Quality and Reporting Accuracy of Phase 1 Drug Radiation Clinical Trials

Phase 1 trials are necessary to evaluate the safety and toxicity of novel multimodality treatment (MMT) prior to routine clinical use. In the era of targeted therapies, phase 1 trials are increasingly used to justify clinical application in selected patient populations, without prospective assessment of outcomes.¹ Despite several published guidelines^2,3 regarding the reporting of phase 1 trials, details regarding therapeutic intent, median follow-up time, and late toxic effects remain poorly documented. In particular, late radiation-related complications are often underreported, despite late toxic effects significantly impacting patient morbidity and mortality.⁴ We performed a systematic review of phase 1 trials involving radiation plus systemic therapeutics (novel agent or chemotherapy) to evaluate the frequency with which clinical efficacy is prospectively identified as an end point, median follow-up time reported, and late toxic effects delineated in published articles.

We systematically identified modern phase 1 MMT trials in the PubMed database with selection and exclusion criteria as delineated in the Figure. Each trial was evaluated for details regarding clinical efficacy as a prespecified end point (ie, prospective therapeutic intent), reporting of clinical outcomes (with or without prespecified end points), median follow-up time, reported grade 3 or higher and dose-limiting toxic effects, and distinction of acute and late radiation toxic effects.

Associations of study characteristics with these end points were evaluated using exact χ² tests (P < .05) and pairwise exact χ² test comparisons; associations with follow-up time were evaluated using Kruskal-Wallis and pairwise comparisons using Dunn procedure⁵ using SAS statistical software (version 9.4; SAS Institute).

The PubMed query methodology and results are outlined in the Figure. The percentage of trials that reported prospectively identified clinical end points, clinical outcomes, median follow-up time, and late toxic effects are delineated in the Table. Trials in which radiation dosing was investigated were associated with a higher frequency of prospective clinical end points compared with novel drug and/or chemotherapy trials (P = .04). No trials prospectively determined the number of patients needed to detect late toxic events.

Our data indicate that prospective therapeutic intent, follow-up time, and late toxic events are significantly underreported in MMT phase 1 trials. Despite adequate follow-up time (>90 days) to assess for late toxic effects, only 12% of trials reported late toxic effects. A possible reason is the difficulty in attributing late toxic effects to a specific treatment owing to tumor recurrence and previous or subsequent systemic and local treatments. Only half of trials prospectively identified clinical end points, yet over 90% reported a clinical outcome, possibly a result of post hoc data exploration. To increase the reliability of phase 1 trial results, we propose that (1) reporting median follow-up time become a trial publication requirement. Improving the quality of trial reporting may expedite the transition to subsequent phase trials and decrease the likelihood of discovering unknown toxic effects in larger populations and costly phase 2 and 3 studies. That (2) prospectively delineating therapeutic intent with prespecified clinical end points should be adopted because it may facilitate early identification of efficacy and improve insurance coverage of and patient access to early-phase trials.² The ability to directly attribute late toxic effects to specific drug radiation combinations is complicated by early patient mortality and confounding treatments. However, because investigators are obligated to their patients and peers to evaluate toxic effects from investigational treatments, we propose that (3) assessment and reporting of late toxic effects be included in phase 1 trials.

Limitations of this study include the time period evaluated, that it is a single database query, and that it only includes published trials and toxic effects. ClinicalTrials.gov was not included because trial registration is not mandated as a condition of publication for phase 1 trials.⁶ However, these data identify 3 important measures, which if addressed, will improve the quality, availability, and interpretability of phase 1 trials.

Corresponding Author: Hyun Kim, MD, Department of Radiation Oncology, Sidney Kimmel Medical College and Cancer Center, Thomas Jefferson University, 111 S 11th St, Philadelphia, PA 19107 (hyun.kim@jefferson.edu).

Published Online: December 17, 2015. doi:10.1001/jamaoncol.2015.4833.

Author Contributions: Dr Kim had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Study concept and design: Kim, Dan, Palmer, Lawrence, Dicker.

Acquisition, analysis, or interpretation of data: Kim, Dan, Palmer, Leiby, Dicker.

Drafting of the manuscript: Kim, Dan.

Critical revision of the manuscript for important intellectual content: All authors.

Statistical analysis: Dan, Leiby, Lawrence.

Administrative, technical, or material support: Kim, Dicker.

Study supervision: Kim, Palmer.

Conflict of Interest Disclosures: None reported.

Funding/Support: Dr Leiby is supported in part by award No. P30 CA056036-16 from the National Cancer Institute (NCI). Dr Lawrence is supported in part by EU FP7 Marie-Curie funding (CIG PCIG10-GA-2011-303795).

Role of the Funder/Sponsor: The NCI and Marie-Curie funding had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.

Previous Presentation: Data were presented in part at the American Association for Cancer Research Radiation Oncology Think Tank, Inaugural Meeting; January 11-13, 2015; Fort Myers, Florida.