Objective: The efficacy of the capecitabine/temozolomide (CAPTEM) regimen has been demonstrated in metastatic neuroendocrine neoplasms (NENs), but because of varying response rates among the patients, biomarkers to predict its response are greatly needed. Here, we investigated the clinical utility of a Ki-67 index to predict the CAPTEM regimen objective responses and select patients who could benefit from this regimen. Methods: Metastatic NENs patients treated with the CAPTEM regimen from 4 high-volume medical centers were selected and grouped in a training and validation cohort. The classification and regression tree (CART) was generated to identify the optimal threshold of Ki-67 for stratifying the patients into different Ki-67 range groups based on their response to the CAPTEM regimen. Results and Conclusions: The overall response rate (ORR) and disease control rate of the entire cohort (N = 151) were 26.5 and 76.2%, respectively, with a median progression-free survival (PFS) of 12.0 months. CART analysis showed that patients in the Ki-67 range group 10-40% demonstrated a significantly higher ORR than those in Ki-67 >40 and <10% groups (p < 0.001 in the training cohort and p = 0.036 in the validation cohort). Response to the CAPTEM regimen was not influenced by the expression of O-6-methylguanine-DNA methyltransferase or primary tumor location. Multivariate analysis identified the Ki-67 index as the only independent prognostic factor for overall survival (p = 0.031) and PFS (p = 0.006). The proposed Ki-67 index was externally validated and could be used to clinically identify suitable metastatic NENs patients who could achieve an optimal cytoreduction using the CAPTEM regimen.