Prediction of the Development of Hepato-Cellular-Carcinoma in Patients with Liver Cirrhosis by the Serial Determinations of Serum Alpha-L-fucosidase Activity

Suzuki**, Objective Evaluation of the usefulness of the serial daterminations of serum alpha-L-fucosidase (AFU) activity for prediction of the development of hepatocellular carcinoma (HCC) was performed. Methods and Patients Serum AFUactivity was determined monthly for 42 months in 73 patients with liver cirrhosis (LC). Results HCCwas diagnosed in 27 patients by means of ultrasonography during this observation period. In 23 (85%) of the 27 patients, serum AFUactivity was found to exceed 700 nmole/ml/h during the LC stage. HCCdeveloped within a few years in 23 (82%) of 28 LC patients with AFUactivity exceeding 700 nmole/ml/h, in contrast, it developed in only 4 (9%) of 45 LC patients with AFUactivity below 700 nmole/ml/h. AFUactivity was already elevated in 23 (85%) of 27 patients at least 6 months before the detection of HCCby ultrasonography. Conclusion It is conceivable that the development of HCCcan be predicted by means of serial determinations of serum AFUactivity in patients with LC. (Internal Medicine 927-931, 1999)


Introduction
Elevated levels of alpha-L-fucosidase (AFU) have been reported in patients with hepatocellular carcinoma (HCC) (1)(2)(3)(4)(5)(6)(7)(8). Wehave reported the usefulness of the determination of serum AFUactivity as a new tumor marker in the diagnosis of HCC (9). Previously, patients with an AFUactivity exceeding 700 nmole/ml/h who had not developed HCCwere regarded as false-positives; it later became evident that HCCdeveloped at a high incidence in these patients. Therefore, it is conceivable that such cases are not false-positives but actually are in the very early stages of HCC.To substantiate this assumption, a prospective study is needed. The present study therefore, was conducted to evaluate the usefulness of serial determinations of serum AFUactivity in LCpatients for the prognostication of developing HCC. Weattempted to clarify whether the development of HCCcan be predicted from the dynaics of serum AFUactivities in patients with LC, and if so, whether prediction by this method can be madeearlier than the detection of HCCwith ultrasonography.

Subjects and Methods
Serum AFUactivities were determined at monthly intervals in 73 LC patients(HCC was not yet detectable by ultrasonographies performed by 2 physicians with at least 10 years of experience with abdominal ultrasonogrraphical diagnostic techiniques). Determinations were performed in the Clinic of Internal Medicine of Surugadai Nihon University Hospital, Tokyo, from April 1994 to September 1997. These patients were followed by ultrasonography (Aloca) every 3 months for from 8 months to 3.5 years. Subjects consisted of58 patients with C type LC, 12 with B type LC, and 3 with non-B non-C type LC. Serum AFUactivity was assayed by a modified method of Zielke et al (10). In brief, 10 |Lil of serum was added to 100 jLtl of 1 mmole/L p-nitrophenyl-oc-Lfucopyranoside dissolved in 150 mmole// citrate-phosphate buffer (pH 5) containing 0.05 mg/ml BSA. This solution was incubated at 37°C for 1 hour, and the reaction was stopped by the addition of 100 |il of 0.2 N NaOH. The absorbance of pnitrophenol was read at 405 nm, and the enzymeactivity was expressed as nano-moles of p-nitrophenyl-a-L-fuco-pyrano-side cleaved per milliliter per hour at 37°C. Statis tical analysis Data were analyzed by Student's unpaired t-test and differences were considered significant at p<0.05.

Results
During the observation period, HCCdeveloped in 27 out of 73 patients with LC. In 23 (85%) of those 27 cases, AFUactivities were already elevated up to 700 nmole/ml/h, during the LC stage. Though the incidence of HCCwas very high in LC patients with an AFU activity exceeding 700 nmole/ml/h (23 (82%) of 28 patients), HCCdeveloped in only 4 (9%) of45 LC patients with AFU activity below 700 nmole/ml/h (Fig. 1). Figure 2 shows the chronological changes in AFUactivity in patients with LC who developed HCC, during the 42-month observation period: In 23 (85%) of 27 LC patients, AFUactivity increased to 700 nmole/ml/h at least 6 months before HCC Figure 1. Relation between the incidence of development of HCCand AFUactivity during the LC stage. The incidence of development of HCCwas significantly higher in LC patients with an AFUactivity exceeding 700 nmole/ml/h than in those with an AFU activity below 700 nmole/ml/h. à": Accompanying HCC, O: No accompanying HCC. AFU: alpha-L-fucosidase, HCC:hepatocellular carcinoma, LC: liver cirrhosis *: p<0.01 was detected by ultrasonography. In the remaining 4 patients, AFUactivity did not rise until HCCwas detected. Elevation of AFUactivity occurred 3 years or more before the detection of HCCin 4 patients, 2 years or more before in 7 patients, 1 year or more before in 17 patients, and 6 months or more before in 23 patients. AFUactivity stayed at high levels after the initial elevation in 17 of 23 patients, but in the remaining 6 cases fell to less than 700 nmole/ml/h, and then rose again before the detection of HCC.In 3 of4 patients who maintained low levels of AFUactivity and accompanying HCC, AFUactivity was below 700 nmole/ml/h, when the HCCwas detected. The dynamics ofAFU activity in the LC patients who did not develop HCCduring the observation period are shown in Fig. 3. Except for 3 of the 45 cases (indicated by #-#), the AFUactivities were maintained at low levels below 700 nmole/ ml/h. Figure 4 shows the ultrasonographic images of HCC which had high levels ofAFU before being diagnosed as HCC. Eleven (48%) of the 23 HCCswere solitary and less than 20 mmin diameter.

Discussion
AFUis an exoglucosidase which exists in the liver, kidney, and epididymis, of mammalianspecies and catabolizes the fucose-containing sugar proteins ( 1 1).
The mechanism of the elevation of serum AFUactivity in patients with HCCis still obscure. One possibility is an increase in protein synthesis by the neoplasm with a consequent increase in fucose turnover (1). This possibility, however, is not supported by recent reports showing significantly low AFU levels in HCCtissue compared with nontumoral liver tissue (5), and similar levels ofAFUactivity in the culture supernatants of HCCcell lines and Chang liver cells (8). It is well knownthat sugar chains of sugar proteins vary in the course of carcinogenesis. In hepatocarcinogenesis, for example, fucosylation of some sugar proteins increases, and this increase of fucosylation maylead to the elevated alpha-Lfucosidase activity.
Whether this speculation is true or not, serum AFUactivity is significantly higher in patients with HCCthan in those with LC and chronic hepatitis, and we have reported the sensitivity of serum AFUactivity in the detection of HCCand its specificity of 96%, and 75% (cut-off level=700 nmole/ml/h). (unpublished personal data), this sensitivity is not sufficiently high to serve as a marker for HCC,however, HCCdid develop in LC patients with a high AFUactivity at a high incidence rate following the above report. These findings led us to add a prospective study for clarifying the relationship between serum AFUactivity and the development of HCCin LC patients. Oneof the purposes of this study was to detect whenserum AFUactivity starts to become elevate in patients with LC, a high risk group for HCC. Computed tomography, magnetic resonance imaging and ultrasonography have been used to detect HCC,with ultrasonography being the most practical and reliable technique to detect early HCC. However, it is still difficult to detect small isoechogenic HCC (12)(13)(14)(15). Although al-Early Diagnosis of HCC by AFU pha-fetoprotein (AFP) is the most useful marker for HCC, this marker is not very sensitive for the detection of early HCC (16,17). It has been reported that there are someisoforms of AFP differentiated by its difference of sugar chains (1 8-20). It was recently reported that the AFPisoform, as analyzed by lectin affinity immunoblotting, is useful for preclinical diagnosis of HCC (21,22). An AFP level exceeding 100 ng/ml is required to detect AFP isoforms, but AFPis usually not too highly elevated in the early stages of HCC.
In contrast, AFUactivity is not related to the tumor size (8). Further, in this study, AFUactivity rose before the HCCwas detectable by ultrasonography in about 85% of the patients with LC associated HCC. LC patients with AFUactivity exceeding 700 nmole/ml/h tended to develop HCCat a high incidence and in a short interval. Conversely, patients with LCwhomaintained low levels of AFUactivity infrequently developed HCC. It is unclear whyAFUactivity remained low in 3 cases of V\CCTt mav be due to changes of sugar chain of suear Drotein accompanying the development of HCCnot occurring in only fucose.
The fact that 48% of the HCCswhich were discovered by elevation ofAFUin the LCstage were the so-called small hepatic cancer, strongly suggests the usefulness of monitoring AFUlevels in LC patients in order to establish an early diagnosis of HCC. In conclusion, serial determinations of serum AFUactivity in patients with LC were helpful to predict the development of HCC.Physicians should recognize that patients with LCwith high AFUactivities are a super-high risk group for HCCand should perform diagnostic imaging studies at least once a month. If HCCis suspected in these patients, further examination (i.e. hepatic angiography, tumor biopsy etc.) must be performed. By doing so, HCCis likely to be detected at a very early stage, and it would be expected that the prognosis of HCC would improve.