Re: Adrenal pheochromocytoma: correlation of MRI appearances with histology and function

Abstract

Dear Editor, I read with great interest the article by Dr. Jacques and colleagues entitled “Adrenal pheochromocytoma: correlation of MRI appearances with histology and function” in the December 2008 issue of European Radiology [1]. The authors concluded that hyperintense pheochromocytomas on T2weighted MR images were relatively uncommon compared to findings of previously described studies [2, 3]. I would like to comment on the “Materials and Methods” and “Results” as several questions have arisen after reading the article. The authors did not consider lesion size for analyzing MR imaging features. The size of pheochromocytomas seemed to be larger than that described in previously reported studies. The mean lesion size shown in their study was 5.5 cm (range, 1.2–15 cm) [1], while that reported in previous studies has ranged from 3.6–4.2 cm (range, 1.1–10.7 cm) [2, 3]. Smaller pheochromocytomas tend to be homogenous, but larger lesions tend to be heterogeneous [2] since hemorrhage, necrosis, cystic change and calcification are more frequent as the lesions increase in size. Therefore, the authors should stratify the lesions in terms of size (for example, ≤ 3 cm, 3–6 cm, > 6 cm) and compare the signal intensities of the lesions as depicted on T2weighted images. If stratification of lesions is carried out, hyperintense pheochromocytomas on T2-weighted MR images might be more common in smaller lesions than in larger lesions. The authors also documented that homogeneous pheochromocytomas were significantly smaller than heterogeneous lesions. Subsequently, MR imaging features may be much different according to the proportion or distribution of lesion size. Investigators should be cautious in analyzing or interpreting their data because of selection bias possibly leading to a very different conclusion in a retrospective study. The authors correlated T2 appearances with pathological findings by means of reviewing histopathology reports. If a tumor is surgically removed, the orientation of the lesion is quite different from the in situ location. For this reason, a lesion-bylesion correlation between MRI and pathological examinations is not so easy to perform solely relying on the findings of a histopathological report. Another factor that may limit the radiological-pathological correlation includes the time interval between MRI examination and surgery. If the time is substantially long, the correlation does not seem to be meaningful. Nevertheless, the authors did not provide the length of the time interval. It is hoped that the authors’ responses may help readers who want to better understand the study, and be useful to those wishing to perform a similar study. A reply to this letter can be found at doi:10.1007/s00330-009-1386-6.