Patterns of Failure Observed in the 2-Step Institution Credentialing Process for NRG Oncology/Radiation Therapy Oncology Group 1005 (NCT01349322) and Lessons Learned.

Author(s): Li XA,  Moughan J,  White JR,  Freedman GM,  Arthur DW,  Galvin J,  Xiao Y,  McNulty S,  Lyons JA,  Kavadi VS,  Fields MT,  Mitchell MP,  Anderson BM,  Lock MI,  Kokeny KE,  Bazan JG,  Currey AD,  Hijal T,  Cheston SB,  Vicini FA

Journal: Pract Radiat Oncol

Date: 2019 Nov 29

Major Program(s) or Research Group(s): NCORP

PubMed ID: 31790823

PMC ID: PMC7255922

Abstract: PURPOSE: To investigate patterns of failure in institutional credentialing submissions to NRG/RTOG 1005 with the aim of improving the quality and consistency for future breast cancer protocols. METHODS AND MATERIALS: NRG/RTOG 1005 allowed the submission of 3-dimensional conformal radiation therapy (3DCRT), intensity-modulated radiation therapy (IMRT), and simultaneous integrated boost (SIB) breast plans. Credentialing required institutions to pass a 2-step quality assurance (QA) process: (1) benchmark, requiring institutions to create a plan with no unacceptable deviations and ≤1 acceptable variation among the dose volume (DV) criteria, and (2) rapid review, requiring each institution's first protocol submission to have no unacceptable deviations among the DV criteria or contours. Overall rates, number of resubmissions, and reasons for resubmission were analyzed for each QA step. RESULTS: In total, 352 institutions participated in benchmark QA and 280 patients enrolled had rapid review QA. Benchmark initial failure rates were similar for 3DCRT (18%), IMRT (17%), and SIB (18%) plans. For 3DCRT and IMRT benchmark plans, ipsilateral lung most frequently failed the DV criteria, and SIB DV failures were seen most frequently for the heart. Rapid review contour initial failures (35%) were due to target rather than organs at risk. For 29% of the rapid review initial failures, the planning target volume boost eval volume was deemed an unacceptable deviation. CONCLUSIONS: The review of the benchmark and rapid review QA submissions indicates that acceptable variations or unacceptable deviations for the ipsilateral lung and heart dose constraints were the most commonly observed cause of benchmark QA failure, and unacceptable deviations in target contouring, rather than normal structure contouring, were the most common cause of rapid review QA failure. These findings suggest that a rigorous QA process is necessary for high quality and homogeneity in radiation therapy in multi-institutional trials of breast cancer to ensure that the benefits of radiation therapy far outweigh the risks.