Background. Several risk factors predict clinical outcome in gastro-entero-pancreatic neuroendocrine neoplasms (GEPNENs); however, the impact of their combination has not been investigated so far. Patients and Methods. A retrospective analysis of stage IV GEPNENs was performed. Multivariate analysis for progression of disease (PD) was performed by Cox proportional hazards method to obtain a risk score. Area under the curve obtained by receiver operating characteristic analysis was used to assess the score performance. Progression-free survival analysis was performed by Kaplan-Meier method. Results. Two hundred eighty-three stage IV GEP-NENs were evaluated, including 93 grade 1 neuroendocrine tumors (32.9%), 153 grade 2 neuroendocrine tumors (54%), and 37 grade 3 neuroendocrine carcinomas (13.1%). Independent risk factors for PD were Ki67, proportion of metastatic liver involvement, and presence of extra-abdominal metastases. The risk score was calculated as follows: (0.025 × Ki67)+[(0 if no liver metastases or liver involvement 50%)]+[(0 if no extra-abdominal metastases) OR (0.528 if extra-abdominal metastases present)]. The risk score accuracy to predict PD was superior compared with the G grading system (area under the curve: 0.705 and 0.622, respectively). Three subgroups of patients with low, intermediate, and high risk of PD according to risk score were identified, median progression-free survival being 26 months, 19 months, and 12 months, respectively. Conclusion. In stage IV GEP-NENs, a risk score able to predict PD was obtained by combining Ki67, proportion of metastatic liver involvement, and presence of extra-abdominal metastases. The score may help to discriminate patients with different progression risk level to plan tailored therapeutic approaches and follow-up programs. © AlphaMed Press 2017.