Opportunities for Statistics in Radiology Clinical Trials

Diana Miglioretti, Nancy Obuchowski, and Todd Alonzo

The ASA Biometrics Section will fund a biostatistician to attend the annual Clinical Trials Methodology Workshop for imagers taking place in January 2015 in Scottsdale, Arizona. The workshop presents an opportunity for a biostatistician interested in imaging clinical trials to learn the relevant methodology and gain experience collaborating with radiologists and imaging specialists.

Medical imaging is playing a much larger role in the diagnosis and treatment of disease. Imaging clinical trials are designed to evaluate how imaging can detect early disease, predict patient outcome to different therapies, and guide less invasive interventions. These trials can be more challenging than therapeutic trials because they often require a reference standard, must account for subjective image interpretations, and are mediated by events that take place after imaging.

Yet radiology has been underserved by biostatisticians. Studies in radiology are often ill-designed, fail to include multiple institutions or readers, focus on outcomes that are not clinically relevant, and do not properly account for clustering in the data. Radiology research requires the use of unique statistical methods, including those used for the evaluation of the accuracy of diagnostic imaging tests (e.g., ROC or fROC curves), for developing prognostic models based on imaging results and assessing the performance of new quantitative imaging biomarker algorithms. Research in radiology would benefit greatly from the attention of, and collaboration with, methodological biostatisticians with appropriate training.

Each year, the Radiology Society of North America (RSNA) holds an intensive one-week workshop on clinical trial methodology for faculty members and fellows in radiology, radiation oncology, and nuclear medicine academic departments. The workshop provides hands-on training in the design, conduct, analysis, and interpretation of clinical trials in radiology and imaging through didactic sessions, one-on-one mentoring, small discussion sessions, self-study, and protocol development workgroups. A curriculum infused with statistical and methodological concepts and examples was developed for this course. This workshop has a distinguished group of 20 MD and PhD faculty members (including five PhD biostatisticians) with extensive experience in radiology clinical trials, providing a student-to-faculty ratio close to 1:1.

The ASA-funded biostatistician trainee will shadow each of the biostatistics faculty members to observe the collaborative process of developing clinical trials in radiology, in addition to learning the methodological aspects of this field. Eligible applicants will hold a PhD degree in biostatistics or a related field (e.g., statistics) and have an interest in the design, conduct, and analysis of imaging clinical trials. Interested applicants should email a one-page letter of interest, CV, and letter of support from their department head to Diana Miglioretti at dmiglioretti@ucdavis.edu with “RSNA CTMW Biostatistics” in the subject line. The deadline for receipt of applications is August 15. Announcement of the award will be made by October 1.