The resident's research project that reuses clinical images
A second-year nuclear medicine resident wants to present a retrospective study at a national meeting on the diagnostic performance of 18F-FDG PET/CT for detecting mediastinal nodal involvement in lung cancer. She plans to pull 220 studies from the department's PACS from the past 3 years, record SUVmax values, and correlate them with pathology from the electronic medical record. She has not submitted an IRB application. She tells the physicist, "It's retrospective and I'm de-identifying it, so it doesn't need IRB review. I just need you to help me export the SUV data." The physicist is a co-investigator on the project.
1. Is the resident correct that de-identified retrospective research is exempt from IRB review? What is the correct analysis?
Read the worked answer
The resident is conflating two separate regulatory concepts.
- Exempt vs. not human-subjects research. Research on fully de-identified data may not constitute human-subjects research under the Common Rule, but that determination is made by the IRB (or an institutional official), not by the investigator. The investigator cannot self-determine exemption.
- HIPAA de-identification. Removing the 18 HIPAA identifiers is not the same as de-identification for research purposes, and the export must actually strip dates, medical record numbers, and other identifiers. SUVmax values alone are not identifiers, but the linked pathology and dates can be.
- Retrospective does not mean exempt. Retrospective chart review is still human-subjects research if it involves identifiable private information. It may qualify for expedited or exempt review, but it still goes through the IRB.
- The physicist's role. As a co-investigator, the physicist shares responsibility. Helping export data from PACS for an unapproved study is participating in unapproved human-subjects research.
The correct answer: the resident must submit an IRB application (or an exempt-determination request) before data extraction. The physicist should decline to export until approval is documented.
2. What should the physicist do next, and what are the professional risks of handling this informally?
Read the worked answer
Next steps:
- Decline the export until an IRB protocol number or exempt determination is on file. This is not obstruction; it is the condition of participation.
- Help the resident draft the application (this is collegial and appropriate): specify the data elements, the de-identification plan, the storage location, and the data-use agreement.
- Confirm the institution's policy on retrospective imaging research, including whether a limited data set or honest-broker pathway is available.
- Document the conversation and the decision to wait.
Professional risks of handling it informally:
- Institutional and regulatory exposure: unapproved human-subjects research can trigger IRB findings, loss of funding, and publication retraction.
- The physicist's own record: as co-investigator, the physicist would be named in any finding. "The resident did it" is not a defense for a co-investigator.
- Patient trust: a data breach or misuse of clinical images damages the department's credibility.
- The resident's career: a retraction or IRB finding early in training is far more damaging than a two-week delay.
The judgment being tested: the physicist must be willing to say no to a trainee and to help them do it correctly, rather than being helpful in the wrong direction.