Open AccessDevelopment and validation of a prognostic nomogram for ambulatory patients with advanced cancer
Author(s) -
Paiva Carlos Eduardo,
Paiva Bianca Sakamoto Ribeiro,
Paula Pântano Naitielle,
Preto Daniel D’Almeida,
Oliveira Cleyton Zanardo,
Yennurajalingam Sriram,
Hui David,
Bruera Eduardo
Publication year - 2018
Publication title -
cancer medicine
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.403
H-Index - 53
ISSN - 2045-7634
DOI - 10.1002/cam4.1582
Subject(s) - nomogram , medicine , percentile , receiver operating characteristic , proportional hazards model , oncology , area under the curve , prognostic variable , cancer , statistics , mathematics
Abstract Predicting survival of advanced cancer patients (ACPs) is a difficult task. We aimed at developing and testing a new prognostic tool in ACPs when they were first referred to palliative care (PC). A total of 497 patients were analyzed in this study (development sample, n = 221; validation sample, n = 276). From 35 initial putative prognostic variables, 14 of them were selected for multivariable Cox regression analyses; the most accurate final model was identified by backward variable elimination. Parameters were built into a nomogram to estimate the probability of patient survival at 30, 90, and 180 days. Calibration and discrimination properties of the Barretos Prognostic Nomogram (BPN) were evaluated in the validation phase of the study. The BPN was composed of 5 parameters: sex, presence of distant metastasis, Karnofsky Performance Status (KPS), white blood cell (WBC) count, and serum albumin concentration. The C‐index was 0.71. The values of the area under the curve (AUC) of the receiver operating characteristic (ROC) curve were 0.84, 0.74, and 0.74 at 30, 90, and 180 days, respectively. There were good calibration results according to the Hosmer‐Lemeshow test. The median survival times were 313, 129, and 37 days for the BPN scores <25th percentile (<125), 25th to 75th percentile (125‐175), and >75th percentile (>175), respectively ( P < .001). The BPN is a new prognostic tool with adequate calibration and discrimination properties. It is now available to assist oncologists and palliative care physicians in estimating the survival of adult patients with advanced solid tumors.