This Article Abstract Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Karakousis, G. C. Articles by Spitz, F. R. Search for Related Content PubMed PubMed Citation Articles by Karakousis, G. C. Articles by Spitz, F. R.

Predictors of Regional Nodal Disease in Patients With Thin Melanomas

¹ Department of Surgery, University of Pennsylvania Health System, Abramson Cancer Center, Philadelphia, Pennsylvania
² Department of Biostatistics and Epidemiology, University of Pennsylvania Health System, Abramson Cancer Center, Philadelphia, Pennsylvania
³ Department of Pathology and Laboratory Medicine, University of Pennsylvania Health System, Abramson Cancer Center, Philadelphia, Pennsylvania
⁴ Department of Dermatology, University of Pennsylvania Health System, Abramson Cancer Center, Philadelphia, Pennsylvania
⁵ Department of Medicine, Division of Hematology and Oncology, University of Pennsylvania Health System, Abramson Cancer Center, Philadelphia, Pennsylvania
⁶ The Pigmented Lesion Group and the Melanoma Program, University of Pennsylvania Health System, Abramson Cancer Center, Philadelphia, Pennsylvania

Correspondence: Address correspondence and reprint requests to: Francis R. Spitz, MD, Division of Endocrine and Oncologic Surgery, Hospital of the University of Pennsylvania, 4th Floor Silverstein Building, 3400 Spruce Street, Philadelphia, PA 19104; E-mail: francis.spitz{at}uphs.upenn.edu.

	ABSTRACT

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION CONCLUSIONS REFERENCES

 
Background: Most melanoma patients present with thin ( 1.0 mm) lesions. Indications for sentinel lymph node (SLN) biopsy are not well defined for this group. Previously, we reported an association between mitotic rate (MR) and SLN positivity in these patients. The study was limited by a relatively small sample size and low statistical power. In this study, we evaluated a large population of patients with thin melanoma from the pre-SLN era to identify predictors of regional nodal disease (RND) that may serve as a surrogate for SLN positivity.

Methods: Eight hundred eighty-two patients evaluated between 1972 and 1991 were included in the study. Univariate and multivariate regression analyses were performed by using clinical and histological data to identify factors associated with RND. A multivariate logistic regression model was developed and applied to the previously reported group of patients with thin melanomas who underwent SLN biopsy between 1996 and 2004 for validation.

Results: Thirty-eight patients (4.3%) had evidence of RND. In the multivariate analysis, a MR >0, vertical growth phase (VGP), male sex, and ulceration were statistically significant predictors of RND. Patients at the highest risk according to a classification tree analysis (VGP and MR >0) had an RND rate of 11.9%. The regression model developed predicted well the SLN status in the validation sample.

Conclusions: Investigation of a large pre-SLN population identified MR >0, ulceration, VGP, and male sex as independently predictive of RND in patients with thin melanomas. These factors may help to identify subgroups of these patients that have clinically significant risks of SLN positivity.

Key Words: Thin melanomas • Regional nodal disease • Mitotic rate • Ulceration • Vertical growth phase

	INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION CONCLUSIONS REFERENCES

 
Thin melanomas ( 1.0 mm thick) represent most invasive melanoma lesions diagnosed, comprising nearly 65% of such lesions reported to the Surveillance, Epidemiology, and End Results program in the past decade.¹ These lesions usually portend a favorable prognosis, with a 5 year survival rate after definitive excision alone of >95%.² Despite their generally benign course, these lesions’ potential for metastases has long been recognized. A metastatic rate of 3%³ for lesions <.76 mm and a more recent rate from our own institution of 6.5%⁴ ( 1.0 mm) have been reported.

Lymphatic mapping (LM) and sentinel lymph node (SLN) biopsy, first reported by Morton et al.⁵ in 1990, has become a well-established technique for accurately identifying, with relatively minimal morbidity, micrometastatic disease in the regional nodal basins.⁶ The method, which offers important prognostic information,⁷ particularly for many early-stage melanoma patients, is routinely used in patients with lesions >1.0 mm. Indications for its use in the increasingly large group of patients with thin lesions, however, are not well established. The reported incidence of SLN positivity in patients with thin melanomas is generally <5%,^7–11 and the reported incidence of SLN positivity with lesions <.76 mm is <2%.^8,9,11 With such low incidence rates, the routine use of LM/SLN in these patients has therefore not been thought to be justified, given the morbidity, albeit small, of the procedure, as well as its associated costs.¹² Identification of clinical and/or histological predictors of SLN positivity in these patients will help in determining subgroups for which the incidence of SLN positivity is clinically significant.

Ulceration and a higher Clark level of invasion were identified as negative prognostic factors for survival among patients with thin lesions in the most current American Joint Committee on Cancer staging system.¹³ We recently reported that the presence of any dermal mitoses (mitotic rate [MR] >0) in such patients may identify those at an increased risk of SLN positivity.¹¹ The evaluation and identification of multiple prognostic variables in our previously reported group was statistically limited by a relatively small number of events: only 9 of the 181 patients had a positive SLN.

In an attempt to overcome some of the limitations in statistical power inherent in evaluating SLN positivity in a population with a relatively low incidence in which the procedure is not routinely performed, we used a large population of patients in the pre-SLN era (1972–1991). More than 800 patients with melanomas 1.0 mm and who had at least 10 years of follow-up were identified from a large prospective database of >6500 patients maintained by the Pigmented Lesion Group at the University of Pennsylvania. We recently reported predictors of metastatic disease overall in this population.⁴ In this study, we considered regional nodal disease (RND) as a surrogate for SLN positivity and developed a multivariate model to predict RND. To validate this model, we then applied it to a group of 181 patients with thin melanomas who actually underwent SLN biopsy between 1996 and 2004. This approach, in addition to permitting greater statistical power, potentially decreases the selection bias that may occur when a population of thin-melanoma patients undergoing SLN biopsy is studied.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION CONCLUSIONS REFERENCES

 
Patients
Clinical and histological information on patients was compiled from the University of Pennsylvania’s Pigmented Lesion Clinic database and clinical charts in accordance with the study protocol approved by the Institutional Review Board at the same institution. Between 1972 and 1991, 1232 patients were evaluated who had invasive thin melanomas (Breslow thickness 1.00 mm), and 882 were included in this study. Inclusion/exclusion criteria for patients were as described elsewhere.⁴ Specifically, eligible patients (n = 1112) included those who underwent definitive treatment (wide local excision) of their lesions with or without prophylactic lymphadenectomy and who were without evidence of metastatic disease at the time of initial diagnosis. In addition, they had to be seen in the Pigmented Lesion Clinic within 1 year of definitive excision. Patients were excluded (n = 230) if complete histopathologic data were unavailable, if their death was unrelated to melanoma, or they were lost to follow-up before 10 years of follow-up.

Clinical and Histopathologic Variables
Six histopathologic variables derived from the primary specimen and three clinical variables were included in the analyses. Pathologic data were obtained from review by one of three pathologists (W. H. Clark Jr., R.E., and D.E.E.). Clinical variables included age (<60 or 60 years), sex, and the anatomical location of the primary lesion (axial vs. extremity). Included in the category of axial lesions were head and neck lesions, truncal lesions, subungual lesions, and lesions of the palms and soles. Histopathologic variables were also considered in a binary fashion and included tumor thickness ( .75 or >.75 mm), Clark level (II/III or IV), MR (0 or >0), growth phase (radial growth phase [RGP] only or vertical growth phase [VGP]), ulceration (present or absent), and RGP regression (present or absent). Commonly used definitions were applied for the histopathologic characteristics considered. MR was reported as the number of mitoses per square millimeter and was converted to a binary variable; VGP was defined by aggregates of tumor in the dermis larger than in the epidermis and/or the presence of any dermal tumor cell mitoses. Ulceration was noted in specimens with an area of complete loss of epidermis over the tumor not related to prior surgical procedures.

Regional Nodal Disease
RND was identified in patients with either pathologically confirmed or clinically evident metastatic disease in the immediate regional nodal basin draining the primary lesion. Included in the group of patients with RND were (1) patients who had apparent RND either by clinical examination or by pathological confirmation subsequent to wide excision of their primary lesion, (2) patients who underwent prophylactic lymphadenectomy at or around the time of their definitive excision with malignant melanoma in the pathologic specimen of the nodal basin, and (3) patients who developed recurrent systemic disease (visceral or distant metastases), if there was either concurrent pathologic or clinical evidence (identified within 6 months) of regional nodal involvement. Patients who underwent prophylactic lymphadenectomies were excluded from the recurrence distribution analyses.

Statistical Analyses
Significant associations between clinical or histologic variables and RND were identified by using Fisher’s exact test. A classification and regression tree analysis was performed by using a recursive portioning algorithm (Salford Systems, San Diego, CA).¹⁴ Multivariate logistic regression analyses were performed with SAS software (SAS Institute Inc., Cary, NC).¹⁵ A reduced multivariate analysis was performed by successively eliminating the variables that were not significant in the multivariate analysis in a stepwise fashion, and the accuracy of the model was determined by using the area under the receiver operator characteristic curve. Predicted probabilities computed from the logistic regression model were compared with the proportions observed in the validation sample by using the ² goodness-of fit statistic. The log-rank statistic was used to compare the Kaplan-Meier survival curves for time to metastasis for those with RND and those without RND. Patients who developed nonlocal disease without evidence of RND or patients who died of a cause unrelated to melanoma were censored at the time of diagnosis or death, respectively.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION CONCLUSIONS REFERENCES

 
Categorization of Patients With the Primary End Points of RND and Time to Development of Disease
With a mean and median follow-up time of 17.2 and 16.4 years, respectively (with a minimum of 10 years’ follow-up), 72 (8.2%) of the 882 thin-melanoma patients were noted to develop nonlocal metastatic disease. Of these, 38 patients (4.3%) had RND, and 34 patients (3.9%) developed nonlocal disease without evidence of RND. In the group of patients with RND, 31 patients (81.6% of the RND group) had RND alone, whereas 7 patients had concurrent RND with systemic disease. Most (92%) cases of RND were pathologically confirmed, and the remainder were clinically defined. All of the clinically defined patients were in the group with both RND and systemic disease. Two patients were included in the RND group because of positive prophylactic regional lymphadenectomies. These results are summarized in Fig. 1.

View larger version (18K): [in this window] [in a new window]   FIG. 1. Categorization of patients with RND. The minimal follow-up (f/u) time of patients was 10 years. RND, regional nodal disease; SD, systemic disease; PC, pathologically confirmed; CD, clinically defined; pts, patients.

Clinical Variables Associated With RND
Most patients who developed RND (n = 38) were male (60.5%; n = 23) and <60 years of age (86.8%; n = 33). The mean age of RND patients was 45.6 ± 13.9 years. Twenty-seven (71.1%) of the 38 patients had axial lesions. The associations of the three clinical variables with RND are listed in Table 1. Only male sex was found to be significantly associated with RND in the univariate analysis (P = .03). This association remained significant in the multivariate analysis (P = .03), and sex was one of the variables ultimately included in the reduced logistic regression model discussed below (P value of the reduced model = .01).

Classification and Regression Tree Analysis for the Development of Prognostic Risk Groups
A prognostic tree was developed to further stratify patients with thin melanomas into different risk groups for the development of RND (Fig. 2). The first partition in the tree was made by using growth phase. Patients with pure RGP lesions, who comprised 46.8% of the population, had only a .5% incidence of RND (95% confidence interval [CI] = .1%–1.7%) and were considered a low-risk group. The second partition in the tree was made among the patients with VGP lesions by using MR. Patients with VGP and mitogenic lesions had an 11.9% incidence of RND (95% CI, 8.0%–16.9%) and made up the group with the highest risk for RND. This group was 25.6% of the overall population. The third partition in the tree was made among the group of patients with VGP and MR = 0 by using sex. Whereas the incidence of RND among all patients with MR = 0 lesions was 3.7%, the incidence in this group among men was 6.6% (95% CI, 2.7%–13.1%). This group may also have a clinically substantial risk of RND, although the CI overlaps between the sex groups.

View larger version (19K): [in this window] [in a new window]   FIG. 2. Classification and regression tree analysis for the development of prognostic risk groups for RND in thin-melanoma patients. RND, regional nodal disease; RGP, radial growth phase; VGP, vertical growth phase; MR, mitotic rate.

In the equation, p reflects the probability of RND in groups and is defined by the values of the covariates (the presence of each poor-prognostic factor is assigned the value 1, and its absence, the value 0).

We hypothesized that the probabilities of RND identified in our pre–SLN-era population would predict SLN positivity in our SLN-era population. To test how well our model predicted SLN positivity, we applied our model to our recently reported (Kesmodel et al.¹¹) cohort of 181 thin-melanoma patients with VGP lesions who actually underwent SLN biopsy (Fig. 3). To obtain the predicted number of patients in this validation sample for each set of covariate values, the probabilities from our prognostic model were multiplied by 181 to estimate the expected number of patients with a positive SLN for each set. There were nine patients who were actually identified with a positive SLN. For the two groups with the highest number of actual positive SLNs—namely, male patients with VGP lesions and MR >0 without ulceration (n = 6 positive SLNs) and female patients with VGP lesions and MR >0 without ulceration (n = 3 positive SLNs)—our model predicted 7.1 and 3.7 patients with positive SLNs, respectively, closely approximating the observed numbers. Furthermore, in groups that had negative SLNs, our model predicted few patients (0–2.5 patients) to have SLN positivity. No significant difference (P = .28; ² goodness-of-fit statistic) was noted between the predicted and observed number of patients with SLN positivity, thus indicating that the model predicted well the SLN outcome in this independent group of patients with thin melanomas. The greatest discordance between the actual and predicted number of patients by using the model was in the group of male patients with VGP lesions and MR = 0 without ulceration. There were no patients with lesional MR = 0 observed to have positive SLNs in the validation group, whereas our prognostic tree analysis identified a 6.6% incidence of RND in male patients with VGP lesions and MR = 0 in the pre-SLN era.

View larger version (25K): [in this window] [in a new window]   FIG. 3. Validation of a logistic regression prognostic model. Bar graphs display the number of SLN-positive patients (actual vs. predicted) in our previously reported modern cohort of thin-melanoma patients11; actual numbers (empty bars) and predicted numbers from a linear regression model using RND as the outcome (shaded bars) are shown. Actual and predicted numbers were compared by using the 2 goodness-of-fit statistic (6 df). SLN, sentinel lymph node; MR, mitotic rate; VGP, vertical growth phase.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION CONCLUSIONS REFERENCES

Several investigators have identified prognostic factors for disease recurrence and disease-specific mortality among patients with thin melanomas. Among these factors are age,¹⁷ sex,^4,16,17 increasing tumor thickness,^13,17–19 a higher level of invasion,^13,16,19 presence of ulceration,^13,18 anatomical site (axial),^16,17,19 an absent or attenuated tumor-infiltrating lymphocyte response,⁴ regression,^16,19 VGP,⁴ and MR.^4,20 An analysis of 17,600 patients performed as part of the most recent revision of the American Joint Committee on Cancer Staging System for melanoma identified ulceration, tumor thickness, and level of invasion in the subgroup of thin-lesion patients as independent negative prognostic factors for survival.¹³ In another study of 681 patients with melanomas <.76 mm, the incidence of metastasis was 4.8%, with negative prognostic factors including male sex, the axial location of the primary tumor, regression, and Clark level IV.¹⁶ Corsetti et al.²¹ corroborated the effect of a higher level on recurrence, reporting in their retrospective series that 18.4% of patients with level III and IV lesions developed metastatic disease. We recently reported on >800 patients with thin melanomas and identified an MR >0 as the prognostic factor associated with metastasis that first split this population sample into two groups by tree analysis.⁴ In patients with an MR >0, the 10-year risk of metastasis was 19.9%, whereas for patients with an MR of 0, the risk was 1.8%. When considering male patients alone with an MR >0, the 10-year metastatic rate increased to 31.1%. The applicability of these analyses that identified prognostic factors for metastatic disease and disease-specific mortality to predicting SLN outcome is unclear.

Many studies have focused on identifying factors associated with positive SLNs in the melanoma population at large and among patients with thin primary lesions specifically. Increasing tumor thickness has consistently been found to be a predictive factor for SLN positivity.^8,15,22–25 Other factors have also been reported to be associated with SLN positivity, with more variability. These include age,^22,24,25 ulceration,^24–26 level,²⁴ anatomical location,²⁵ lymphovascular invasion,²⁶ and microsatellitosis.²⁶ MR has been identified as an independent predictor of SLN positivity in several studies.^15,22,26 In a study evaluating SLNs in 275 patients, an MR of >5 per high-powered field was found to be a significant predictor of SLN positivity, in addition to ulceration and thickness 1.25 mm.¹⁵ Sondak et al.²² similarly identified MR, thickness, and, additionally, age as factors that were significantly associated with positive SLNs. The study noted that younger patients (<35 years) with an increased MR had a clinically significant incidence of positive SLNs. Results from a study from the Sydney Melanoma Unit corroborated the importance of MR as a predictor of outcome in melanoma patients, although the investigation evaluated survival and not SLN positivity as the outcome.²⁰

Studies evaluating SLN status in patients with thin melanomas have yielded somewhat conflicting results. Our group has reported that VGP is a risk factor for nodal metastasis, with an overall SLN positivity rate of 5.6%.¹⁰ For patients with tumors of .76 to 1 mm in thickness, the incidence of SLN positivity increased to 9.7%. Oliveira Filho et al.²⁷ similarly noted the importance of VGP as a predictor of SLN positivity in a select group of patients with thin melanomas. More recently, we reported MR as an important predictor of SLN positivity in a study of 181 patients with thin ( 1.00-mm) melanoma lesions, noting a 12.3% incidence of SLN positivity among patients with MR >0 and tumor thickness .76 mm and 1.00 mm.¹¹ In a study of 134 patients with lesions 2.0 mm, tumor regression was reported to be associated with an increased incidence of SLN positivity.²⁸ Other studies did not identify subgroups of patients with thin melanomas whose incidence of SLN positivity was high enough to justify the use of the procedure. Jacobs et al.⁹ reported on the SLN status in 65 patients with thin melanomas. The incidence of positive SLN in this group was 3%. In patients with tumor thickness <.75 mm, it was 0%. These investigators indicated that the surgical procedure may not be justified in this population. Similar conclusions were reached by Bleicher et al.,⁸ in a study of patients with <1.5 mm lesions, who concluded that SLN biopsy in patients with <.76 mm-thick lesions is rarely indicated. In a review of 146 patients with thin lesions, Stitzenberg et al.²⁹ did not identify any significant independent predictors of SLN positivity among the clinical and histopathologic variables examined, although they reported an incidence of SLN positivity in this group of 4% and indicated that further studies to identify such factors were therefore warranted.

In this study, we investigated a large prospective database of melanoma patients with thin primary lesions in the pre-SLN era to identify predictors of RND, which we hypothesized could serve as a surrogate for SLN positivity. This assumption, based on the notion that microscopic disease would be expected to precede clinically evident disease, was corroborated by our findings that the incidence of RND was at least equivalent to the SLN positivity rates reported for more recent cohorts of patients. In addition, a model using prognostic factors of RND predicted well the SLN outcome of a distinct population that actually underwent SLN biopsy. By evaluating RND as the outcome in a large group of patients with long-term follow-up, we were able to achieve statistical power that would otherwise be difficult to obtain when SLN positivity is studied in this population.

A strength of our study design is that there may be less likelihood of selection bias, which could be introduced when SLN positivity is studied in a population in which the technique is not routinely performed. The incidence of RND in our study group was 4.3%, which is in accordance with results we and others have reported for the SLN positivity rate in patients with thin lesions.^7,10,11,29 If there were substantial selection bias in the choosing of patients with thin lesions for SLN biopsy, we would have expected to find a lower rate in this study of RND.

By univariate analysis, we identified VGP, tumor thickness .76 mm, Clark level IV, presence of ulceration, presence of mitoses (MR >0), and male sex as significantly associated with RND. Multivariate analysis revealed only VGP, presence of mitoses, ulceration, and male sex as independent predictors of RND. These findings are congruent with our previously published findings that the presence of mitoses and VGP are predictors of SLN positivity.^10,11 The clinical significance of mitogenic melanomas being associated with SLN positivity is also in accordance with several other recently published articles.^20,22,30 It is interesting to note that tumor thickness, which has consistently been associated with SLN positivity, was not identified as an independent predictor of RND by multivariate analysis. This was not the result of using the cutoff of .76 mm, because when thickness was treated as a continuous variable, there was still no significant association with RND by multivariate analysis. This finding may be attributable to the strong predictive effect of MR and/or VGP and the cosegregation of these variables with tumor thickness.

Our prognostic tree indicated that in patients with VGP and mitogenic lesions, the incidence of RND was 11.9%. This is a group of patients whose incidence of nodal metastases approximates that of patients with thicker melanomas^22,31,32 in which SLN biopsy is routinely performed. Approximately 26% of our population fell into this group. Male patients with VGP lesions and without evidence of mitoses had a risk of RND of 6.6%, with 12% of the study population falling into this group. In our previous study, there was a 0% incidence of SLN positivity in patients with an MR of 0. The incongruence of these results may be explained by the relatively small number of positive events in the previous study. Differences in patient study populations, from either biological differences or selection bias, could also be present.

If one combines the two higher-risk groups of patients identified in this study, the incidence of RND is 10.2% (37.6% of the entire study group). If RND is a good surrogate marker for SLN positivity, performing SLN biopsy on just over a third of all melanoma patients with thin lesions would yield a 10% SLN positivity rate and would capture 89% of all patients with a positive SLN. It should be noted that although ulceration was not identified in our prognostic tree, it was a significant independent predictor of RND. In fact, nearly 36% of patients with ulceration had RND. On the basis of these results, one should consider the presence of ulceration as potentially justifying the use of SLN biopsy. In our population, however, ulceration was present in only approximately 1.6% of the entire study group, and, therefore, only approximately .6% of the entire thin-melanoma population investigated displayed ulceration and developed RND. Although it is a significant predictor, ulceration may yield little effect on clinical decision making for the vast proportion of thin-melanoma patients.

By treating prognostic variables in a binary fashion, we attempted to eliminate some of the variability that may exist between institutions in the pathologic assessment of specimens (perhaps at the expense of detecting finer nuances) and thereby foster general-izability. For instance, by treating MR in a binary manner (MR of 0), a lesion could be classified simply according to the presence or absence of mitoses, and the variability that may arise in the precise manner that MR is defined (e.g., per high-powered field or per square millimeter) becomes less of a consideration.

	CONCLUSIONS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION CONCLUSIONS REFERENCES

 
Male sex, presence of mitoses, presence of ulceration, and VGP were identified as independent predictors of RND in patients with thin melanomas. A prognostic model using these variables predicted well the SLN positivity in a separate validation group of patients undergoing SLN biopsy. These results may help to further stratify patients with thin melanomas into subgroups with appreciable risks of SLN positivity. Large prospective studies are needed to validate these findings and address the clinical utility of SLN biopsy in this group of patients who make up the majority of patients currently presenting with melanoma.

	ACKNOWLEDGMENTS

 
This study was supported in part by the SPORE on Skin Cancer (CA-093372; M. Herlyn, principal investigator). The authors thank all of the patients who were seen at the Pigmented Lesion Clinic (PLC) and gave their consent for use of their data for research studies, as well as the investigators (W. H. Clark, Jr, MD [deceased], E. E. Bondi, MD, L. P. Bucky, MD, L. S. Callans, MD, B. Chang, MD, K. T. Flaherty, MD, A. C. Halpern, MD, R. Hamilton, MD, D. Hershock, MD, D. D. Larossa, MD, S. R. Lessin, MD, D. Low, MD, P. Van Belle, MD, and J. Wolfe, MD) and staff (R. Holmes, S. Hotz, N. Lowden, I. Matozzo, M. Price, M. Synnestvedt, and J. Thompson) of the PLC for their contributions over the last three decades to the Melanoma Core Database on which this report is based.

Received for publication June 8, 2005. Accepted for publication October 12, 2005.

	REFERENCES

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION CONCLUSIONS REFERENCES

 

1. Surveillance, Epidemiology, and End Results (SEER) Program. Public-usedata (1973–2000), National Cancer Institute, DCCPS, Surveillance Research Program, Cancer Statistics Branch. Released April 2004. Based on the November 2003 submission. Available at: http://www.seer.cancer.gov/publicdata.
2. Reintgen D, Balch CM, Kirkwood J, Ross M. Recent advances in the care of the patient with malignant melanoma. Ann Surg 1997; 225:1–14.[CrossRef][Medline]
3. Woods JE, Soule EH, Creagan ET. Metastasis and death in patients with thin melanomas (less than 0.76 mm). Ann Surg 1983; 198:63–4.[Medline]
4. Gimotty PA, Guerry D, Ming ME, et al. Thin primary cutaneous malignant melanoma: a prognostic tree for 10 year metastasis is more accurate than American Joint Committee on Cancer staging. J Clin Oncol 2004; 22:3668–76.[Abstract/Free Full Text]
5. Morton DL, Wen DR, Wong JH, et al. Technical details of intraoperative lymphatic mapping for early stage melanoma. Arch Surg 1992; 127: 392–9.[Abstract/Free Full Text]
6. Morton DL, Thompson JF, Essner R, et al. Validation of the accuracy of intraoperative lymphatic mapping and sentinel lymphadenectomy for early-stage melanoma: a multicenter trial. Multicenter Selective Lymphadenectomy Trial Group. Ann Surg 1999; 230:453–63; discussion 463–5.[CrossRef][Medline]
7. Gershenwald JE, Thompson W, Mansfield PF, 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 1999; 17:976–83.[Abstract/Free Full Text]
8. Bleicher RJ, Essner R, Foshag LJ, et al. Role of sentinel lymphadenectomy in thin invasive cutaneous melanomas. J Clin Oncol 2003; 21:1326–31.[Abstract/Free Full Text]
9. Jacobs IA, Chang CK, DasGupta TK, Salti GI. Role of sentinel lymph node biopsy in patients with thin (<1 mm) primary melanoma. Ann Surg Oncol 2003; 10:558–61.[Abstract/Free Full Text]
10. Bedrosian I, Faries MB, Guerry DT, et al. Incidence of sentinel node metastasis in patientswith thinprimary melanoma (< or = 1 mm) with vertical growth phase. Ann Surg Oncol 2000; 7:262–7.[Abstract]
11. Kesmodel SB, Karakousis GC, Botbyl JB, al. Mitotic rate as a predictor of sentinel lymph node positivity in patients with thin melanomas. Ann Surg Oncol 2005; 12:449–58.[Abstract/Free Full Text]
12. Agnese DM, Abdessalam SF, Burak WE Jr, et al. Cost-effectiveness of sentinel lymph node biopsy in thin melanomas. Surgery 2003; 134:542–7; discussion 547–8.[CrossRef][Medline]
13. Balch CM, Soong SJ, Gershenwald JE, et al. Prognostic factors analysis of 17,600 melanoma patients: validation of the American Joint Committee on Cancer melanoma staging system. J Clin Oncol 2001; 19:3622–34.[Abstract/Free Full Text]
14. Steinberg D, Colla P. CART: Tree-Structured Non-Parametric Data Analysis. San Diego: Salford Systems, 1995.
15. Wagner JD, Gordon MS, Chuang TY, et al. Predicting sentinel and residual lymph node basin disease after sentinel lymph node biopsy for melanoma. Cancer 2000; 89:453–62.[CrossRef][Medline]
16. Slingluff CL Jr, Vollmer RT, Reintgen DS, Seigler HF. Lethal "thin" malignant melanoma. Identifying patients at risk. Ann Surg 1988; 208:150–61.[Medline]
17. Leiter U, Buettner PG, Eigentler TK, Garbe C. Prognostic factors of thin cutaneous melanoma: an analysis of the central malignant melanoma registry of the German Dermatological Society. J Clin Oncol 2004; 22:3660–7.[Abstract/Free Full Text]
18. McKinnon JG, Yu XQ, McCarthy WH, Thompson JF. Prognosis for patients with thin cutaneous melanoma: long-term survival data from New South Wales Central Cancer Registry and the Sydney Melanoma Unit. Cancer 2003; 98:1223–31.[CrossRef][Medline]
19. Mansson-Brahme E, Carstensen J, Erhardt K, et al. Prognostic factors in thin cutaneous malignant melanoma. Cancer 1994; 73:2324–32.[CrossRef][Medline]
20. Azzola MF, Shaw HM, Thompson JF, et al. Tumor mitotic rate is a more powerful prognostic indicator than ulceration in patients with primary cutaneous melanoma: an analysis of 3661 patients from a single center. Cancer 2003; 97:1488–98.[CrossRef][Medline]
21. Corsetti RL, Allen HM, Wanebo HJ. Thin < or = 1 mm level III and IV melanomas are higher risk lesions for regional failure and warrant sentinel lymph node biopsy. Ann Surg Oncol 2000; 7:456–60.[Abstract]
22. Sondak VK, Taylor JM, Sabel MS, et al. Mitotic rate and younger age are predictors of sentinel lymph node positivity: lessons learned from the generation of a probabilistic model. Ann Surg Oncol 2004; 11:247–58.[Abstract/Free Full Text]
23. Nguyen CL, McClay EF, Cole DJ, et al. Melanoma thickness and histology predict sentinel lymph node status. Am J Surg 2001; 181:8–11.[CrossRef][Medline]
24. McMasters KM, Wong SL, Edwards MJ, et al. Factors that predict the presence of sentinel lymph node metastasis in patients with melanoma. Surgery 2001; 130:151–6.[CrossRef][Medline]
25. Rousseau DL Jr, Ross MI, Johnson MM, et al. Revised American Joint Committee on Cancer staging criteria accurately predict sentinel lymph node positivity in clinically node-negative melanoma patients. Ann Surg Oncol 2003; 10:569–74.[Abstract/Free Full Text]
26. Mraz-Gernhard S, Sagebiel RW, Kashani-Sabet M, et al. Prediction of sentinel lymph node micrometastasis by histological features in primary cutaneous malignant melanoma. Arch Dermatol 1998; 134:983–7.[Abstract/Free Full Text]
27. Oliveira Filho RS, Ferreira LM, Biasi LJ, et al. Vertical growth phase and positive sentinel node in thin melanoma. Braz J Med Biol Res 2003; 36:347–50.[Medline]
28. Olah J, Gyulai R, Korom I, et al. Tumour regression predicts higher risk of sentinel node involvement in thin cutaneous melanomas. Br J Dermatol 2003; 149:662–3.[CrossRef][Medline]
29. Stitzenberg KB, Groben PA, Stern SL, et al. Indications for lymphatic mapping and sentinel lymphadenectomy in patients with thin melanoma (Breslow thickness < or =1.0 mm). Ann Surg Oncol 2004; 11:900–6.[Abstract/Free Full Text]
30. Thompson JF, Shaw HM. Should tumor mitotic rate and patient age, as well as tumor thickness, be used to select melanoma patients for sentinel node biopsy? Ann Surg Oncol 2004; 11:233–5.[Free Full Text]
31. Cascinelli N, Belli F, Santinami M, et al. Sentinel lymph node biopsy in cutaneous melanoma: the WHO Melanoma Program experience. Ann Surg Oncol 2000; 7:469–74.[Abstract]
32. Joseph E, Brobeil A, Glass F, et al. Results of complete lymph node dissection in 83 melanoma patients with positive sentinel nodes. Ann Surg Oncol 1998; 5:119–25.[Abstract]

This article has been cited by other articles:

				B. E. Wright, R. P. Scheri, X. Ye, M. B. Faries, R. R. Turner, R. Essner, and D. L. Morton Importance of Sentinel Lymph Node Biopsy in Patients With Thin Melanoma Arch Surg, September 1, 2008; 143(9): 892 - 900. [Abstract] [Full Text] [PDF]

				G. C. Karakousis, P. A. Gimotty, B. J. Czerniecki, D. E. Elder, R. Elenitsas, M. E. Ming, D. L. Fraker, D. Guerry, and F. R. Spitz Regional Nodal Metastatic Disease Is the Strongest Predictor of Survival in Patients with Thin Vertical Growth Phase Melanomas: A Case for SLN Staging Biopsy in These Patients Ann. Surg. Oncol., May 1, 2007; 14(5): 1596 - 1603. [Abstract] [Full Text] [PDF]

				S. N. Markovic, L. A. Erickson, R. D. Rao, R. H. Weenig, B. A. Pockaj, A. Bardia, C. M. Vachon, S. E. Schild, R. R. McWilliams, J. L. Hand, et al. Malignant Melanoma in the 21st Century, Part 2: Staging, Prognosis, and Treatment Mayo Clin. Proc., April 1, 2007; 82(4): 490 - 513. [Abstract] [Full Text] [PDF]

This Article Abstract Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Karakousis, G. C. Articles by Spitz, F. R. Search for Related Content PubMed PubMed Citation Articles by Karakousis, G. C. Articles by Spitz, F. R.