Excision margins in high-risk malignant melanoma: 1 cm versus 3 cm
N = 900
- Population
- Patients with a primary cutaneous melanoma of at least 2 mm thickness (first report) or greater than 2 mm Breslow thickness (long-term report), on the trunk or limbs.
- Design
- Randomised, open-label multicentre trial at 59 hospitals (57 in the UK, one in Poland, one in South Africa); patients with a single primary localised cutaneous melanoma thicker than 2 mm on the trunk or limbs (excluding palms and soles) were randomised 1:1 centrally to a 1 cm or 3 cm excision margin after initial surgery, between 16 December 1992 and 22 May 2001. Not registered (predated mandatory registration). First report median follow-up 60 months. In the long-term analysis the 8.8-year median is the follow-up among patients not known to have died (censoring at death); median follow-up for all patients was 5.7 years (68 months), per the full text of the 2016 report. Planned sample size 600, increased to 900 because of a lower recurrence rate than expected.
- Primary endpoint
- Melanoma locoregional recurrence, DFS
- Arms
- 1 cm margin (experimental, n = 453): 1 cm excision margin
- 3 cm margin (control, n = 447): 3 cm excision margin
- People
- Thomas JM, senior author Bliss JM
Melanoma locoregional recurrenceHR 1.26 (95% CI 1.00–1.59)MSSHR 1.24 (95% CI 0.96–1.61)OSHR 1.07 (95% CI 0.85–1.36)MSSHR 1.24 (95% CI 1.01–1.53)OSHR 1.14 (95% CI 0.96–1.36)Δ Surgical complications−7 points
Show the numbers
| Endpoint | Analysis | 1 cm margin | 3 cm margin | Effect |
|---|---|---|---|---|
| Melanoma locoregional recurrenceRandomised patients (900) | primarymedian follow-up 60 monthsThomas 2004 | — | — | HR 1.26 (95% CI 1.00–1.59)P=0.05Hazard ratio for 1 cm vs 3 cm margins; 168 locoregional recurrences as first events with 1 cm margins versus 142 with 3 cm margins. Primary endpoint of the original trial (per the 2016 report); the 2016 report quotes a multivariable-adjusted HR of 1.34 (1.06-1.71, p=0.02) for locoregional relapse-free survival, which is not recorded here. |
| MSSRandomised patients (900) | primarymedian follow-up 60 monthsThomas 2004 | — | — | HR 1.24 (95% CI 0.96–1.61)P=0.1Hazard ratio for death attributable to melanoma, 1 cm vs 3 cm margins; 128 vs 105 deaths attributable to melanoma. Original secondary endpoint (per the 2016 report). |
| OSRandomised patients (900) | primarymedian follow-up 60 monthsThomas 2004 | — | — | HR 1.07 (95% CI 0.85–1.36)P=0.6Hazard ratio for death, 1 cm vs 3 cm margins; the abstract describes overall survival as similar in the two groups. Original secondary endpoint (per the 2016 report). |
| MSSRandomised patients (900), intention-to-treat | long term-updatemedian follow-up 8.8 yearsHayes 2016 | — | — | HR 1.24 (95% CI 1.01–1.53)P=0.041Unadjusted hazard ratio for deaths attributed to melanoma, 1 cm vs 3 cm margins; 194 vs 165 such deaths among 494 deaths overall (359 attributed to melanoma). The 8.8-year median follow-up is among patients not known to have died; for all patients it was 5.7 years. Secondary endpoint of the original trial. |
| OSRandomised patients (900), intention-to-treat | long term-updatemedian follow-up 8.8 yearsHayes 2016 | — | — | HR 1.14 (95% CI 0.96–1.36)P=0.14Unadjusted hazard ratio for death from any cause, 1 cm vs 3 cm margins; 253 vs 241 deaths overall, a difference the abstract describes as not significant. The 8.8-year median follow-up is among patients not known to have died; for all patients it was 5.7 years. Secondary endpoint of the original trial. |
| Surgical complicationsRandomised patients (900) | long term-updateHayes 2016 | 8% | 15% | Surgical complications were reported in 35 patients with 1 cm margins and 65 with 3 cm margins; no p-value is given in the abstract. |
| Endpoint | Analysis | 1 cm margin | 3 cm margin | Effect |
|---|---|---|---|---|
| Melanoma locoregional recurrenceRandomised patients (900) | primarymedian follow-up 60 monthsThomas 2004 | — | — | HR 1.26 (95% CI 1.00–1.59)P=0.05Hazard ratio for 1 cm vs 3 cm margins; 168 locoregional recurrences as first events with 1 cm margins versus 142 with 3 cm margins. Primary endpoint of the original trial (per the 2016 report); the 2016 report quotes a multivariable-adjusted HR of 1.34 (1.06-1.71, p=0.02) for locoregional relapse-free survival, which is not recorded here. |
| MSSRandomised patients (900) | primarymedian follow-up 60 monthsThomas 2004 | — | — | HR 1.24 (95% CI 0.96–1.61)P=0.1Hazard ratio for death attributable to melanoma, 1 cm vs 3 cm margins; 128 vs 105 deaths attributable to melanoma. Original secondary endpoint (per the 2016 report). |
| OSRandomised patients (900) | primarymedian follow-up 60 monthsThomas 2004 | — | — | HR 1.07 (95% CI 0.85–1.36)P=0.6Hazard ratio for death, 1 cm vs 3 cm margins; the abstract describes overall survival as similar in the two groups. Original secondary endpoint (per the 2016 report). |
| MSSRandomised patients (900), intention-to-treat | long term-updatemedian follow-up 8.8 yearsHayes 2016 | — | — | HR 1.24 (95% CI 1.01–1.53)P=0.041Unadjusted hazard ratio for deaths attributed to melanoma, 1 cm vs 3 cm margins; 194 vs 165 such deaths among 494 deaths overall (359 attributed to melanoma). The 8.8-year median follow-up is among patients not known to have died; for all patients it was 5.7 years. Secondary endpoint of the original trial. |
| OSRandomised patients (900), intention-to-treat | long term-updatemedian follow-up 8.8 yearsHayes 2016 | — | — | HR 1.14 (95% CI 0.96–1.36)P=0.14Unadjusted hazard ratio for death from any cause, 1 cm vs 3 cm margins; 253 vs 241 deaths overall, a difference the abstract describes as not significant. The 8.8-year median follow-up is among patients not known to have died; for all patients it was 5.7 years. Secondary endpoint of the original trial. |
| Surgical complicationsRandomised patients (900) | long term-updateHayes 2016 | 8% | 15% | Surgical complications were reported in 35 patients with 1 cm margins and 65 with 3 cm margins; no p-value is given in the abstract. |
Data table
| Endpoint | HR | 95% CI |
|---|---|---|
| Melanoma locoregional recurrence | 1.26 | 1–1.59 |
| MSS | 1.24 | 0.96–1.61 |
| OS | 1.07 | 0.85–1.36 |
| MSS | 1.24 | 1.01–1.53 |
| OS | 1.14 | 0.96–1.36 |
Appears in
Sources
Thomas JM, Newton-Bishop J, A'Hern R, Coombes G, Timmons M, Evans J, et al. Excision margins in high-risk malignant melanoma. N Engl J Med. 2004;350(8):757-66.
Hayes AJ, Maynard L, Coombes G, Newton-Bishop J, Timmons M, Cook M, et al. Wide versus narrow excision margins for high-risk, primary cutaneous melanomas: long-term follow-up of survival in a randomised trial. Lancet Oncol. 2016;17(2):184-192.