Purpose: To design an alternative workflow for cancer-risk assessment to predict distant metastasis (DM) of nasopharyngeal carcinoma (NPC). Methods: We enrolled 234 patients with non-disseminated NPC and a family history of cancer who underwent intensity-modulated radiotherapy and concurrent chemo-radiotherapy with/without induction chemotherapy in our primary cohort. We conducted univariate and multivariate analyses of the associated prognostic factors, built a nomogram model for distant-metastasis-free survival (DMFS), and confirmed the prognostic value of weight-loss ratio (WTratio). The secondary cohort included 97 patients with available pre-DNA levels who were treated at our cancer center. We performed internal validation with the primary cohort and external validation with the secondary cohort, and compared the new DMFS model with the current 7th TNM staging system. Results: In the primary cohort, 95.9% patients experienced weight loss. The N group (N2-3 vs. N0-1, P = 0.037) and pre-DNA level (P = 0.02) were independent prognostic factors for DMFS in NPC patients. Smoking (P = 0.051) and WTratio (P = 0.052) showed a significant trend for DMFS. WTratio was an independent prognostic factor for DMFS (P = 0.03). Smoking, WTratio, N group, and pre-DNA level were merged to build a risk-score model for DMFS using a nomogram, which could predict survival after internal and external validation. Conclusions: Maintaining body weight during treatment is essential to prevent DM of NPC. Compared with the current 7th TNM staging system, the new DMFS model might better predict DM of NPC. The alternative workflow designed could be applied for prognostic analysis of other cancers.