Clinical Investigation
Optimizing Radiation Therapy Quality Assurance in Clinical Trials: A TROG 08.03 RAVES Substudy

https://doi.org/10.1016/j.ijrobp.2015.08.029Get rights and content

Purpose

To explore site- and clinician-level factors associated with protocol violations requiring real-time-review (RTR) resubmission in a multicenter clinical trial to help tailor future quality assurance (QA) protocols.

Methods and Materials

RAVES (Radiation Therapy–Adjuvant vs Early Salvage) (Trans-Tasman Radiation Oncology Group 08.03) is a randomized trial comparing adjuvant with early salvage radiation therapy in men with positive surgical margins or pT3 disease after prostatectomy. Quality assurance in RAVES required each clinician and site to submit a credentialing dummy run (DR) and for each patient's radiation therapy plan to undergo external RTR before treatment. Prospectively defined major violations from trial protocol required remedy and resubmission. Site and clinician factors associated with RTR resubmission were examined using hierarchical modeling.

Results

Data were collected from 171 consecutive patients, treated by 46 clinicians at 32 hospitals. There were 47 RTR resubmissions (27%) due to 65 major violations. The relative rate of resubmission decreased by 29% per year as the study progressed (odds ratio OR. 0.71, P=.02). The majority of resubmissions were due to contouring violations (39 of 65) and dosimetric violations (22 of 65). For each additional patient accrued, significant decreases in RTR resubmission were seen at both clinician level (OR 0.75, P=.02) and site level (OR 0.72, P=.01). The rate of resubmission due to dosimetric violations was only 1.6% after the first 5 patients. Use of IMRT was associated with lower rates of resubmission compared with 3-dimensional conformal radiation therapy (OR 0.38, P=.05).

Conclusion

Several low- and high-risk factors that may assist with tailoring future clinical trial QA were identified. Because the real-time resubmission rate was largely independent of the credentialing exercise, some form of RTR QA is recommended. The greatest benefit from QA was derived early in trial activation and clinician experience.

Introduction

An important part of modern multicenter radiation therapy clinical trials is treatment plan quality assurance (QA). This ensures that participating institutions are delivering comparable and consistent treatment to the target volumes and normal structures. A QA program defines an acceptable range of deviations from trial protocol, with mechanisms to correct violations that could affect trial outcomes. The importance of good QA is highlighted in the Trans-Tasman Radiation Oncology Group (TROG) 02.02 trial, where the quality of radiation therapy delivered had a greater influence on overall survival than the intervention studied (1). Clinical trial protocol deviations have been shown to reduce efficacy and increase normal tissue complications in numerous other studies 2, 3, 4. As such, QA has become a standard part of modern radiation therapy clinical trials.

The use of complex modern radiation therapy techniques has resulted in more rigorous QA programs. Intensive QA results in fewer deviations from protocol; however, limited evidence exists to support the optimal amount and type of trial-specific QA activities 5, 6. The current challenge in clinical trial QA is to reduce the complexity and costs without reducing the efficacy and safety using risk-adaptation strategies.

RAVES (Radiation Therapy–Adjuvant vs Early Salvage)–TROG 08.03 is a multicenter, randomized clinical trial comparing adjuvant with early salvage radiation therapy in men with positive surgical margins or pT3 disease after radical prostatectomy (7). Quality assurance in RAVES requires each clinician and each site to submit a credentialing dummy-run case and for each patient's radiation therapy plan to undergo external real-time review before commencing treatment. Prospectively defined major violations from trial protocol require remedy and resubmission before starting treatment. Our study aimed to explore site- and clinician-level factors that are associated with protocol violations and real-time-review resubmissions to help tailor future QA protocols.

Section snippets

Patient demographics

RAVES is a phase 3, multicenter, controlled trial led by the TROG collaborative group in Australia and New Zealand. Eligible patients had undergone radical prostatectomy for prostate adenocarcinoma and had at least 1 of the following risk factors: positive surgical margins, extra prostatic extension, or seminal vesicle involvement. Patients were randomized 1:1 to either adjuvant radiation therapy commenced within 4 to 6 months of radical prostatectomy or close observation with salvage radiation

Cohort characteristics

Between June 2009 and October 2014, data from 171 patients out of 174 consecutively treated patients (98%) were reviewed. A total of 46 ROs from 32 institutions around Australia and New Zealand took part in the trial. All RO and site credentialing dummy-run and real-time-review submissions were available for analysis.

The median total number of patients recruited by each RO was 7 (range, 1-20), with a mean of 2 patients per year, resulting in a submission every 175 days (median). The median

Discussion

Clinical trial radiation therapy QA programs have become more comprehensive and labor-intensive and can be a barrier to trial accrual (6). The current challenge is to optimize the intensity of QA directed by good-quality evidence and to evaluate which components of a QA program deliver the most benefit.

Our study is the first analysis of a prospectively collected real-time-review QA in postoperative prostate cancer radiation therapy. We have observed a resubmission rate of 27%. Despite dummy-run

Conclusion

The RAVES QA protocol is a useful template for future multi-institute radiation therapy studies. The majority of resubmissions (89%) passed, indicating feasibility to achieve trial protocol requirements while removing clinically significant protocol violations.

Several low- and high-risk factors were identified that may assist with tailoring future clinical trial QA. Because the real-time resubmission rate was largely independent of the credentialing exercise, some form of real-time-review QA is

References (23)

Cited by (0)

C.F.-B.'s institution received grant support from the New Zealand Health Research Council for this project.

Conflict of interest: none.

View full text