PURPOSE We aimed to investigate the prognostic factors that can predict terminal stage survival (TSS) at the time of terminal cancer diagnosis.

METHODS We prospectively evaluated 141 patients immediately after the diagnosis of terminal cancer by their attending oncologists. A total of 32 factors, including performance status, clinical prediction of survival, time to terminal cancer (TTC), clinical symptoms, signs, and laboratory tests including the neutrophil-lymphocyte ratio (NLR), were analyzed. TSS was defined as the time from the diagnosis of terminal cancer to death.

RESULTS The mean age of the 141 patients studied was 58.7 years, and 53 were female (38 %). The median TSS was 1.7 months (95 % confidence interval [CI] 1.43-1.97). In the univariate analyses, the TSS was significantly associated with 16 of the 32 factors tested. In the multivariate analysis, a lower Karnofsky performance status (KPS), a shorter TTC (<24 months), a high NLR (≥5), and a high C-reactive protein (CRP) level (≥10 mg/dL) were independently associated with a poorer prognosis. A scoring system (scale, 0-6) developed based on the multivariate analysis could be used to classify terminal cancer patients into better (0-2 points; TSS 3.9 months), intermediate (3-4 points; TSS 1.7 months), or worse (5-6 points; TSS 0.9 month, P < 0.001) prognosis.

CONCLUSION The median TSS after the diagnosis of terminal cancer in advanced cancer patients was 1.7 months. The scoring system using KPS, TTC, NLR, and CRP could predict TSS in these patients.