A new, simple and objective prognostic score for phase I cancer patients

Abstract Aim To ensure safety and optimise efficacy, careful patient selection for participation in oncologic phase I trials is warranted. Therefore, we did a validation study on existing phase I prognostic scores, and subsequently aimed to make an even more simple prognostic score. Patients and methods We retrospectively analysed characteristics and clinical outcome of 122 patients who participated in eight different phase I studies in our centre. A literature search was performed for existing prognostic scores which were validated in our dataset. Additionally, a simple prognostic score able to predict overall survival (OS), progression free survival (PFS), 90-d mortality and duration of participation was developed. Results In our population the median OS was 31 (range 4–241) weeks, median PFS was 10 (range 3–119) weeks. Thirteen patients (11%) died within 90-d and median participation duration was 11 (range 1–119) weeks. Two out of five existing prognostic scores could be validated in this dataset for OS. Based on multivariate analyses a new, objective and simple score, consisting of LDH, sodium and haemoglobin, was tested. High score (2–3 points) compared to low score (0–1) significantly predicted OS (HR 1.878, p = 0.003), PFS (HR 1.156, p = 0.038), participation duration (HR 1.858, p = 0.004) and 90-d mortality (OR 4.200, p = 0.022). Conclusion We propose a new prognostic model, using LDH, sodium and haemoglobin, helpful to predict OS, PFS, participation duration and 90-d mortality. Larger multicentre studies are needed to validate this score.