MSLT-II
Completion dissection or observation for sentinel-node metastasis in melanoma
Phase 3N = 1,939NCT00297895
- Population
- Patients with melanoma and sentinel-node metastases; 1939 randomised, 1934 evaluable in the intention-to-treat analysis and 1755 in the per-protocol analysis.
- Design
- International, multicentre randomised phase 3 trial (the full text calls it phase 3; PubMed lists the publication type as phase III; ClinicalTrials.gov lists the phase as NA): patients with sentinel-node metastases (detected by standard pathological assessment or a multimarker molecular assay) were assigned 1:1 to immediate completion lymph-node dissection or nodal observation with ultrasonography. The primary endpoint was melanoma-specific survival; secondary endpoints included disease-free survival and the cumulative rate of nonsentinel-node metastasis. 1939 patients underwent randomisation (matching the registry enrolment); 1934 had data evaluable in the intention-to-treat analysis and 1755 in the per-protocol analysis, which the authors report because it is the most clinically pertinent.
- Primary endpoint
- MSS
- Arms
- Completion lymph-node dissection (control): Immediate completion lymph-node dissection after sentinel-node metastasis
- Nodal observation (experimental): Nodal observation with ultrasonography
- People
- Faries MB, senior author Elashoff RM
Δ 36-month MSS0 pointsΔ 36-month DFS−5 pointsΔ 36-month Regional nodal disease control−15 pointsΔ Lymphedema−17.8 points
Show the numbers
| Endpoint | Analysis | Nodal observation | Completion lymph-node dissection | Effect |
|---|---|---|---|---|
| MSSPer-protocol analysis (1755 patients) | primarymedian follow-up 43 monthsFaries 2017 | 86% at 3 y | 86% at 3 y | P=0.42 (log-rank)Mean 3-year rates reported with standard errors (plus or minus 1.3 and 1.2), not confidence intervals. The abstract states that dissection was not associated with increased melanoma-specific survival in the intention-to-treat analysis (1934 patients) either, without giving rates. |
| DFSPer-protocol analysis (1755 patients) | primarymedian follow-up 43 monthsFaries 2017 | 63% at 3 y | 68% at 3 y | P=0.05 (log-rank)Disease-free survival was a secondary endpoint. Mean 3-year rates with standard errors (plus or minus 1.7 for each arm). The authors state the difference rests on better disease control in the regional nodes and must be interpreted with caution. |
| Regional nodal disease controlPer-protocol analysis (1755 patients) | primaryFaries 2017 | 77% at 3 y | 92% at 3 y | P<0.001 (log-rank)Mean 3-year rates with standard errors (plus or minus 1.0 and 1.5). |
| LymphedemaDissection and observation groups as reported (full text: 'patients in the dissection group' and 'in the observation group'; analysis population not further specified) | primaryFaries 2017 | 6.3% | 24.1% | P<0.001Lymphedema at the most recent follow-up (30 April 2016); the full text reports P<0.001 and that, among patients with lymphedema, it was mild in 64%, moderate in 33% and severe in 3% (all patients combined). |
| Endpoint | Analysis | Nodal observation | Completion lymph-node dissection | Effect |
|---|---|---|---|---|
| MSSPer-protocol analysis (1755 patients) | primarymedian follow-up 43 monthsFaries 2017 | 86% at 3 y | 86% at 3 y | P=0.42 (log-rank)Mean 3-year rates reported with standard errors (plus or minus 1.3 and 1.2), not confidence intervals. The abstract states that dissection was not associated with increased melanoma-specific survival in the intention-to-treat analysis (1934 patients) either, without giving rates. |
| DFSPer-protocol analysis (1755 patients) | primarymedian follow-up 43 monthsFaries 2017 | 63% at 3 y | 68% at 3 y | P=0.05 (log-rank)Disease-free survival was a secondary endpoint. Mean 3-year rates with standard errors (plus or minus 1.7 for each arm). The authors state the difference rests on better disease control in the regional nodes and must be interpreted with caution. |
| Regional nodal disease controlPer-protocol analysis (1755 patients) | primaryFaries 2017 | 77% at 3 y | 92% at 3 y | P<0.001 (log-rank)Mean 3-year rates with standard errors (plus or minus 1.0 and 1.5). |
| LymphedemaDissection and observation groups as reported (full text: 'patients in the dissection group' and 'in the observation group'; analysis population not further specified) | primaryFaries 2017 | 6.3% | 24.1% | P<0.001Lymphedema at the most recent follow-up (30 April 2016); the full text reports P<0.001 and that, among patients with lymphedema, it was mild in 64%, moderate in 33% and severe in 3% (all patients combined). |