CN2 - ROPRO – Real-World Data Prognostic Score: A novel tool to assess patients' performance status

Background

The assessment of our patients' performance status (PS) to predict survival has always been important for toxicity monitoring, treatment selection and clinical trial eligibility. Some existing tools to determine PS may not be suitable for decision-making due to their reliance on subjective assessment, leading to limited reliability and restricted predictive value in patients with better PS.

Methods

By mining the data of more than 120,000 cancer patients from Flatiron Health, we derived the Real-World Data Prognostic Score (ROPRO) for overall survival in cancer patients [1]. ROPRO comprises 27 independently associated, routinely collected clinical parameters such as lactate dehydrogenase, albumin or neutrophil/lymphocyte ratio. ROPRO is a pan-cancer score derived from 17 cancer cohorts and has been shown to be applicable across multiple other tumor types. We have validated ROPRO in over 20 independent clinical studies, demonstrating that ROPRO outperforms other prognostic scores across cancer types and treatment modalities.

Results

We present new data regarding two applications of the ROPRO in clinical development. First, we are using baseline ROPRO for patient selection. ROPRO is a strong predictor of short-term survival [3-months ROC-AUC=82.5] and thus provides quantitative and un-biased decision support to investigators to determine 12-week life expectancy. In this context, ROPRO outperforms standard scores such as Royal Marsden Hospital Score [3-months ROC-AUC=59.3] and ECOG [3-months ROC-AUC=70.4]. Second, we are presenting new data on delta ROPRO analysis over time which may help clinicians detect treatment benefits or progression early.

Conclusions

There is an unmet need for improved and data-driven decision making for clinical development and clinical practise. We believe ROPRO may address that need by being more objective and discriminatory. We show that by using a large amount of Real-World Data (RWD) of oncology patients in combination with artificial intelligence methods, innovative tools aiding clinical decision making, can be developed. A wider adoption will help validating ROPRO further as a quantitative tool to assess patients performance status, an aspect particularly relevant for cancer immunotherapy.

Clinical trial identification

Editorial acknowledgement

Legal entity responsible for the study

Roche Diagnostics GmbH.

Funding

F. Hoffmann-La Roche Ltd.

Disclosure

J. Desai: Financial Interests, Personal, Advisory Role: Eisai; Financial Interests, Personal, Advisory Role: BeiGene; Financial Interests, Personal, Advisory Role: Pierre Fabre; Financial Interests, Personal, Advisory Role: Bayer; Financial Interests, Institutional, Advisory Role: Amgen; Financial Interests, Institutional, Research Grant: Roche; Financial Interests, Institutional, Research Grant: GlaxoSmithKline; Financial Interests, Institutional, Research Grant: Novartis; Financial Interests, Institutional, Research Grant: Bionomics; Financial Interests, Institutional, Research Grant: BeiGene; Financial Interests, Institutional, Research Grant: Lilly; Financial Interests, Institutional, Research Grant: Bristol Myers Squibb; Financial Interests, Institutional, Research Grant: AstraZeneca/MedImmune. A. Bauer-Mehren: Financial Interests, Personal, Stocks/Shares: Roche. All other authors have declared no conflicts of interest.