PSMA immunohistochemistry as a diagnostic biomarker of hepatocellular carcinoma.
[BACKGROUND & AIMS] A combination of three immunohistological markers (Glypican 3, heat shock protein 70 [HSP70], and glutamine synthetase [GS]) is routinely used to differentiate hepatocellular carci
- 표본수 (n) 39
APA
Véron K, Becht E, et al. (2025). PSMA immunohistochemistry as a diagnostic biomarker of hepatocellular carcinoma.. JHEP reports : innovation in hepatology, 7(11), 101542. https://doi.org/10.1016/j.jhepr.2025.101542
MLA
Véron K, et al.. "PSMA immunohistochemistry as a diagnostic biomarker of hepatocellular carcinoma.." JHEP reports : innovation in hepatology, vol. 7, no. 11, 2025, pp. 101542.
PMID
41113122
Abstract
[BACKGROUND & AIMS] A combination of three immunohistological markers (Glypican 3, heat shock protein 70 [HSP70], and glutamine synthetase [GS]) is routinely used to differentiate hepatocellular carcinoma (HCC), but this panel's sensitivity is suboptimal. Our aim was to assess the diagnostic value of prostate-specific membrane antigen (PSMA) expression for diagnosing HCC in a series of hepatocellular nodules and compare its performance with that of routinely used markers.
[METHODS] We included 320 hepatocellular nodules from 188 patients in a test cohort and 87 hepatocellular nodules from 48 patients in an external validation cohort distributed as follows: regenerative nodules (RN, n = 39+22), low-grade dysplastic nodules (LGDN, n = 38+16), high-grade dysplastic nodules (HGDN, n = 30+8), early HCC (≤2-cm nodules) (n = 107+24), HCC (n = 106+17), and corresponding non-tumour livers (NTL, n = 152+37). PSMA, HSP70, Glypican 3, and GS expression was assessed by immunohistochemistry on tissue microarrays. For each marker or combination of markers, sensitivity, specificity, and accuracy were calculated.
[RESULTS] In the test cohort, PSMA was expressed in 83% of HCC (n = 88/106), 77% of early HCC (n = 82/107), 27% of HGDN (n = 8/30), 21% of LGDN (n = 8/38), 18% of RN (n = 7/39), and 3% of NTL (n = 5/152). In the validation cohort, the sensitivity and specificity of PSMA for HCC diagnosis were 0.95 and 0.77, respectively, and its accuracy was 0.83. The sensitivity and the specificity of the Glypican 3-HSP70-GS (≥2 positive markers) combination for HCC diagnosis were 0.41 and 0.99, respectively, and its accuracy was 0.80. Adding PSMA to this combination increased the sensitivity and accuracy to 0.85 and 0.86, respectively.
[CONCLUSIONS] PSMA alone has shown good performance in diagnosing HCC, outperforming the combination of the three routinely used markers. When sufficient material is available, adding Glypican 3, HSP70, and GS to PSMA could be recommended.
[IMPACT AND IMPLICATIONS] Differentiating hepatocellular nodules, particularly high-grade dysplastic nodules and hepatocellular carcinoma (HCC), based on histologic criteria remains challenging. In this study, we assess the diagnostic value of a new immunohistochemical marker, prostate-specific membrane antigen (PSMA), for diagnosing HCC in two independent series of hepatocellular nodules and compare its performance with that of routinely used markers (Glypican 3, heat shock protein 70 [HSP70], and glutamine synthetase [GS]). PSMA alone has demonstrated good performance in diagnosing HCC, superior to the combination of the three routinely used markers, and could be useful in practice for differentiating difficult-to-classify hepatocellular nodules. When the material is sparse, using PSMA alone could be recommended, whereas when sufficient material is available, adding PSMA to Glypican 3, HSP70, and GS may be advised, as this combination has shown the best performance for HCC diagnosis.
[METHODS] We included 320 hepatocellular nodules from 188 patients in a test cohort and 87 hepatocellular nodules from 48 patients in an external validation cohort distributed as follows: regenerative nodules (RN, n = 39+22), low-grade dysplastic nodules (LGDN, n = 38+16), high-grade dysplastic nodules (HGDN, n = 30+8), early HCC (≤2-cm nodules) (n = 107+24), HCC (n = 106+17), and corresponding non-tumour livers (NTL, n = 152+37). PSMA, HSP70, Glypican 3, and GS expression was assessed by immunohistochemistry on tissue microarrays. For each marker or combination of markers, sensitivity, specificity, and accuracy were calculated.
[RESULTS] In the test cohort, PSMA was expressed in 83% of HCC (n = 88/106), 77% of early HCC (n = 82/107), 27% of HGDN (n = 8/30), 21% of LGDN (n = 8/38), 18% of RN (n = 7/39), and 3% of NTL (n = 5/152). In the validation cohort, the sensitivity and specificity of PSMA for HCC diagnosis were 0.95 and 0.77, respectively, and its accuracy was 0.83. The sensitivity and the specificity of the Glypican 3-HSP70-GS (≥2 positive markers) combination for HCC diagnosis were 0.41 and 0.99, respectively, and its accuracy was 0.80. Adding PSMA to this combination increased the sensitivity and accuracy to 0.85 and 0.86, respectively.
[CONCLUSIONS] PSMA alone has shown good performance in diagnosing HCC, outperforming the combination of the three routinely used markers. When sufficient material is available, adding Glypican 3, HSP70, and GS to PSMA could be recommended.
[IMPACT AND IMPLICATIONS] Differentiating hepatocellular nodules, particularly high-grade dysplastic nodules and hepatocellular carcinoma (HCC), based on histologic criteria remains challenging. In this study, we assess the diagnostic value of a new immunohistochemical marker, prostate-specific membrane antigen (PSMA), for diagnosing HCC in two independent series of hepatocellular nodules and compare its performance with that of routinely used markers (Glypican 3, heat shock protein 70 [HSP70], and glutamine synthetase [GS]). PSMA alone has demonstrated good performance in diagnosing HCC, superior to the combination of the three routinely used markers, and could be useful in practice for differentiating difficult-to-classify hepatocellular nodules. When the material is sparse, using PSMA alone could be recommended, whereas when sufficient material is available, adding PSMA to Glypican 3, HSP70, and GS may be advised, as this combination has shown the best performance for HCC diagnosis.