LAUSR

Nasal-type, extranodal nature killer (NK)/T-cell lymphoma-associated hemophagocytic syndrome (NK/T-LAHS) is a rare and life-threatening disease, requiring investigation of risk stratification. We conducted a retrospective study and proposed nomograms to predict NK/T-LAHS. The discriminative ability and calibration of the nomograms for prediction were tested using C statistics and calibration plots. We analyzed 533 patients with extranodal NK/T-cell lymphoma (ENKTL), out of which 71 were diagnosed with hemophagocytic syndrome (HPS), with a cumulative incidence of 13.3%. Significant difference for 2-year survival was found between patients with and without HPS (14.7% vs. 77.5%). Analyses showed that Eastern Cooperative Oncology Group (ECOG) performance status (PS) ≥2, B symptoms, and bone marrow (BM) invasion were significantly associated with NK/T-LAHS. We used these data as the basis to establish a nomogram of risk index for ENKTL (RINK). In 335 patients with available data for Epstein-Barr virus DNA (EBV-DNA), we found high viral copies (≥4,450 copies/ml) were correlated with NK/T-LAHS. When these data were added to RINK, we developed another nomogram that included EBV-DNA data (RINK-E). The nomograms displayed good accuracy in predicting NK/T-LAHS with a C-statistics of 0.919 for RINK and a C-statistics of 0.946 for RINK-E, respectively. The calibration chart also showed an excellent consistency between the predicted and observed probabilities. The proposed nomograms provided individualized risk estimate of HPS in patients with ENKTL.