C mented association with primary sclerosing

Transcription

1 CA19-9 Does Not Predict Cholangiocarcinoma in Patients With Primary Sclerosing Cholangitis Undergoing Liver transplantion Adrian Fisher, Neil D. Theise", Albert Minf, Eytan Mor, Suhm Emre, Adam Pearl, Myron E. Schwartz, Charles M. Miller, and Patricia A. Sheiner The results of liver transplantation in patients with cholangiocarcinoma have been poor. It has been suggested that elevated serum CA19-9 levels predict cholangiocarcinoma in patients with primary sclerosing cholangitis. We analyzed the predictive value of CAl9-9 antigen as a marker of cholangiocarcinoma in patients with primary sclerosing cholangitis evaluated for liver transplantation. We reviewed the charts of 26 patients with primary sclerosing cholangitis (stage IV) in whom preoperative serum CA19-9 levels were determined; 22 of 26 underwent liver transplant. Explant specimens were serially sectioned and examined for tumor. In 3 of the 26 patients, cholangiocarcinoma was diagnosed during pretransplantation evaluation; exploratory laparotomy on the last patient showed no evidence of cholangiocarcinoma, and this patient is awaiting trans- plantation. Twelve of the 26 patients had CA19-9 levels more than double the laboratory reference range (-37 U/mL) (mean f 13 U/mL, range U/mL). Two of the 12 patients with elevated CA19-9 levels had cholangiocarcinoma. Of the 14 patients with normal levels, two had cholangiocarcinoma. No correlation between elevated CA19-9 and bile duct dysplasia was noted. Sensitivity for serum CA19-9 levels more than twice the reference range is 5%, specificity is 54.5%, positive predictive value is 16.6%. An elevated serum CA19-9 level in a patient with stage IV primary sclerosing cholangitis does not reliably predict coexisting cholangiocarcinoma. Persistently high or rising serum CA19-9 levels do not indicate more urgent need for liver transplantation. Copyright 1995 by the American Association for the Study of Liver Diseases holangiocarcinoma (CCA) has a well-docu- C mented association with primary sclerosing cholangitis (PSC); 5% to 15% of patients with PSC are found to have CCA as well.' Accurate and timely diagnosis of cholangiocarcinoma in the presence of PSC can be difficult because of the lack of pathognomonic signs, laboratory data, or radiological findings. Detection of CCA is frequently delayed until an advanced stage, when a patient with previously stable disease begins to deteriorate rapidly.2 Earlier diagnosis via a sensitive and specific serum marker may allow surgical therapy and improved long-term survival. Liver transplantation in patients with cholangiocar- From the Department of Surgery, Division of Liver Transplantation; Department of Pathology*; and the Department of Medicine, Division of Hepatologyf; The Mount Sinai Medical Center, New York, NY. Address reprint requests to: Patricia A. Sheiner, MD, Division of Liver Transplantation, Mount Sinai Medical Center, Box 114, One Gustave L. Levy Place, New York, NY 129. Copyright 1995 by the American Association for the Study of Liver Diseases /95/12-2$3./ cinoma has been associated with rapid recurrence and poor survival, both in recipients with overt CCA and in those in whom the lesion is found incidentally in the explant ~pecimen.~-~ Because individuals with PSC comprise 1% to 24% of patients undergoing liver transplantation in the United States,' a reliable marker of CCA in such patients would be immensely useful. lo CA19-9 antigen, an easily assayed biochemical marker, is present in many gastrointestinal epithelial tumors. Several authors have reported the sensitivity and specificity of serum CA19-9 levels in detecting pancreatic cancer and de novo cholangiocarcinoma.l1-l6 Although its behavior in PSC is not well described, it has recently been suggested that elevated serum CA19-9 levels in patients with PSC may signify the presence of CCA.'J7J8 We analyzed the predictive value of CA19-9 antigen as a marker of CCA in patients with PSC evaluated for liver transplant. Materials and Methods We reviewed the charts of 26 consecutive adult patients with PSC who were evaluated for liver transplantation at the Mount Sinai 94 Liver Transplantation and Surgep, Vol 1, No 2 (March), 1995: pp 9498

2 CAI9-9 Does not Predict CCA in PSC 95 Medical Center between March 1992 and April These 26 patients either underwent liver transplantation or exploratoty laparotomy or were found to have CCA during pretransplantation evaluation. One additional patient with PSC was evaluated and listed for transplant at Mt. Sinai during the study period but subsequently underwent transplantation elsewhere and is not included in our analysis. All 26 patients had at least one preoperative serum CA19-9 level determination. Serum samples were drawn during outpatient evaluation; none were drawn during an episode of acute bacterial cholangitis. When multiple preoperative CAI 9-9 determinations had been made, serum levels were classified as normal or elevated based on the last determination before transplantation or diagnostic intervention. Serum CA19-9 levels did not affect clinical management. Serum CA19-9 levels were measured by enzyme immunoassay (Roche Biomedical Laboratories, Raritan, New Jersey); the normal range of this assay correlates closely with the standard radioimmunoassay (Centocor, Inc, Malvern, PA) (M. Haider, Roche Biomedical Laboratories, Raritan, New Jersey, personal communication, June 1994). Serum CA19-9 levels between and 37 U/mL constitute the normal reference range. In order to exclude the statistical distribution of high-normal CA19-9 values, we selected twice the reference range as the lower limit for abnormal elevation (74 U/mL); this value is greater than six standard deviations above the laboratory mean. PSC was diagnosed on the basis of liver enzyme profile, cholangiography, and often, liver biopsy specimen. Because many patients underwent cholangiography before referral to our center, uniform retrospective review of all studies for dominant stricture was not possible. Histological findings consistent with PSC were observed in all liver explant specimens. No patient with acute bacterial cholangitis underwent transplantation. No other gastrointestinal malignancies were discovered in our cohort. Colonoscopy was performed in all patients to identify or evaluate inflammatory bowel disease (IBD). To rule out incidental CCA, the entire explanted liver was serially sectioned at.5- to 1.-cm intervals and examined for tumor. Additional sectioning of the larger bile ducts was performed. Further random sampling of each liver explant was performed for histological analysis; five sections each from the right and left lobes, one section of the caudate lobe, and at least one cross-section of the extrahepatic portal vein, hepatic artery, and common bile duct were examined. Additional sections of focal lesions were also taken. All sections were stained wth hematoxylineosin. Results are expressed as mean? SD. Statistical significance of group mean values was analyzed with either x2 or Student s t-test, as appropriate. Sensitivity, specificity and positive predictive value were calculated according to standard definitions. Results Of the 26 patients, 22 underwent liver transplantation; three were discovered to have CCA in the pretransplantation evaluation. One additional pa- tient was explored for presumptive CCA, but all biopsies were negative; he is currently well, with no evidence of CCA, and awaiting liver transplant 13 months after exploration. In this patient, because no liver explant specimen was available for analysis, the possibility of occult CCA remains. In 12 of 26 patients, preoperative serum CA19-9 levels were more than double the laboratory reference range (mean, U/mL; range U/mL) (Fig 1). Two of the 12 patients were found to have CCA. In one patient, diagnosis was made by ultrasound-guided biopsy during the pretransplantation evaluation; in the other, CCA was found incidentally in the explant specimen. Their serum CA19-9 levels were 353 and 227 U/mL, respectively. Neither patient had evidence of extrahepatic spread of CCA, as determined by computed tomography (CT) scan and endoscopic retrograde cholangiopancreatography (ERCP) in the first patient and exploratory laparotomy in the second. In 1 of the 12 (83.3%) patients, including seven patients in whom CA19-9 levels were persistently elevated or rose on sequential measurements, there was no evidence of CCA. In 14 patients, serum CA19-9 levels were normal at one or more determinations (mean U/mL; range < U/mL). In 1 of these 14, CCA was found in a single periportal lymph node at exploratory laparotomy, whereas in another fineneedle aspiration showed CCA (with no evidence of extrahepatic spread on CT scan). Both patients had serum CA19-9 levels <5 U/mL. Thus, 2 of 12 patients with elevated serum CA19-9 levels had CCA, versus 2 of 14 patients whose serum CA19-9 levels were within the reference range (P = ), Sensitivity for serum ( levels greater than twice the.+; :, 2 ;./:::,,.,,!,;!:;?:,;, 1 88 eo Patients Figure 1. Pretransplantation CA19-9 levels at all determinations in the study population., Final pretransplantation CA19-9 level;, other pretransplantation CA19-9 levels; *, patient with CCA.

3 96 Fisher et al reference range is SO%, specificity is 54.5%. Positive predictive value for this cut-off value is 16.6%. There was no significant gender difference between the two groups (Table 1). Concomitant IBD was found in six patients with elevated serum CA19-9 levels and in eight patients with normal levels (P = ). Three patients with CCA had associated IBD. Mean total bilirubin levels obtained at the time of determination of serum (319-9 levels were not significantly different (Table 1). Among patients with high serum CA19-9 levels, the mean age was 34.5 years, versus 51 years among those with normal levels (P <.5). All patients showed advanced PSC (stage IV), that is, biliary cirrhosis. One patient with an elevated serum CA19-9 level showed mild dysplasia on histological examination of an extrahepatic bile duct but no CCA. No other patient had histological evidence of dysplasia. No evidence of viral hepatitis, PBC, or other associated liver pathology was identified in explant specimens. Follow-up of study patients was 1% (range, 1-25 months). The patient with incidental CCA in the hepatectomy specimen died of recurrent disease at 16 months after transplantation. A marked elevation of serum CA19-9 was noted at recurrence. No other patient has developed CCA postoperatively. The three patients who were excluded from transplantation because of CCA died from progression of their disease. The patient who received a transplant elsewhere had a serum CA19-9 level of U/ml. There was no evidence of CCA in the explanted liver (although the other center s protocol for examination Table 1. Characteristics of Patients With Normal and Elevated Serum CA19-9 Levels CA U/mL Variable (n = 14) Serum CA19-9 (U/mL) 2.5 f 17 Age (Y) Gender (male/ female) 1/4 Bilirubin (mg/dl) 11.8? 8.5 IBD (n) 8 Cholangiocarcinoma (n) 2 Abbreviation: NA, not applicable. CA19-9 > 74 U/mL (n = 12) t t f P NA <.5 of explant specimens is unknown), and the patient is alive and well 16 months after transplantation. Discussion In normal individuals, serum levels of CA19-9 are rarely elevated, although concentrations in gastrointestinal secretions are often high. Benign biliary tract processes uncommonly elevate serum CA19-9 levels above 37 U/mL. Investigators have shown sensitivity > 7% and specificity > 9% for such processes employing a value greater than 37 U/mL as the abnormal reference range. 13,14 In patients with cholangitis, cirrhosis, or pancreatitis, however, elevated serum levels of CA19-9 are reported. l1 The biliary tree in PSC is markedly distorted. Strictures, beading, dilatations, and diverticulae of the biliary tract on cholangiography confirm the diagnosis. These findings, however, may mask the presence of CCA. The sensitivity of invasive diagnostic procedures (exfoliative cytology and bile duct brushing) is disappointing, rangmg from 3% to 8%. Because histological severity of PSC, disease duration, and biochemical abnormalities do not adequately predict occurrence of CCA, identification of PSC patients at risk for CCA has not been p~ssible. ~ Serum CA19-9 levels have been useful in detection of de novo CCA,2 but their value in documenting CCA in conjunction with PSC is unclear. In a recent series of 37 patients with PSC (9 of whom had concomitant CCA), Nichols et a1 found that serum CA19-9 levels > 1 U/mL predicted CCA with a sensitivity of 89% and a specificity of 86%.2 Our own series does not support these findings. Furthermore, raising our abnormal reference value to > 1 U/mL as in Nichol s study still leaves 7 of 9 patients with high serum CA19-9 levels but no CCA-that is, a 77.8% false-positive rate. Specificity and positive predictive value are then 68.1% and 22.2%, respectively. These rates are clearly inadequate for clinical utility of a putative tumor marker. Our series suffers from a small number of patients with CCA (n = 4). Our rate of CCA occurrence, however, is similar to that reported by other investigators. Also of note is the fact that we were able to examine the total hepatectomy specimen in all patients who have received transplants and thus confirm the absence of CCA. Occult CCA in patients with normal serum levels of CA19-9 remains possible in Nichols cohort. Such false negatives would reduce the sensitivity of the serum CA19-9 in Nichols study

4 CA19-9 Does not Predict CCA in PSC 97 population. In fact, Marsh et a1 reported that 5 of 55 patients undergoing transplantation for PSC had occult CCA found in the explant specimens6 whereas Ramage et a1 found four occult CCA in 8 patients.i8 Given that cirrhosis is known to elevate CA19-9 levels and that CA19-9 levels reflect the degree of inflammatory activity in the biliary tree, high levels in our study patients without evidence of CCA may well represent severe, end-stage PSC (ie, biliary cirrhosis or stage IV disease). Patients with end-stage PSC may still show normal CA19-9 levels if their PSC is relatively quiescent. Nichols reports that 21 of 28 patients with isolated PSC showed stage 111 or IV disease and that 5 of 9 patients with CCA also had such advanced disease. In contrast, Miros et al reported on 11 patients with PSC, 4 of whom had associated CCA; histological examination of the hepatectomy specimen showed that patients with CCA had less advanced PSC than those with PSC alone.4 It is thus possible that the combination of elevated serum CA19-9 levels and clinical parameters or cholangiography showing early-stage PSC should heighten suspicion of coexistent CCA. There was a significant age difference between our patients with elevated serum CA19-9 levels (mean age, 34.5 years) and those with normal levels (mean age, 51 years). Although the importance of this finding is unclear, it may indicate that our younger patients had a more rapidly progressive and inflammatory form of PSC. Possibly, the older patients represent a group with a slowly progressive variant and therefore present at a later age, with lower serum levels of CA19-9. Nichols notes that in patients with CCA, the malignancy was suspected on the basis of clinical deterioration. These patients all had regional or distant metastatic disease.2 Thus, his patients represent a subgroup with advanced metastatic CCA that is likely to be associated with high serum levels of CA19-9 because of volume and extent of tumor. In our program, such patients are excluded from consideration before formal referral to the transplant service. Patients with CCA missed by the screening process likely have early or small volume disease and might well present with normal CA19-9 levels. In our cohort, therefore, elevated serum CA19-9 levels are most probably due to advanced PSC. Ramage et a1 report that use of two criteria, serum CA19-9 levels > 2 U/mL and carcinoembryonic antigen (CEA) levels > 7 U/mL, predicts CCA in patients with PSC with increased sensitivity (75%) and specificity (87%).18 Two of our patients with CCA had normal serum CA19-9 levels; preoperative CEA data was within normal range in one patient and not available in the other. Of the two patients with elevated serum CA19-9 and CCA, CEA was within normal range in one and markedly elevated in the other. We can therefore make no judgment regarding increased sensitivity or specificity of the combined criteria. In summary, an elevated serum CA19-9 level in a patient with stage IV primary sclerosing cholangitis does not reliably predict coexisting cholangiocarcinoma. Excluding patients with PSC and elevated serum CA19-9 levels from liver transplantation on the basis of presumptive CCA would preclude a marked improvement in quality of life and longevity for a group of patients with end-stage PSC but without CCA. Persistently high or rising serum CA19-9 levels do not indicate a more urgent need for liver transplantation; thus, accelerating transplantation is not warranted in this population. Nevertheless, an elevated serum CA19-9 level, especially in patients with clinical or histological evidence of early-stage PSC, should prompt an intensive effort to exclude malignancy. Acknowledgment The authors thank Nancy Ehrlich for her editorial assistance. References 1. Lindor KD, Wiesner RH, MacCarty RL, Ludwig J, La- Russo NF. Advances in primary sclerosing cholangitis. Am J Med 199;89: Nichols JC, Gores GJ, LaRusso NF, Wiesner RH, Nagorney DM, Ritts RE. Diagnostic role of serum CAI 9-9 for cholangiocarcinoma in patients with primary sclerosing cholangitis. Mayo Clin Proc 1993;68: Goldstein RM, Stone M, Tillery GW, Senzer N, Levy M, Husberg BS, et al. Is liver transplantation indicated for cholangiocarcinoma? Am J Surg 1993; 166: Miros M, Kerlin P, Walker N, Harper J, Lynch S, Strong R. Predicting cholangiocarcinoma in patients with primary sclerosing cholangitis before transplantation. Gut 1991 ; Rosen CB, Nagorney DM, Wiesner RH, Coffey RJ Jr, LaRusso NF. Cholangiocarcinoma complicating primary sclerosing cholangitis. Ann Surg 1991 ;213: Marsh JW, lwatsuki S, Makowka L, Esquivel CO, Gordon RD, Todo S, et al. Orthotopic liver transplantation for primary sclerosing cholangitis. Ann Surg 1988;27:

5 98 Fisher et al 7. Stieber AC, Marino IR, lwatsuki S, Starzl TE. Cholangiocarcinoma in sclerosing cholangitis: the role of liver transplantation. Int Surg 1989;74: lwatsuki S, Gordon RD, Shaw BW, Starzl TE. Role of liver transplantation in cancer therapy. Ann Surg 1985; 22: Mor E, Schwartz ME, Sheiner P, Hytiroglou P, Katz E, Miller CM. Incidental cholangiocarcinoma associated with primary sclerosing cholangitis in a patient undergoing liver transplantation: strategy for management. Transplant Sci 1993;3: Langnas AN, Grazi GL, Stratta RJ, Wood RP, Wagner M, Markin RS, et al. Primary sclerosing cholangitis: the emerging role for liver transplantation. Am J Gastroenterol 199;85: Steinberg W. The clinical utility of the CA19-9 tumorassociated antigen. Am J Gastroenterol 199;85: Pleskow DK, Berger HJ, Gyves J, Allen E, McLean A, Podolsky DK. Evaluation of a serologic marker, CAI 9-9, in the diagnosis of pancreatic cancer. Ann Intern Med 1989;11: Ritts RE, Del Villano BC, Go VLW, Herberman RB, Klug TL, Zurawski VR. Initial clinical evaluation of an immuno- radiometric assay for CA19-9 using the NCI serum bank. Int J Cancer 1984;33: Jalanko H, Kuusela P, Roberts P, Sipponen P, Haglund C, Makela. Comparison of a new tumor marker, CAI 9-9, with alphafetoprotein and carcinoembryonic antigen in patients with upper gastrointestinal diseases. J Clin Pathol 1984;37: Beretta E, Malesci A, Zerbi A, Mariani A, Carlucci M, Bonato C, et al. Serum CA19-9 in the postsurgical follow-up of patients with pancreatic cancer. Cancer 1987;6: Glenn J, Steinberg WM, Kurtzman SH, Steinberg SM, Sindelar WF. Evaluation of the utility of a radioimmunoassay for serum CA19-9 levels in patients before and after treatment of carcinoma of the pancreas. J Clin Oncol 1988;6: Rogers SA, Podolsky DK. Predicting cholangiocarcinoma in patients with primary sclerosing cholangitis: an analysis of the serological marker CAI 9-9. Hepatology 1994;19: Ramage JK, Farrant JM, lorns R, Williams R. Serum tumor markers for diagnosis of cholangiocarcinoma in patients with primary sclerosing cholangitis [abstract]. Gut 1992;33(supp):Sl.