Toxicity of Phase I Radiation Oncology Trials: Worldwide Experience

Reviewer: Lara Bonner Millar, MD
The Abramson Cancer Center of the University of Pennsylvania
Last Modified: November 2, 2010

Authors: C. Glass1, R. B. Den2, A. P. Dicker2, Y. R. Lawrence2
1 Jefferson Medical College, Philadelphia, PA
2 Department of Radiation Oncology, Jefferson Medical College of Thomas Jefferson University, Philadelphia, PA


  • Informed consent involves understanding the risks and benefits of trial enrollment, but in the phase I setting, true quantitative information is rarely known. The authors performed an analysis of published radiation oncology (RO) phase I trials emphasizing patient outcomes with attention to toxicities of treatment.
  • The medical oncology literature documents a 0.5% risk of death and a response rate of 5-10% for phase I trials.
  • RO and chemotherapy trials differ because chemotherapy trials are often performed in patients with metastatic disease, who have been heavily pre-treated.
  • This is in contrast to phase I/II trials using radiation, which may examine dose escalation or addition of a novel agent as part of definitive treatment. Furthermore, in RO, qualitative predictable toxicity may exist depending on the site treated.

Materials and Methods

  • All phase I and I/IIRO clinical trials published in the years 2001, 2005 and 2009 were identified via a PubMed search: “trial OR study OR phase” AND “radiation.”
  • All trials utilized radiation therapy, and many also included cytotoxic chemotherapy and/or biologic agents. Exclusion criteria: articles only published in abstract form, strict phase II trials, non-English studies, and non-cancer pathologies.
  • Variables extracted from the full-text of manuscripts included demographics, treatment site, trial design, toxicity, ethical and legal considerations, and response rate.
  • When multiple adverse events occurred in a single subject, these were calculated as per-study, not per-subject.


  • 119 trials were identified. 17 trials were excluded since the manuscript was unobtainable, leaving 102 trials for review (2001, 27; 2005, 45; and 2009, 30).
  • 48% were performed in the USA, and 20% in Europe. 66% were multi-institutional, and 17% were conducted by a cooperative group.
  • 68% of trials were phase I and 32% were combination phase I/II trials.
  • The most frequent treatment sites were lung 22%, brain 12%, rectum 11%, and head-and-neck 11%.
  • 68% involved dose escalating a chemotherapeutic or biologic agent, 17% dose-escalated radiation alone, and 1% of trials escalated both radiation and/or a chemotherapeutic agent.
  • 4553 subjects were enrolled; this amounted to 1300-1400 pts enrolled per year. 65% were male, and mean age was 62.
  • Median or mean follow-up was 25.2 months.
  • 68% reached MTD (maximum tolerated dose) and 26% were treated below the MTD; 92% of pts completed RT.
  • There were 39 treatment related deaths, and 1881 major acute toxicities (defined as grade 3 or worse excluding death according to CTCAE criteria or equivalent). The risk of death was 0.1% and there was a 41% risk of grade 3+ toxicity per patient.
  • Neither radiation dose, nor gender, nor average age affected the number of events.
  • The authors calculated a “toxicity ratio” for the trials as the # of grade 3-5 events/ # patients on the trial. High toxicity trials had about one event per patient. Pancreatic and head and neck trials were likely to be high toxicity trials.
  • 50% of trials reported late radiation toxicity data.
  • The reporting of benefit was inconsistent, preventing meaningful analysis.

Author's Conclusions

  • Of the 6 million patients who receive radiation therapy annually worldwide, only 1500 patients are enrolled onto phase I trials.
  • The risk of death is low (0.1%), but the risk of major toxicity is significant (40%). These figures have to be considered in the context of the expected benefit/toxicity with standard of care treatments. If high toxicity risk features are present, gentler dose escalation, increased scrutiny, and a need to ensure than toxicity will not compromise completion of radiation therapy are important.

Clinical Implications

  • This is the first systemic analysis of toxicity on phase I radiation trials. The data was gathered from trials in disease sites which are representative of the sites seen in RO practice, so the results have broad applicability.
  • Despite the low risk of death, the risk of major acute toxicity in phase I radiation trials is high. Patient concerns about toxicity probably impacts on the (low) numbers accrual onto clinical trials. Minorities tend to have lower than average accrual on clinical trials, but whether this is also true in the context of phase I RO trials is not clear as this demographic information was not reported.
  • Some of the trials included were re-irradiation trials, so in theory, the additional toxicity from re-irradiation could have affected the overall toxicity rate (although the presenter commented that the trials where toxicity was high were not re-irradiation trials).
  • The toxicity ratio is useful, but requires that toxicity data be accurately collected and reported. This is challenging since the reliability of toxicity data can questionable when there are numerous steps in the reporting process. Likewise, the lack of reporting on late toxicity in half the trials indicates the difficulties in capturing and scoring this data. In phase I RO trials, late complications are often not taken into account, which is problematic considering that certain late complications may have worse impact on patients than early toxicities. Furthermore, late toxicity is variously defined or may not be defined at all in some studies. In clinical trials, late toxicity should be consistently defined and scoring consistently applied.