Dear Sir,

Isolated limb perfusion (ILP) was introduced in clinical practice in the 1950s and earlier studies using hyperthermic ILP with melphalan for limb melanoma reported impressive results with overall response (OR) rates of 80% and complete response (CR) rates of 30 and−50% [1]. More recent reports indicate that CR rates exceeding 50% are now obtained in most melanoma treatment centres where ILP is undertaken [2] and even higher CR rates have been reported when tumour necrosis factor (TNF) has been used with melphalan [3].

We read with interest the paper by Romics in your June 2011 issue utilising melphalan ± TNF alpha in patients with unresectable melanoma of the limb and reporting an OR rate of 93% and CR rate of 67%. ILP is also undertaken in our unit and we recently audited our results which are summarised below.

From 1980 to 2001, 378 patients with limb malignant melanoma underwent therapeutic ILPs at St Mary’s Hospital, London. 41 patients had ILP of the upper limb while 337 had ILP of the lower limb (236 iliac ILP/101 femoral ILP). We employed the melphalan and mitomycin C drug regimen. 284/378 patients had complete follow-up and of these 32% had a CR, 38% had >50% response, 25% had <50% response and 6% had no response. There was a 13% leak rate in responders and 21% in non-responders. There was no postoperative mortality, while perioperative morbidity was seen in 50/284 patients (18%) and included limb oedema (34), deep venous thrombosis (8), vascular damage (4) and transient nerve/muscle damage (2). 60% of our patients survived more than a year on 5-year follow-up.

ILP is a technically demanding, costly and time-consuming technique which should be undertaken in tertiary referral centres by experts. Limb amputation for nonresectable disease can lead to stump recurrences which are difficult to control and make the fitting of a prosthesis challenging. However, with good patient selection, ILP is a valuable therapeutic option and in our experience it is preferable to amputation as an effective therapeutic modality for nonresectable limb melanoma despite the significant risk of regional and systemic complications and lack of an effect on survival.

We, the undersigned, are pleased for an opportunity to report our results to augment the ever-increasing pool of knowledge relevant to this technique.