Trends in the characteristics, dose-limiting toxicities and efficacy of phase I oncology trials: The Cancer Research UK experience.

Introduction: Phase I oncology trials have evolved over the years, and these changes could have implications for future studies and patients.

Methods: Adult trials sponsored by Cancer Research UK Centre for Drug Development between 1995 and 2013 were analysed. Forty-nine trials were divided into two groups based on the starting date for recruitment: 1995-2003 (24 trials, n = 603) and 2004-2013 (25 trials, n = 750) for comparative purposes.

Results: From 1995-2003 to 2004-2013, there was a shift towards studying non-cytotoxic agents that are administered orally. In later trials, patients tended to have better performance status, were older, had greater disease burden, and were more likely to have received prior treatment. In 2004-2013, wider variety of dose escalation designs were used, and studies were more likely to be multicentre, target/disease specific, conducted in first-/any-line setting and to require tumour biopsy. The overall incidence of dose-limiting toxicities (DLTs) was unchanged (10.9%; risk of death 0.4%), but DLTs such as neuropathy, stomatitis and thrombocytopaenia were less frequent in the more recent trials, while elevated liver enzymes were more frequent. Non-classical DLTs emerged in the later trials, including hypertension, hypophosphataemia, cardiac and ophthalmic toxicities. Disease control rate (DCR) increased from 27.9% (1995-2003) to 36.0% (2004-2013; P = 0.0033) due to higher rates of disease stabilisation.

Conclusion: Changes in trial designs, therapeutic agents, patient characteristics and DLTs were observed. Although the nature of DLTs changed, the incidence was similar in the two time periods and DCR improved, suggesting that the benefit-risk balance for patients participating in early-phase trials remains acceptable.