Arising from: Thomas MJ (2008) Prognostic false-positivity of the sentinel node in melanoma. Nat Clin Pract Oncol 5: 18–23 doi:10.1038/ncponc1014

We would like to take this opportunity to make some comments regarding the paper by Thomas.1 From the period 1994 to 2002, the Multicenter Selective Lymphadenectomy Trial-1 (MSLT-1) was conducted by Morton and collaborators to study the potential therapeutic effects of the sentinel node (SN) procedure, followed by a completion lymph-node dissection (CLND) in instances of SN-positivity, versus the standard therapy, which consisted of wide local excision (WLE) followed by a therapeutic lymph-node dissection (TLND) in case of regional nodal recurrence.2

In his analysis of the MSLT-1 data, Thomas points out that there was no overall survival benefit associated with the conduct of a SN procedure in melanoma patients.1 However, post-randomization subgroup analyses were conducted, and one of the major conclusions of the authors of the MSLT-1 was that there is a survival benefit for SN-positive patients (who all underwent CLND) versus patients who had undergone a WLE and a TLND for clinical regional lymph-node recurrence that emerged during follow-up.

Thomas presents a model that is based on the premise that there must be a proportion of 'prognostic false-positivity' associated with SN biopsy. At a median follow up of more than 5 years, only 15.6% of patients in the WLE arm developed nodal recurrence. Based on the rationale that an equal number of patients in the SN group would have developed a nodal recurrence and thus should also have a positive SN to begin with, the number of SN-positive patients should be 119 (15.6% of 764), which is virtually identical to the presented value of 122. Since there is also a proportion of false-negative patients who develop nodal recurrence after a negative SN procedure, the total number of node positive patients was 148 (19.4%), which is a significantly higher than the 15.6% nodal recurrence rate at 5 years in the WLE-only group. Thomas concludes that approximately 24% (i.e. 148–122 = 29) of the SN-positive patients can be considered false-positive. Similar to other medical tests, the SN procedure is associated with false negative and false positive test results.3

There seems to be a striking similarity with our institutional reported rate of 22% of sub-micrometastases, defined as metastatic clusters of cells with a diameter <0.1mm, compared to the false-positive rate of 24%, as calculated by the model reported by Thomas.1 Moreover, at the 2007 ECCO14 meeting on an EORTC Melanoma Group multicenter study, we reported that in 388 SN-positive patients, those with sub-micrometastases in the SN had an estimated 5-year overall survival rate of 91%, which is the same as that observed in SN-negative patients.4 Furthermore, we reported on the influence of the sub-micrometastases in a study comparing SN-positive patients with those treated with WLE and TLND. The results showed that there seemed to be a 13% survival benefit (P = 0.12) for SN patients, but in those without sub-micrometastases the difference was only 6% (P = 0.42).5 Our studies, therefore, support the concept that SN-positivity in patients with sub-micrometastases can equate to SN-negativity.

A word of caution, however, seems justified. Our observation and the calculation model proposed by Thomas seem to confirm the possibility of “clinical false-positive SN” in melanoma. Nonetheless, we would like to point out that the model described by Thomas is only a model and our histopathologic tumor load classification observations would not have been possible without the surgical removal of the SN in the first place. Our data have the drawback that the median follow-up is limited to 3–3.5 years. It is therefore possible that patients with sub-micrometastases in the SN, are precisely those patients who will develop late tumor recurrence; for example, after 8, 10 or even 15 years. We will therefore expand our studies within the EORTC Melanoma Group network to establish long-term follow-up data on the prognosis of these SN-positive patients, whose tumor load is classified by the Rotterdam Criteria.6

Another important issue discussed by Thomas is that the 5.2% difference in disease-free survival observed in the MSLT-1 is merely a delay instead of a prevention of metastasis, and that this is achieved at the cost of a surgical procedure in all patients. We agree that the main endpoint of such a study (immediate versus delayed treatment) should be distant metastasis-free survival or overall survival. This is reminiscent of the observations in the EORTC 18832 trial, whereby wide excision of the primary plus extra-corporal isolated limb perfusion (ILP) was performed versus wide excision only for stage I high-risk melanoma patients.7 Despite a borderline significant impact effect of the ILP on disease-free survival (P = 0.03), this did not translate into any difference in distant metastasis-free survival or overall survival. Therefore, adjuvant ILP is no longer recommended. Finally, we would like to point out that SN-biopsy can be avoided by preoperative ultrasound of the sentinel lymph nodes in a significant percentage of melanoma patients. This is particularly true for those with stage T3, T4 primary tumors, where micrometastasis can be identified in up to 65% of SN-positive patients, as recently reported in a series of 400 consecutive patients.8

In conclusion, the SN status and moreover the amount of SN tumor burden are the most important prognostic factors for stage I and II melanoma patients. SN status and the amount of SN tumor burden reflect tumor biology and, therefore, are a crucial stratification criterion for adjuvant therapy trials striving to identify new effective adjuvant therapy strategies, as was reported at the ASCO annual meeting in 2007.9