The application of markers (HSP70 GPC3 and GS) in liver biopsies is useful for detection of hepatocellular carcinoma

Background/Aims

Liver biopsy for hepatocellular carcinoma (HCC) detection is largely restricted to small hepatocellular lesions, which are often morphologically challenging, requiring careful distinction between dysplastic nodules (high-grade) and well-differentiated HCC.

Methods

We investigated the diagnostic accuracy of a panel of markers (HSP70 GPC3 and GS), previously tested in resection specimens, in a series of liver biopsies of large regenerative nodules (n = 13), low-grade dysplastic nodules (n = 21), high-grade dysplastic nodules (n = 50), very well-differentiated (VWD) (n = 17), well-differentiated (WD-G1) (n = 40) and G2-3 (n = 35) HCC.

Results

Almost all cases of large regenerative and low-grade dysplastic nodules did not stain while high-grade dysplastic nodules showed 1 marker (22%) but never 2 or 3. For HCC detection the overall accuracy of marker combination was 60.8% (3 markers) and 78.4% (2 markers) with 100% specificity. When restricted to VWD + WD-G1 HCC the accuracy was 57% (3 markers) and 72.9% (2 markers) with 100% specificity.

Conclusions

This panel proved useful to detect well-differentiated HCC in biopsy. Two immunoreactive markers (out of 3) are recommended as the most valuable diagnostic combination for HCC detection. The diagnostic accuracy of the panel could be improved using additional markers, as suggested by studies of expression profiling in other human models.

Introduction

An increasing number of small hepatocellular nodules (<3 cm) are currently detected by imaging during the follow-up of cirrhotic patients. To identify small and early hepatocellular carcinoma (HCC), recent international guidelines have established the need to biopsy hepatic nodules 1–2 cm in size with controversial features (after 2 imaging tests) and nodules larger than 2 cm with dubious imaging (after 1 imaging test) [1]. In other words, today liver biopsy is recommended only for challenging diagnoses of small hepatocellular nodules not satisfactorily addressed by imaging, whose sensitivity for small HCC detection is 33% [2], [3]. Histologic diagnosis is often very difficult because it is based solely on the analysis of tiny fragments to distinguish between early/well-differentiated HCC and dysplastic nodules (both low- and high-grade) [4], [5], [6], [7]. It has been suggested that this is a grey area likely characterized by the progressive merging of malignancy within regenerative and dysplastic nodules [7]. The morpho-phenotypical details of the process of human hepatocarcinogenesis are mostly unknown because they are largely based on the analysis of explanted HCC, an advanced and end-stage setting of cirrhosis, which is likely different from the cirrhotic background of patients undergoing US surveillance.

In addition to the above indications, oncologists have started to encourage a histological diagnosis even in obvious cases of HCC so as to identify those patients who are candidates for treatment because of the recent availability of a putatively effective drug against HCC [8]. The current scenario is therefore characterized by a growing number of liver biopsies performed to gather evidence for starting treatment in individual patients. While the diagnosis is sometimes very straightforward, not requiring anything other than H&E, in most cases it requires the proper evaluation of several fine histological features, the reticulin framework, and the type of (neo)vascularization [4], [6], [7]. A number of cases (especially for general pathologists) still remain very difficult and result in diagnostic delays or the need to re-biopsy. Therefore, there is a current need for markers of HCC to aid in making the pathological diagnosis.

Several markers have been proposed to be used for the differential diagnosis of hepatocellular nodules, including Glypican 3, Heat Shock Protein 70 and Glutamyne Synthetase [9], [10], [11], [12], [13]. We previously tested this panel in surgically resected specimens [14]. When at least 2 of these markers were positive the sensitivity and specificity for early/very well-differentiated HCC detection was very good (72% and 100%, respectively) [14]. However, given the focal distribution of hepatocellular immunoreactivity in several cases, we wondered whether the diagnostic accuracy would also be retained in liver biopsies, where, as recently emphasized [2], no validation studies are available.

The present work was undertaken to test the diagnostic accuracy of this panel (HSP70 GPC3 and GS) in a large series of liver biopsies of small hepatocellular nodules gathered from different institutions. The aim was not merely to compare the results obtained from surgical specimens and from liver biopsies, but rather to evaluate whether the application of this panel on liver biopsies can actually increase its diagnostic power for the detection of HCC in the daily practice of diagnostic pathology.