Ww. Mayosmith et al., CHARACTERIZATION OF ADRENAL MASSES (LESS-THAN-5 CM) BY USE OF CHEMICAL-SHIFT MR-IMAGING - OBSERVER PERFORMANCE VERSUS QUANTITATIVE MEASURES, American journal of roentgenology, 165(1), 1995, pp. 91-95
OBJECTIVE. The purposes of this study were to evaluate the ability of
chemical shift MR imaging to differentiate 1- to 5-cm adrenal adenomas
from metastases and to compare subjective interpretation with several
different quantitative measures. SUBJECTS AND METHODS. Forty-three pa
tients with 46 proved adrenal lesions (28 adenomas and 18 metastases)
had MR imaging with a gradient-echo breath-hold technique and echo tim
e varied to obtain in-phase and out-of-phase images. Qualitative analy
sis of the MR images was done by three experienced observers, who repo
rted their confidence in diagnosing the benignity versus the malignanc
y of the adrenal masses on the basis of signal loss on out-of-phase im
ages. Quantitative analysis was performed by calculating the differenc
e in signal intensity between in-phase and out-of-phase images by use
of regions of interest (signal intensity index) and by use of adrenal-
liver, adrenal-spleen, and adrenal-muscle signal intensity ratios. Res
ults between quantitative and qualitative measures were compared by us
e of receiver operating characteristic (ROC) analysis. RESULTS. The me
an signal intensity was significantly different between adenomas and m
etastases on out-of-phase images (64 versus 98) (p < .0005) but not in
-phase images (130 versus 122) (p = .47). The adrenal-spleen ratio dis
criminated between adenomas and metastases better than did the adrenal
-liver ratio, the adrenal-muscle ratio, or the signal intensity index.
No significant difference in interpretation among the three observers
was evident (areas under the ROC curves, 0.93, 0.95, and 0.96). The p
erformance of the observers was comparable to the results obtained wit
h the adrenal-spleen ratio measurement (area under the ROC curve, 0.97
). CONCLUSION. Experienced observers were able to differentiate adrena
l adenomas from metastases just as well as quantitative measures. The
adrenal-spleen ratio is the best quantitative means of distinguishing
benign from malignant adrenal masses. Chemical shift MR imaging is a g
ood but imperfect discriminator of adrenal adenomas.