PURPOSE: It is commonly believed that the right lobe of the liver is t
he most frequent site of colorectal liver metastases. However direct e
vidence for this is lacking in the literature. This study was designed
to document and evaluate the pattern of liver metastases according to
liver segments. METHODS: A retrospective medical records review of 27
0 patients with hepatic metastases from colorectal adenocarcinoma was
performed to determine the pattern of metastases according to Couinaud
's segments. There was operative confirmation of liver metastases in 2
02 patients (75 percent). A total of 1,166 segments were involved with
liver metastases and were used in this analysis. Segment I was exclud
ed from calculations because of its dual blood supply. When groups con
tained different numbers of Couinaud's segments an appropriate correct
ion was made in the statistical comparison. RESULTS: There were 736 (6
3 percent) right lobe segments (V, VI, VII, VIII) and 430 (37 percent)
left lobe segments (II, III, IV) involved with metastases (P < 0.001)
. Of 430 left lobe segments, 298 (69 percent) involved the left latera
l segment (II, III) and 132 (31 percent) involved the medial segment (
IV) of the left lobe (P = 0.25). When evaluating patients with 4 or le
ss segments involved, there were 258 right lobe segments (75 percent)
and 88 left lobe segments (25 percent) involved with metastases (P < 0
.0001). Of 88 left lobe segments with 4 or less segments involved, the
re were 68 segments (77 percent) involving the left lateral segment an
d 20 segments (23 percent) involving the medial segment of the left lo
be (P = 0.035). All. of these relationships were statistically indepen
dent of the primary tumor site. CONCLUSIONS: We have documented an ove
rall predominance of right hepatic lobe metastases independent of site
of the primary colorectal carcinoma. However, when metastases occurre
d in the left lobe, the left lateral segment (II, III) was more common
ly affected than the medial segment (TV), but this difference was stat
istically significant only when there were four or less segments invol
ved with liver metastases.