Preoperative Assessment of Resectability of Hepatic Hilar ...

Note: This copy is for your personal non-commercial use only. To order presentation-ready copies for distribution to your colleagues or clients, contact us at www.rsna.org/rsnarights.

Purpose:

To retrospectively assess the accuracy of combined multiphasic computed tomography (CT) and direct cholangiography for evaluation of the resectability of hilar cholangiocarcinoma, on the basis of revised criteria for unresectability, by using surgery as the reference standard.

Materials and Methods:

Institutional review board approval was obtained, and informed consent was waived. From 1998 to 2003, 55 patients (37 men, 18 women; mean age ⫾ standard deviation, 59 years ⫾ 12) with surgically proved hilar cholangiocarcinomas who underwent preoperative CT (single– detector row CT, n ⫽ 26; multi– detector row CT, n ⫽ 29) and cholangiography were included for study. The authors’ revised criteria for unresectable tumor were contralateral hepatic artery invasion; main or contralateral portal vein invasion longer than 2 cm; biliary extension to the contralateral secondary confluence, farther than 2 cm from hepatic hilum; enlarged lymph nodes at the celiac, portacaval, and paraaortic area; and other ancillary findings. Tumor resectability based on these parameters was determined at imaging by two radiologists in consensus. MannWhitney U test and weighted ␬ coefficient of agreement were used for accuracy determination.

Results:

For depiction of portal vein invasion (in 26 patients), CT yielded an accuracy of 85.5%. Arterial invasion was found at surgery in 19 patients, with CT providing an accuracy of 92.7%. For prediction of node involvement (15 patients, 27%), CT yielded an accuracy of 83.6%. The extent of ductal involvement could be accurately predicted in 46 patients (84%) (weighted ␬ ⫽ 0.767). In 30 of 42 patients with disease classified as resectable according to revised criteria, disease was found to be resectable at surgery (71.4% positive predictive value). In 11 of 13 patients with disease classified as unresectable according to revised criteria, unresectable disease was confirmed (84.6% negative predictive value). Overall accuracy of resectability was 74.5%.

Conclusion:

Combined interpretation of CT and direct cholangiographic images by using our revised criteria resulted in overall accuracy of 74.5% for prediction of resectability for hilar cholangiocarcinoma.

1

From the Department of Radiology and Institute of Radiation Medicine, Seoul National University College of Medicine, Clinical Research Institute, Seoul National University Hospital, 28 Yongon-dong, Chongno-gu, Seoul, Korea (H.Y.L., S.H.K., J.M.L., J.K.H., B.I.C.); and Department of Surgery, Seoul National University, College of Medicine, Seoul, Korea (S.W.K., J.Y.J.). Received March 12, 2005; revision requested May 10; revision received June 7; accepted June 21; final version accepted June 21. Address correspondence to J.K.H. (e-mail: hanjk @radcom.snu.ac.kr).

娀 RSNA, 2006

姝 RSNA, 2006 Radiology: Volume 239: Number 1—April 2006

113

䡲 GASTROINTESTINAL IMAGING

Ho Yun Lee, MD Se Hyung Kim, MD Jeong Min Lee, MD Sun-Whe Kim, MD Jin-Young Jang, MD Joon Koo Han, MD Byung Ihn Choi, MD

ORIGINAL RESEARCH

Preoperative Assessment of Resectability of Hepatic Hilar Cholangiocarcinoma: Combined CT and Cholangiography with Revised Criteria1

GASTROINTESTINAL IMAGING: Assessment of Hepatic Hilar Cholangiocarcinoma

O

n the basis of the classification scheme for cholangiocarcinoma proposed by the Japanese Liver Cancer Group (1), cholangiocarcinomas are classified into three types according to the macroscopic appearance of tumor: mass forming, intraductal growing, and periductal infiltrating; the last type is most prevalent in the hilar portion of the biliary tree. In the case of an infiltrative growth pattern along the biliary tree and anatomic proximity of vessels to the bile duct at the hepatic hilum, thorough preoperative evaluation of the extent of tumor with regard to these structures is necessary. With improved imaging techniques and knowledge on this type of tumor, surgical approaches for the resection of hilar cholangiocarcinoma have evolved. Investigators have demonstrated that complete resection inclusive of partial hepatectomy and regional lymphadenectomy may be the surgical treatment of choice for select patients (2–4). With recent surgical techniques and devices, the results of surgery for hilar cholangiocarcinoma have improved during the past few years (2,5–9). Since there is great variability in bile duct anatomy— that is, the secondary confluence of the bile duct is only a few millimeters apart from the hepatic hilum in some patients, while it is several centimeters apart from the hepatic hilum in others— not all Bismuth type IV cholangiocarcinomas (Fig 1) (7) are surgically unresectable. Simultaneous extended hepatectomy, removal of extrahepatic bile ducts with or without resection, and anastomosis of the portal vein are considered potentially curative even for Bismuth type IV hilar cholangiocarcinoma (8,10–13).

Advances in Knowledge 䡲 Combined CT and cholangiographic findings can be used in the preoperative assessment of hilar cholangiocarcinoma. 䡲 Our revised Bismuth classification criteria of unresectable tumor include vascular invasion, lymph node metastasis, and biliary extension. 114

The purpose of this study was to retrospectively assess the accuracy of combined multiphasic computed tomography (CT) and direct cholangiography in the evaluation of resectability of hilar cholangiocarcinoma, on the basis of revised criteria for unresectability, by using surgery as the standard of reference.

Materials and Methods Patient Selection Our institutional review board approved this retrospective study and waived the requirement for informed consent. From January 1998 to December 2003, we identified patients with a diagnosis of hilar cholangiocarcinoma who underwent surgical treatment at Seoul National University Hospital; patients were identified from a database maintained by the hepatobiliary section of the Department of Surgery. The inclusion criteria for this study were as follows: patients with hilar cholangiocarcinoma who underwent surgical exploration at our hospital, complete preoperative imaging that included multiphasic CT and cholangiography, and a diagnosis of hilar cholangiocarcinoma at pathologic examination of a surgically resected tumor. Of the 63 patients initially selected, eight were excluded from analysis for one of the following reasons: (a) an interval of more than 6 weeks between CT imaging and surgery (n ⫽ 3), (b) an incomplete pathologic examination (ie, incomplete lymph node or vascular assessment at pathologic evaluation) (n ⫽ 3), or (c) unusual pathologic results of papilloma or high-grade papillary adenoma (n ⫽ 2). The remaining 55 patients comprised our study population. The mean age of the patients was 59 years ⫾ 12 (standard deviation), and the age range was 29 –76 years. There were 37 men and 18 women. No patients had known risk factors such as a history of Clonorchis sinensis infestation, intrahepatic stone diseases, or primary sclerosing cholangitis (14). All CT scans were obtained within a 6-week period before surgery (mean interval, 16 days ⫾ 1.8).

Lee et al

All cholangiograms were also obtained within a 6-week period before surgery (mean interval, 16 days ⫾ 2).

CT Technique All patients underwent triple-phase CT, which consisted of precontrast, arterial dominant, and portal dominant phases. Single– detector row CT (Somatom Plus; Siemens, Erlangen, Germany) was performed in 26 patients, and multi– detector row CT (LightSpeed, GE Medical Systems, Milwaukee, Wis, or MX8000, Marconi Medical Systems, Cleveland, Ohio) was performed in 29 patients. For precontrast imaging, 5-mm-thick sections were acquired in 43 patients, and 8-mm-thick sections were acquired in the remaining 12 patients. After 120 mL of nonionic contrast material (iopromide, Ultravist 370; Schering, Berlin, Germany) was administered at a rate of 2.5 mL/sec with a power injector (Stellant Dual; Medrad, Indianola, Pa), arterial and venous phase helical CT scans were obtained. The scanning parameters for single– detector row CT were 5-mm section thickness, pitch of 1.5, reconstruction intervals of 3 mm (arterial phase) and 5 mm (portal venous phase), 150 mAs, 120 kVp, and a matrix of 512 ⫻ 512. Scanning parameters for multi– detector row CT included a gantry rotation time of 0.5– 0.8 second with a 4 ⫻ 1.25 or 8 ⫻ 1.25-mm detector configuration, 2.5-mm section thickness, pitch of 1.0 –1.5, 3-mm reconstruction interval for both phases, 150 mAs, 120 kVp, and a matrix of 512 ⫻ 512. For arterial phase scanning, a delay of 11 seconds

Published online before print 10.1148/radiol.2383050419 Radiology 2006; 239:113–121 Author contributions: Guarantors of integrity of entire study, H.Y.L., J.K.H.; study concepts/study design or data acquisition or data analysis/interpretation, all authors; manuscript drafting or manuscript revision for important intellectual content, all authors; manuscript final version approval, all authors; literature research, H.Y.L.; clinical studies, all authors; statistical analysis, H.Y.L., S.H.K.; and manuscript editing, all authors Authors stated no financial relationship to disclose. Radiology: Volume 239: Number 1—April 2006


(for single– detector row CT) or 13 seconds (for multi– detector row CT) was used after the maximal attenuation of the aorta reached 100 HU with bolus tracking. After completion of arterial phase scanning, a delay of 15 seconds (for single– detector row CT) or 30 seconds (for multi– detector row CT) was used for portal venous phase scanning.

Cholangiography Technique Percutaneous transhepatic cholangiography and biliary drainage procedures were performed by two experienced radiologists (J.K.H., J.M.L.) in an angiography suite and were guided by means of both ultrasonography and fluoroscopy. To induce sedation and analgesia, midazolam hydrochloride (Dormicum; Bukwang Pharmaceuticals, Seoul, Korea) and fentanyl citrate (Sublimaze; Keukdong Pharmaceuticals, Seoul, Korea) or meperidine hydrochloride (Demerol; Keukdong Pharmaceuticals) were used when necessary. A 21-gauge Chiba needle (Cook, Bloomington, Ind) was used for percutaneous cholangiography. After obtaining the cholangiogram, an 0.018-inch guidewire (Cook) was inserted into the duct through the Chiba needle, and then, by using the standard guidewire technique, a 5-F yellow sheath and a 0.035-inch guidewire (Terumo, Tokyo, Japan) were placed. Finally, an 8.5-F biliary drainage catheter (Cook) was inserted into the bile duct. Image Analysis The multiphasic CT images in 55 patients were retrospectively reviewed in consensus by two board-certified radiologists (J.K.H., S.H.K.) who were blinded to the final surgical findings; these authors have served mainly as gastrointestinal radiologists for 10 (J.K.H.) and 20 (S.H.K.) years and have interpreted abdominal CT images of the liver as part of their daily clinical and research practice. The radiologists reviewed the images by using the cine mode on the picture archiving and communication system viewer. Helical CT images were viewed in the transverse plane and were assessed for local parenchymal extension of tumor, arterial involvement, venous involvement, metastatic lymphadenopathy, peritoneal spread, and liver parenchymal changes, such as lobar atrophy. Radiology: Volume 239: Number 1—April 2006

Diagnostic criteria for vascular involvement on the transverse images included vessel occlusion, stenosis or contour deformity associated with tumor contact, or greater than 50% perimeter contact with the tumor (15–17). Nodal status was evaluated by using the TNM classification system of the International Union Against Cancer (18). The regional lymph nodes were categorized into two groups, N1 and N2. N1 lymph nodes were defined as the hilar, cystic duct, and pericholedochal nodes. N2 lymph nodes were defined as the periportal, periduodenal, peripancreatic, celiac, and superior mesenteric nodes. Regional lymph nodes were considered positive if they were larger than 10 mm in short-axis diameter; if they had central necrosis, even if they were smaller than 10 mm; or if they were hyperattenuating compared with liver parenchyma in the portal phase (19). Biliary involvement was evaluated with CT and direct cholangiography at the same time by the consensus of the same two radiologists (J.K.H., S.H.K.); cholangiography included percutaneous transhepatic cholangiography and percutaneous transhepatic biliary drainage tube cholangiography. If there was a discrepancy, results at direct cholangiography had priority to those at CT. On CT images, the bile duct was considered to be involved if the ductal wall was irregularly thickened, generally with asymmetric upstream dilatation of the intrahepatic ducts; if it was hyperattenuating relative to the liver; or if it had an intraductal soft-tissue mass or ductal wall obliterating the lumen (20). On direct cholangiograms, diagnostic criteria of biliary duct invasion included irregularity of the bile duct obliterating the lumen, with or without asymmetric upstream dilatation of the intrahepatic duct (21).

Tumor Resectability: Revised Criteria The observers determined tumor resectability on the basis of the analysis described in the previous section. We modified the previous criteria for unresectability on the basis of local tumorrelated factors and metastasis (2,22) by the consensus of surgeons in our hospital (Table 1). Our revised criteria for

Lee et al

resectability include contralateral hepatic artery invasion, segmental main or contralateral portal vein invasion longer than 2 cm, biliary extension to the contralateral secondary confluence farther than 2 cm from the hepatic hilum, enlarged lymph nodes at the right side of the celiac axis and portacaval area, and other ancillary findings, such as peritoneal seeding and liver parenchymal changes of lobar atrophy.

Surgery In all patients, a standardized surgical approach was used. Surgery was performed by two board-certified surgeons who have served mainly as hepatobiliary specialists for 10 (S.W.K.) and 20 (J.Y.J.) years. Exposure of the biliary confluence and assessment for vascular involvement were accomplished, and histologic assessment of the resection margin was performed. Full exploration was followed by curative or palliative surgical treatment (ie, extended hepatectomy, including excision of involved vessels, and extended lymph node dissection or palliative hepaticojejunostomy). Postoperative Follow-up Incidence of tumor recurrence after the curative surgery, mode of recurrence, mortality rate, and median disease-free survival period were followed up by using the total clinical information of radiologic, pathologic, and laboratory Figure 1

Figure 1: Schematic drawing of Bismuth classification of biliary obstruction: type I, confined to common hepatic duct; type II, confined to bifurcation of common hepatic duct; type III, affecting bifurcation of common hepatic duct with spread of tumor to right (IIIa) or left (IIIb) duct; and type IV, affecting bifurcation of common hepatic duct with spread of tumor to both hepatic ducts. 115


findings. In addition, the median survival period after surgery in patients who underwent palliative surgery was determined.

Statistical Analysis Preoperative assessment of vascular, biliary, or nodal invasion and overall resectability were retrospectively compared with findings at surgery and the final pathologic results (by H.Y.L.). The sensitivity, specificity, positive predictive value, negative predictive value, and accuracy were calculated. Vascular assessments performed by using single– and mulit– detector row CT were also analyzed separately. In the comparison between single– and multi– detector row CT, the Mann-Whitney U test was used to obtain P values. We employed the weighted ␬ coefficient of agreement by using GraphPad Instat software (version 2.0; GraphPad Software, San Diego, Calif) to analyze the agreement between the Bismuth type assessed at CT and the pathologic extent of ductal involvement. All continuous variables are

reported as the mean ⫾ standard deviation.

Results Histopathologic Findings Adenocarcinomas were histopathologically classified as sclerosing-infiltrating, nodular, and polypoid tumors in 36, 13, and six patients, respectively. Surgical and pathologic findings showed portal invasion in 26 patients and arterial invasion in 22 patients. Histologically proved metastases to N2 lymph nodes were present in 15 (27%) of 55 patients and including one metastasis to paraaortic nodes. Vascular Invasion Arterial involvement was described on CT images in 20 patients (Fig 2); one case of arterial involvement ultimately proving to be false-positive (Fig 3). Among the remaining 35 patients, arterial involvement was missed at CT in three patients (Fig 4). Sensitivity, spec-

Table 1 Previous and Revised Criteria for Unresectability of Hilar Cholangiocarcinoma Previous Criteria Bismuth type IV

Invasion of main portal vein or proper hepatic artery Atrophy of one hepatic lobe with contralateral vascular invasion Atrophy of one hepatic lobe with contralateral tumor extension to second biliary confluence Invasion of second biliary confluence on one lobe and contralateral vascular invasion Metastasis to N2 lymph nodes or paraaortic lymph nodes* Distant metastasis

Our Revised Criteria

Evaluation Tool

Type IV and tumor extends farther than 2 cm from the hilum Same, plus involved segment is longer than 2 cm Same

Cholangiography

Same

CT and cholangiography

Same

CT and cholangiography

Metastasis to celiac, portacaval, or paraaortic lymph nodes Same

CT

ificity, positive predictive value, negative predictive value, and overall accuracy were 86.4%, 97.0%, 95.0%, 91.4%, and 92.7%, respectively. Arterial invasion was diagnosed in nine of the 26 patients evaluated with single– detector row CT; this included two false-negative cases and one false-positive case. Sensitivity, specificity, positive predictive value, negative predictive value, and overall accuracy were 77.8%, 94.1%, 87.5%, 88.9%, 88.5%, respectively, at single– detector row CT and 92.3%, 100%, 100%, 94.1%, 96.6%, respectively, at multi– detector row CT. The difference between single– and multi– detector row CT was not statistically significant (P ⫽ .406). Portal invasion was detected at surgery in 26 patients, and CT accurately depicted portal invasion in 20 of these patients (Fig 5). The six false-negative findings involved the main portal trunk in five cases and the right branch in two cases. Among the 29 patients in whom there was an intact portal vein, two false-positive cases were noted. This produced a sensitivity, specificity, positive predictive value, negative predictive value, and overall accuracy of 76.9%, 93.1%, 90.9%, 81.8%, and 85.5%, respectively. Single– detector

Figure 2

CT

CT

CT

* N2 lymph nodes ⫽ peripancreatic, periduodenal, celiac, superior mesenteric, and posterior pancreaticoduodenal lymph nodes.

116

Lee et al

Figure 2: Transverse arterial phase helical CT scan in 24-year-old man with unresectable hilar cholangiocarcinoma encasing the hepatic artery shows ill-defined low-attenuating mass (white arrows) in hilar area and abrupt cutoff of proper and right hepatic arteries (black arrow) within the soft-tissue lesion. Radiology: Volume 239: Number 1—April 2006


row CT scans were obtained in 13 patients who were determined to have portal vein invasion at surgery. Among these patients, single– detector row CT failed to depict the vascular invasion in four. The two false-positive cases were both determined at single– detector row CT. Sensitivity, specificity, positive predictive value, negative predictive value, and overall accuracy were 64.3%, 84.6%, 81.8%, 73.3%, 76.9%, respectively, at single– detector row CT and 84.6%, 100%, 100%, 88.9%, 93.1%, respectively, at multi–detector row CT. The difference was not significant (P ⫽ .143).

Lymph Node Metastasis The incidence of N2 regional metastasis was 27% (15 of 55 patients). CT accurately depicted N2 regional metastasis in eight of these 15 patients. Among the 40 patients without lymph node metastasis, there were false-positive findings noted in two patients. The positive predictive value was 80.0%, and the negative predictive value was 84.4%. Sensitivity and specificity were 53.3% and 95%, respectively. Overall, nodal involvement was correctly assessed with CT in 46 of 55 patients (accuracy, 83.6%). The mean size of metastatic nodes was 12 mm, and central necrosis was seen in three patients. Biliary Duct CT and cholangiography accurately depicted ductal extension in 46 (84%) of 55 patients (Table 2). The weighted ␬ value was 0.767. Underestimation was noted in seven patients (Fig 6), and overestimation was noted in two patients. Resectability Of 42 patients with tumors that were thought to be resectable on the basis of CT and cholangiographic findings, 30 underwent curative surgical treatment, while the other 12 were retrospectively found to have unresectable tumors (positive predictive value, 71.4%). Sources of inaccuracies involved (a) underestimation of bile duct involvement (n ⫽ 6), (b) underestimation of diffuse tumor infiltration extending to the heRadiology: Volume 239: Number 1—April 2006

Lee et al

Figure 3

Figure 3: Transverse CT scans in 58-year-old man with exophytic mass invading portal vein. (a) Arterial phase scan shows exophytic mass (black arrows) at the hilum, encasing proper and right hepatic arteries (white arrow). (b) On portal phase scan, fat plane between mass and right portal vein, as well as main portal vein, is obliterated (arrows). However, the patient underwent surgery and curative left lobectomy. The main portal vein was invaded by tumor, but anastomosis was possible. Proper hepatic artery was abutting the tumor without invasion.

Figure 4

Figure 4: (a, b) Transverse portal phase helical CT scans at two levels in 66-year-old man with unresectable tumor because of hepatic artery invasion show hyperattenuating tumor (white arrow) at the level of the hilum (type II) and the right intrahepatic ducts moderately dilated. We predicted the tumor was unresectable because of conglomerated malignant lymph nodes at the pericholedochal area (black arrow). The proper hepatic artery (arrowheads), however, appeared intact. The patient underwent palliative hepaticojejunostomy for huge necrotic nodes infiltrating to the pancreas with proper hepatic artery involvement.

paticoduodenal ligament (n ⫽ 3) (Fig 7), and (c) inaccurate assessment of vascular involvement (n ⫽ 3) (Fig 8). Tumors were predicted to be unresectable in 13 patients, and this assessment proved to be correct in 11 patients (negative predictive value, 84.6%). Vascular involvement was overestimated in two patients

(Fig 3b). Sensitivity and specificity were 93.8% and 47.8%, respectively. Overall accuracy of CT for determination of resectability was 74.5%.

Postoperative Survival In 32 patients who underwent curative surgery, the mortality rate was 18.8% 117


(n ⫽ 6; range, 1–26 months; mean, 10.4 months). Death occurred as a result of postoperative liver failure (n ⫽ 1), cholangitis or wound infection (n ⫽ 2), and tumor recurrence or seeding (n ⫽ 3). Among the 26 long-term survivors, tumor recurred in 15 patients. The

Figure 5

mode of recurrence was peritoneal seeding in two patients, metastasis in four patients, and local recurrence in nine patients. The median disease-free survival time in the other 11 patients was 36.3 months (range, 18 – 69 months) at the conclusion of our study in February 2005. The median survival period after surgery in the 23 patients who underwent palliative surgery was 12.7 months (range, 0 –36 months) at the conclusion of the study in February 2005.

Discussion

Figure 5: Transverse portal phase CT scan in 59-year-old woman with resectable tumor despite invasion of main portal vein. Scan shows the straight margin of the left portal vein and proximal right portal vein (white arrows) along the tumor (black arrows). The patient underwent right hepatectomy with hepaticojejunostomy and portal vein resection and anastomosis.

In the evaluation of hepatic arterial and portal venous invasion, our results are consistent with those of Cha et al (22) and with those in other reports (20,23), which suggests that CT is accurate in depicting vascular invasion. Moreover, a detailed assessment of vascular structures can be achieved with thin-collimation data and high-spatial-resolution multi– detector row CT, especially with CT scanners with 16 detector rows (24). As expected, our results with multi– detector row CT are superior to those with single– detector row CT. For nodal metastasis, the results of our study showed that the incidence of N2 regional metastasis was 27.3%. This is similar to results given in other surgi-

Figure 6

Figure 6: (a, b) Contiguous portal phase cholangiograms in 76-year-old woman with type IV hilar cholangiocarcinoma show the tumor involving the hepatic duct bifurcation (arrow). We estimated bile duct involvement as type II at imaging. At surgery, ductal invasion was found to extend beyond the bilateral second confluence (ⴱ), and the patient underwent palliative common bile duct resection with hepaticojejunostomy. 118

Lee et al

cal reports (10,11,25–30), which show the incidence ranging from 14% to more than 50%. Our result regarding nodal assessment with CT was considered acceptable, given the inherent limitations of lymph node evaluation with CT. We could accurately assess the extent of biliary ductal involvement in 46 (84%) of 55 patients, which exceeds the accuracies reported for several prior studies (13,20,31). This result could be attributed to our combined use of CT and cholangiography. Results of one previous study (26) showed that CT tends to cause underestimation of the intraductal extent of tumor. When combined with direct cholangiography, however, more accurate assessment was possible (20,32). Obstructive jaundice is usually the manifesting symptom of hilar cholangiocarcinoma, and complete relief of jaundice by means of biliary drainage is needed before surgery. Therefore, the combined use of CT and direct cholangiography for preoperative assessment is usual. Despite the improved assessment achieved by using both imaging modalities, tumor invasion of the intrahepatic bile ducts was underestimated in seven patients. This could be explained by the infiltrative growth pattern and the propensity of cholangiocarcinoma to grow axially along segmental ducts. In two patients, extensive tumor infiltration to the hepaticoduodenal ligament was missed. For intraductal papillary type cholangiocarcinoma, cholangiography can lead to overestimation of bile duct involvement because of intraductal necrotic material. At contrast material– enhanced CT, infiltrating tumors are seen as a focally thickened ductal wall obliterating the lumen. About 80% of these tumors are hyperattenuating relative to the liver, and those tumors are caused by sclerotic lesions with abundant fibrous tissue (20). The tumor can be better contrast-enhanced on delayed images. Delayed phase CT was not performed in our study because of the added radiation, scanning time, and poor quality for the evaluation of small anatomic structures. Small anatomic structures are poorly demarcated and Radiology: Volume 239: Number 1—April 2006


Lee et al

Figure 7

Figure 8

Figure 7: (a, b) Contiguous transverse portal phase CT scans in 60-year-old woman with unresectable hilar cholangiocarcinoma show focal wall thickening and enhancement in the hepatic duct (arrow). We predicted the tumor to be operable, but during surgery extensive tumor infiltration was noted along the hepaticoduodenal ligament, and the patient underwent palliative hilar resection with hepaticojejunostomy. Retrospectively, we could find the subtle soft-tissue infiltration along the hepaticoduodenal ligament (arrowheads).

Table 2 Evaluation of Type of Biliary Duct Obstruction Seen at CT and Cholangiography Type at CT and Cholangiography I II IIIa IIIb IV Total

Type at Pathologic Evaluation IIIa IIIb

I

II

3 0 0 0 0

0 4 1 0 1

0 0 18 0 0

3

6

18

IV

Total (n ⫽ 55)

0 0 0 6 0

0 1 4 2 15

3 5 23 8 16

6

22

55

Note.—Data are numbers of cases of each type of biliary duct obstruction according to the Bismuth classification system.

detailed evaluation of vascular structures is impossible on delayed phase images. In our experience, portal phase images are optimal in the evaluation of both tumor and vessels. Our study results validate helical CT as a highly effective modality in determining unresectability. The main causes of inaccuracies were underestimation of biliary and vascular invasion. Underestimation of diffuse tumor infiltration extending to the hepaticoduodenal ligament (n ⫽ 2) was another source of prediction failure. In one case, we could find the subtle soft-tissue infiltration along the hepaticoduodenal ligament on retrospective review. Atrophy of one hepatic lobe with contralateral vascular or biliary invaRadiology: Volume 239: Number 1—April 2006

sion was considered an unresectable situation because compensation of insufficient liver function by the residual atrophic lobe would be impossible after resection of the contralateral lobe containing the tumor. After curative surgery, the mortality rate was 18.8%, and tumor recurred in 56% of patients. These results are similar to the results previously reported (33,34). Magnetic resonance (MR) imaging in conjunction with MR cholangiopancreatography has proved helpful in diagnosing hilar cholangiocarcinoma and in determining resectability in recent years (21,23,35–38). However, we could not include such imaging modalities because MR cholangiopancreatog-

Figure 8: (a, b) Contiguous transverse portal phase CT scans in 68-year-old man with hilar cholangiocarcinoma (arrowheads) with invasion of main portal vein (arrows) show right portal vein invasion in long segment, together with 1.2-cmlong segmental invasion of left portal vein. We predicted right lobectomy with left portal vein resection and anastomosis would be possible. However, the patient underwent palliative hepatic duct resection and hepaticojejunostomy because of severe main portal vein involvement.

raphy was not the routine preoperative evaluation method in our hospital in most patients. Since preoperative biliary drainage was the routine procedure, we used direct cholangiography as a reference standard because it had higher spatial resolution (39). Recently, MR cholangiopancreatography has been performed before attempting percutaneous transhepatic biliary drainage to choose the entry site for the drainage tube. In one recent study (13), despite 119


good visualization of the ducts at MR imaging, the exactness and reliability of MR imaging for assessment of the borders of tumor growth were clearly inferior to those of percutaneous transhepatic cholangiography. Courbie`re et al (40) reported that MR cholangiography caused underestimation of type III and IV strictures compared with direct cholangiography because of a failure to recognize a noncommunicating hepatic ductal branch. The main limitation of MR cholangiography in comparison with direct cholangiography is its inability to offer therapeutic intervention, and replacement of direct cholangiography with MR cholangiography still seems a distant prospect. The entire upper abdomen can be covered with use of a collimation of less than 1 mm in one breath hold. With the thin-section images, high-quality multiplanar reconstructions can be acquired, which are helpful for assessing the anatomy of the biliary system (23,41). Unfortunately, we could not use multiplanar reconstructions because our study included single– detector row CT data, and multi– detector row CT with four detector rows for larger case collection and multi– detector row CT with 16 detector rows have been performed only recently. In the future, further improvement in biliary evaluation might be achieved with the use of such additional imaging modalities. Our study had several limitations. First, the removed lymph nodes were not mapped in a one-to-one site correspondence for correlation with imaging. Second, prior biliary drainage could have affected subsequent bile duct evaluation with CT scanning. In such cases, however, the use of direct cholangiography for bile duct involvement helped alleviate this limitation. Third, variation in the degree of aggressiveness among the surgeons may be another limitation. Finally, we could not consider the fine and complicated aspects of surgery, which caused the small discrepancy in the assessment of resectability. For example, portal vein reconstruction is clearly much easier on the left than it is on the right because of the angulation of the left portal vein (39). 120

In conclusion, combined interpretation with multiphasic helical CT and direct cholangiography by using our revised criteria allows accurate prediction of resectability for hilar cholangiocarcinoma.

References 1. Nakanuma Y, Minato H, Kida T, Terada T. Pathology of cholangiocellular carcinoma. In: Tobe T, Kameda H, Okudaira M, Ohto M, eds. Primary liver cancer in Japan. Tokyo, Japan: Springer-Verlag, 1994;39 –50. 2. Jarnagin WR, Fong Y, DeMatteo RP, et al. Staging, resectability, and outcome in 225 patients with hilar cholangiocarcinoma. Ann Surg 2001;234:507–517. 3. Su CH, Tsay SH, Wu CC, et al. Factors influencing postoperative morbidity, mortality, and survival after resection for hilar cholangiocarcinoma. Ann Surg 1996;223:384 –394. 4. Launois B, Reding R, Lebeau G, Buard JL. Surgery for hilar cholangiocarcinoma: French experience in a collective survey of 552 extrahepatic bile duct cancers. J Hepatobiliary Pancreat Surg 2000;7:128 –134. 5. Lee SG, Lee YJ, Park KM, Hwang S, Min PC. One hundred and eleven liver resections for hilar bile duct cancer. J Hepatobiliary Pancreat Surg 2000;7:135–141. 6. Lillemoe KD, Cameron JL. Surgery for hilar cholangiocarcinoma: the Johns Hopkins approach. J Hepatobiliary Pancreat Surg 2000; 7:115–121. 7. Lygidakis NJ, Sgourakis GJ, Dedemadi GV, Vlachos L, Safioleas M. Long-term results following resectional surgery for Klatskin tumors: a twenty-year personal experience. Hepatogastroenterology 2001;48:95–101. 8. Munoz L, Roayaie S, Maman D, et al. Hilar cholangiocarcinoma involving the portal vein bifurcation: long-term results after resection. J Hepatobiliary Pancreat Surg 2002;9: 237–241. 9. Lygidakis NJ, Singh G, Bardaxoglou E, et al. Changing trends in the management of Klatskin tumor. Hepatogastroenterology 2004;51:689 – 696. 10. Neuhaus P, Jonas S, Bechstein WO, et al. Extended resections for hilar cholangiocarcinoma. Ann Surg 1999;230:808 – 818. 11. Ogura Y, Kawarada Y. Surgical strategies for carcinoma of the hepatic duct confluence. Br J Surg 1998;85:20 –24. 12. Tashiro S, Tsuji T, Kanemitsu K, Kamimoto Y, Hiraoka T, Miyauchi Y. Prolongation of survival for carcinoma at the hepatic duct confluence. Surgery 1993;113:270 –278.

Lee et al

13. Otto G, Romaneehsen B, Hoppe-Lotichius M, Bittinger F. Hilar cholangiocarcinoma: resectability and radicality after routine diagnostic imaging. J Hepatobiliary Pancreat Surg 2004;11:310 –318. 14. de Groen PC, Gores GJ, LaRusso NF, Gunderson LL, Nagorney DM. Biliary tract cancers. N Engl J Med 1999;341:1368 –1378. 15. Lu DS, Reber HA, Krasny RM, Kadell BM, Sayre J. Local staging of pancreatic cancer: criteria for unresectability of major vessels as revealed by pancreatic-phase, thin-section helical CT. AJR Am J Roentgenol 1997; 168:1439 –1443. 16. O’Malley ME, Boland GW, Wood BJ, Fernandez-del Castillo C, Warshaw AL, Mueller PR. Adenocarcinoma of the head of the pancreas: determination of surgical unresectability with thin-section pancreaticphase helical CT. AJR Am J Roentgenol 1999;173:1513–1518. 17. Hann LE, Getrajdman GI, Brown KT, et al. Hepatic lobar atrophy: association with ipsilateral portal vein obstruction. AJR Am J Roentgenol 1996;167:1017–1021. 18. Sobin LH, Wittekind CH, eds. International Union Against Cancer (UICC): TNM classification of malignant tumors. New York, NY: Wiley-Liss, 1997;170 –173. 19. Engels JT, Balfe DM, Lee JK. Biliary carcinoma: CT evaluation of extrahepatic spread. Radiology 1989;172:35– 40. 20. Han JK, Choi BI, Kim AY, et al. Cholangiocarcinoma: pictorial essay of CT and cholangiographic findings. RadioGraphics 2002;22:173–187. 21. Manfredi R, Masselli G, Maresca G, Brizi MG, Vecchioli A, Marano P. MR imaging and MRCP of hilar cholangiocarcinoma. Abdom Imaging 2003;28:319 –325. 22. Cha JH, Han JK, Kim TK, et al. Preoperative evaluation of Klatskin tumor: accuracy of spiral CT in determining vascular invasion as a sign of unresectability. Abdom Imaging 2000;25:500 –507. 23. Zech CJ, Schoenberg SO, Reiser M, Helmberger T. Cross-sectional imaging of biliary tumors: current clinical status and future developments. Eur Radiol 2004;14:1174 – 1187. 24. Fletcher JG, Wiersema MJ, Farrell MA, et al. Pancreatic malignancy: value of arterial, pancreatic, and hepatic phase imaging with multi– detector row CT. Radiology 2003; 229:81–90. 25. Tillich M, Mischinger HJ, Preisegger KH, Rabl H, Szolar DH. Multiphasic helical CT in diagnosis and staging of hilar cholangiocarciRadiology: Volume 239: Number 1—April 2006


noma. AJR Am J Roentgenol 1998;171:651– 658. 26. Feydy A, Vilgrain V, Denys A, et al. Helical CT assessment in hilar cholangiocarcinoma: correlation with surgical and pathologic findings. AJR Am J Roentgenol 1999;172:73–77. 27. Kitagawa Y, Nagino M, Kamiya J, et al. Lymph node metastasis from hilar cholangiocarcinoma: audit of 110 patients who underwent regional and paraaortic node dissection. Ann Surg 2001;233:385–392. 28. Tojima Y, Nagino M, Ebata T, Uesaka K, Kamiya J, Nimura Y. Immunohistochemically demonstrated lymph node micrometastasis and prognosis in patients with otherwise node-negative hilar cholangiocarcinoma. Ann Surg 2003;237:201–207. 29. Kurosaki I, Tsukada K, Hatakeyama K, Muto T. The mode of lymphatic spread in carcinoma of the bile duct. Am J Surg 1996; 172:239 –243. 30. Kayahara M, Nagakawa T, Ohta T, Kitagawa H, Tajima H, Miwa K. Role of nodal involvement and the periductal soft-tissue margin in middle and distal bile duct cancer. Ann Surg 1999;229:76 – 83.

Radiology: Volume 239: Number 1—April 2006

31. Choi BI, Lee JH, Han MC, Kim SH, Yi JG, Kim CW. Hilar cholangiocarcinoma: comparative study with sonography and CT. Radiology 1989;172:689 – 692. 32. Choi BI, Lee JM, Han JK. Imaging of intrahepatic and hilar cholangiocarcinoma. Abdom Imaging 2004;29:548 –557. 33. Kondo S, Hirano S, Ambo Y, et al. Forty consecutive resections of hilar cholangiocarcinoma with no postoperative mortality and no positive ductal margins: results of a prospective study. Ann Surg 2004;240:95–101. 34. Weber SM, Jarnagin WR, Klimstra D, DeMatteo RP, Fong Y, Blumgart LH. Intrahepatic cholangiocarcinoma: resectability, recurrence pattern, and outcomes. J Am Coll Surg 2001;193(4):384 –391. 35. Reinhold C, Bret PM. Current status of MR cholangiopancreatography. AJR Am J Roentgenol 1996;166:1285–1295. 36. Fulcher AS, Turner MA. HASTE MR cholangiography in the evaluation of hilar cholangiocarcinoma. AJR Am J Roentgenol 1997; 169:1501–1505. 37. Guthrie JA, Ward J, Robinson PJ. Hilar

Lee et al

cholangiocarcinoma: T2-weighted spin-echo and gadolinium enhanced FLASH MR imaging. Radiology 1996;201:347–351. 38. Lopera JE, Soto JA, Munera F. Malignant hilar and perihilar biliary obstruction: use of MR cholangiography to define the extent of biliary ductal involvement and plan percutaneous interventions. Radiology 2001;220: 90 –96. 39. Rea DJ, Munoz-Juarez M, Farnell MB, et al. Major hepatic resection for hilar cholangiocarcinoma: analysis of 46 patients. Arch Surg 2004;139:514 –523. 40. Courbie`re M, Pilleul F, Henry L, Ponchon T, Touzet S, Valette PJ. Value of magnetic resonance cholangiography in benign and malignant biliary stenosis: comparative study with direct cholangiography. J Comput Assist Tomogr 2003;27:315–320. 41. Itoh S, Ikeda M, Ota T, Satake H, Takai K, Ishigaki T. Assessment of the pancreatic and intrapancreatic bile ducts using 0.5-mm collimation and multiplanar reformatted images in multislice CT. Eur Radiol 2003;13:277– 285.

121