Abstract
Background
The 31-gene expression profile test (Class 1A: low-risk; 1B/2A: intermediate-risk; 2B: high-risk) is validated to identify patients with cutaneous melanoma who can safely forego sentinel lymph node biopsy (SLNB). The objective of the current study is to quantify SLNB reduction by clinicians using 31-GEP.
Methods
Patients with T1-T2 tumors eligible for SLNB were seen by surgical oncologists (89.1%), dermatologists (7.8%), and medical oncologists (3.1%). After receiving 31-GEP results but before SLNB, clinicians were asked which clinical and pathological features influenced SLNB decisions (n = 191). The Exact binomial test was used to compare SLNB procedure rates to a contemporary study (78% SLNB baseline rate). Logistic regression modeling (odds ratio [OR], 95% CI) was used to identify features associated with SLNB procedure rates.
Results
One hundred clinical decisions (52.4%) were influenced by the 31-GEP to forego SLNB and 70% (70/100) were not performed. Of the 30 performed, 0% (0/30) were positive. The 31-GEP influenced sixty-three clinical decisions (33.0%) to perform SLNB, and 92.1% (58/63) were performed. There was a clinically meaningful 29.4% reduction of SLNBs performed in patients with a Class 1A result relative to the baseline rate of 78.0% (p < .01). In patients ≥55 or ≥65-year-old, SLNB reduction was 32.3% (p < .01), 28.3% (p < .01), respectively. Overall, 85.3% of decisions relating to SLNB were influenced by 31-GEP results.
Conclusion
In this prospective, multicenter study, clinicians demonstrated clinically meaningful use of the 31-GEP test to forego or pursue SLNB in patients with T1-T2 tumors resulting in a significant, risk appropriate decrease in SLNBs.
Introduction
Sentinel lymph node biopsy (SLNB) is endorsed by current guidelines from the American Joint Committee on Cancer (AJCC) and the National Comprehensive Cancer Network (NCCN) as a staging procedureCitation1,Citation2. Patients with T1a tumors (assumed to have an SLNB positivity rate <5%) are not recommended for SLNB without additional high-risk features, while patients with T1b tumors should discuss and consider an SLNB (assumed positivity rate 5–10%), and patients with T2-T4 tumors should be offered an SLNB (assumed positivity rate >10%)Citation2. However, the SLNB procedure does not affect melanoma-specific survival (MSS rate is no different between patients undergoing an SLNB with a wide local excision [WLE] vs. WLE alone), a positive SLN is not predictive of response to therapy, and one study using the Surveillance, Epidemiology, and End Results (SEER) database showed that up to 88% of SLNB are negative, which provides little information about survival, as two-thirds of patients who progress and die from melanoma are SLNB negativeCitation3,Citation4.
These data have led to recent studies showing the benefit of adjuvant therapy in patients with stage IIB-IIC melanoma (e.g. pathologically SLN negative with T3b-T4 tumors), leading to FDA approval of pembrolizumab for patients with stage IIB-IIC melanomaCitation5. A high SLN negativity rate, particularly in patients with stage I melanoma (e.g. T1a-T2a), has led to questions about SLNB recommendations using traditional population-based clinicopathology features aloneCitation6,Citation7. Most recently, Moncrieff and colleagues showed in a large, multicenter study that patients with T1b-T2a tumors have an 11.4% SLN positivity rate, and 99.3% of those with a positive SLN were classified as stage IIIA, for which 5-year MSS is similar to that of stage IIACitation8. The SLNB procedure also carries a healthcare cost burden with costs reported between $6700–$15,223Citation9,Citation10, and one study demonstrated that for SLNB to have a positive impact, 142 SLNBs need to be performed to identify a single patient that would die from melanomaCitation11.
These clinical issues demonstrate that an approach to SLNB decision-making that relies on AJCC 8th edition staging (Breslow thickness and ulceration status) may not be the optimal route to determine which patients get an SLNB. Understanding the molecular basis for tumor progression and identifying the signatures that, once validated, predict this progression, is important for continued improvements to patient careCitation12–14. Recent studies have demonstrated that the incorporation of molecular information (gene expression profile [GEP] testing) with patient-specific clinicopathologic factors can improve the accuracy of identifying patients at low risk of SLN positivity that can safely forego the procedureCitation15. We previously showed that the 31-gene expression profile test (31-GEP; DecisionDx-Melanoma, Castle Biosciences, Inc., Friendswood, TX) is validated to identify patients with T1-T2 melanoma who have <5% risk of a positive SLN and retain high MSS ratesCitation15–19. Using the 31-GEP in conjunction with current staging factors can help reduce the number of unnecessary SLNBs, associated complications, and impact healthcare costs.
A prospective, multicenter study was designed to directly test the hypothesis that clinicians use the 31-GEP test in the context of other clinical and pathologic features to pursue or forego SLNB in patients with T1-T2 tumors and that incorporating 31-GEP testing into clinical decision-making can reduce the number of SLNBs performed.
Methods
The DecisionDx-Melanoma Impact on Sentinel Lymph Node Biopsy Decisions and Clinical Outcomes (DECIDE) prospective study was designed to identify the effect of physician use of 31-GEP testing on the conductance of SLNB procedures in dermatologic, medical, and surgical oncology centers. Eligible patients newly diagnosed with cutaneous melanoma from 22 institutions (including academic, hospital, and community practices) were enrolled in the study from March 2020–May 2022 if the treating clinician was considering using the 31-GEP to guide SLNB decision-making, the patients were at least 18 years old, could provide written informed consent, and were reasonably able to follow-up with the enrolling physician at regular intervals. Patients were excluded if they were not able to give written informed consent, had stage III or IV melanoma at diagnosis, had a previous diagnosis in the same anatomical location as the current primary, if the clinician ordered the 31-GEP test for something exclusively other than SLNB guidance, or the patient already had an SLNB at the time of consent. Moreover, patients were removed from analysis if there was a failure of the 31-GEP to provide a result or if there was a failure of SLNB to map (Supplemental Figure 1). Institutional review board approval was obtained for each institution. A secure web-based portal allowed clinicians to input study results. During study visit one, informed consent was obtained, and the 31-GEP test was ordered. All 31-GEP testing was performed at a CAP/CLIA-certified lab. At visit two, the 31-GEP result was received. At visit three, the clinician recorded preference for or against SLNB based on each of the following variables: 31-GEP result, Breslow thickness, ulceration, mitotic rate, lymphovascular invasion (LVI), transected base, presence of regression, presence of tumor-infiltrating lymphocytes (TILs), microsatellites, patient preference, patient age, and patient comorbidities. SLNB status (e.g. performed or not) and SLNB result (if performed) were recorded (Supplemental Figure 2).
The DECIDE study aims to answer two primary objectives: (1) determine the impact of 31-GEP test results specifically on SLNB surgical decisions in patients with SLNB-eligible T1-T2 melanoma, and (2) track and evaluate 5-year clinical outcomes for patients in each 31-GEP subclass, including those who did and did not undergo SLNB and those with T3-T4 melanoma. Patients with T1a tumors were included in the study if they had one of the following high-risk features: mitotic rate ≥2/mm2, lymphovascular invasion, absence of TILs, microsatellites, regression, age <40, or a transected based. Primary aim 2 of the study analyzing 5-year outcomes is still underway and will be reported once sufficient follow-up has been amassed. To assess primary aim 1 and determine if clinicians incorporate 31-GEP testing into SLNB decision making in a clinically impactful manner, power calculations were performed when expecting a 20% reduction in SLNB based on the influence of 31-GEP testing compared to a 90% baseline surgery rate in SLNB-eligible patients based on previous studies in primarily surgical centersCitation15,Citation18. Based on these assumptions, 172 patients were needed for SLNB analysis. The analysis used an alpha spending of 0.01 so that p < .01 represents a significant difference between groups for all statistical tests. As the rate of SLNB may vary between institutions and T-stages, we performed a sensitivity analysis by comparing the SLNB rate in our cohort to varying baseline SLNB rates. Sensitivity analysis was accomplished by varying the assumed SLNB base rate from 90% to 50% in 5% intervals. Differences between SLNB rates were compared using the Exact binomial test using the R statistical package (v4.2.0) The association between tumor, clinical, and pathological features with SLNB performance was studied by using stepwise selection on a logistic regression model incorporating the 31-GEP result, Breslow thickness, ulceration, mitotic rate, LVI, transected base, presence of regression, presence of TILs, microsatellites, patient preference, patient age, and patient comorbidities. Significant variables from the logistic regression model are reported as odds ratio (OR) and 95% confidence intervals (CI). SLNB reduction rates were analyzed based on a previous algorithm indicating that patients with a 31-GEP Class 1A result who are either ≥65 years old (primary analysis) or ≥55 years old (secondary analysis) with T1-T2 tumors have a < 5% risk of SLN positivityCitation18. For two patients, the influence of the 31-GEP on the SLNB decision was not reported (both received a Class 1A result). Therefore, calculations on the influence of 31-GEP testing on SLNB rates are done using n = 191.
Results
The clinical and pathological characteristics of patients are shown in . One hundred and ninety-three patients met the inclusion criteria and were included in the analysis. The median age was 65 years old, most had a T1b tumor (43.0%), a T1a-HR tumor (26.4%), or a T2a tumor (23.3%), while 7.3% had a T2b tumor. Ulceration was present in 11.4% of tumors. Most had mitotic rate <2/mm2 (71.0%), and no regression (81.8%), no microsatellites (99.4%), and no lymphovascular invasion (98.7%). A transected base was observed in 53.3% of cases. The median Breslow thickness was 0.9 mm and 89.1% of clinicians were surgical oncologists, 7.8% were dermatologists, and 3.1% were medical oncologists.
Table 1. Patient demographics.
All patients
One hundred clinical decisions (52.4%; 100/191) were influenced by the 31-GEP to forego SLNB with 70% (70/100) of biopsies not performed (). In the remaining 30% (30/100) of cases where the clinician decision was influenced by the 31-GEP to forego SLNB, but the biopsy was performed, 83% (25/30) were influenced by patient preference, and 0% (0/30) of the SLNBs were positive (Supplemental Figure 1). Sixty-three clinical decisions (32.6%; 63/191) were influenced by the 31-GEP result to perform SLNB with 92.1% (58/63) of biopsies performed (). Of the 5 cases where biopsies were not performed, 80% (4/5) were influenced by patient preference. The SLNB positivity rate for the group influenced by the 31-GEP was 13.8% (8/58) (Supplemental Figure 1). The overall influence of 31-GEP on SLNB decisions was 85.3% (163/191) and concordance between performing an SLNB and 31-GEP influence was 78.5% (128/163) ().
Table 2. Influence of the 31-GEP on performing SLNB in T1a-HR – T2 tumors.
The overall SLNB performance rate was 59.1% (95% CI 51.8–66.1%), significantly lower than the SLNB contemporary baseline rate of 78.0% (95% CI 75.5–80.5%, p < .01) seen in Whitman et al. for the same tumor categoriesCitation15. In patients with a Class 1A result, the SLNB rate was 48.6% (95% CI 40.0–57.2%) (67/138), significantly lower than the contemporary baseline rate of 78.0% (p < .01). Sensitivity analysis showed that an SLNB rate of 48.6% for Class 1A was significantly lower than any baseline SLNB rates ≥60% (Supplemental Table 1). The SLNB positivity rate in those with a Class 1A result was 3.0% (2/67). In contrast, patients with an intermediate Class 1B/2A result had an SLNB performance rate of 80.6% and an SLN positivity rate of 13.8% (4/29), and patients with a high-risk Class 2B result had an SLNB performance rate of 94.7%, and an SLN positivity rate of 22.2% (4/18) ().
Table 3. SLNB performance rates in the DECIDE study compared to previous study.
Next, stepwise selection on a logistic regression model was performed to determine which features were associated with the likelihood of performing an SLNB. Variables included in the model are shown in . Variables that remained in the model after the stepwise selection as being associated with SLNB performance were the 31-GEP (OR = 14.83, 95% CI 6.06–46.79), Breslow thickness (OR = 2.63, 95% CI 1.20–6.18), tumor-infiltrating lymphocytes (OR = 4.71, 95% CI 1.10–22.54), age (OR = 2.12, 95% CI 0.93–5.14), and patient preference (OR = 32.73, 95% CI 11.53–126.53). Because p < .01 was considered statistically significant, only the 31-GEP and patient preference reached statistical significance. Variables dropped from the model due to no association with performing SLNB were ulceration, mitotic rate, LVI, regression, transected base, microsatellites, and patient comorbidities.
Table 4. Stepwise selection on logistic regression model identifying factors associated with performing an SLNB.
Patients ≥55 years old
In patients ≥55 years old, the SLNB performance rate in patients with a Class 1A result was 44.6% (95% CI 34.7–54.8%) (45/101), significantly lower than the contemporary baseline rate 76.9% (95% CI 73.8–79.9%) (p < .01) for the subgroup ≥55 years old in Whitman et al. Sensitivity analysis showed that an SLNB rate of 44.6% for Class 1A/≥55 years old was significantly lower than baseline SLNB rates ≥60% (Supplemental Table 1). The SLNB positivity rate in those with a Class 1A result was 2.2% (1/45). In contrast, patients with an intermediate Class 1B/2A result had an SLNB performance rate of 77.4% and an SLN positivity rate of 16.7% (4/24), and patients with a high-risk Class 2B result had an SLNB performance rate of 100% and an SLNB positivity rate of 26.7% (4/15) ().
Patients ≥65 years old
In patients ≥65 years old, the SLNB performance rate in patients with a Class 1A result was 48.0% (95% CI 36.3–59.9%) (36/75), significantly lower than the contemporary baseline rate 76.3% (95% CI 72.5–79.9%) (p < .01) for the subgroup ≥65 years old in Whitman et al. Sensitivity analysis showed that an SLNB rate of 48.0% for Class 1A/≥65 years old was significantly lower than baseline SLNB rates ≥65% (Supplemental Table 1). The SLNB positivity rate in those with a Class 1A result was 0% (0/36). In contrast, patients with an intermediate Class 1B/2A result had an SLNB performance rate of 71.4% and an SLN positivity rate of 26.7% (4/15), and patients with a high-risk Class 2B result had an SLNB performance rate of 100% and an SLN positivity rate of 33.3% (4/12) ().
Discussion
While SLN positivity is a staging factor for patients with cutaneous melanoma, 88% of SLNBs are negativeCitation4. Although a negative SLNB suggests better population survival than a positive result, two out of three patients who undergo an SLNB and go on to die from melanoma are SLNB negativeCitation3. Moreover, SLNB is associated with complication rates up to 11.3%Citation20. The 31-GEP is validated for risk of recurrence and metastasis prognosis and has also been validated to identify patients with T1-T2 melanoma who have <5% risk of a positive SLN and retain high MSS ratesCitation16,Citation18,Citation21. This study was initiated to document if and how decisions regarding SLNB performance are impacted by the 31-GEP, and, in particular, to address the rate at which clinicians reduce SLNBs in patients with low-risk 31-GEP results. The findings add to the body of evidence demonstrating the clinical use of the 31-GEP (Supplemental Table 2)Citation18,Citation22–26.
We showed that patients with a low-risk result received SLNB at a reduced rate compared to patients with an intermediate or high-risk result and a comparable contemporary group of patients where 31-GEP testing was not used to guide SLNB decisionsCitation15,Citation18. However, as the rate of SLNB may vary between institutions and T-stages, we performed a sensitivity analysis by comparing the SLNB rate in our cohort to varying baseline SLNB rates (that is, SLNB rates where 31-GEP was not available to guide SLNB decisions). This analysis demonstrated that dermatologists (7.8% of study clinicians), medical oncologists (3.1% of study clinicians), and surgical oncologists (89.1% of clinicians) using the 31-GEP in their decisions to perform SLNB reduced the number of biopsies significantly compared to baseline rates ranging between 90%-60%. In our study, the comparison between the study cohort was made using a separate contemporary cohort also seen at primarily surgical centers that demonstrated a 78% SLNB rate in patients with T1a-HR-T2 tumorsCitation15,Citation18, and likely provided an accurate comparison in the intended use population. Clinicians using the 31-GEP in SLNB decision-making performed 29.4% fewer biopsies than in current practice (48.6% in Class 1A vs. 78.0%), with a false-negative rate of 3.0% (2/67) for all cases, 0% (0/36) in patients ≥65-years-old and 2.2% (1/45) in those ≥55-years-old.
An interesting, but not unexpected, finding was that patient preference played such an outsized role in SLNB decision-making, evidenced by the large odds ratio in the logistic regression analysis. We have recently shown that patients with cutaneous melanoma have a strong desire to receive the results of gene expression testing as part of a shared decision on the treatment plan for their cancerCitation27, and the impact of patient involvement is consistent with an earlier report in breast cancerCitation28. In this study, in 83.3% (25/30) of cases where the clinician was influenced by the 31-GEP to forego SLNB but performed the procedure, patient preference was recorded as favoring the procedure. Similarly, in 80% (4/5) of cases for which the clinician was influenced by the 31-GEP to perform an SLNB but did not, patient preference favored foregoing SLNB. While it is likely that patient preference encompasses many other clinical and pathological features (i.e. comorbidities, age, etc.), these results highlight the importance of shared decision-making between patient and clinician. Although limitations exist in understanding what leads to patient preferences, by including patient preference in our study, we can begin to analyze which clinical and pathological features may influence patient decisions.
The study has some limitations. First, tumor location was not included as a question about influencing SLNB performance rates. Studies have shown that tumors on the head and neck have lower SLNB rates and may be a confounding factor in the analysisCitation29. However, in this study 19.7% of melanomas were on the head and neck compared to 21.0% in the contemporary baseline cohortCitation15. Next, the current analysis only looked at aim 1 of the DECIDE study regarding the influence of 31-GEP testing on SLNB performance rates, and outcomes data are still accruing. However, multiple previous studies have shown that the 31-GEP independently stratifies the risk of recurrence, metastasis, and death regardless of SLN statusCitation16,Citation17,Citation19,Citation30. Finally, systematic collection of different clinical and pathological factors likely varies by site and may contribute bias to the results. Despite these limitations, these data demonstrate that using the 31-GEP as a comprehensive molecular risk-stratification tool can help identify patients at low and high risk of SLN positivity and those at low and high risk of recurrence regardless of SLN status. Moreover, since the initiation of this study, the 31-GEP has been integrated with clinical and pathologic features for precise SLNB risk prediction (i31-GEP SLNB)Citation15 and recurrence risk (i31-GEP ROR)Citation31. As the i31-GEP SLNB was published in late 2021, there were not enough patients in the current study to perform analyses using this tool.
Conclusion
Results of the DECIDE study support previous reports that clinicians use the 31-GEP to guide SLNBCitation25,Citation26,Citation32 and provide compelling evidence that clinicians who utilize the 31-GEP as part of their routine clinical care perform fewer SLNBs relative to those who do not use the 31-GEP to guide SLNB decision-making. Finally, previous studies have shown that clinicians use the 31-GEP to guide office visits, referrals, and surveillance imaging strategiesCitation22–24, and an expert panel of dermatologists have given recommendations for implementing 31-GEP testing into clinical practiceCitation33. Taken together, the 31-GEP has demonstrated clinical use in managing patients for both SLNB guidance and long-term risk management.
Transparency
Declaration of funding
This study was funded by Castle Biosciences.
Author contributions
MY, BSS, TB, RE, JMG enrolled patients in the study and reviewed the data and manuscript. CB and BM analyzed data and wrote the manuscript.
Supplemental Material
Download MS Word (142.1 KB)Acknowledgements
None.
Declaration of financial/other relationships
M.Y., B.S., and T.B. have no conflicts of interest to declare. R.E. and J.M.G are on the speaker’s bureau for Castle Biosciences, Inc. R.E. is an advisor for Castle Biosciences, Inc. B.M. and C.N.B. are employees and stock/option holders at Castle Biosciences, Inc. Peer reviewers on this manuscript have no relevant financial or other relationships to disclose.
Data availability statement
No data will be made publicly available.
References
- Gershenwald JE, Scolyer RA, Hess KR, et al. Melanoma staging: evidence-based changes in the American joint committee on cancer eighth edition cancer staging manual. CA Cancer J Clin. 2017;67:472–492.
- National Comprehensive Cancer. NCCN guidelines melanoma: cutaneous. Version 2.2022. 2022.
- Morton DL, Thompson JF, Cochran AJ, et al. Final trial report of sentinel-node biopsy versus nodal observation in melanoma. N Engl J Med. 2014;370:599–609.
- Chen J, Xu Y, Zhou Y, et al. Prognostic role of sentinel lymph node biopsy for patients with cutaneous melanoma: a retrospective study of surveillance, epidemiology, and end-result population-based data. Oncotarget. 2016;7:45671–45677.
- Luke JJ, Rutkowski P, Queirolo P, et al. Pembrolizumab versus placebo as adjuvant therapy in completely resected stage IIB or IIC melanoma (KEYNOTE-716): a randomised, double-blind, phase 3 trial. Lancet. 2022;399(10336):1718–1729.
- Weitemeyer MB, Helvind NM, Brinck AM, et al. More sentinel lymph node biopsies for thin melanomas after transition to AJCC 8th edition do not increase positivity rate: a Danish population-based study of 7148 patients. J Surg Oncol. 2022;125:498–508.
- Egger ME, Stevenson M, Bhutiani N, et al. Should sentinel lymph node biopsy be performed for all T1b melanomas in the new 8th edition American Joint Committee on cancer staging system? Journal of the American College of Surgeons. 2019;228:466–472.
- Moncrieff MD, Lo SN, Scolyer RA, et al. ASO visual abstract: evaluation of indications for sentinel node biopsy in early-stage melanoma with the advent of adjuvant systemic therapy: an international multicenter study. Ann Surg Oncol. 2022;29:5946–5947.
- Agnese DM, Abdessalam SF, Burak WE, et al. Cost-effectiveness of sentinel lymph node biopsy in thin melanomas. Surgery. 2003;134:542–547. discussion 547–548.
- Aiken TJ, Stahl CC, Schwartz PB, et al. Sentinel lymph node biopsy is associated with increased cost in higher risk thin melanoma. J Surg Oncol. 2021;123:104–109.
- Marchetti MA, Bartlett EK. Sentinel lymph node biopsy in cutaneous melanoma-where do we stand? JAMA Dermatol. 2021;157:1159–1160.
- Dika E, Riefolo M, Porcellini E, et al. Defining the prognostic role of microRNAs in cutaneous melanoma. J Invest Dermatol. 2020;140:2260–2267.
- Riefolo M, Porcellini E, Dika E, et al. Interplay between small and long non-coding RNAs in cutaneous melanoma: a complex jigsaw puzzle with missing pieces. Mol Oncol. 2019;13(1):74–98.
- Bittner M, Meltzer P, Chen Y, et al. Molecular classification of cutaneous malignant melanoma by gene expression profiling. Nature. 2000;406(6795):536–540.
- Whitman ED, Koshenkov VP, Gastman BR, et al. Integrating 31-gene expression profiling with clinicopathologic features to optimize cutaneous melanoma sentinel lymph node metastasis prediction. J Clin Oncol Precision Oncology. 2021;5:1466–1479.
- Hsueh EC, DeBloom JR, Lee JH, et al. Long-term outcomes in a multicenter, prospective cohort evaluating the prognostic 31-gene expression profile for cutaneous melanoma. J Clin Oncol Precision Oncology. 2021;5:589–601.
- Jarell A, Skenderis B, Dillon LD, et al. The 31-gene expression profile stratifies recurrence and metastasis risk in patients with cutaneous melanoma. Future Oncol. 2021;17(36):5023–5031.
- Vetto JT, Hsueh EC, Gastman BR, et al. Guidance of sentinel lymph node biopsy decisions in patients with T1–T2 melanoma using gene expression profiling. Future Oncol. 2019;15(11):1207–1217.
- Arnot SP, Han G, Fortino J, et al. Utility of a 31-gene expression profile for predicting outcomes in patients with primary cutaneous melanoma referred for sentinel node biopsy. Am J Surg. 2021;221(6):1195–1199.
- Moody JA, Ali RF, Carbone AC, et al. Complications of sentinel lymph node biopsy for melanoma – a systematic review of the literature. Eur J Surg Oncol. 2017;43(2):270–277.
- Gerami P, Cook RW, Wilkinson J, et al. Development of a prognostic genetic signature to predict the metastatic risk associated with cutaneous melanoma. Clin Cancer Res. 2015;21(1):175–183.
- Dillon LD, McPhee M, Davidson RS, et al. Expanded evidence that the 31-gene expression profile test provides clinical utility for melanoma management in a multicenter study. Curr Med Res Opin. 2022;38(8):1267–1274.
- Berger AC, Davidson RS, Poitras JK, et al. Clinical impact of a 31-gene expression profile test for cutaneous melanoma in 156 prospectively and consecutively tested patients. Curr Med Res Opin. 2016;32(9):1599–1604.
- Schuitevoerder D, Heath M, Cook RW, et al. Impact of gene expression profiling on decision-making in clinically node negative melanoma patients after surgical staging. J Drugs Dermatol. 2018;17(2):196–199.
- Marson J, Litchman G, Svoboda R, et al. Assessment of the 31-gene expression profile test by dermatologists: a cross-sectional survey from national dermatology conferences. J of Skin. 2021;5(2):101–107.
- Scott AM, Dale PS, Conforti A, et al. Integration of a 31-gene expression profile into clinical Decision-Making in the treatment of cutaneous melanoma. Am Surg. 2020;86(11):1561–1564.
- Ahmed K, Siegel JJ, Morgan-Linnell SK, et al. Attitudes of patients with cutaneous melanoma toward prognostic testing using the 31-gene expression profile test. Cancer Med. 2022. DOI:10.1002/cam4.5047
- O'Neill SC, Brewer NT, Lillie SE, et al. Women’s interest in gene expression analysis for breast cancer recurrence risk. J Clin Oncol. 2007;25(29):4628–4634.
- Ollek S, Minkova S, Taqi K, et al. Population-based assessment of sentinel lymph node biopsy in the management of cutaneous melanoma. cjs. 2022;65(3):E394–E403.
- Gerami P, Cook RW, Russell MC, et al. Gene expression profiling for molecular staging of cutaneous melanoma in patients undergoing sentinel lymph node biopsy. J Am Acad Dermatol. 2015;72(5):780–785.e3.
- Jarell A, Gastman BR, Dillon LD, et al. Optimizing treatment approaches for patients with cutaneous melanoma by integrating clinical and pathologic features with the 31-gene expression profile test. J Am Acad Dermatol. 2022;87(6):1312–1320.
- Farberg AS, Glazer AM, White R, et al. Impact of a 31-gene expression profiling test for cutaneous melanoma on dermatologists’ clinical management decisions. J Drugs Dermatol. 2017;16(5):428–431.
- Kwatra SG, Hines H, Semenov YR, et al. A dermatologist’s guide to implementation of gene expression profiling in the management of melanoma. J Clin Aesthet Dermatol. 2020;13:s3–s14.