ABSTRACT
<p>Objectives: We developed versions 2 and 3 of the Biological Prognostic Score (BPS) for advanced cancer patients and confirmed the prediction accuracy. Methods: We conducted a parametric survival analysis using blood test data, performance status (PS), clinical symptoms, age, sex, and cancer type as variables for advanced cancer patients who completed or suspended cancer treatment, in the development of BPS2 and BPS3. We then prospectively compared the accuracy between BPS2/BPS3 and the Palliative Prognostic Index (PPI). Results: We developed the BPS2 and BPS3 based on the data from 589 patients in a development cohort. While the former version was calculated based on the cholinesterase, blood urea nitrogen, and white blood cell counts, the latter was calculated based on the BPS2, ECOG PS and edema. For 206 patients in a validation cohort, the overall accuracy in prediction of survival for 3 and 6 weeks using the BPS2 and BPS3 were significantly higher than those for the PPI. Conclusion: The usefulness of BPS2 and BPS3 was suggested.</p>
ABSTRACT
Objective: To develop and confirm the validity of a Biological Prognostic Score using only blood test results for prediction of prognosis in patients with advanced cancer. Methods: We conducted parametric survival time analysis using blood test results, age, sex, and primary diagnosis as independent variables, and event of death as a dependent variable, among patients in a palliative care unit (a development cohort). We then developed the Biological Prognostic Score (BPS). Thereafter, we confirmed the accuracy of the BPS and the Palliative Prognostic Index (PPI) prospectively among patients, who withdrew or withheld further curative or life-prolonging therapies, in other facilities (a validation cohort). Results: We developed the BPS, which consists of cholinesterase, blood urea nitrogen, and total iron-binding capacity, from 122 patients in a development cohort. We then examined 195 patients in a validation cohort and found that the area under the receiver operating characteristic curve for 1-9 week survival prediction was BPS=0.76-0.86 and PPI=0.69-0.73. Discussion: Our results suggest that the BPS was valid. It will be necessary to perform further examinations in multiple facilities and to explore more generalized parameters that could replace total iron-binding capacity in our BPS.