Estimating the Population-based Overall Survival Benefits of First-course Chemotherapy in Cancers

Abstract

Background: Randomised clinical trials describe the benefit of chemotherapy (CT) for cancer patients with selected patient and disease characteristics. The overall survival (OS) benefits for the whole population of cancer patients in Australia, if evidence-based guidelines for CT were implemented, are unknown. The purpose of this thesis was to estimate the absolute improvements in overall survival that might result from systematically applying evidence‐based guidelines about the optimal use of a first course of chemotherapy to the whole population of people with cancer in Australia.

Methods and Materials: Decision trees with evidence-based indications for CT have been previously defined. Each branch of the tree corresponds to a specific cohort who have, or do not have, defined indication for curative and palliative CT. CT OS benefit was defined as the absolute incremental OS of first-course CT over best supportive care for palliative indications or over radiotherapy or surgery with or without radiotherapy in curative indications. Multiple electronic citation databases were systematically queried, including Medline and Cochrane library. In cases where there were multiple sources of the same level of evidence, then hierarchical meta-analysis was performed. A literature search was updated with a cut-off of December 2017. The OS benefits of CT were estimated for one year, five years and ten years. To assess the robustness of our estimates, sensitivity analyses were performed.

Results: The estimated one-year, five-year, ten-year absolute population-based OS benefits of optimally-used first-course CT for cancer were 5.9% (95% Confidence Interval, CI 4.8%-7.0%), 4.5% (95% CI, 4.4%-4.5%) and 3.3% (95% CI, 3.3%-3.4%) respectively. The additional benefit of CT over other modality for five-year OS ranged from 0% (malignant pleural mesothelioma) to 26.7% (testes). Sixty one percent of OS benefit was attributed to first-course palliative CT at one year, and 24% at five years.

Conclusion: This is the first comprehensive study, using robust modelling methodologies, to estimate the population-based overall survival benefits from first-course CT for all cancers in Australia. The model can be adapted for other jurisdictions readily. First-course CT offers vital benefits to the cancer population. The population survival benefit can serve as an evidenced-based benchmark as part of the consensus quality standard in the treatment of cancer in Australia. Consequently, it can further aid decision making for cancer treatment. As a result, future studies may utilise this approach to estimate the population benefits for surgery and other endpoints. This model is flexible, enabling the consideration of any future changes to the population benefits of CT in the management of cancer within Australia or can be altered to address other jurisdictions where differing prevalence and stage data may exist.