SmmaryIn 14 consecutive patients with recurrent melanoma of the lower limb a total of 35 biopsies were taken at the end of perfusion treatment to assess melphalan tissue concentrations in tumour, skin/subcutis and muscle tissue. In tumour tissue (n = 12) the mean melphalan concentration was 6.8 pg g-', which was significantly higher than that of healthy skin/subcutis (3.2 zg g-'; n = 10), but equal to that of muscle tissue (6.5 pg g-'; n = 13). The correlation between melphalan concentration in the tissues and the concentration in the perfusate was studied. The latter was assessed in the form of melphalan peak concentration and the area under the curve (AUCO,60) of the melphalan concentration-time curve. Tumour concentration proved to be correlated linearly with AUCO-.w (R = 0.6, P = 0.002) and muscle concentration with melphalan peak concentration (R = 0.8, P = 0.04). There was no relation between skin/subeutis concentrations and the perfusate parameters. Further research is warranted to study the relationship between melphalan tissue concentration, tumour response and regional toxicity.With regional isolated perfusion, high levels of melphalan can be achieved in the vasculature of a limb with no or negligible leakage to the systemic circulation (Kroon, 1988 (Kroon, 1988). Melphalan as single dose was gradually injected (over one circulation time of the circuit, i.e. within about 2-3 min) into the arterial line. The total perfusate volume measured a median of 750 ml with a haematocrit of about 0.25.During the perfusion perfusate samples were taken from the venous line at 5 min intervals and analysed for melphalan concentration by high-performance liquid chromatography (HPLC) assay (Chang et al., 1978). Melphalan peak concentration and AUCO.60 were assessed. Tumour, skin/subcutis and muscle biopsies were taken at the end of perfusion. The method of analysing the tissue samples was as described by Scott et al. (1990). The specimens were snap frozen and stored at -20'C for batch analysis. They were later thawed, weighed and homogenised in a known volume of acidic buffer. Duplicate specimens were then assayed by HPLC. Differences in melphalan concentration between tumour, skin/subcutis and muscle tissue were pairwise analysed using the Wilcoxon signed-rank test. Mean values are given with the standard deviation.
ResultsA total of 35 tissue biopsies were taken, consisting of 12 tumour, 10 skin/subcutis and 13 muscle specimens (Table I). From eight patients three tissue biopsies were available for analysis. The mean melphalan concentration was 6.8 (4.8) lg g-' for tumour biopsies, 3.2 (2.5) pg gI for skin/ subcutis biopsies and 6.5 (3.7)1Lggg' for muscle biopsies. There was a significant difference in melphalan concentration between tumour and skin/subcutis biopsies (P = 0.01) as well as between muscle and skin/subcutis biopsies (P = 0.01).The mean melphalan peak concentration was 48.9 (18.2)ILgml-' and the mean AUCO.