Dr. Bailie recognized early in his career that he was seeing many patients who were referred to him because they had already obtained an MRI and the radiologist had reported an abnormality. This often resulted in a referral to Dr. Bailie with the expectation for surgery. However, after a careful history and physical examination, it was clear that the MRI report was not a reflection of the entire problem. This led Dr. Bailie to critically review each patient MRI personally and correlate the findings with the patient’s complaints and physical findings, not relying on the report to make a treatment decision.
In 2002, he studied over 500 of his knee and shoulder arthroscopy surgical patients comparing his interpretation of the pre-operative MRI with that of the radiologist and to the findings observed at surgery. What he found was as follows:
- Radiologists in the Phoenix area missed over 48% of clinically important problems and over-interpret other age related changes on the MRI scans as being clinically important.
- The rate of missed readings increased proportionally with patient age (i.e normal degenerative findings on MRI interpreted as being abnormal and pathologic)
- Although the radiologists may see the same structures on the images, they interpret the findings differently and, thus express these in the report as being “abnormal” when, if fact, the findings are normal for age.
- Although, given the limitations of imaging technology (i.e. MRI is not an actual picture but a computer generated image created by manipulating hydrogen molecules in tissues with a magnet), Dr. Bailie was over 93% accurate in interpreting clinical relevance when correlating his findings to those found at surgery (gold standard).
- The most commonly over-interpreted problem noted by the radiologist in the shoulder was SLAP labral tears (stating a tear exists when it really did not). This is similar to what has been reported around the U.S. and has led to significant inappropriate rise in SLAP labral repairs, according to the ABOS.
- The most commonly missed treatable problem in the shoulder, was a concealed high grade partial rotator cuff tear (i.e stated a rear did not exist when in fact it was present).
- In the knee, the most common mistake made by radiologists was failing to provide the clinically important details associated with degenerative meniscus tears (such as extrusion, subchondral edema ect), leading to over-interpretation of normal age-related changes as “pathological tears”. The orthopedic literature has shown that all degenerative knees have some degree of meniscus tearing and not all tears are clinically significant or contribute to pain.
Multiple published studies, performed in the U.S. and other countries, have since supported his findings in both the knee and shoulder. For example, Susan van Grinsven, et. al., published a study out of The Netherlands recently in the Journal of Shoulder and Elbow Surgery. They found that EXPERIENCED Orthopedic Surgeons, who took the time to carefully review MRI scans and correlate the images with clinical and surgical findings, interpreted MRI scans of the shoulder for instability related findings more accurately than their radiology counterparts. The important part of this study was that, this was only true of experienced surgeons and not all surgeons.
Dr. Bailie concluded that without clinical correlation (e.g. MRI findings should match the symptoms and complaints of the patient), surgery was much less likely to be successful. It is VERY IMPORTANT that Dr. Bailie personally have all imaging studies available for review during your appointment and not just a report of the study.
Dr. Bailie firmly believes it is important that physicians continually learn and critically evaluate themselves, including their ability to interpret MRI scans so they may provide the best treatment options for their patients.