Sentinel Node Biopsy in Melanoma: Lessons Learned From Different Positivity Rates From Different Hospitals
Kinnier et al1 recently evaluated the positivity rates of sentinel lymph node biopsy (SLNB) in 33,639 cases of Stage IA-III melanoma across 646 hospitals in the United States. The ratio between the observed positivity and the expected positivity (O:E ratio) was adopted to divide hospitals into terciles, with an “as expected” category (middle tercicle) having an O:E ratio between 0.61 and 1.17. The overall median SLNB positivity was 8%. Hospitals with a lower than expected (low tercicle) and hospitals with a higher than expected (high tercicle) SLNB positivity were more likely to be low-volume hospitals. In addition, Stage I melanoma patients treated at hospitals of the low tercile had a slightly but significantly worse 5-year survival (90.0%) than patients treated at hospitals of both the middle (91.9%) and the high tercicle (91.6%). The authors concluded that SLNB positivity rate could be a measure for internal quality reporting and that surgery at hospitals with lower than expected SLNB positivity rates was associated with decreased survival.
We agree with the authors that some interinstitutional discrepancies in SLNB positivity rates are due to different pathology protocols.2 Nevertheless, in our opinion, false positives due to immunohistochemistry, as speculated by the authors, are highly unlikely or, at least, are highly unlikely to be the unique reason for higher than expected SLNB positivity rates. Some low-volume hospitals often adopt more thorough pathology protocols because these have an acceptable impact on their routine workload. Riber-Hansen et al3 demonstrated that a complete step-sectioning of the SLNB samples allowed to detect 28% of nodal metastases that would have been missed by examining only 6 cut levels as per the EANM-EORTC protocol.4 However, detecting a greater number of metastases might be not always truly relevant, because a proportion of positive SLNB in melanoma patients might be “prognostic false positives” (percent difference between the SLNB-positive cases and the nodal recurrences in an observation group in which SLNB is not performed).5 Kinnier et al1 found that the difference in the SLNB positivity rate between hospitals of the middle tercicle and hospitals of the high tercicle had no prognostic relevance; in our opinion, Kinnier et al's finding is indeed a strong argument: i) in favor of the existence of “prognostic false positives”;5 ii) against the implementation of exceedingly accurate pathology protocols (ie, complete step sectioning of the node2; molecular upstaging6) in SLNB of melanoma.
On the other hand, the worse prognosis of pT1 melanoma patients treated at hospitals with lower than expected SLNB positivity might reflect a problem of false negative histopathology. This assumption, however, cannot be set forth, because the unique criterion that must be adopted to label SLNB negative cases as false negatives are nodal recurrences7 and not survival rates. Indications to SLNB in clinical stage I melanoma are yet to be clearly defined: some variations in patient selection across different institutions make a comparison of the data not reliable. By following the NCCN guidelines,8 some cases of pT1a melanoma, >0.75 mm in Breslow thickness, will be finally labelled as clinically occult stage III; these cases are prone to be understaged not because of the “low quality” of some institutions but simply because some institutions follow the criteria detailed by the American Joint Committee on Cancer 7 staging system.9
For all the above, we think that Kinnier et al's1 conclusive statement must be viewed with caution because of the lack of at least 2 prerequisites, namely standardized protocols for the histolopathologic examination of SLNB specimens and homogeneous indications to SLNB in clinical stage I melanoma.