This Article Full Text (PDF) Purchase Article View Shopping Cart Alert me when this article is cited Alert me if a correction is posted Services Email this article to a colleague Similar articles in this journal Alert me to new issues of the journal Save to my personal folders Download to citation manager Rights & Permissions Citing Articles Citing Articles via Google Scholar Google Scholar Articles by Thomas, J. M. Search for Related Content PubMed Articles by Thomas, J. M. Social Bookmarking                            What's this?

In Reply

Royal Marsden Hopsital, Fulham Road, London, United Kingdom

The letter from Drs Kretschmer and Hilgers defines the four most important controversies concerning sentinel node biopsy (SNB) in melanoma.

We begin with statistical accuracy relating to survival. Kretschmer and Hilgers, like Morton¹ and Thompson,^2,3 have argued that it is valid and legitimate within the Multicenter Selective Lymphadenectomy Trial (MSLT)-I to compare the survival of patients in the control arm (wide local excision and observation) who developed palpable nodal recurrence, with patients in the test arm (wide local excision, SNB with or without complete lymphadenectomy) who are SN-positive. In other words, making the assumption that "most if not all occult SN-metastases will eventually become palpable nodal recurrences";⁴ meaning that these two groups of patients are biologically comparable. That is almost certainly not the case and furthermore, a survival comparison of the two groups of patients in question is statistically invalid. In a randomized controlled trial, survival comparisons must be made on an intention-to-treat basis and from the point of random assignment. Patients can be identified and stratified before random assignment. Within MSLT-I, for example, this could apply to prognostic factors such as primary tumor thickness and site. However, it was not known before random assignment which patients would develop palpable nodal recurrences and which patients would be SN-positive. Therefore, these patients represent selected subgroups, which invalidate any statistical comparison. This error is an elementary and fundamental misunderstanding of the rules of statistical methodology.

Next we examine the argument relating to false-positivity within the SN. This means small clumps of melanoma cells, which are not destined to progress to palpable nodal recurrence, a possibility discussed in the original protocol of MSLT-I. The results published up to this point⁴ state that 144 of 800 patients in the control arm (18.1%) developed palpable nodal recurrence while 229 of 1,173 patients (19.5%) were SN-positive in the test arm and proceeded to completion lymphadenectomy (CL). The assumption that these two groups of patients are biologically comparable is made because these percentages are similar. However, that assumption ignores the fact that there were 59 false-negative patients among the 944 SN-negative patients in the test arm who also must have been SN-positive, but this was not identified by the SNB procedure. Therefore, it follows that if there is no overall survival difference at the point of random assignment, there must be 59 good-prognosis patients among the 229 SN-positive patients to counterbalance the 59 false-negative patients who would have had a poor prognosis. Therefore, the approximate incidence of false-positivity is 59 of 229 (26%). This would account for the prognostic difference between the two selected sub-groups within MSLT-I. In 2005, in his presentation of these results at the 41st Annual Meeting of the American Society of Clinical Oncology, Morton¹ claimed that the 5-year survival was significantly higher in SN-positive patients after CL when compared with patients with palpable nodal recurrence (71% v 55%; P = .0033). This apparent survival difference could simply be the results of a difference in prognosis rather than the result of a therapeutic benefit from CL. This author’s hypothesis is entirely consistent with the lack of any overall survival benefit at the point of random assignment whereas the hypothesis of biologic equivalence, with one group gaining a therapeutic benefit from CL, is not. To draw comparisons and conclusions between postrandom assignment subgroups is at best conjecture and at worst scientific manipulation. Therefore, it is likely that approximately 26% of SN-positive patients are given inaccurate information about their prognosis, undergo unnecessary CL, and possibly unnecessary adjuvant therapy.⁵ False-positivity within the SN is rarely discussed in the literature and this author predicts that it will eventually be accepted as one of the limiting factors of the SNB procedure.

Next, we discuss our hypothesis⁶ that SN-positive patients who undergo CL have an increased incidence of in-transit metastases (ITM). This possibility has been rejected by Pawlik et al⁷ who state that the increased incidence of ITM after CL is entirely due to more aggressive disease in that group of patients. However, apart from the articles listed in our literature overview,⁶ others studies have come to the same conclusion. For example, Borgstein et al⁸ and de Vries et al⁹ describe the incidence of ITM as four and five times greater (respectively) in SN-positive patients as compared with SN-negative patients. Also, Pawlik et al⁷ failed to quote Stehlin et al,¹⁰ their predecessors at M.D. Anderson Hospital (Houston, TX), who described ITM as a complication of lymph node dissection and stated that "a causal relationship between lymphatic obstruction and in-transit metastases is unequestionable." In his presentations of the results of MSLT-I, Morton has stated¹ that CL does not increase the incidence of ITM, but in the published results⁴ we are told that 52 of 944 SN-negative patients (5.5%) developed ITM. Unfortunately, no figure is given for the incidence in SN-positive patients nor for patients in the control arm. Therefore, this question remains unresolved and CL may be associated with an iatrogenic risk. This possibility should be explained to patients during the consent process in preparation for SNB.

Finally, we address the prognostic importance of SN-status and the claim that the procedure is patient-driven in order to gain more information about their individual prognosis. It has been shown that SN status is a better prognostic factor than other primary tumor characteristics when compared individually,¹¹ but it is unlikely that algorithms of such prognostic factors,¹² combined with ultrasound assessment of the SN basin(s)¹³ will not provide similar information. Therefore, if the SNB procedure is indeed patient driven, this may be on the basis of inaccurate information and data interpretation relating to the importance and reliability of SN status. For example, we know that approximately 13% of SN-negative patients will recur somewhere by 5 years.¹¹

For the present, the SNB procedure can only benefit doctors and industry. In the absence of effective adjuvant therapy, it cannot benefit patients and may carry an iatrogenic risk.

Author's Disclosures of Potential Conflicts of Interest

The author indicated no potential conflicts of interest.

REFERENCES

1. Morton DL, Thompson JF, Cochran AJ, et al: Interim results of the Multicenter Selective Lymphadenectomy Trial (MSLT-1) in clinical stage I melanoma. J Clin Oncol 23:710s, 2005 (suppl; abstr 7500)

2. Thomas JM: Time for comprehensive reporting of MSLT-I-Authors’ reply. Lancet Oncol 7:9-11, 2006[CrossRef][Medline]

3. Thompson JF, Uren RF, Shaw HM: Time for comprehensive reporting of MSLT-I–Authors’ reply. Lancet Oncol 7:11-12, 2006[CrossRef]

4. Morton DL, Cochran AJ, Thompson JF, et al: Sentinel node biopsy for early-stage melanoma: Accuracy and morbidity in MSLT-I. Ann Surg 242:302-313, 2005[Medline]

5. Thomas JM. Caution with sentinel node biopsy in cutaneous melanoma. Br J Surg 2006; 93:129-130.[CrossRef][Medline]

6. Thomas JM, Clark MA: Selective lymphadenectomy in sentinel node-positive patients may increase the risk of local/in-transit recurrence in malignant melanoma. Eur J Surg Oncol 30:686-691, 2004[CrossRef][Medline]

7. Pawlik TM, Ross MI, Thompson JF, et al: Low risk of in-transit metastasis in patients with cutaneous melanoma undergoing sentinel lymph node biopsy. J Clin Oncol 23:4588-4590, 2005[Free Full Text]

8. Borgstein PJ, Meijer S, van Diest PJ: Are locoregional cutaneous metastases in melanoma predictable? Ann Surg Oncol 6:315-321, 1999[Abstract]

9. de Vries M, Jager PL, Suurmeijer AJH, et al: Schildwachtklierbiopsie bij het melanoom: Prognostische betekenis en nadelen bij 300 patienten. Ned Tijdschr Geneeskd 149:1845-1851, 2005 (English Translation)[Medline]

10. Stehlin JS, Smith JL, Jing B-S, et al: Melanomas of the extremities complicated by in-transit metastases. Surg Gyn Obstet 122:3-14, 1966

11. Gershenwald JE, Thompson W, Mansfield PE, et al: Multi-institutional melanoma lymphatic mapping experience: The prognostic value of sentinel lymph node status in 612 Stage I or II melanoma patients. J Clin Oncol 17:976-983, 1999[Abstract/Free Full Text]

12. Kruper L, Botbyl B, Czerniecki B, et al: Predicting sentinel lymph node status in stage I/II melanoma. J Clin Oncol 23:710s, 2005 (suppl; abstr 7501)

13. Rossi CR, Mocellin S, Scagnet B, et al: The role of pre-operative ultrasound scan in detecting lymph node metastasis before sentinel node biopsy in melanoma patients. J Surg Oncol 83:80-84, 2003[CrossRef][Medline]

CiteULike    Complore    Connotea    Del.icio.us    Digg    Facebook    Reddit    Technorati    Twitter    What's this?

This Article Full Text (PDF) Purchase Article View Shopping Cart Alert me when this article is cited Alert me if a correction is posted Services Email this article to a colleague Similar articles in this journal Alert me to new issues of the journal Save to my personal folders Download to citation manager Rights & Permissions Citing Articles Citing Articles via Google Scholar Google Scholar Articles by Thomas, J. M. Search for Related Content PubMed Articles by Thomas, J. M. Social Bookmarking                            What's this?