Utility of clinico-biological data for long-term prognosis of head and neck terminal cancer

Objective: The prognosis of terminal head and neck cancer is traditionally determined using indices such as the palliative prognosis index. We aimed to develop an alternative prognosis index using clinico-biological data. Methods: This retrospective case-series study included 33 head and neck cancer patients whose cancer recurred despite receiving radiation therapy between April 2010 and April 2014. Clinico-biological data were collected the day patients were diagnosed as terminal. Bivariate correlation analyses were performed on survival times and clinico-biological data. For multivariate regression analyses, patients were divided into two groups: (1) patients who survived > 120 days, and (2) those who survived < 120 days. Group clinico-biological data were used to determine survival-time cutoff points for the prognosis index. Results: Bivariate analyses revealed significant correlations between survival time and BMI, hemoglobin, albumin, C-reactive protein (CRP), Onodera's Prognostic Nutritional Index (O-PNI), modified Glasgow Prognostic Score (mGPS), and Performance Status (PS). Multivariate analyses showed a strong correlation between survival time and BMI. The two groups differed significantly in BMI, albumin, CRP,O-PNI, and mGPS. In multivariate analyses BMI, CRP, and O-PNI differed significantly in the two groups. Conclusion: BMI <= 16.4, CRP >= 1.01, and O-PNI <= 33.4 are significant predictors of long-term survival in terminal cancer patients.