Trial Information

A Randomised Trial of Surgical Excision Margins for Thick Primary Cutaneous Melanoma (>2 mm). A Multicenter Trial Comparing 2-cm vs 4-cm

Traditionally CMM have been excised with wide resection margins of 5-cm (sometimes with
10-cm towards the local lymph node basin) with the radical removal of lymph nodes. This
treatment emerged from a recommendation from Handley in 1907 based on a single pathological
specimen. This "radical" surgical management resulted in bad cosmetic results, lymph oedema,
long hospital inpatient stay, frequent skin grafting and/or complicated skin flap
reconstructions. Not until some 60-plus years later did questions arise in clinical
practices whether the need for this extensive surgery was mandated and clinical practice was
not substantially changed until the late 1980's. Retrospective studies published in the
1980s suggested that narrower excision margins may be appropriate for treatment of some
CMMs, especially thinner lesions.

Breslow tumour thickness of the CMM is the most important prognostic indicator of localised
disease and is therefore the information upon which today's surgical strategies are founded.
However, recommendations vary over the world especially for thicker tumors. For CMM of ≤ 1
mm thickness most centers use a 1 cm margin, but for tumours 1.01 - 4 mm the margins of
resection are 1-3 cm depending on the country. Most patients with CMM > 4 mm are operated on
with a margin of 2-cm today. The different national guidelines are thus, somewhat confusing
and in a report from 2004 Thomas showed that a 1-cm margin for CMM with a poor prognosis (≥2
mm) is associated with a greater risk of regional recurrence than in a 3-cm margin, but with
a similar survival rate.

Still, quite sound evidence now exists to state that narrower excisions - for thinner
tumors - is as safe as more wide surgery. To date, five published randomised trials (11
reports) have been published to access what type of surgery to recommend in the different
prognostic groups but despite this effort there are still controversies. In a report by Lens
based on 4 randomised trials the authors concluded that current evidence was not sufficient
to address the optimal surgical margins for all CMM. Furthermore the Cochrane report from
2009 states; "Current randomised trial evidence is insufficient to address optimal excision
margins for primary cutaneous melanoma". However, the studies were not designed to access
"optimal" surgery, they were designed to compare one surgical strategy with another, where
after the results have been interpreted into clinical guidelines.

In conclusion there is a need for additional studies and further research is required
especially for the patients with poorer prognoses, i.e. with tumors >2 mm.

In 1992 a multicenter trial was launched from the Swedish Melanoma Study Group. The 936
patients in the study were included from January 22 1992 to May 19 2004. Patients were
recruited from Sweden ( 6 centres with 644 pat), Denmark (180 pat), Estonia (80 pat) and
Norway (32 pat). Randomisation routines were set up by the steering committee and eligible
patients were randomised locally by telephone calls to national and international cancer
centres (upon a histologically proven diagnoses and signed patient consent form). Only
patients with a CMM >2 mm and with localised disease (who fulfilled the in- and exclusion
criteria) were eligible for study inclusion. Patients with CMM on the hands, feet, head-neck
and ano-genital region were excluded. Final surgery must had been planned within 8 weeks
after date of diagnosis. All analyses were conducted according to the intention-to-treat
principle.

Patients were followed clinically every 3 months for 2 years and thereafter every 6 months
up to 5 years. Follow-up data was thus collected from cancer registries, cause of death
registries and medical records. The overall mean follow-up time was 6 years and 9 months (6
years and 7 months vs. 6 years and 10 months).

Statistical analyses were made by Kaplan Meier life-table curves. Prognostic factors was
assessed with the use of a uni- and multivariate Cox regression analysis.