Liver function tests in metastatic uveal melanoma

Liver Function Tests in Metastatic Uveal Melanoma IGOR KAISERMAN, MD, MSC, MPA, RADGONDE AMER, MD, AND JACOB PE’ER, MD

● PURPOSE: To evaluate trends in liver function tests before detection of liver metastases from uveal melanoma. ● DESIGN: Retrospective, comparative, observational case control study. ● METHODS: Setting: The Israel uveal melanoma center at the Hadassah University Hospital. Patient Population: A total of 307 uveal melanoma patients who were diagnosed with uveal melanoma and followed between the years 1988 and 1998. Of them, 30 metastatic patients who had regular follow-up by liver function tests (LFTs) and liver imaging were included in this study. Eighty nonmetastatic patients were randomly chosen as controls. Observation Procedure: The medical records of the metastatic and control groups were reviewed documenting LFTs and liver imaging results. Main Outcome Measures: The mean level of each LFT, its sensitivity, specificity, and likelihood ratio at various time periods before the detection of metastases by liver imaging. ● RESULTS: At the time of diagnosis of liver metastases by imaging, 50% of patients had at least one abnormal LFT (compared with only 5% of the control group). While no change was noted in the mean serum levels of bilirubin, mean lactate dehydrogenase (LDH), alkalinephosphatase, gamma glutamyl transpeptidase (␥GTP) aspartate-aminotrasferase, and alanine-aminotrasferase levels seem to rise, even within normal limits, during the 6 months before the detection of metastases. Based on likelihood ratios, alkaline-phosphatase and lactate dehydrogenase were the most predictive tests. Lactate dehydrogenase and aspartate-aminotransferase were already predictive at 80% of the upper normal limit, whereas alkaline-phosphatase and ␥-glutamyltransferase were most predictive at the upper normal limit.

Biosketches and/or additional material at www.ajo.com Accepted for publication Aug 14, 2003. InternetAdvance publication at ajo.com August 28, 2003. From the Department of Ophthalmology, Hadassah University Hospital, Jerusalem, Israel. Inquiries to Igor Kaiserman, MD, MSc, MPA, Department of Ophthalmology, Hadassah University Hospital, P.O.B. 12000, IL-91120 Jerusalem, Israel; fax: (⫹972) 26416242; e-mail: [email protected]

236

©

2004 BY

● CONCLUSIONS:

Monitoring the changes in selected LFTs, even within normal limits, can help predict metastatic uveal melanoma. (Am J Ophthalmol 2004;137: 236 –243. © 2004 by Elsevier Inc. All rights reserved.)

U

VEAL MELANOMA IS THE MOST COMMON PRIMARY

intraocular tumor. Its annual incidence is about six cases/million.1 The reported rate of liver metastases from posterior uveal melanoma ranges between 10% at 5 years2 and 46% at 15 years3 after diagnosis. Despite a high accuracy of diagnosis and availability of various methods of treatment, the mortality due to metastatic uveal melanoma has remained high.4 Once metastases are detected, the median survival time is less than 5 months.5 The primary site of metastasis of uveal melanoma is the liver.6 Among patients who died metastatic uveal melanoma, the liver was found to almost be always involved.7 Thus, the liver should be the main target for screening for uveal melanoma metastases. This can be achieved by either liver function tests (LFTs) or by liver imaging (ultrasonography [US] or computed tomography [CT]). There is considerable debate regarding the diagnostic value of the LFTs. Most reports have found them to have a relatively low sensitivity.8,9 Some reported the most sensitive LFTs to be gamma glutamyl transpeptidase (␥GT) (sensitivity, 21%) and alkaline-phosphatase (ALK) (sensitivity, 25%).9 Others reported lactate dehydrogenase (LDH) to be more sensitive (sensitivity, 67%).8,10,11 Liver imaging is considered the gold standard for detecting metastases. However, these tests are not performed routinely in all patients with uveal melanoma, and they cannot detect lesions smaller than a certain size. Thus, the optimal screening method for metastatic uveal melanoma is still debatable. While some centers use primarily LFTs and clinical examination,12,13 others prefer liver imaging as the primary screening tool.5 Recently, the importance of early detection of metastatic uveal melanoma has come into renewed focus as new treatment avenues are being explored.14,15 In this study we looked at the sensitivity, specificity, and likelihood ratio of liver function tests in detecting this fatal condition.

ELSEVIER INC. ALL

RIGHTS RESERVED.

0002-9394/04/$30.00 doi:10.1016/j.ajo.2003.08.045

METHODS OF 307 UVEAL MELANOMA PATIENTS (MEAN AGE, 60 YEARS;

range, 29 –92 years) who were diagnosed with uveal melanoma and followed in our department between the years 1988 and 1998, 35 developed liver metastasis (mean age, 59 years; range, 29 – 84 years). The diagnosis of metastatic uveal melanoma was confirmed either by excisional liver biopsy or by fine-needle aspiration biopsy. Of those 35 patients, five patients were excluded due to insufficient follow-up (two or more missed visits, three patients) or having had metastasis present at the first visit (two patients). Thirty patients (mean age, 58.2 ⫾ 14.2 years; range, 29 – 84 years) who developed liver metastases while undergoing regular (every 6 months) follow-up with LFTs (including bilirubin, aspartate-aminotrasferase [AST], alanine-aminotrasferase [ALT], LDH, ALK, and ␥GT) and liver imaging (US in all cases and CT in 87% of cases) were included in this study. Eighty (mean age, 59.0 ⫾ 16.2 years; range, 22–90 years) uveal melanoma patients who were followed for at least 5 years were randomly chosen as controls from the 272 patients who did not develop metastasis. The mean age of the controls did not differ significantly from the mean age of the metastatic group nor did the gender ratio (62.5% females in the metastatic group vs 60% females in the controls). The control group was not matched to the study group by other prognostic parameters of uveal melanoma (such as largest tumor diameter or tumor location) because those parameters are not expected to influence LFTs without the presence of metastases. All available medical records of the metastatic group and control group were reviewed, and the LFTs and liver imaging results were documented. None of the patients in either group had elevated LFTs or other known liver disease at entry into the study. The date of detection of liver metastasis was defined as the date that a liver lesion (later proven to be a metastasis) was first detected by imaging. The levels of the LFTs before and after this date were grouped and averaged at ⬎36, 36, 24, 12, and 6 months before detection of metastasis and 6 and ⬎6 months afterward. Each LFT was grouped into its closest time window. For each time period, only one set of LFTs per patient was included. If an LFT was missing for a certain patient for a certain time period, no extrapolation was performed. In the control group, all the available LFTs were used. The sensitivity, specificity, and likelihood ratio (LR) for detecting metastases were calculated for each LFT at various time periods. The LR was calculated by LR ⫽ sensitivity/(1 ⫺ specificity). The LR is the likelihood that, at a certain time period, a given LFT result would be expected in a patient with metastasis compared with the likelihood that the same result would be expected in a patient without metastasis. It has advantages over sensiVOL. 137, NO. 2

LIVER TESTS

IN

tivity and specificity because it is less likely to change with the prevalence of the disorder. For these calculations, the LFT levels of nonmetastatic uveal melanoma patients were used as controls (all LFTs of the control group were included). To determine at what level should an LFT be considered indicative of metastases, four different cutpoints were defined at 60%, 80%, 100%, and 120% of the LFT’s upper normal limit, and the sensitivity, specificity, and likelihood ratio of prediction at each cut-point was plotted.

RESULTS THE 30 METASTATIC PATIENTS INCLUDED IN THIS STUDY

were followed every 8.9 ⫾ 3.1 month (mean ⫾ standard deviation) for a mean period of 30.1 ⫾ 10.7 months, whereas the control group was followed every 8.2 ⫾ 2.4 months for a mean period of 75.4 ⫾ 20.2 months. At the time of diagnosis of liver metastasis, 15 (50%) patients had at least one abnormal LFT (compared with only 4 [5%] in the control group throughout the whole follow-up period). None of these were symptomatic. All patients (in both groups) had at least one US examination. Liver CT was performed in 26 (87%) metastatic patients (for confirmation of the US findings) and was in agreement with the US findings in all cases. Four patients (13%) had pulmonary metastases that were diagnosed by either chest X-ray or chest CT. Six metastatic patients (20%) had at least one abnormal LFT at a time when liver US was still interpreted as normal. All those six patients developed US-proven liver metastases within 6 months of the abnormal LFT detection. Figure 1 presents the mean value of each LFT (⫾ standard error [SE]) at various time periods prior and subsequent to the detection of metastases by liver imaging. Figure 2 presents the mean change (⫾ SE) in each LFT’s value, relative to a preceding test, at the various time periods. While no statistically significant changes were noted in mean serum levels of bilirubin pre- and postdetection of metastases, mean AST started to rise (within the normal limits) 6 months before metastases detection and peaked 6 months after the detection. Mean ALT levels were above normal limits only 6 months after the detection, although a nonstatistically significant increase in mean ALT (of 17.5 U/l) was already detected 6 months before metastases. The LDH, ALK, and ␥GT seem to rise 6 months before the detection of metastases to reach a peak at the time of detection. While 6 months before metastases detection mean ALK levels were increased but still within normal limits, mean LDH and ␥GT were both above the upper normal limit. This is also expressed in Figure 2, where ALK shows the largest increase at the time of diagnosis of metastases, whereas ␥GT and LDH present their largest increase during the preceding 6 months.

METASTATIC UVEAL MELANOMA

237

FIGURE 1. Mean level of each liver function test (ⴞ standard error) at various time periods prior and subsequent to the detection of metastases by liver imaging. The laboratory upper normal limit (continuous line) and the mean level in the control group (dashed line) are marked. ALK ⴝ alkaline-phosphatase; ALT ⴝ alanine-aminotrasferase; AST ⴝ aspartate-aminotrasferase; LDH ⴝ lactate dehydrogenase; ␥GT ⴝ ␥-glutamyltransferase.

Figure 3 presents the sensitivity of each LFT in detecting metastases at various time periods. The sensitivity was calculated at various cut-points ranging from 60% to 120% of the upper normal limit. As expected, the lower the cut-point the higher the test’s sensitivity. However, for a test to forecast metastases, it has to show a rise in sensitivity before the detection of liver metastases. 238

AMERICAN JOURNAL

Whereas the sensitivity of bilirubin did not change throughout the follow-up period, the sensitivity of the other LFTs (at all cut-points except 60%) showed a constant rise towards a peak at the time (ALK, ␥GT, and LDH) or 6 months after (AST and ALT) the detection of metastases. Whereas ALK and ␥GT sensitivity started rising 3 years before metastases detection, LDH sensitivity OF

OPHTHALMOLOGY

FEBRUARY 2004

FIGURE 2. Mean change (ⴞ standard error) in each liver function test level, relative to the preceding test, at various time periods prior and subsequent to the detection of metastases by liver imaging. ALK ⴝ alkaline-phosphatase; ALT ⴝ alanine-aminotrasferase; AST ⴝ aspartate-aminotrasferase; LDH ⴝ lactate dehydrogenase; ␥GT ⴝ ␥-glutamyltransferase.

started rising only 6 months before. Figure 4 presents the LR of forecasting liver metastases for each LFT; ALK and LDH presented the best LR six months before metastases detection (ALK: LR ⫽ 27 and LDH: LR ⫽ 26) and at the VOL. 137, NO. 2

LIVER TESTS

IN

time of metastases detection (ALK: LR ⫽ 70 and LDH: LR ⫽ 42). Whereas AST and LDH reached their best LR at a cut-point of 80% of the upper normal limit, the other LFTs reached their best LR at the upper normal limit.

METASTATIC UVEAL MELANOMA

239

FIGURE 3. The sensitivity of each liver function test in forecasting metastases at various time periods. The sensitivity was evaluated at four different cut-points: 60% (black circles), 80% (white circles), 100% (black triangles), and 120% (white triangles) of the liver function tests upper normal limit. ALK ⴝ alkaline-phosphatase; ALT ⴝ alanine-aminotrasferase; AST ⴝ aspartate-aminotrasferase; LDH ⴝ lactate dehydrogenase; ␥GT ⴝ ␥-glutamyltransferase.

240

AMERICAN JOURNAL

OF

OPHTHALMOLOGY

FEBRUARY 2004

FIGURE 4. The likelihood ratio of each liver function test in forecasting metastases at various time periods. The likelihood ratio was evaluated at four different cut-points: 60% (black circles), 80% (white circles), 100% (black triangles), and 120% (white triangles) of the liver function tests laboratory upper normal limit. ALK ⴝ alkaline-phosphatase; ALT ⴝ alanine-aminotrasferase; AST ⴝ aspartate-aminotrasferase; LDH ⴝ lactate dehydrogenase; ␥GT ⴝ ␥-glutamyltransferase.

DISCUSSION ROUTINE SCREENING OF ASYMPTOMATIC PATIENTS WITH

uveal melanoma is necessary for early diagnosis and treatment of the metastatic disease. In recent years, the VOL. 137, NO. 2

LIVER TESTS

IN

importance of early diagnosis has come into focus because new treatment strategies, such as hepatic intra-arterial chemotherapy,16 are being explored. Recent reports have shown that only 26% of patients are diagnosed due to symptoms, between screening visits, and this is on median

METASTATIC UVEAL MELANOMA

241

We reported here that only three patients had metastases detected by routine chest X-ray. Because the yield of chest X-ray was found to be very low as well, in other studies, some centers have already abandoned it as a screening test.10 In conclusion, despite the limitations of this study, such as the small number of cases and the retrospective analysis, we showed that mean AST, ALT, ALK, LDH, and ␥GT rose 6 months before detection of metastases. Monitoring the changes in these enzymes relative to previous examinations (even within normal limits), and not relying solely on their absolute values, could help predict metastases. Based on likelihood ratios, ALK and LDH are the best predictive tests; LDH and AST should be considered predictive when their levels rise above 80% of the laboratory upper normal limit, whereas ALK and ␥GT are predictive only above the upper normal limit. As 50% of patients with known metastasis had normal LFTs and 20% of patients with metastasis had at least one abnormal LFT at a time when liver US was still interpreted as normal, the best method of screening for hepatic metastasis would probably be a combination of LFT and liver imaging. In absence of effective therapy for metastatic uveal melanoma, the role of screening might be controversial. However, as in recent years new treatment avenues are being explored,14,15 an effective screening program for uveal melanoma might be imperative in the future.

10.7 months after a previous screening visit.10 Thus, in the majority of patients with metastatic uveal melanoma, the presence of metastases is discovered by routine screening before the appearance of overt symptomatology, making an efficient screening strategy most important for early detection. Hicks and associates9 reported a general low yield of all liver screening tests for detecting uveal melanoma metastases. They found the most sensitive liver function tests to be ␥GT (sensitivity 21%) and ALK (sensitivity 25%). However, Eskelin and associates10 have found that at least one LFT was abnormal in 70% of the metastatic patients. In comparison, the ability of liver US to diagnose all existing liver metastases was reported earlier as being 37%.5 Because many patients have more than one liver metastasis, liver US was shown to detect more than 78% of the metastatic uveal melanoma cases.10 We have shown here that 20% (6 of 30) of the metastatic uveal melanoma patients had at least one abnormal LFT at the time that liver US was still interpreted as normal. Thus, a correct follow-up and interpretation of the LFTs could help in early suspicion of metastatic uveal melanoma before it reached the minimal size detectable by liver US. In such cases additional imaging with CT scan or magnetic resonance imaging must be considered, as there is high possibility of metastasis. We described here the mean levels and the changes in each LFT at different time points before detection of metastases by imaging. None of the mean LFT levels showed any abnormality more than 6 months before detection of metastases. However, during the half-year before metastases were detected, some LFTs (ALK, ␥GT, LDH, AST, and ALT) seemed to rise, mostly within normal limits. Mean LDH and ␥GT were even above the upper normal limit. These findings correspond with previous reports of the higher sensitivity of these enzymes.8 –11 Because those enzymes signify cholestasis, one might assume that diffuse micrometastases are present in the liver up to 6 months before their detection, causing a restriction of the bile outflow. Because ␥GT, LDH, and ALK can also reflect the tumor burden, they might have not only a diagnostic but also a prognostic value.5 Although increased LFTs in our patients could also be the result of other common factors, such as cholelithiasis, viral or alcoholic hepatitis, cirrhosis or drug-induced, the fact that the metastatic patients did not have increased mean LFTs more than 6 months before development of liver metastases together with the fact that in the control group only 5% of the LFTs were abnormal would suggest that the increased mean LFTs levels noted 6 months before detection of liver metastases are probably related to the metastatic process. The apparent decrease in the mean serum levels of some LFTs 6 to 12 months after the detection of metastases might symbolize either liver failure or a selection bias induced by the death of the more severe cases. 242

AMERICAN JOURNAL

REFERENCES 1. Mooy CM, De Jong PT. Prognostic parameters in uveal melanoma: A review. Surv Ophthalmol 1996;41:215–228. 2. Diener West M, Earle JD, Fine SL, et al. The COMS randomized trial of iodine 125 brachytherapy for choroidal melanoma, III: initial mortality findings. COMS Report No 18. Arch Ophthalmol 2001;119:969 –982. 3. McLean IW, Foster WD, Zimmerman LE. Uveal melanoma: Location, size, cell type, and enucleation as risk factors in metastasis. Hum Pathol 1982;13:123–132. 4. Singh AD, Shields CL, Shields JA. Prognostic factors in uveal melanoma. Melanoma Res 2001;11:255–263. 5. Leyvraz S, Spataro V, Bauer J, et al. Treatment of ocular melanoma metastatic to the liver by hepatic arterial chemotherapy. J Clin Oncol 1997;15:2589 –2595. 6. Einhorn LH, Burgess MA, Gottlieb JA. Metastatic patterns of choroidal melanoma. Cancer 1974;34:1001–1004. 7. Zakka KA, Foos RY, Omphroy CA, Straatsma BR. Malignant melanoma: Analysis of an autopsy population. Ophthalmology 1980;87:549 –556. 8. Donoso LA, Berd D, Augsburger JJ, et al. Metastatic uveal melanoma: Pretherapy serum liver enzyme and liver scan abnormalities. Arch Ophthalmol 1985;103:796 –798. 9. Hicks C, Foss AJ, Hungerford JL. Predictive power of screening tests for metastasis in uveal melanoma. Eye 1998; 12:945–948. 10. Eskelin S, Pyrhonen S, Summanen P, et al. Screening for metastatic malignant melanoma of the uvea revisited. Cancer 1999;85:1151–1159. OF

OPHTHALMOLOGY

FEBRUARY 2004

11. Char DH. Metastatic choroidal melanoma. Am J Ophthalmol 1978;86:76 –80. 12. Sutherland CM, Chmiel JS, Haik BG, et al. Patient characteristics, methods of evaluation and treatment of ocular melanoma in the United States for the years 1981 and 1987. Surg Gynecol Obstet 1993;177:497–503. 13. Collaborative Ocular Melanoma Study Group. Design and methods of a clinical trial for a rare condition: The Collaborative Ocular Melanoma Study. COMS Report No. 3. Control Clin Trials 1993;14:362–391.

14. Aoyama T, Mastrangelo MJ, Berd D, et al. Protracted survival after resection of metastatic uveal melanoma. Cancer 2000;89:1561–1568. 15. Woll E, Bedikian A, Legha SS. Uveal melanoma: Natural history and treatment options for metastatic disease. Melanoma Res 1999;9:575–581. 16. Egerer G, Lehnert T, Max R, et al. Pilot study of hepatic intraarterial fotemustine chemotherapy for liver metastases from uveal melanoma: A single-center experience with seven patients. Int J Clin Oncol 2001;6:25–28.

REPORTING VISUAL ACUITIES The AJO encourages authors to report the visual acuity in the manuscript using the same nomenclature that was used in gathering the data provided they were recorded in one of the methods listed here. This table of equivalent visual acuities is provided to the readers as an aid to interpret visual acuity findings in familiar units.

Table of Equivalent Visual Acuity Measurements Snellen Visual Acuities 4 Meters

6 Meters

20 Feet

Decimal Fraction

LogMAR

4/40 4/32 4/25 4/20 4/16 4/12.6 4/10 4/8 4/6.3 4/5 4/4 4/3.2 4/2.5 4/2

6/60 6/48 6/38 6/30 6/24 6/20 6/15 6/12 6/10 6/7.5 6/6 6/5 6/3.75 6/3

20/200 20/160 20/125 20/100 20/80 20/63 20/50 20/40 20/32 20/25 20/20 20/16 20/12.5 20/10

0.10 0.125 0.16 0.20 0.25 0.32 0.40 0.50 0.63 0.80 1.00 1.25 1.60 2.00

⫹1.0 ⫹0.9 ⫹0.8 ⫹0.7 ⫹0.6 ⫹0.5 ⫹0.4 ⫹0.3 ⫹0.2 ⫹0.1 0.0 ⫺0.1 ⫺0.2 ⫺0.3

From Ferris FL III, Kassoff A, Bresnick GH, Bailey I. New visual acuity charts for clinical research. Am J Ophthalmol 1982;94:91–96.

VOL. 137, NO. 2

LIVER TESTS

IN

METASTATIC UVEAL MELANOMA

243