Does gamma-glutamyltransferase correlate with liver tumor burden in neuroendocrine tumors?

The pathophysiology behind the rise in GGT levels in patients with liver metastasis has not been described yet. Interestingly, we did not observe a significant effect of radiologically detectable cholestasis on GGT levels. However, detectability of cholestasis on imaging depends on the imaging modality and is much lower on CT and PET-CT than on MRI. Furthermore, the rise of GGT could rather be driven by cholestasis in small bile ducts, undetectable in radiographic imaging. The tumor microenvironment of NETs has already been intensively researched. It could be shown that infiltrating immune cells mediate an immunosuppressive microenvironment [15, 16]. Therefore, peritumoral inflammation seems to be a less likely explanation for GGT elevations in NET liver metastases. A connection between GGT activity in the tumor’s membrane and hepatic tumor growth has been reported for melanoma cells [7]. As GGT expression in NET cells has not been investigated yet, increased cell turnover in large GGT-positive metastases would pose an explanation for the fact that GGT elevations are so common in our NET cohort. We therefore conducted a PCR analysis on GGT expression in four different NET cell lines compared to HEPG2-cells as a control sample. NET cells showed very low expression levels of GGT compared to HEPG2-cells (data not shown). Thus, NET cells do not seem to be the source of GGT elevations, it rather the surrounding liver tissue. Hence, NET metastases to other organs should not lead to elevated GGT.

Considering GGT as a biomarker for liver metastasis, no data for NETs were found. For renal cell carcinoma [9], ovarian cancer [6], and endometrial cancer [10], elevated serum GGT was shown to be of negative prognostic value in general. In patients with colorectal cancer, an initial decrease in GGT under therapy was associated with improved overall response and progression-free survival [17]. Yet, in the mentioned studies, there were no analyses on the presence or size of liver metastases. A correlation between GGT and liver tumor diameter and volume has so far only been demonstrated for hepatocellular carcinoma [8]. Liver metastases are common in a variety of solid neoplasms, particularly in GI cancer, and monitoring them is usually of great clinical importance. Our study demonstrates a significant association between GGT and liver tumor burden caused by metastases. This should also prompt further research in other oncologic entities. The clinical utility of GGT as a marker for liver tumor burden is favored by the fact that it is an established, easy-to-access and commonly performed test for various indications. It should be noted, however, that only half of our patients, whom all had liver metastases, showed an increase of GGT at the time of evaluation. Hence, GGT is not suited for ruling out liver metastases. If liver metastases are known though, normal GGT values might rule out a tumor burden >50%, according to our data. Greater clinical utility of GGT testing might, however, be achieved in serial testing over the course of treatment.

Due to the slow growth dynamics of NETs and consecutively long treatment and follow-up periods, there is a clinical need for therapy monitoring with laboratory markers. A whole range of biomarkers are known for NETs, although some are only applicable to specific entities [18]. Measurement of 5-hydroxyindoleacetic acid in either urine or plasma may be of use as a biomarker in functionally active NETs of the small intestine, yet it has not shown to be a reliable prognostic marker [19]. Chromogranin A (CgA) is a protein found in cells of neuroendocrine origin and has been the most widely used biomarker for neuroendocrine neoplasms in general to date [20]. It has shown to be predictive of disease-free survival after surgery as well as therapeutic response and is associated with a high liver tumor burden [21‐23]. A recent study found that CgA is associated with disease progression in pancreatic NETs and predictive of negative outcome in patients with small intestine or cecum NETs, however, the results were limited to these subgroups [24]. Regarding its role as a follow-up parameter, a review and meta-analysis showed that it has sufficient accuracy, especially when baseline values are impaired [25] Other authors conclude that the sensitivity and specificity of CgA are insufficient for its use as a clinical biomarker [26]. For example, it has been shown that CgA is also elevated in chronic liver diseases such as cirrhosis, hepatitis and hepatocellular carcinoma [27]. Furthermore, not all NETs reliably express and secrete CgA, limiting its use in routine clinical practice [28]. NETest is a novel diagnostic tool based on mRNA detection in the patient’s blood [29]. Recent studies have shown it to be of high diagnostic accuracy and predictive of disease progression or stable disease, respectively [30]. In comparison to CgA, it was found to be far more accurate in predicting therapy response or progression free survival [31]. However, NETest is still not in routine clinical use and costs are estimated to be very high (3000–4000$/year). Like GGT, alkaline phosphatase (AP) is a widely used laboratory marker for cholestasis. Studies have shown it to be a negative predictor of survival in patients treated for NETs [32, 33]. Another recent study retrospectively analyzing 49 NET patients confirmed the negative prognostic value of AP, but it found no correlation between AP levels and the quantity or size of metastases [34]. Since AP elevation was detected in only one in three patients and no correlation with the disease extent was found, it may be less suitable as a laboratory follow-up marker for NET patients.

To date, no study has focused on the potential role of GGT for therapy monitoring in patients with neuroendocrine liver metastases. However, since GGT testing is often performed, clinicians tended to attribute a rise in GGT with hepatic progression. Hence, our study poses an evaluation of an until then common clinical practice. The results support this assumption, showing that there is indeed an association between GGT and liver tumor burden. It should be noted, though, that low or moderate liver tumor burden may not be detected by GGT testing. In patients with a low liver tumor burden, the serial testing of GGT, therefore, seems to be helpful only insofar as an increase in GGT can indicate progression of the disease. Normal values do not exclude progression, as in our analyses the GGT values only rise reliably above a tumor burden of 50%. Accordingly, serial GGT tests may be of less utility in these patients. However, from a clinical point of view, close surveillance is of greater importance in patients with high tumor burden, as these patients have a worse prognosis [35]. In case of a sudden increase in GGT, liver tissue damage due to other conditions should be considered. Change dynamics might help to distinguish between liver tissue damage or malignant progression, as GGT due to NET progression appeared to increase slowly and steadily, matching the clinical course of the patients.

When directly comparing GGT and the current standard CgA as biomarkers for neuroendocrine tumor disease, there are some important points to consider. Whereas CgA is of prognostic value for intra- and extra-hepatic disease, GGT has only been evaluated for liver tumor burden. Using CgA as a biomarker is only possible in tumors expressing and secreting CgA. In contrast, GGT is not dependent on tumor-specific features and might be of value in all different kinds of NETs with liver involvement. However, a larger study addressing this point is warranted. As GGT determination is part of routine laboratory diagnostic, results are often immediately available, whereas CgA determination is restricted to specialized laboratories and is often performed only once a week or even less frequently, thus causing a delay in response to changing values.

There were several limitations to this study. This was a single-center study with relatively small sample sizes, especially in the longitudinal cohort. Due to the retrospective design, no causal conclusions can be drawn from the data. In addition, retrospective studies carry the risk of selection bias. Patients usually had multiple GGT testing with matching radiographic evaluation. The time of evaluation was therefore chosen individually for each patient, avoiding confounding factors such as liver or biliary duct interventions, operations, or other causes for GGT increase. However, we did not collect data on patients’ concomitant diseases (e.g., diabetes, metabolic syndrome) or the use of medication, which might have confounded the results. In addition, as GGT is not a specific lab test, there may have been unknown confounders facilitating GGT elevations. Imaging evaluation was done on a visual scale, a method which has been established in similar studies but to which a certain degree of subjectivity is inherent. This approach was chosen because a very large number of liver metastases were present in our patient collective, so volumetry of each one would not have been feasible. However, since readers were blinded to clinical information and laboratory findings, the subjectivity of the method does not create any systematic error. Regarding the longitudinal analysis, which was conducted exploratively, no direct conclusions can be drawn for clinical practice due to the small sample size. However, to our knowledge, this study is the first to investigate GGT as a biomarker for clinical follow-up of liver metastases.

The association between GGT elevations and liver metastases demonstrated in this study should raise physicians’ awareness of possible disease progression when detected in a routine examination. Further research in larger longitudinal series is required to assess the utility of GGT as a follow-up parameter. If confirmed in future studies, GGT can be implemented in clinical practice as a very cost-effective tool for therapy monitoring of liver metastases under systemic treatment for NETs to trigger radiographic evaluation, thereby allowing timely detection of disease progression and adaptation of therapy.