Competing risk nomograms for nasopharyngeal carcinoma in the intensity-modulated radiotherapy era: A big-data, intelligence platform-based analysis

Abstract

Background and purposeLacking quantitative evaluations of competing risk data of nasopharyngeal carcinoma (NPC), we aimed to evaluate the probability of NPC- and other cause-specific mortality (NPC-SM; OCSM) and develop competing risk nomograms to quantify survival differences. Material and methodUsing the institutional big-data intelligence platform, 7251 NPC patients undergoing intensity-modulated radiotherapy between 2009–2014 were identified to establish nomograms based on Fine and Gray’s competing risk analysis. ResultsThe 5-year NPC-SM and OCSM of the cohort were 13.1% and 1.2%, respectively, and elevated 5-year OCSMs were observed in patients aged ≥65 years (5.5%) or with severe comorbidities (4.3%). Age was most predictive of OCSM: patients aged 55–64 and ≥65 years exhibited subdistribution hazard ratios (SHRs) of 2.70 (95% confidence interval [CI], 1.64–4.4; P < .001) and 5.78 (95% CI, 3.32–10.08; P < .001), respectively. Comorbidity measured using the Charlson Comorbidity Index (CCI) was also strongly predictive of OCSM: patients with CCI scores of 1 and ≥2 exhibited SHRs of 2.33 (95% CI, 1.46–3.71; P < .001) and 2.58 (95% CI, 1.16–5.73; P = .020), respectively. All validated factors were integrated into the competing nomograms: age, sex, histology type, tumor and node stages, plasma Epstein–Barr virus-DNA level, lactate dehydrogenase level, and C-reactive protein (CRP) level into the NPC-SM model (concordance [c]-index = 0.743); and age, CCI, Albumin level, and CRP level into the OCSM model (c-index = 0.793). ConclusionOCSM represents a significant competing event for NPC-SM in elderly patients and patients with comorbidities. We present the first prognostic nomograms to quantify competing risks, which may help to tailor individualized treatment.