Article Text

Download PDFPDF

Assessment of proliferation markers in metastatic melanoma in sentinel lymph nodes
  1. André Joannou-Coetzee1,
  2. Noranette Villena1,
  3. Barry W E M Powell2,
  4. Martin G Cook1,3
  1. 1Department of Histopathology, Royal Surrey County Hospital, Guilford, UK
  2. 2Melanoma Unit, St George's Hospital, London, UK
  3. 3Division of Clinical Medicine, University of Surrey, Guildford, Surrey
  1. Correspondence to Professor Martin G Cook, Department of Histopathology, Royal Surrey County Hospital, Guildford, Surrey GU2 7XX, UK; m.cook{at}nhs.net

Abstract

Aim Some views on sentinel nodes for melanoma seem to cast doubt on the relevance of micrometastases in the sentinel nodes of patients with melanoma, suggesting that small metastases or isolated tumour cells can be ignored. Tumour dormancy has been proposed for their postulated lack of progression. The implication of the argument seems to be that minute metastases are inactive and therefore non-threatening, whereas larger ones are proliferative and therefore have aggressive potential.

Methods 54 sentinel lymph nodes were studied with histologically identified micrometastatic melanoma using the protocol accepted by the European Organisation for Research and Treatment of Cancer melanoma group. These were studied with respect to metastasis size and by use of immunohistochemical markers of proliferation (MIB-1) and dormancy (p16).

Results The authors have demonstrated no correlation between the size of metastases and their proliferative activity. Very small metastases may not show proliferative activity, but this may be a reflection of the small number of assessable cells rather than a genuine reflection of the tumoural characteristics. Furthermore, the minute size of some of these metastases resulted in no residual tumour being present in adjacent sections. Where further sections did show more tumour, these small metastases were invariably p16 negative, suggesting dormancy was not the explanation for the lack of measurable proliferation. Occasionally, larger metastases, clearly not clinically insignificant, showed no proliferative activity presumably, considering their size, a transient phenomenon.

Conclusion These findings suggest that variable phases in proliferation occur in metastases, and no conclusion of clinical insignificance can be made on the basis of small size.

  • Histopathology
  • autopsy pathology
  • encological surgical pathology
  • forensic pathology
  • surgical pathology
  • melanoma
  • sentinel node
  • dermatopathology
  • diagnostics
  • proliferation
  • immunohistochemistry
  • metastasis

Statistics from Altmetric.com

Introduction

Histological examination of the sentinel lymph node (SLN) is a well-established and widely practiced technique used in staging melanoma in patients who have clinically negative lymph nodes. It is usually recommended for patients with primary melanomas of >1 mm in thickness.1 It may also be considered in patients with thin melanomas (<1 mm) that have adverse prognostic factors such as ulceration or a high mitotic count.2 The status of the SLN is a guide to whether a complete lymph node dissection should be performed. It may also be used to select patients who could benefit from trials of adjuvant therapy.

In most studies, around 75% to 80% of SLN do not contain melanoma. More extensive protocols have been developed to increase the detection rate of tumour deposits and thereby reduce the risk of false-negative cases.3 4 These have achieved detection rates up to 33%. However, recently, it has been suggested that there is a possibility of prognostic false positivity.5–7 In other words, the presence of a small tumour deposit of <0.1 mm may indicate that the patient would have the same prognosis as those whose SLN was found to be negative.6 Furthermore, it has been proposed that high-resolution ultrasound of regional lymph node basins could replace the SLN biopsy.8 It is maintained that deposits from 3 to 4 mm in size could be identified by using this technique and that if there were indeed any deposits smaller than this, they could be regarded as insignificant.8 9

The presumption of this argument relies on the premise that small tumour deposits within the SLNs are destined for dormancy or destruction by the host.8 9 However, dormancy is also a proposed mechanism to explain why long disease-free periods exist.10 It is not an indication of permanent inactivity but rather a temporary, albeit sometimes of long duration, cessation of proliferation resulting in a slowing down of the rate of clinical presentation of metastases. These metastases are occult but may prove fatal if not removed before they resume proliferation and further spread. The sentinel node procedure is both a means of accomplishing removal of occult metastases and a staging procedure.

We aimed to assess whether this assumption is valid by measuring the proliferation rates and dormancy in SLN metastases.

Methods

The material consisted of formalin-fixed paraffin-embedded tissue obtained from the histology archives of the Royal Surrey County Hospital. The initial identification of metastatic melanoma in SLN was made by using the protocol accepted by the European Organization for Research and Treatment of Cancer.11 The protocol involves the production and examination of six pairs of sections stained with H&E and S-100 (figure 1A), at 50 μm intervals with spare sections cut at each level.

Figure 1

(A) S-100 positive staining metastatic melanoma (×50 magnification) and (B) MIB-1 staining of metastatic melanoma (×50 magnification) in adjacent section to that in (A).

The cases chosen represented a wide range of size of metastases but with particular emphasis on those that would be considered clinically insignificant in ultrasonographic detection of such metastases in the sense that they were <3 mm.

After confirmation of the diagnosis, immunostaining with MIB-1 was performed on the adjacent section to the one containing metastasis (figure 1B).

The proliferation marker Ki67 was demonstrated using mouse monoclonal biotinylated secondary antibody (clone: MIB-1 (8), Isotype: IgG1κ from Dako, Glostrup, Denmark) followed by peroxidase, strepavidin peroxidase and diaminobenzidin washing and rinsing followed by chromogen development and counterstaining with haematoxylin. Staining for p16 as an indicator of possible dormancy was also performed in cases where tissue containing metastases was still available. We chose p16, monoclonal antihuman p16INK4a antibody, clone E6H4 (MTM Laboratories, Heidelberg, Germany).

The micromorphometrical parameters of a total of 54 SLNs with metastases and one with naevus cells only were identified. The maximum dimensions of the metastases were measured. The proliferation marker, MIB-1, and the dormancy marker, p16, were used in staining the lymph nodes and the metastatic tumour deposits. The total number of cells present within the metastasis was counted along with the number of MIB-1-positive nuclei. No mitoses were identifiable on the standard H&E slides. Lymphocytes were distinguished from melanocytes based on the morphology (cell size and nuclear characteristics) of the respective cells, which were examined at ×640 magnification. The percentages were calculated as the number of MIB-1-positive or p16-positive cells divided by the number of total cells in the metastasis.

Ethical approval was obtained via generic consent using the OMICS technology ethical approval with subsequent approval from the Surrey Research Ethics Committee and the Royal Surrey County Hospital Foundation Trust. The tissue used was archived before 2004 and therefore not covered by the current Human Tissue Act.

Results

The S-100 positively stained micrometastases were identified as in figure 1A and the MIB-1 staining seen as nuclear as in figure 1B. The metastases ranged in maximum dimension from 0.02 to 16 mm (figure 2A). The proportion of cells showing MIB-1 staining ranged from 0% to 75% positive overall (figure 2B).

Figure 2

(A) Range of maximum dimension of metastatic tumour deposits (in millimetres) and (B) percentage MIB-1 positivity arranged in order from smallest to largest metastatic tumour deposit.

The percentage of positive staining showed no correlation with the size of the metastasis. Apart from large metastases (>3 mm), the counts of the whole metastases were assessed and percentage positivity calculated. In the large metastases, a proportion of the tumour was assessed ranging from 30% to 50%.

Some of the cases consisted of no more than five cells, and in three of these, MIB-1-positive cells were present. There were 23 cases where the micrometastases were <0.4 mm. The MIB-1 positivity in these smaller micrometastases ranged from 0% to 60%.

The MIB-1 proliferative index was not significantly related to the maximum dimension of the metastasis or to the microanatomic site of the metastasis.12 One patient had two SLNs removed, both of which contained metastases. One showed 10% MIB-1 positivity, whereas the other 0% positivity.

p16 showed negativity in all but one of the available cases, namely those in which metastases were still present in the residual paraffin blocks. Because we were studying very small micrometastases, there were occasions when the entire tumour burden was represented in the diagnostic sections, leaving none for further assessments. Of the 54 cases, 10 were found to still contain an appropriate amount of tumour to survive further examination in adjacent sections. From these 10 cases, only 6 were found to have tumour in the p16 sections. Focal p16 positivity was found within one metastasis (maximum dimension 0.59 mm), which showed marked pleomorphism, and the proportion of MIB-1 positivity was assessed to be 29%. Its status as a metastasis was not in doubt. The positive staining in this case was cytoplasmic with negative nuclear staining. This pattern of staining has been described in rare cases of melanoma where the staining pattern shows differing patterns of cytoplasmic and intensity of nuclear staining.13 Nuclear p16 staining is nevertheless that which is usually assessed as a potential reflection of dormancy.13 The remaining five cases were completely negative (figures 3 and 4).

Figure 3

Negative p16 staining in metastatic melanoma tumour deposit (×50 magnification).

Figure 4

p16 staining in a control (a benign intradermal melanocytic naevus) and H&E (×50 magnification).

Discussion

MIB-1 is a nuclear protein that is defined by its reactivity with the monoclonal antibody from the MIB-1 clone. Two isoforms of 345 and 395 kDa have been identified. The antigen is preferentially expressed during all active phases of the cell cycle, such as the G1, S, G2 and M phases, and is absent in resting cells (ie, those in G0).14 It is uncertain how many cells that show MIB-1 nuclear positivity will progress to mitosis, but it was chosen, as it has been shown to be superior to mitotic count and other markers of proliferation.14 p16 is the protein product of the CDKN2A gene. It normally regulates the cell cycle by binding and deactivating the cyclin-dependent kinase 4/6 complex (CDK4/6) bound to cyclin D. Loss of p16 results in loss of opposition to the CDK4/6 cyclin D complex at the transition from the G1 to the S phase. The cell therefore cycles faster, and the mitotic rate increases. It is thought that loss of the p16 protein may play a role in melanomagenesis by dysregulation of the cell cycle. p16 is expressed in benign naevi, sun-exposed melanocytes, in situ melanomas and subtypes of melanomas with superficial invasion. Melanoma subtypes associated with vertical growth have consistently shown loss of p16 in most cases. Most importantly, it has been established that lymph node melanoma metastases show loss of p16.15

The staging of melanoma is performed in an attempt to predict the probability of systemic dissemination of melanoma and therefore the likelihood of early death. When tumour deposits are found within the SLN, it suggests that the tumour may have a potential for a wider dissemination. It is evident that tumour seen in an SLN must have passed through a lymphovascular channel to reach the node. These lymphatic emboli are often no more than what some would term isolated tumour cells (<0.2 mm) where they are found in a lymph node. There are occasions when the number of cells in a metastasis is <10. Studies have been performed that suggest sentinel node metastases <0.1 mm in maximum dimension are an insignificant finding in that they are not usually associated with further progression of disease.6

We have demonstrated that metastases of any size in a lymph node usually comprise cells that are MIB-1 positive and are therefore replicating. Very occasionally, deposits are so small that they consist of such a small number of cells and that failure to detect MIB-1 positivity is a shortcoming of the technique or the statistical analysis used to study the deposits and not a confirmation of dormancy. Occasionally, finding a MIB-1 staining percentage of 0% in such small metastases is a statistical inevitability and more likely a reflection of the numbers of cells available for study rather than a true negative.

The technique of sectioning provides only a selective two-dimensional assessment of a three-dimensional structure. This is illustrated by a case in this study. A patient had metastases in two SLNs removed at the same procedure from one primary melanoma. Both contained metastases. One lymph node had a metastasis of 0.15 mm in maximum dimension, amounting to approximately 50 cells in the section of which five cells were MIB-1 positive (10% positive). The second lymph node had a metastasis of 0.12 mm and consisted of 33 cells. None of them were MIB-1 positive (0% positive). Given that the primary melanoma was the same for both, we would have assumed that they would have the same growth or proliferative characteristics rather than that each metastasis had a proliferative potential of its own, although this probably cannot be ruled out. A MIB-1 staining of 10% in the one node would therefore suggest that a proliferation index of 0% in the other node was not a true reflection of this melanoma's proliferative potential. Indeed, all the percentages calculated in this study might be a reflection of the minimal proliferative potential of these metastases. A true reflection of the proliferative potential of the metastases would require assessment of the entire deposit, which is not practicable.

It also suggests that the proliferation index may show a temporal and a spatial variation. It is probable that MIB-1-positive cells would have been detected if further levels had been examined in the cases where 0% was found. Nevertheless, six deposits, all <0.4 mm in maximum dimension, showed no MIB-1 positivity. There were, however, 22 other cases that were ≤0.4 mm in maximum diameter and where the MIB-1 positivity ranged from 2% to as much as 60%. In addition, p16 could only be performed on selected cases where tissue was indeed still available to study. p16 was negative in all but one case where there was focal cytoplasmic but no nuclear positivity. This aberrant staining pattern was similar to that reported in metastatic melanoma occasionally.13

There are previous studies reported in the literature where dormancy of melanoma micrometastases was assessed by using similar methods as in our study by looking at the proliferation fraction.16 These studies, compared with ours, looked at the equilibrium maintained between the tumour proliferating fraction versus the fraction of tumour undergoing elimination through apoptosis. If the two fractions were similar, that would imply that the tumour mass remained static, and this would suggest that the micrometastases were dormant.16 A possible marker of dormancy, such as p16, has to our knowledge not been assessed in such cases. Prolonged quiescence of metastatic melanoma would also require angiogenic dormancy, the existence of which is still currently unknown.10

Metastases of the size detected by sonography (minimum 3 mm in diameter) could not have reached the lymph node as that size. Afferent lymphatic vessels enter the lymph node on its convex surface and are significantly smaller than the efferent lymphatic vessels. Their calibre is approximately 0.5 mm—much smaller than that detectable by sonography. The tumour must have arrived as emboli <0.5 mm in diameter and then grown within the lymph node. We have shown that melanoma cells in metastases have a variable proliferation rate, but of those studied, none were found to be dormant.

The assessment of clinical outcome in SLN-negative cases is subject to the limitations of the techniques, and there is a possibility of false negativity. The more favourable prognosis seen in negative SLN may also include cases where small foci of metastatic tumour had been removed in totality (ie, therapeutic effect of the sentinel node biopsy itself). Furthermore, in these studies of small metastases, assessing the difference in clinical prognosis includes a lead time bias in that the smaller the metastatic deposit, the longer it will take for recurrence of the disease to be detected.2 17 There is some evidence that micrometastases are indeed clinically relevant in that patients with isolated tumour cells (<0.2 mm) in SLNs have been reported to have a significantly higher melanoma-specific mortality than those with SN-negative status.2

Conclusion

The question of whether all melanoma metastases, including very small tumour deposits, are clinically relevant would need to be answered with careful and prolonged follow-up such as in the Multicentre Selective Lymphadenectomy 2 trial, which will include melanoma deposits detected by reverse transcriptase PCR or the MINITUB trial.3 Our results suggest that small deposits cannot be assumed to be dormant or clinically insignificant. It seems unjustified to presume that any micrometastatic deposit of melanoma, regardless of how small it is, is without potential to develop at some stage. Furthermore, high-resolution ultrasonography cannot detect metastases <3 mm, and 87% of metastases of that size were shown to have proliferative activity. We have argued that even the 13% without demonstrable proliferation cannot be presumed to be insignificant because of the limitations of the assessments used. Again, the average maximum dimension of the metastases in our cases is <3 mm, which is below the lower limit of detection by ultrasonography, and we suggest that any melanoma metastasis cannot be said to be prognostically insignificant. We would argue that it remains justified to detect and remove all metastases in SLN.

Take-home messages

Any melanoma metastasis when detected and no matter how small cannot be ignored. There is no way to predict its dormancy.

Acknowledgments

The authors would like to thank Lisa Baker and Stuart Holmes for their valued assistance in preparation of the final manuscript.

References

Footnotes

  • Competing interests None declared.

  • Ethics approval This study was conducted with the approval of the Surrey Research Ethics Committee.

  • Provenance and peer review Not commissioned; externally peer reviewed.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.