Clinical Features and Long-Term Survival of Metastatic Hepatic Neuroendocrine Neoplasms Secondary to Gastroenteropancreatic Site: An Analysis by Applying the Grading Classification

Abstract

Background and Purpose. Neuroendocrine neoplasms occurring in the liver are very rare, in which metastatic hepatic neuroendocrine neoplasms [(MH)-NENs] secondary to gastroenteropancreatic NENs [(GEP)-NENs] account for their majority. The clinical features and long-term survival of (MH)-NENs secondary to (GEP)-NENs were not clear, especially for each grading group of G1 neuroendocrine tumors (NETs), G2 NETs, and G3 NETs and G3 neuroendocrine carcinomas (G3 NECs). Method. Data of patients who were surgically treated and clinicopathologically diagnosed as (MH)-NENs secondary to (GEP)-NENs at West China Hospital of Sichuan University from January 2006 to December 2018 were retrospectively collected and analyzed by the grading classification for (GEP)-NENs. Results. We identified 150 patients with (MH)-NENs secondary to (GEP)-NENs, including 10 patients with G1 NETs, 26 with G2 NETs, 33 with G3 NETs, and 81 with G3 NECs. There were significant differences between patients with G1/G2/G3 NETs and those with G3 NECs, such as age at diagnosis (P=0.041), synchronous liver lesion (P=0.032), incidental diagnosis (P=0.014), tumor largest diameter (P=0.047), vascular invasion (P=0.017), and extrahepatic metastatic disease (P=0.029). The estimated 3-year overall survival for patients with G1 NETs, G2 NETs, G3 NETs, and G3 NECs was 100%, 79.4%, 49.5%, and 20.7%, respectively (P<0.001). The survival of G1 NETs or G2 NETs was significantly better than that of G3 NETs (P=0.013, P=0.037, respectively) and G3 NECs (P=0.001, P<0.001; respectively). Patients with G3 NECs present notably worse survival than those with G3 NETs (P=0.012), while survival comparison between G1 NETs and G2 NETs was not statistically different (P=0.131). The grading classification for (GEP)-NENs was an effective independent predictor of survival for (MH)-NENs secondary to (GEP)-NENs (hazard ratio: 4.234; 95% confidence intervals: 1.984–6.763; P=0.003). Conclusion. Our demonstration revealed that the grading classification for (GEP)-NENs could well stratify (MH)-NENs secondary to (GEP)-NENs into prognostic groups and supported its wide use in clinical practice.