John C Gore
MR imaging and spectroscopy entered clinical practice around 1980 and have since generated one of the largest literatures in diagnostic medicine. That literature is conventionally described as self-correcting. This editorial examines the record of medical MR since its clinical advent and finds that description only partly accurate. Documented episodes failing that standard span a spectrum from honest error through pathological science to research misconduct. Examples include a plasma proton MR spectroscopy test for cancer that failed independent replication and preceded a federal misconduct finding; an influential conclusion on optimal field strength from a laboratory with a commercial position in the field strength it favored; the fabrication of numerous papers on cardiac MR imaging by an investigator with 93 coauthors; four decades of non-reproducible claims to detect neuronal currents directly with MRI; a spectroscopic biomarker of neural progenitor cells that was misinterpreted; a disturbing report that an MR examination improves mood in bipolar disorder; repeated assurances regarding the safety profile of gadolinium-based contrast agents; claims of detecting the migration of silicone in vivo that relied on inadequate and manipulated MR spectra which influenced legal hearings and court actions; and methodological failures affecting large parts of the functional MR imaging literature. In nearly every instance the problem was identified promptly by competent contemporaries, yet the original reports were seldom retracted, corrected, or formally refuted. Structural reasons for this failure are examined, together with suggestions for remedies to reduce the dissemination of bogus science.