Autopsy as a quality control measure for radiology, and vice versa

Radiologic studies are anatomic studies lacking the natural full-color, 3D, and microscopic-level examination of autopsies, suggesting that autopsies might be able to serve as quality control for radiology.

Cases in which complete or near-complete autopsies were performed at a university hospital in 2008 were reviewed, and antemortem radiologic diagnoses were compared with corresponding autopsy findings. Discrepancies between antemortem radiologic diagnoses and autopsy findings were categorized.

For 729 of the 828 diagnoses reviewed in the study, the pathologic condition in question was thought to be present at the time that a radiologic study of the relevant anatomic region was performed. Of these 729 radiologic diagnoses, 201 (27.6%) were determined to be discrepant from the corresponding autopsy diagnoses (i.e., autopsy deemed correct), but many of these radiologic discrepancies were not of clinical significance. The radiologic error rate considers only the clinically relevant discrepancies categorized as "missed diagnosis" or "misinterpretation"; it was calculated to be 3.3%. Interestingly, 32 autopsy discrepancies (i.e., radiology deemed correct) were also identified in the study.

The results of this study suggest that even in 2008 patients sometimes died with undiagnosed or misdiagnosed diseases. Radiologic diagnoses discrepant from autopsy findings were consistently identified in this study and show that autopsies can help radiologists sharpen their skills in interpreting radiologic studies and can perhaps serve as quality control for radiology. The results also suggest that radiology can serve as quality control for autopsy.