“…In fact, only 29.2% of patients with benign or most likely benign lesions were included in this group. This small patient number was due to several factors: 1) most of the patients had been referred and had already received CT exams, which explains the small number of MRI performed as the initial examination in the casuistry; this situation contrasts with the recommendations of many authors, who judge the MRI as the exam of choice in the diagnostic evaluation of focal liver lesions
2
,
10
,
12
,
15
,
28
,
25
; the MRI performed in the follow-up examinations in the present study were able to clarify the diagnoses in at least six patients allocated into groups 2 and 3 who had not been diagnosed through CT and who had small hemangiomas, preserved areas of parenchyma in fatty liver and FNH; 2) a significant portion of patients (28.4%) had small lesions between 1 and 2 cm in size; the central scar of FNH and the heterogeneous areas of necrosis or hemorrhage of HA are infrequent in small lesions, which hampers the differential diagnosis of such lesions; 3) the definition of HA was based on rigid criteria, only including cases in which intra- or perilesional hemorrhaging was present; this feature was chosen because bleeding, although uncommon (21 - 40% of cases)
1
, is a reliable finding in the differentiation of FNH; 4) MRI with hepatic-specific contrast was not part of the early stage of the study, having been used in only a few cases in the last two years; the use of hepatic-specific contrast in MRI has been considered as the best way to establish a differential diagnosis between HA and FNH. In a recent meta-analysis that included 10 studies and 304 patients with FNH subjected to MRI with gadoxetic acid (Gd-EOB-DTPA), Suh et al
26
concluded that High/Iso signal intensity on the hepatobiliary phase of the examination occurs in most patients (94-97%) with FNH.…”