in the Radiodiagnosis Department, at our university hospitals. The data were collected from 69 patients (39 males and 30 females) with the majority of cases belonged to the 5th to 6th decade age which clinically presented with obstructive jaundice. Written consent was obtained from all patients included in our study (except for a young patient, the consent was obtained from her parents). The consent and exam protocols were approved by the international review board (IRB), faculty of medicine with reference number ZU-IRB#4539. More than one clinical presentation could be found in the same patient; the clinical presentations were jaundice (63 patients), biliary colic (52 patients), dark urine (45 patients), pale stool (32 patients), pruritus (24 patients), fatty dyspepsia (23 patients), vomiting (22 patients), fever (16 patients), cachexia (12 patients), and abdominal swelling (10 patients).

Patient inclusion criteria
The following are the inclusion criteria: 1. All cases with clinically suspected biliary obstruction and abdominal US revealing any degree of biliary obstruction 2. Elevated serum levels of indirect bilirubin or alkaline phosphatase enzyme

Patient exclusion criteria
The following are the exclusion criteria: 1. Case contraindications for contrast-enhanced computed tomography scanning, for example, thyroid disease, pheochromocytoma, and contrast hypersensitivity 2. Patients with raised renal functions not on dialysis 3. Pregnant females (mainly in the first trimester) 4. Morbid obesity (>150 kg), due to the couch weight limitation, besides those patients who get dyspenic when lying in a supine position, an obstacle that hinders the image quality

Scan protocol and parameters
All cases were scanned by 128™ detector rows scanner (Ingenuity core, v3.5.7.25001, Philips Healthcare Systems, The Netherlands). Scanning parameters were 350 mAs, 120 kVp, reconstruction interval of 5 mm, section collimation of 2.5-5 mm, and table speed of 7.5-10 mm per rotation within one breath-hold attainment of 8-10 s. Contrast administration is by automatic power injector (model mark V: medrad, Indianola, PA).

Patient preparation
Fasting for 6 h prior to scanning and vigorous oral hydration 2 h earlier to the examination are required. Water-soluble oral contrast was used only in those patients with suspected portahepatis lesions. Tencentimeter contrast diluted in 1 L of water should be drinked by the patient during a period of 1 h prior to the exam with the last 250 mL drinked while sitting on the couch. An intravenous catheter was introduced at the arm vein for intravenous contrast administration. Creatinine clearance must be >30 mL/min, GFR >45 mL/min/1.73 m 2 , or serum creatinine level <1.5 mg/dL.

Patient position
The patient lies supine, head first; the scanning begins from the lung bases down to the pelvic inlet during suspended complete expiration or complete inspiration.

CT scan protocol
Non-contrast phase It is used to make a baseline for deciding whether an identifiable lesion enhances and helps in identifying biliary stones or pancreatic calcification which may be concealed by contrast material.
Contrast-enhanced phases It comprises arterial, portovenous, and late phases. Administration of a non-ionic contrast agent (300 mg of iodine per milliliter) was done intravenously (2 mL/kg), through the intravenous catheter, at a rate of 3-5 mL/s for 30 s via a programmed powerful automatic injector, at that time, the helical acquisition began (8 s) after a threshold level of 140 HU was extended in the abdominal aorta then the arterial phase attained (20-30 s), after injection of contrast media. The porto-venous phase developed (50-60 s) after the injection of contrast media; finally, the delayed phase assimilated (5-10 min) after the injection of the contrast media.
Post-procedure care The following are the post-procedure care: excess oral hydration and follow-up for 30 min for any reactions or side effects after contrast injection, such as inflammation, rash, itching, or trouble in breathing.

Image analysis
Post-processing technique The 3D reestablishment with thin planar slicing (1 mm) and MPR was achieved in coronal and sagittal planes for well description of the biliary tract wall and intraluminal lesions. Such technical development permits thinner slices to be gained in shorter scan periods, with good patient agreement and lower motion artifacts.

Standard reference examinations
MDCT cholangiography findings were correlated with operative and endoscopic interference (17 calcular group patients), ERCP (12 patients of the remaining benign nature lesions and the negative case), PET/CT for portahepatis lesions (4 patients), and biopsy and histopathology in malignant masses (36 patients).

Statistical analysis
The data analysis was performed using SPSS 20 (Chicago SPSS, SPSS Inc., Chicago, IL). Qualitative data were expressed as number and percent. Quantitative data were displayed as mean and SD. Through the comparison of MDCT diagnosis with the last diagnosis acquired from standard reference inspections, the diagnostic precision of MDCT cholangiography in the assessment of the reasons for biliary obstruction will be accessed in terms of accuracy, specificity, sensitivity, PPV, and NPV.

Results
Our cross-sectional prospective study included 69 patients, 39 males and 30 females, with their ages ranging from 13:82 years; the mean age for males is 55.61 years ± 13.68 SD and for female is 54.23 years ± 18.01 SD, making the overall mean age of the total patients 55.01 years ± 15.6 SD. The degree of biliary radical dilatation was assessed as follows: mild-localized to the central radicals; moderate-extending to the first-and second-order branches; and severe-involving the most peripheral branches as well. Biliary dilatation of mild degree was the most observed finding, and distal obstruction was the most observed level.
Regarding the nature of the cause of biliary obstruction, malignant causes (36 cases 52.2%) were more predominant than benign causes.
Our patients were categorized according to the cause of the biliary obstruction into malignant lesions in 36 patients (cholangiocarcinoma in 14 patients (20.3%), pancreatic carcinoma in 14 patients ( Fig. 1) (20.3%), periampullary carcinoma in 3 patients, hepatic focal lesions in 3 patients, gall bladder carcinoma in 1 patient, and gastric carcinoma in 1 patient), calcular biliary obstruction ( Fig. 2) in 17 patients 24.6% (due to the presence of common bile duct stone), iatrogenic lesions in 4 patients 5.8% (due to the presence of fluid collection and aerobilia in patient with a history of recent biliary surgery or intervention), portahepatis lesions in 4 patients 5.8% (due to the presence of enlarged portahepatis LNs; 2 cases proved to be lymphoma and two cases metastatic portahepatis LNs), benign stricture in 3 patients 4.3% (two cases due to recurrent cholangitis and one case attributed to sludge), inflammatory lesions in 3 patients 4.3% (due to the presence of signs of inflammation: enhancement in arterial phase, haziness or stranding of surrounding the fat planes, thickening of the common bile duct wall), congenital lesion in one patient 1.4% (presented as a cystic lesion that communicates to the bile duct and is discrete from the GB; classified according to modified Todani classified as type IA), (Fig. 3) and negative in 1 patient 1.4%; from our results, the most frequent group was malignant stricture, and the least frequent group was congenital (Table 1). Table 1 shows the differences between the MDCT cholangiography and the gold standard method regarding the malignant stricture findings that can be summarized as follows: cholangiocarcinoma in 14 (20.3%) cases by MDCT (Fig. 4) compared to 15 cases (21.7%) by gold standard, pancreatic carcinoma in 14 cases in both (20.3%), periampullary carcinoma in 3 (4.3%) in both, hepatic focal lesions in 3 cases (4.3%) by MDCT compared to 2 cases (2.9%) by gold standard cases, and GB and gastric carcinoma (Fig. 5) in one case at each of them (2.9%). Other causes of biliary obstruction show no differences between both methods except for portahepatis lesions (Fig. 6), with fewer cases detected by the latter method (3 cases 4.3%); benign stricture was detected more at the latter method in 5 cases (7.2%), as well as no negative cases detected at the latter method.
Comparing benign and malignant causes regarding the demographic data, our study showed higher male affection in the malignant group by 64% while females showed higher affection in the benign group by 53%; the most affected age group was 60:<70 years in both the malignant and benign groups 17 (43.6%) and 8 (26.7%), respectively. The mean age for the affection in the malignant group is 58.4 years compared to 50.9 years in the benign group. Both sex and age differences between the malignant and benign groups showed non-significant difference statistically through calculating their p values using the chi-square test ( Table 2). Our study showed a statistically significant difference only for the prevalence of cachexia (considering the clinical presentation) towards the malignant group and fever towards the benign group through MDCT cholangiography findings at the benign and malignant groups showed high statistically significant difference regarding the presence of mass (towards the malignant group at 37 cases 90.3%) and also regarding the presence of calcification (towards the benign group at 6 cases 21.4%) while the findings showed a statistically non-significant difference for the degree of biliary dilatation and level of obstruction by calculating their p values using the chi-square test (Table 3); mild degree of obstruction was dominant at either category, at 20 cases (71.4%) and at 20 cases (48.8%) for the benign and malignant groups, respectively.  showing heterogeneous pattern as the dominant pat-tern}) for the latter phase while it showed a statistically non-significant difference for the delayed phase enhancement by calculating their p values using the chi-square test (Table 4), which was absent at 40 cases (97%) for the malignant group while at 27 cases (96.4%) for the benign group.
There was a very good agreement between the standard method for the detection of the site (Table 5) and cause (Table 6) of biliary obstruction and MDCT by getting a p value of kappa equal to 1 (the highest possible score for agreement) and 0.87, respectively, using the Stuart-Maxwell test and high significance in the detection of the site of obstruction Table 7 shows very good agreement between MDCT and standard method regarding the detection of malignant stricture type by getting a p value of kappa 0.91 using the Stuart-Maxwell test and high significance in the detection of the type of malignant lesion.
The results of this study presented high diagnostic performance for the MDCT cholangiography regarding the level and cause of obstruction that was highest in malignant (100% sensitivity), iatrogenic (Fig. 7), inflammatory, and congenital causes while it yields the lowest diagnostic performance in the detection of benign stricture 40% sensitivity (Table 8).

Discussion
Imaging evaluation of the biliary tree includes ultrasound, CT, CT cholangiography with a biliary contrast agent, MRCP, ERCP, and PTC. Both ultrasound and MRCP do not imply an ionizing radiation. Both do not utilize contrast agents. Sonography is not superior to evaluate the main biliary tract, with a reported sensitivity ranging from 20 to 80%, at biliary stone detection [26].
Though MRCP is a reliable non-invasive technique, its use is hindered in those with pacemakers or aneurysm clips and those who are claustrophobic [27].
ERCP has the privilege of providing diagnostic and therapeutic intervention in the same session (endoscopic sphincterotomy, stone extraction, and endoscopic guided biopsy). However, it yields a little information about solid abdominal organs, is invasive as well, and poses a 0.5-5.0% complication rate (diverse reaction to sedatives, cardio-respiratory dysfunction, pancreatitis, perforation of the gut, bleeding, cholangitis, sepsis, and death) [28].
MDCT pre-contrast series is useful in those with an elevated bilirubin level or abnormal liver or renal functions. By using CPR/MPR, all intra-hepatic ducts can be viewed, a feature that is not possible at ERCP or PTC. Contrast series data set is used as well to reproduce angiography images to delineate tumoral vascular invasion [25]. MDCT cholangiography assesses globally the biliary obstruction regarding the biliary tree, vessels, and solid abdominal organs (liver, pancreas, and duodenum), a feature that does not exist at ERCP and PTC [29,30].
In our series, all our patients were complaining of yellowish discoloration of the skin and sclera as the most common clinical complaint presented in all patients except for 6 cases. Our result agrees with Mathew et al. [31].
In this study, MDCT identified the level of biliary obstruction in 69 patients with 100% accuracy which agreed with Mohamed et al. [16].
Our study showed a higher prevalence for male affection to biliary obstruction than for females by 56.5 to 43.5%. Rishi et al. [20] showed different prevalence which is equal to male to female affection.
Regarding the most affected age group, our study showed that 60-70 years is the most affected group, while Rishi et al. [20] stated that the majority of cases are in the age group 41-60 years.
Moderate degree of biliary duct dilatation initiated by malignant sources was more severe (41.5%) than that of  [16]. In our study, it was found that calcular obstructive jaundice was the main cause of benign OJ (24.6%); these were as those of Mathew et al. [31] who showed calcular cause was responsible for 22% of cases of biliary obstruction.
Regarding the nature of the cause of biliary obstruction, our study showed malignant dominance by 59.4% compared to 40.6% for benign which agrees with Narayanaswamy et al. [21] which showed 66.7% for malignancy while 33.3 for benign causes. But Mathew et al. [31] stated benign causes were 56% while malignant was only 44%.
Malignant stricture was identified by MDCT in 36 cases; 34 cases were confirmed by standard examination; a case had a slight left-sided intrahepatic dilatation with hyperenhancement of the duct wall and abrupt duct narrowing, identified as cholangiocarcinoma. ERCP identification was acute cholangitis. The second case was a well-known case of colon cancer on follow-up, a hepatic focal lesion identified as necrotic metastasis but cytology confirmed an abscess. Alternatively, a case of metastatic portahepatis LN was confirmed to be cholangiocarcinoma. The SN, SP, PPV, NPV, and ACC of MDCT cholangiography in the detection of malignant stricture were 94.4%, 93.94, 94.44%, 93.93%, and 94.2%, respectively, that come to an agreement with Mohamed et al. [16] which the results were 96.8%, 94.9, 98.2%, 96.7%, and 95.6%.
In our series, pancreatic carcinoma is present in 14 patients responsible for 20.3% of the total cases; in the study of Mohamed et al. [16], pancreatic carcinoma was responsible for 17.8% of the causes of biliary obstruction. Mathew et al. [31] claimed that it affected 20% of cases. MDCT identified all the cases of pancreatic carcinoma with 100% accuracy. Most studies agree with the basis of diagnosing as ours: a hypodense mass relative to enhanced pancreatic tissue that may show faint peripheral enhancement.
In other malignancies, there were 5 cases, 3 cases diagnosed as hepatic focal lesion, 1 case as GB carcinoma and gastric carcinoma, and one case was wrongly diagnosed as metastatic focal lesion but proven as an abscess. Mohamed et al.'s [16] results included 8 cases of intrahepatic malignant masses, one case was diagnosed as benign stricture but pathologically proven to be metastasis from non-Hodgkin lymphoma. Tummala et al. [32] reported extrinsic compression by focal lesion adjacent to the bile duct in 3.2% of their patients (compared to 4.3% in our study).
Periampullary carcinoma was found in 3 cases, all were proven pathologically to be periampullary carcinoma; it represented 4.3% of the cases. Mohamed et al. [16] showed a higher affection rate (14%) with 100% accuracy that agreed with our results while Narayanaswamy et al. [21] was 96%.
Calcular etiology was diagnosed in 17 cases, 1 case was diagnosed by MDCT as calcular but proven to be a benign stricture (sludge) and another case which was diagnosed by MDCT as negative (no detected cause) proven to be a black cholesterol stone that was not visible. It was the most common cause of benign biliary obstruction in our study by contributing to 24.6% of the cases with SN 94.1%, SP 98%, PPV 94.1%, NPV 98% and, ACC 97.1%. Our results agree with Rishi et al. [20] that revealed in their study ACC of MDCT in the detection of choledocholithiasis is 98%, SN of 100%, and SP of 97.4% showed also was the most frequently diagnosed cause of obstruction by 24%. You et al. [9] mentioned intraductal high attenuating focal lesions were detected in enhanced or unenhanced scans subsequent to the course of CBD as criteria for diagnosis which agreed with our findings. Four patients with post-operative iatrogenic bile duct injury were included which represented 5.8% of the total causes. Heller et al. [4] imply that iatrogenic is the second most common reason for benign biliary tract strictures after calcular which agreed with ours. The study showed free fluid in most cases which was noted in the abdomen accompanying fluid at the GB bed. MDCT diagnosed all the patients correctly based on the history and findings making statistical measures for the diagnostic performance of MDCT in detecting bile duct injury as follows: SN 100%, SP 100%, PPV 100%, NPV 100%, and ACC 100%. El-gerby et al. [33] and Meng et al.'s [34] results showed the same SN, SP, PPV, NPV, and ACC as ours for biliary leakage detection.   Four patients had portahepatis lesions in our study. It contributed by 5.8% of the causes which showed different results from Mohamed et al. [16] which was 4.4% only. Two cases were diagnosed as lymphoma involving portahepatis LN and other two cases as malignant metastatic portahepatis LN. Rishi et al. [20] stated that lymphoma represented 6% of the causes of obstruction compared to 2.9% in our study and 100% ACC in detecting lymphoma which was the same finding as Mathew et al. [31]. Our diagnostic performance was as follows: SN 100%, SP 98.48%, PPV 75%, NPV 100%, and ACC 98.55%.
Benign stricture was diagnosed by MDCT in three cases; two cases were proven to be due to recurrent cholangitis which agrees with Heller et al. [4] who stated cholangitis is considered one of the commonest causes of benign strictures. Mohamed et al. [16] mentioned the MDCT criterion for identifying a benign stricture which includes a smooth and gradually elongated narrowing of CBD in a short section measuring less than 1 cm without the presence of a mass. Our diagnostic performance for the detection of the benign strictures was as follows: SN 40%, SP 98.44%, PPV 66.67%, NPV 95.45%, and ACC 94.20%. Mohamed et al.'s [16] statistical results were SN 66.7%, SP 80.2%, PPV 80.2%, NPV 91.7%, and ACC 92.6% which were near to our results except for sensitivity as one case was diagnosed by MDCT as benign stricture proven by biopsy as a small cholangiocarcinoma, while Mathew et al. [31] showed higher percentages as follows: SN 100%, SP 97.8%, PPV 83.3%, NPV 100%, and ACC 98%.
Inflammatory was diagnosed by MDCT in three cases, two cases were due to acute pancreatitis and one case was due to acute cholecystitis. Patel et al. [35] stated that  One of the limitations of this work includes the small number of cases, particularly in the benign stricture group. The main limitations of the use of intravenous agents are the relatively high rate of allergic reactions and the risk of renal and/or hepatic toxicity so patients with biliary obstruction and high renal functions could not be assessed and risk of radiation and anesthesia for young patients.