Customize your JAMA Network experience by selecting one or more topics from the list below.
Isom C, Wheless L, Hooks MA, Kauffmann RM. Early Melanoma Nodal Positivity and Biopsy Rates Before and After Implementation of the 7th Edition of the AJCC Cancer Staging Manual. JAMA Dermatol. 2019;155(5):572–577. doi:10.1001/jamadermatol.2018.5902
Are the staging criteria in the 7th edition of the AJCC Cancer Staging Manual (AJCC 7) associated with changes in the distribution of thin melanomas for patients undergoing nodal surgery and for patients with positive nodes?
In this cross-sectional study of 86 846 patients diagnosed before AJCC 7 (2004-2009) and of 54 434 patients diagnosed after AJCC 7 (2010-2013), there was a decrease in the number of nodal surgeries performed and an increase in the number of patients with T1b melanomas undergoing nodal surgery.
Our results suggest that more patients were upgraded to T1b melanomas and fewer nodal biopsies were performed without clinically significant change in nodal positivity rates.
There has been a continued increase in the incidence of newly diagnosed melanomas, most of which are T1 melanomas. The associations between changes in tumor staging, implemented with the 7th edition of the AJCC Cancer Staging Manual (AJCC 7), and sentinel lymph node biopsy rates and nodal positivity rates remain to be seen.
To evaluate the change that the implementation of the AJCC 7 had on staging criteria and the distribution of thin melanomas requiring nodal surgery and nodal positivity rates.
Design, Setting, and Participants
Retrospective cross-sectional study from 2004 through 2013 of all adults (≥18 years) diagnosed with a T1 (Breslow depth ≤1.0 mm) melanoma using The National Cancer Database that captures 70% of all newly diagnosed cancers from accredited Commission on Cancer organizations, including both academic and community settings. Data were analyzed in May 2017.
Patients were grouped together based on year of diagnosis, before and after 2009.
Main Outcomes and Measures
To determine the sentinel lymph node biopsy rate before and after the implementation of the AJCC 7.
A total of 141 280 patients met inclusion criteria. Of 86 846 patients diagnosed from 2004 through 2009, 53.7% (49 644) were male and had a mean (SD) age of 57.7 (16.4) years. Of 54 434 patients diagnosed from 2010 through 2013, 54.3% (31 086) were male and had a mean (SD) age of 59.5 (15.9) years. After 2010, there was a 3.8% decrease in the number of nodal surgeries performed (32 485 of 86 846 patients [37.6%] vs 18 379 of 54 434 patients [33.8%]; P < .001). The nodal positivity rate decreased 1.0% from (9.8% [3166 of 86 846] to 8.8% [1618 of 54 434]) (P < .001). An increase in the proportion of T1b melanomas being evaluated, from 48.8% to 62.2%, was seen (P < .001). Of T1b melanomas that underwent nodal evaluation from 2004 through 2009, 74.0% had Clark level IV (invasion of the reticular dermis) or Clark level V (invasion of the deep, subcutaneous tissue) and 9.5% were ulcerated. From 2010 through 2013, of the T1b melanomas undergoing nodal evaluation, 82.6% had an elevated mitotic rate only, 3.7% were ulcerated, and 13.7% had both ulceration and an elevated mitotic rate.
Conclusions and Relevance
It appears that after the institution of AJCC 7, there was an overall decrease in the number of T1 melanomas undergoing nodal biopsy without a clinically relevant change in sentinel lymph node positivity, with an increase in the number of T1b melanomas undergoing nodal evaluation.
The 7th edition of the AJCC Staging Manual (AJCC 7)1 was published in 2009, which changed the staging of thin melanomas. After 2009, a melanoma with a Breslow depth of 1 mm or smaller (defined as thin melanoma) was considered T1b if it had either ulceration present or an elevated mitotic rate (defined as ≥1 mitosis/mm2).1 This staging change affected the large subset of newly diagnosed melanomas known as thin melanomas.2
The National Comprehensive Cancer Network guidelines during the AJCC 7 stated that for melanomas ranging from 0.75 to 1.0 mm, sentinel lymph node biopsy (SLNB) should be considered if the melanoma had high-risk features.3 While there is no consensus on the definition of high-risk features, the presence of ulceration or a mitotic rate of 1 or more is commonly used.4,5 There is controversy regarding whether an elevated mitotic rate has predictive associations with a positive SLNB in thin melanomas. The SLNB–positivity rate in thin melanomas varies from 5% to 11.0%.6-19
Melanoma is the sixth most common cancer diagnosis in the United States, with an estimated 87 110 new cases diagnosed in 2017.20 The incidence of melanoma is predicted to increase each year and thin melanomas are the largest subset of newly diagnosed melanomas.2,21 The increasing incidence of thin melanomas leads to a large population of patients who may potentially harbor nodal disease, and for whom SLNB may be recommended.2 The current AJCC 83 melanoma staging guidelines eliminated mitotic rate as a differentiator between T1a and T1b melanomas. With this change, we sought to evaluate associations between implementation of the change in tumor staging in the AJCC 7 and the distribution of thin melanomas undergoing nodal surgery and their nodal positivity rates.22
Data were abstracted from the National Cancer Database from years 2004 through 2013. Adults (≥18 years) with T1 primary melanoma tumor depths (Breslow depth ≤1 mm) were included. Patients were excluded if the Breslow depth was larger than 1.0 mm or they only had melanoma in situ. Melanomas were classified as T1a or T1b based on the AJCC guidelines at time of their diagnosis. Patients who were missing ulceration status or Clark level were excluded. Only those patients diagnosed after 2009 were excluded if the mitotic rate was missing. Patients were also excluded if they had lymph nodes removed but did not have pathologic data reported. We divided patients into 2 groups (2004-2009 and 2010-2013) based on when the AJCC 7 went into effect. Data were analyzed in May 2017.
The National Cancer Database collects all lymph node data of patients in an aggregate form. This means that nodes removed from SLNB and completion lymphadenectomy procedures are all captured together, and there are no codes to distinguish which procedures a patient underwent. This difficulty is similar to that seen in other national databases.19 Because all patients had thin melanomas, we assumed that they all initially had an SLNB and then, if needed, a completion lymphadenectomy. We initially divided patients into 2 groups: those with an SLNB alone (defined as 1-3 nodes recorded) and those with an SLNB followed by a completion dissection or up-front lymphadenectomy (defined as ≥4 nodes recorded).23,24 Preliminary analyses suggested that outcomes between those classified as an SLNB and those with more than 4 nodes were similar, therefore we combined all patients into 1 nodal evaluation group for analysis. We refer to this entire group as having a nodal evaluation.
This research was approved by the institutional review board of Vanderbilt University Medical Center, Nashville, Tennessee; the institution waived informed consent based on de-identified data. All analyses were conducted using SAS, version 9.1 (SAS Institute) and R, version 3.3.1 (R Project for Statistical Computing). A 2-sided P value of <.05 was considered significant. For analysis between groups, the χ2 test was used for categorical variables and the t test, analysis of variance, or Mann-Whitney U test were used for continuous variables. Variables for the logistic regression were defined a priori based on the previous literature and AJCC 7 T1b melanoma criteria.5-7,19 Logistic regression was used to calculate odds ratios and 95% CIs for the association between lymph node–positivity and elevated mitotic rate (≥1), adjusting for ulceration, Breslow depth, age, and sex.
A total of 419 787 melanoma cases were entered into the National Cancer Database from 2004 through 2013. After applying exclusion criteria, 141 280 patients with T1 melanoma remained. Of 86 846 patients diagnosed from 2004 through 2009, 53.7% (49 644) were male and had a mean (SD) age of 57.7 (16.4) years, and of 54 434 patients diagnosed from 2010 through 2013, 54.3% (31 086) were male and had a mean (SD) age of 59.5 (15.9) years (Table 1). Most of the patients in each group were white race/ethnicity (97.9% vs 97.2%; P < .001). There was a 3.8% statistically significant decrease in nodal evaluation rates; 32 485 of 86 846 patients (37.6%) before 2010, and 18 379 of 54 434 patients (33.8%) after 2010 (P < .001). This was accompanied by a 1.0% decrease in nodal positivity rates (3166 vs 1618; P < .001). A statistically significant increase in the proportion of patients with T1b melanomas undergoing nodal evaluation was seen in the 2010 through 2013 group (2004-2009: 15 855 of 86 846 patients [48.8%] vs 2010-2013: 11 427 of 54 434 patients [62.2%]; P < .001).
In the 2004 through 2009 group, 15 855 of 86 846 patients (48.8%) with T1b melanomas had a nodal evaluation, based on staging criteria from AJCC 6. The most common T1b melanoma criterion identified on pathologic findings was Clark level IV (invasion of the reticular dermis) or level V (invasion of the deep, subcutaneous tissue), in 11 736 of 15 855 patients (74.0%) and it was associated with a positive nodal rate of 10.2% (1197 of 2279 patients) (Table 2). The least frequent T1b melanoma criteria noted in the nodal evaluation group was ulceration only in 1499 of 15 855 patients (9.5%). Patients with ulceration only had a higher nodal positivity rate (223 of 2279 patients [14.9%]) than patients with Clark level IV or V (1197 of 2279 [10.2%]). Combining ulceration and Clark level IV or V into 1 group had an additive effect with the nodal positivity rate increasing to 32.8% (859 of 2279 patients).
In the 2010 through 2013 group, 11 427 of 18 379 patients with T1b melanomas (62.2%) underwent nodal evaluation. The proportion of melanomas with ulceration alone decreased from 9.5% (1499 of 15 855 patients) to 3.7% (420 of 11 427) after 2009. The most common T1b melanoma feature present in those undergoing a nodal evaluation was an elevated mitotic rate in 9439 of 11 427 patients (82.6%), leading to a nodal positivity rate of 8.2% (772 of 9439 patients) and 1568 of 11 427 patients (13.7%) had both ulceration and elevated mitotic rate. The patients with positive lymph nodes (772 of 1257 [61.4%]) had only elevated mitotic rate on pathologic findings. Overall, 96.3% of patients (11 007 of 11 427) undergoing a nodal evaluation had either an elevated mitotic rate alone or an elevated mitotic rate in the presence of ulceration. There were 9035 of 54 434 T1 melanomas (16.6%) from the 2010 through 2013 group that were classified as T1b melanoma based just on an elevated mitotic rate alone and they underwent nodal evaluation.
A multivariable model was created to determine what factors were associated with nodal positivity in all patients with T1b melanoma who underwent nodal evaluation from 2010 through 2013 (Table 3). Predictors of nodal positivity included a mitotic rate of 1 or more (odds ratio [OR], 1.93; 95% CI, 1.69-2.19; P < .001) and ulceration (OR, 3.27; 95% CI, 2.86-3.27; P < .001). Older age (OR, 0.992; 95% CI, 0.988-0.996; P < .001), female sex (OR, 0.80; 95% CI, 0.7-0.90; P < .001) and Breslow depth (OR, 0.991; 95% CI, 0.990-0.993; P < .001) were protective against positive nodes.
Within the entire cohort of 141 280 thin melanomas diagnosed from 2004 through 2013, 6901 (4.9%) were ulcerated, 16 238 (11.5%) had a mitotic rate of 1 or more, and 2153 (1.5%) had both elements. The most common pathologic feature was an elevated mitotic rate, which increased in frequency as the depth increased from 7.0% (0.01-0.50 mm) to 22.6% (0.76-1.00 mm) (Table 4). In the ultrathin melanomas (0.01-0.50 mm), 11 476 of 85 142 patients (13.5%) had primary tumor ulceration and/or an elevated mitotic rate. In thin melanomas, 25 291 of 31 345 patients (17.9%) had high-risk features. In the thin melanoma group (0.76-1.0 mm), 7637 of 24 793 patients (30.8%) should be considered for an SLNB per the National Comprehensive Cancer Network guidelines.3
Our data showed a statistically significant decrease (3.8%) in the number of nodal evaluations performed for patients with thin melanomas after 2009. This was accompanied by a 1.0% decrease in the nodal positivity rate. While this decrease in positivity is statistically significant, it is likely not clinically significant.
We also observed that the composition of the nodal evaluation group shifted after the AJCC 7, with more patients with T1b melanomas undergoing nodal evaluations. In agreement with previous studies,6,7 we found that T1b melanomas with an elevated mitotic rate alone were associated with the presence of a positive node and demonstrated a nodal positivity rate of 8.2%. This is a particularly important observation after the publication of the AJCC 8, which removed mitotic rate from the staging guidelines for thin melanomas.22 Our data supported continued attention to the mitotic rates in patients with thin melanomas, when deciding whether a sentinel node biopsy should be performed. We also showed that ultrathin melanomas (Breslow depth <0.51 mm) had a higher than expected rate of ulceration and elevated mitotic rate. Whether the presence of these high-risk features in ultrathin melanomas influences survival, nodal positivity, or need for formal nodal evaluation in melanomas of this depth remains unclear.
A previous study25 using the Surveillance, Epidemiology, and End Results (SEER) database evaluated patients from 2004 through 2010 and found no change in the number of SLNBs after the implementation of the AJCC 7 (13.4% before 2010 vs 14.4% for 2010; P = .056). Those authors noted a decrease in the number of patients with T1b melanomas undergoing SLNB from 40.9% to 33.3% (P < .001) after 2010. One possible reason for the difference in our findings was that the authors of that study included melanomas designated T1 lesion as not otherwise specified. These lesions could have represented either T1a or T1b melanomas, preventing correct classification in their analysis. In our study, we excluded melanomas that could not be classified, ensuring that all patients in the cohort could be accurately assigned as T1a or T1b melanoma. After 2010, the SEER research25reported an increase from 7.7% to 26.3% (P < .001) in the proportion of T1 melanoma (not otherwise specified). This increase could account for the decrease in T1b melanomas. The T1 melanomas (not otherwise specified) may have actually been T1b melanomas. Furthermore, the authors of that study25 also concluded that an elevated mitotic rate did not predict positive nodes. Our data supported previous reports that a mitotic rate of 1 or more in thin melanomas was associated with nodal positivity.6,7 The previous study25 using the SEER database had a total of 37 884 thin melanomas with only 5357 reported after 2010. In our study, we used the National Cancer Database data that captures 70% of newly diagnosed cancers and produced a much larger population for study (n = 141 280 cases). Given the increased scope of the National Cancer Database compared with the SEER database, our study captures a more diverse group of practitioners and may have been more generalizable.26
A study by Oude Ophius et al,19 performed a similar analysis using the Dutch national database of 29 546 patients with T1 melanoma and found an increase in the percentage of nodal biopsies from 2.0% to 4.8% without a change in the nodal positivity rates (10.5% to 8.8%). Their nodal biopsy rate was significantly lower than our study and the SEER study.25 In that study, the authors state that before 2012, nodal biopsy for T1b melanomas was not recommended in their national guidelines, which explained why the biopsy rate reported in their study was lower than reported in our study.19
The pathologic features associated with the performance of nodal evaluation did appear to have been influenced by the change in the staging guidelines. An increase in the number of patients with T1b melanomas undergoing a nodal evaluation was seen in the 2010 through 2013 group. In addition, a change in the composition of the patients with T1b melanomas undergoing nodal evaluation also occurred. We reported a decrease in the number of patients with ulceration alone, as well as a shift of the largest cohort undergoing nodal evaluation from those with Clark level IV or V to those with a mitotic rate of 1 or more. This result supports the conclusion that the introduction of mitotic rate into the staging guidelines was quickly adopted by practitioners who treat melanoma and influenced their decision to perform SLNB. Furthermore, we observed a decrease in the number of patients undergoing nodal evaluation for ulceration only, as well as an overall decrease in the number of melanomas with ulceration. With attention shifting to the presence of elevated mitotic rate, the pathologist may have been less concerned with reporting ulceration and chose instead to focus on reporting mitotic rate. There was also a decrease in the nodal positivity rate from 14.9% to 10.0% among the patients with T1b melanomas and ulceration only in the 2010 through 2013 group. If the pathologist was focused more on accurately reporting mitotic rate, this decrease in nodal positivity rate may be artificial. Patients with ulceration could have been shifted from the ulceration alone group to the T1b melanoma group containing both ulceration and elevated mitotic rate, with increased attention to mitotic rate.
Ulceration remains an important prognostic indicator, with a higher OR for nodal positivity than mitotic rate (3.37 vs 1.94). However, consistent with other studies,6,16 our data support mitotic rate as an independent predictor of positive sentinel nodes. While the presence of ulceration is a stronger predictor for nodal positivity than mitotic rate, an elevated mitotic rate alone was associated with a nodal positivity in 8.2% of patients.
Identification of positive sentinel lymph nodes has important implications for staging, treatment, and survival.9,27 The Multicenter Selective Lymphadenopathy Trial–227 showed that sentinel lymph node positivity provides prognostic information. Furthermore, it is important to identify patients with nodal involvement, as the presence of nodal disease raises the disease designation to stage 3, and may influence decisions regarding the administration of systemic therapy, the performance of routine surveillance scans, and the interval of follow-up.27 While the AJCC 8 has removed mitotic rate from the staging of thin melanomas, we conclude that mitotic rate should continue to be used as a criterion for considering nodal evaluation, based on our analysis that to our knowledge is the largest to date.22
Our study has several limitations. First, all lymph nodes in the National Cancer Database are reported in aggregate, whether from an SLNB or completion lymphadenectomy procedure. We were unable to ascertain whether lymph nodes were evaluated as part of an SLNB, a completion lymph node dissection, or up-front lymphadenectomy. We elected to combine these patients into 1 group because these were the only thin melanomas that may not have had an up-front lymphadenectomy.
Second, the reporting of mitotic rate was not routinely performed before 2010, and few patients in the 2004 through 2009 group had information about this variable. Therefore, we could not make a direct comparison between the groups about this variable. In the 2004 through 2009 group, most patients with a reported mitotic rate were diagnosed in 2009. With so many patients missing data regarding mitotic rate, we were unable to look at the entire cohort (2004-2013) as 1 group using the AJCC 6 and AJCC 7. In a separate analysis using patients in the 2010 through 2013 cohort, we observed similar results, with an increase in T1b melanomas diagnosed, and an increase in nodal evaluations being performed for T1b melanomas. Third, in our multivariable model, the Breslow depth showed an inverse association between tumor depth and nodal positivity.
Fourth, there was a statistically significant difference in age (1.8 years) and a sex distribution (0.6%) between groups. Previous studies have shown that older age was associated with a lower nodal positivity rate, which we also observed.28,29 Although the age difference between our 2 groups was statistically significant, a difference of 1.8 years was not clinically significant and would not be expected to influence the nodal positivity rate in a clinically relevant manner. Similarly, although there was a statistically significant difference in the proportion of men and women in our 2 groups (0.6%), this is also of unlikely clinical relevance. Finally, controversy remains regarding what is considered a high-risk feature in thin melanomas.3 For this article, we defined a high-risk feature as the presence of ulceration and/or a mitotic rate of 1 or more. We acknowledge that these may not be criteria used by all practitioners treating melanoma, and there are other criteria that may have been used to determine when a patient underwent a nodal evaluation.
After the addition of elevated mitotic rate to the AJCC 7, there appeared to be a decrease in the number of nodal evaluations being performed and an increase in the number of patients with T1b melanomas undergoing nodal evaluation. Most of these nodal evaluations were for a mitotic rate of 1 or more, which appeared to be associated with nodal positivity. Our data also showed that high-risk features in melanomas below 0.76 mm depth appear not to be rare (15.2%), and further research is needed to determine the optimal approach to SLNB for these patients.
Accepted for Publication: December 22, 2018.
Corresponding Author: Rondi M. Kauffmann, MPH, MD, Division of Surgical Oncology and Endocrine Surgery, Vanderbilt University Medical Center, 2220 Pierce Ave, 597 Preston Research Building, Nashville, TN 37232 (email@example.com).
Published Online: March 6, 2019. doi:10.1001/jamadermatol.2018.5902
Author Contributions: Drs Isom and Kauffmann had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Concept and design: Isom, Hooks, Kauffmann.
Acquisition, analysis, or interpretation of data: Isom, Wheless, Kauffmann.
Drafting of the manuscript: Isom, Kauffmann.
Critical revision of the manuscript for important intellectual content: All authors.
Statistical analysis: Wheless, Kauffmann.
Administrative, technical, or material support: Isom, Kauffmann.
Supervision: Hooks, Kauffmann.
Conflict of Interest Disclosures: None reported.
Funding/Support: Dr Isom reported support by grant T32 CA106183, a National Cancer Institute Institutional National Research Service Award Postdoctoral Training Grant.
Role of the Funder/Sponsor: The funding organization had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.