Objectives/purpose: To explore the components in head and neck cancer (HNC) care that shape routine practice, identify strategies for mapping the individualised multimodal treatment, and summarise outcomes analysis strategies.
Sample and setting: Narrative review.
Procedures: Publications on HNC care protocols and their delivery in real-world settings were reviewed to establish baseline information on current practices. Building on this foundation, we assessed how oncology data was organised into treatment sequences in published literatures, using the real-world oncology data standards to identify the key aspects of treatment sequencing. The review concludes with example of known real-world treatment sequences and outlines considerations for selecting appropriate analytical methods to measure HNC clinical outcomes.
Results: To better assess the outcomes of routine HNC care, we reviewed 98 observational studies for factors that expand beyond the stereotypical (person, tumour, treatment) triad. Real-world data captures complex and dynamic treatment decisions through many serial observations. However, based on benchmarking with four HNC guidelines, the treatment information reported in real-world settings often lacks the critical parameters outlined in clinical protocols. These details are essential for delivering personalised treatment and evaluating treatment outcomes. As treatment delivery evolves with the growing number of options, its representation as treatment sequences has become increasingly complex, reflecting the delicate decisions involved in adjusting multimodal treatment parameters to manage treatment failures and recurrences. Concerted efforts are needed to systematically collect harmonised real-world oncology data and apply advanced time-to-event analysis methods to effectively capture the patterns in multidimensional treatment decision factors and treatment sequences.
Conclusion and clinical implications: Review of literature has demonstrated that using treatment sequencing approach and scalable time-to-event methods allows to understand treatment patterns, enhance real-world evidence, and estimate clinical outcomes more effectively. This review demonstrates the potential to support data-driven care by reducing uncertainty and improving the adoption of proven personalised treatment plans.