Background This prospective study was designed to test the hypothesis that
abnormal liver blood now is related to poor prognosis in patients with colo
rectal cancer.
Methods The hepatic perfusion index (HPI), measured by dynamic hepatic scin
tigraphy, was assessed in 202 patients with colorectal cancer. Assessment f
or overt hepatic metastasis included liver palpation at laparotomy and peri
operative computed tomography (CT). Follow-up at a dedicated clinic include
d regular abdominal ultrasonography and CT.
Results The HPI was abnormal (greater than 0.37) in 92 (88 per cent) of 105
patients with overt liver metastases. Of 89 patients with no evidence of o
vert metastases or residual tumour after primary resection, 52 had an abnor
mal and 37 a normal HPI. At a median follow-up of 39 (range 13-76) months,
25 of 38 patients with recurrnce had an abnormal HPI. Some 31 of 45 patient
s who died had an abnormal HPI. The HPI predicted overall recurrence (P = 0
.04, log rank test). Multivariate analysis showed the HPI was independent o
f Dukes stage for predicting disease-free survival (P = 0.04, relative risk
1.94 (95 per cent confidence interval (c.i.) 1.03-3.67) failed to attain s
ignificance for overall survival (P = 0.055, relative hazard 1.88 (95 per c
ent c.i. 1.00-3.58)).
Conclusion The HPI predicts a poor outcome in patients with colorectal canc
er and may be useful in patient selection for adjuvant chemotherapy.