Annals of Surgical Oncology Cite Track
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
10.1245/ASO.2005.05.016
Annals of Surgical Oncology 12:282-288 (2005)
© 2005 Society of Surgical Oncology
This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Wong, S. L.
Right arrow Articles by Coit, D. G.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Wong, S. L.
Right arrow Articles by Coit, D. G.

Original Article

A Nomogram That Predicts the Presence of Sentinel Node Metastasis in Melanoma With Better Discrimination Than the American Joint Committee on Cancer Staging System

Sandra L. Wong, MD1, Michael W. Kattan, PhD2, Kelly M. McMasters, MD, PhD3 and Daniel G. Coit, MD1

1 Department of Surgery, Memorial Sloan-Kettering Cancer Center, 1275 York Avenue, New York, New York 10021
2 Department of Biostatistics and Epidemiology, Cleveland Clinic Foundation, 9500 Euclid, Cleveland, Ohio 44195
3 Department of Surgery, Division of Surgical Oncology, University of Louisville School of Medicine, 550 S. Jackson Street, Louisville, Kentucky 40202

Correspondence: Address correspondence and reprint requests to: Daniel G. Coit, MD; E-mail: coitd{at}mskcc.org.


   ABSTRACT
TOP
 ABSTRACT
INTRODUCTION
 METHODS
 RESULTS
 DISCUSSION
 CONCLUSION
 REFERENCES
 
Background: The threshold and indications for sentinel lymph node (SLN) biopsy in patients with melanoma remain somewhat arbitrary. Many variables associated with SLN positivity have previously been identified, including a significant association between the American Joint Committee on Cancer (AJCC) staging system and SLN status. We developed a user-friendly nomogram that takes several characteristics into account simultaneously to more accurately predict the presence of SLN metastasis for an individual patient.

Methods: A total of 979 patients who underwent successful SLN biopsy for cutaneous melanoma at a single institution between February 1991 and November 2003 were included in the analysis. Predictors were used to develop a nomogram, based on logistic regression analysis, to predict the probability of SLN positivity. A large multi-institutional trial with 3108 patients was used to validate the predictive accuracy of the nomogram compared with the AJCC staging system.

Results: The nomogram was developed and found to be accurate and discriminating. The concordance index of the nomogram, a measure of predictive ability, was .694 when evaluated with the validation dataset. In contrast, the concordance index of the AJCC staging system was lower (.663; P < .001).

Conclusions: Using commonly available clinicopathologic information, we developed a nomogram to accurately predict the probability of a positive SLN in patients with melanoma. This tool takes several characteristics into account simultaneously. This model should enable improved patient counseling and treatment selection.

Key Words: Melanoma • Sentinel node biopsy • Nomogram • Prognosis


   INTRODUCTION
TOP
 ABSTRACT
 INTRODUCTION
METHODS
 RESULTS
 DISCUSSION
 CONCLUSION
 REFERENCES
 
In the United States, there will be an estimated 59,580 new cases of melanoma in 2005. Although the 5-year survival rate for those who present with localized disease is 97%, the rate decreases to ≤60% for patients with disease in the regional nodal basins.1 Sentinel lymph node (SLN) biopsy is a minimally invasive outpatient procedure that accurately stages the regional lymph nodes. Involvement of SLNs is thought to be the most important prognostic factor for patients with melanoma.2,3

SLN biopsy can identify the 15% to 20% of patients with clinically node-negative melanoma who have occult regional nodal metastases. The threshold and indications for SLN biopsy are somewhat arbitrary. The most commonly used criterion is the thickness of the primary melanoma. Other factors that have been investigated include the Clark level, the presence of ulceration, a high mitotic rate, evidence of regression, presence of lymphovascular invasion, presence of vertical growth phase, presence of multiple nodal basin drainage, positive deep margins on previous shave biopsy, primary disease site, sex, and age. Many have previously attempted to identify clinicopathologic factors predictive of SLN metastasis as part of the effort to better define groups of patients likely to benefit from SLN biopsy.2,4–6

The newest version of the American Joint Committee on Cancer (AJCC) staging system7 identified factors found to be prognostic of survival in early-stage melanoma. Tumor thickness and the presence of ulceration were found to be important independent predictors of survival on the basis of an analysis of 17,600 patients.3 However, it remains unclear whether the same factors also predict SLN metastasis. Rousseau et al.8 recently examined the relationship between AJCC stage and risk of SLN metastasis. Successive AJCC stage groupings correlated with an increasing risk of positive SLNs: stage IA, 2%; stage IB, 9%; stage IIA, 24%; stage IIB, 34%; and stage IIC, 53%. AJCC stage was found to be a better predictor of SLN metastasis than tumor thickness alone because it incorporated Clark level, ulceration, and thickness.

The purpose of this study was to construct a prediction model for SLN positivity that simultaneously recognizes multiple prognostic factors. We reviewed prospectively collected data on SLN biopsy from a large single-institution experience and created a multivariable nomogram to predict SLN positivity in an individual patient. The nomogram’s validity was then tested by using data from a large multi-institutional trial of SLN biopsy in melanoma patients. Finally, the predictive ability of the nomogram was compared with that of AJCC clinical staging. Improved risk estimation for SLN positivity will allow for improved patient counseling and treatment selection.


   METHODS
TOP
 ABSTRACT
 INTRODUCTION
 METHODS
RESULTS
 DISCUSSION
 CONCLUSION
 REFERENCES
 
Derivation Data
A prospective surgical melanoma database of all patients undergoing SLN biopsy at Memorial Sloan-Kettering Cancer Center was reviewed. Between February 1991 and November 2003, 1032 patients underwent a successful SLN biopsy procedure with the removal of at least 1 SLN. The procedure was performed in patients with ≥1.0-mm-thick melanoma or in thinner melanoma with invasion to Clark level IV or V. Data collected on patients included thickness, age, sex, Clark level, tumor location, presence of ulceration, presence of lymphovascular invasion, mitotic rate, presence of regression, presence of vertical growth phase, presence of satellitosis, presence of tumor-infiltrating lymphocytes, and number of draining nodal basins. Complete clinicopathologic data for the variables of interest were available for 979 patients, and these patients were included in this study. SLN biopsy was performed by using an intradermal injection of technetium sulfur colloid. Preoperative lymphoscintigraphy was performed to identify all draining nodal basins, as well as ectopic or in-transit nodes. During surgery, an intradermal injection of isosulfan blue dye and the handheld gamma probe were used to guide detection of the SLN. An SLN was defined as any blue-stained node or any node with ≥10% of the hottest node’s ex vivo counts. All SLNs were examined by using hematoxylin and eosin (H&E) staining and, if negative by H&E, immunohistochemical staining with S-100 and HMB-45.

Validation Data
Validation data were obtained from the Sunbelt Melanoma Trial, a prospective, randomized multi-institutional study with 3286 patients from 79 institutions accrued between June 1997 and October 2003. The institutional review board–approved trial was undertaken to evaluate the role of adjuvant interferon alfa-2b on disease-free and overall survival for patients with nodal metastasis identified by SLN biopsy. Randomized treatment assignment did not affect the performance of SLN biopsy, and patients were eligible for the current analysis if the procedure was performed successfully. Patients underwent wide excision of their primary melanoma and SLN biopsy. Eligibility criteria included patients aged 18 to 70 years with ≥1.0-mm-thick melanoma and clinically uninvolved regional lymph nodes. SLN biopsy was performed with an intradermal injection of isosulfan blue dye and technetium sulfur colloid. All blue nodes and all nodes with radioactive counts measuring ≥10% of the ex vivo counts from the hottest node were removed and designated as SLNs.9 All nodes underwent histological analysis at multiple levels with H&E staining and immunohistochemical analysis for S-100 protein. Some centers included immunohistochemical staining for HMB-45 and MART-1 as well. A central pathology review committee examined the first 10 cases from every center and any case of SLN metastasis. Metastasis detected by either H&E or immunohistochemistry determined SLN positivity. Patients with incomplete data in any of the categories used in the nomogram were excluded from analysis. Data with complete derivation criteria were available for 3108 patients, and these patients were used to validate the nomogram.

Statistical Methods
We constructed our prediction model by using logistic regression. Routinely available variables included in the nomogram were clinician-selected according to their practical prognostic values. Additional variables were excluded if data were not uniformly collected and obtainable. For ordinal or higher variables, restricted cubic splines were used to accommodate potential nonlinear effects and to improve the model fit. Age and tumor thickness were modeled as continuous variables. The nomogram was developed by using patient data from the Memorial Sloan-Kettering Cancer Center. Bootstrapping is a statistical method in which resampling from the population is repeated many times to simulate the presentation of new patients to the nomogram; such a technique was used to calibrate the nomogram for predictive accuracy.

The nomogram was then validated with patient data from the Sunbelt Melanoma Trial. For each patient in the validation dataset, we computed our nomogram-predicted probability of SLN positivity and the AJCC-based predicted probability of SLN positivity. The predictive ability of these two methods was compared by computing the concordance index for each method and testing for a difference by using the DeLong method.10 The concordance index is the probability that, given two randomly selected patients—one with and one without SLN metastasis—the patient with the positive SLN is, in fact, predicted to have a higher probability of a positive SLN. The concordance index ranges from .5 (i.e., as good as a coin flip) to 1.0 (perfect discrimination). All statistical analyses were performed with S-Plus 2000 Professional (Insightful Corp., Seattle, WA).


   RESULTS
TOP
 ABSTRACT
 INTRODUCTION
 METHODS
 RESULTS
DISCUSSION
 CONCLUSION
 REFERENCES
 
The descriptive statistics for patients from Memorial Sloan-Kettering Cancer Center and the Sunbelt Melanoma Trial are listed in Table 1Go. Tumor thickness, Clark level, presence of ulceration, sex, age, and site of the primary tumor were variables included in the nomogram model. SLNs were positive in 158 patients (16.1%) from the Memorial Sloan-Kettering Cancer Center and in 604 patients (19.4%) from the Sunbelt Melanoma Trial.


View this table:
[in this window]
[in a new window]
  		TABLE 1. Clinicopathologic features of patients in the derivation and validation datasets
 
The nomogram (Fig. 1Go) predicts the probability that an individual patient will have SLN metastasis. To use the nomogram, the patient’s value on each predictor variable axis is located; a line is drawn upward to determine how many points the patient receives for each variable value. The points are summed; that number is located on the "total points" axis. A line is then drawn downward to determine the predicted probability of SLN positivity.



View larger version (14K):
[in this window]
[in a new window]
  		FIG. 1. Nomogram to predict the presence of positive sentinel lymph nodes in melanoma. How to use the nomogram: The patient’s values on each axis are located, and a line is drawn upward to determine how many points the patient receives for each variable value. The points are summed; that number is located on the "total points" axis. A line is drawn downward to then determine the predicted probability of sentinel node positivity. The nomogram software is available for download free of charge at http://www.mskcc.org/predictiontools.

 
The nomogram was first evaluated by using the derivation dataset (from Memorial Sloan-Kettering Cancer Center). Because these data are being used to both develop and initially evaluate the nomogram, boostrapping was performed to obtain bias-corrected estimates of its performance. The bootstrap-corrected concordance index was .704.

The predictive ability of the nomogram was then compared with the AJCC-based predictive probability of SLN positivity by using the validation data from the Sunbelt Melanoma Trial. Calibration of the nomogram was evaluated graphically (Fig. 2Go) by grouping patients according to their nomogram-predicted probabilities and plotting the actual proportions of patients with positive SLNs for each group. A dashed 45° line indicates where an ideal nomogram, one that predicts perfectly, would lie. Our nomogram’s predictive probabilities tended to be lower than what actually occurs.



View larger version (8K):
[in this window]
[in a new window]
  		FIG. 2. Calibration plot. The dotted line is a reference line indicating where an ideal nomogram would lie (perfect prediction). Our nomogram’s performance is shown for each decile of the population.

 
In the validation dataset, the nomogram discriminated better than did the AJCC stage (concordance index, .694 vs. .663, respectively; P < .001). Figure 3Go illustrates the nomogram-predicted probabilities within each AJCC stage to illustrate the heterogeneity of risk within each stage. Between AJCC clinical stages, there is an overlapping risk of SLN positivity with adjacent stages and no clean separation of patients with differing risks. This indicates somewhat poor discrimination.



View larger version (23K):
[in this window]
[in a new window]
  		FIG. 3. Nomogram predictions compared with American Joint Committee on Cancer (AJCC)-based predictions. The data are displayed graphically; the distribution of predicted probabilities of sentinel node positivity within each AJCC stage is shown. Every stage overlaps with subsequent stages, thus illustrating a lack of discrimination.

 

   DISCUSSION
TOP
 ABSTRACT
 INTRODUCTION
 METHODS
 RESULTS
 DISCUSSION
CONCLUSION
 REFERENCES
 
The value of SLN biopsy is its ability to serve as an accurate staging procedure. Patients with evidence of SLN metastasis undergo completion lymphadenectomy of the involved regional nodal basin as part of standard clinical practice. However, we do not know whether there is a benefit in terms of improved survival with completion lymphadenectomy, nor do we know what the natural history is for patients who, for whatever reason, forego completion lymphadenectomy in the face of a positive SLN. An improved ability to predict the presence of SLN metastasis may be helpful in risk estimation and decision making for patients who are uncertain about whether to pursue SLN biopsy.

The indications for SLN biopsy in patients with clinically node-negative melanoma are continually being refined. The goal of many studies has been to identify patients with a significant risk of occult nodal metastases. However, the procedure has the potential to be overused and should be avoided in patients with an exceedingly low risk of regional nodal involvement.

In general, patients with primary tumors at least 1.0 mm thick or with Clark level IV or V (if <1.0 mm thick) are offered this procedure if there is no clinical evidence of regional nodal disease. SLN biopsy for melanoma is widely accepted, but not all surgeons agree on the specific indications for its use. Morton and colleagues’11 original landmark study included all clinical stage I melanoma patients. Although arbitrary, evidence at the time suggested that stage I patients were the ones most likely to benefit from elective lymph node dissection; according to the AJCC staging system at the time, this included patients with tumors <4 mm thick. The incidence of SLN metastases was 20.6% in the entire series but was only 9.8% in patients with tumor thickness <1.5 mm.11 As SLN biopsy for melanoma became a more commonly performed procedure, the indications evolved to melanoma ≥1.0 mm thick. There are data to suggest that SLN positivity is exceedingly rare in melanoma <.76 mm.12,13 However, others have argued that SLN biopsy should be expanded to include certain thin melanomas (<1.0 mm), citing 5% to 7% positivity rates.14

Many studies have attempted to correlate clinico-pathologic factors such as tumor thickness, Clark level, presence of ulceration, sex, age, tumor location, and multiple draining nodal basins with the risk for nodal metastases. There has been variable success in determining stand-alone predictors. Different statistical techniques have been applied, with discordant results. No study on prognostic factors has corroborated the findings of another study and matched each reported factor. Variation in reported results may be due to differences among study populations or to the interactions of risk factors.

Most reports have found tumor thickness to be a consistent predictor of SLN positivity; as the thickness increases, the probability of a positive SLN increases. Tumor thickness is often the only factor used in determining the need for SLN biopsy. However, Sondak et al.15 recently found that thickness as a sole variable had poor sensitivity and specificity in predicting SLN positivity and that a coin flip was only slightly worse. In the same study, thickness was a significant factor in a probabilistic model that also included age and mitotic rate. Although thickness seems to be the most consistently used predictive variable, its value is questionable when tumors are thin (<1 mm) or thick (>4 mm).

There is no consensus as to what variables, or what combination of variables, best predict SLN positivity. Variable selection for the nomogram was largely based on clinician selection. Some factors were excluded because they were not uniformly reported by our institution or by investigators for the Sunbelt Melanoma Trial. The addition of more variables to the nomogram may increase its predictive ability, but limiting the model to commonly available and reproducible factors makes it more uniformly applicable. Certainly, any number of models can be developed and may be found to have a significant association with SLN metastasis. As more factors are added to the model or as markers for disease are discovered, an important factor to consider is whether there is any added predictive ability with additional data. Begg et al.16 have found that many prognostic factors contain little information that is not already available when standard prognostic variables are combined optimally. AJCC clinical staging takes thickness, ulceration, and Clark level (if melanoma is ≤1.0 mm thick) into account; a significant increase in SLN metastases was seen in each successive clinical stage grouping.8 Tumor thickness and the presence of ulceration are the most powerful predictors of survival according to the AJCC multivariate analysis3 and, not surprisingly, were also found to be the most powerful predictors of SLN positivity.

We built our nomogram by using all routinely available variables in a comprehensive database to maximize predictive accuracy. Statistical significance based on univariate or multivariate analyses was not performed, because this produces a less accurate prediction model. In addition, continuous variables were not categorized, because this reduces information and also decreases the predictive ability of the model. Furthermore, it is inherently difficult to combine or "count" risk factors because doing so assumes that each variable carries equal predictive weight. Rather, we generated a model based on the simultaneous interaction of variables, and variables were considered relevant if they improved our ability to predict a positive SLN. We compared the predictive ability of our five-variable nomogram with that of the AJCC classification and found our model to be quantitatively more discriminating. Clinically, improvements in prediction models are marked by a higher concordance index. Although the differences reported for the concordance index between our nomogram and AJCC stages (.694 vs. .663, respectively) are relatively small, the concordance index is the most useful metric for comparing prediction models, and the difference we found was statistically significant.

One of the strengths of this nomogram is its generalizability to a larger patient population. The variables used in the nomogram include five commonly available and reproducible data points. This tool also allows for individualization of prediction because several characteristics are taken into account simultaneously for a given patient. No prior attempt has been made to integrate the various prognostic factors into a precise, user-friendly model.

The major pitfall in the development of clinically relevant nomograms is external validation failure, or lack of generalizability of the model in clinical practice. Using data from a single institution is subject to criticism that similarities in terms of diagnosis and treatment preferences bias the nomogram. Ideally, nomograms are tested by using multiple cohorts of patients before implementation of the tool in clinical practice. We validated our nomogram by using an external dataset from a large multicenter trial and confirmed that the variables are reproducible and that the model is accurate. We expected the concordance index to be slightly lower in the validation dataset than in the derivation dataset (.704 vs. .694).

Because SLN biopsy is performed selectively in patients with very thin tumors (<1 mm thick), retrospective analyses such as this study are confounded with patient selection bias. As a result, the nomogram may not have as much predictive accuracy at the lower ends of the tumor-thickness scale. However, the nomogram may help to identify a higher-risk subgroup of patients within a larger group of patients with borderline-thickness tumors for SLN biopsy.

In examining the visual representation of the nomogram (Fig. 1Go), it seems that invasion to Clark level V yields fewer points than invasion to Clark level IV. Patients with Clark level V tumors may have other features that are, on average, worse than those associated with Clark level IV tumors. Therefore, patients with Clark level V tumors will accrue more points in the nomogram from other prognostic variables. In a multivariable model such as this one, it is highly artificial and potentially misleading to examine one axis at a time while assuming that the other axes are fixed. Although this is intuitively difficult to accept, interactions between the Clark level and other prognostic variables account for this discrepancy. The small number of patients with Clark level V tumors may also contribute to this apparent discrepancy, but combining patients with Clark level IV and V tumors could falsely narrow the confidence intervals and does not improve the nomogram’s predictive ability.

In this analysis, we combined readily available prognostic factors into a predictive model. The resultant nomogram provides useful information for individual patients with melanoma to facilitate informed decisions for the use of SLN biopsy. At a certain threshold, the likelihood of finding a positive SLN will be too low to justify its use; that level may vary from individual to individual. The <10% probability of SLN positivity may be an acceptable risk for some patients who wish to avoid the procedure entirely, but not for others. The nomogram itself is not intended to make treatment decisions. Rather, it is a tool that estimates the likelihood of SLN positivity in an individual patient. We plan to provide handheld and desktop software that implements this nomogram, free of charge, as we do with our other software, at http://www.mskcc.org/predictiontools.


   CONCLUSION
TOP
 ABSTRACT
 INTRODUCTION
 METHODS
 RESULTS
 DISCUSSION
 CONCLUSION
REFERENCES
 
We developed a nomogram for use as a prediction model for SLN positivity in patients with melanoma. The nomogram has been successfully validated on an external dataset. The predictive ability of the nomogram was tested against predictions based on the AJCC clinical staging system and was found to be more accurate and discriminating. Individualized risk estimates for SLN positivity can facilitate informed decision making for melanoma patients.


   ACKNOWLEDGMENTS
 
The Sunbelt Melanoma Trial is supported by a grant from Schering Oncology-Biotech and the Center for Advanced Surgical Technologies of Norton Hospital, Louisville, KY.

Received for publication May 17, 2004. Accepted for publication November 19, 2004.


   REFERENCES
TOP
 ABSTRACT
 INTRODUCTION
 METHODS
 RESULTS
 DISCUSSION
 CONCLUSION
 REFERENCES
 

  1. Jemal A, Murray T, Ward E, et al. Cancer statistics, 2005. CA Cancer J Clin 2005;55:10–30.[Abstract/Free Full Text]
  2. 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]
  3. 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]
  4. 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]
  5. 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]
  6. Porter GA, Ross MI, Berman RS, Lee JE, Mansfield PF, Gershenwald JE. Significance of multiple nodal basin drainage in truncal melanoma patients undergoing sentinel lymph node biopsy. Ann Surg Oncol 2000;7:256–61.[Abstract]
  7. Balch CM, Buzaid AC, Soong SJ, et al. Final version of the American Joint Committee on Cancer staging system for cutaneous melanoma. J Clin Oncol 2001;19:3635–48.[Abstract/Free Full Text]
  8. 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]
  9. McMasters KM, Reintgen DS, Ross MI, et al. Sentinel lymph node biopsy for melanoma: how many radioactive nodes should be removed? Ann Surg Oncol 2001;8:192–7.[Abstract/Free Full Text]
  10. DeLong ER, DeLong DM, Clarke-Pearson DL. Comparing the areas under two or more correlated receiver operating characteristic curves: a nonparametric approach. Biometrics 1988;44:837–45.[CrossRef][Medline]
  11. 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]
  12. Bleicher RJ, Essner R, Foshag LJ, Wanek LA, Morton DL. Role of sentinel lymphadenectomy in thin invasive cutaneous melanomas. J Clin Oncol 2003;21:1326–31.[Abstract/Free Full Text]
  13. 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]
  14. Zapas JL, Coley HC, Beam SL, Brown SD, Jablonski KA, Elias EG. The risk of regional lymph node metastases in patients with melanoma less than 1.0 mm thick: recommendations for sentinel lymph node biopsy. J Am Coll Surg 2003;197:403–7.[CrossRef][Medline]
  15. 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]
  16. Begg CB, Cramer LD, Venkatraman ES, Rosai J. Comparing tumour staging and grading systems: a case study and a review of the issues, using thymoma as a model. Stat Med 2000;19:1997–2014.[CrossRef][Medline]



This article has been cited by other articles:


Home page
 JCOHome page
A. Iasonos, D. Schrag, G. V. Raj, and K. S. Panageas
How To Build and Interpret a Nomogram for Cancer Prognosis
J. Clin. Oncol., March 10, 2008; 26(8): 1364 - 1370.
[Abstract] [Full Text] [PDF]

Home page
 JCOHome page
C. M. Balch and S.-j. Soong
Predicting Outcomes in Metastatic Melanoma
J. Clin. Oncol., January 10, 2008; 26(2): 168 - 169.
[Full Text] [PDF]

Home page
 Clin. Cancer Res.Home page
M. I. Ross
Early-stage melanoma: staging criteria and prognostic modeling.
Clin. Cancer Res., April 1, 2006; 12(7): 2312s - 2319s.
[Abstract] [Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Wong, S. L.
Right arrow Articles by Coit, D. G.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Wong, S. L.
Right arrow Articles by Coit, D. G.

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS