Performance of the 23-gene Expression Profile (23-GEP) Test by Histopathological Evaluation in an Independent, Multi-Center Performance Cohort of Cutaneous Melanocytic Neoplasms

Main Article Content

Matthew S. Goldberg
Kiran Motaparthi
Gregory A. Hosler
Clay J. Cockerell
Sarah I. Estrada
Natalie D. Depcik-Smith
Jose A. Plaza


23-gene expression profile, 23-GEP, melanoma, nevus, ancillary diagnostic test, melanocytic neoplasms


Histopathologic evaluation can effectively diagnose most melanocytic neoplasms; however, lesions considered to be difficult-to-diagnose pose challenges for accurate classification of malignant potential, which can lead to over- or under-treatment. Ancillary testing to provide additional information is available for such cases. The validated diagnostic 23-GEP test provides an objective result of benign, malignant, or intermediate. In this large, independent cohort, the performance of the 23-GEP in its current laboratory is presented. Lesions from patients ≥18 years old were enrolled from eight centers or from melanoma cases clinically submitted for prognostic 31-GEP testing. Lesions were independently reviewed by 3–5 dermatopathologists and included in the study if they were fully concordant or non-concordant without conflicting diagnoses (i.e., both benign and malignant designations) or majority of unknown malignant potential designations, resulting in a cohort (n=2512) of benign nevi (n=1140) and malignant melanomas (n=1372). Accuracy metrics and two-tailed 95% confidence intervals (CIs) were calculated using resampling x10,000 iterations to establish a balanced number of benign versus malignant samples. The 23-GEP performance within this cohort was 91.3% (95% CI, 89.2–93.2%) sensitivity, 91.9% (89.8–93.8%) specificity, 92.2% (9.30–94.0%) positive predictive value, and 91.0% (89.0–92.9%) negative predictive value; 7.8% of lesions received an intermediate result. These metrics do not deviate significantly from previously published studies. These data demonstrate that the 23-GEP has an overall accuracy of 91.6% (90.1–93.0%), further supporting its use as an ancillary test that can be integrated with clinical and histopathologic information to guide final diagnosis.


1. Shoo, B. A. et al. J Am Acad Dermatol 2010. 62 (5) 751–56.

2. Gerami, P. et al. Am J Surg Pathol 2010. 34 (6) 816–21.

3. Haws, B. et al. J Cutan Pathol 2012. 39 (9) 844–49.

4. Elmore, J. G. et al. BMJ 2017. 357 (1) j2813.

5. Clarke, L. E. et al. J Cutan Pathol 2015. 42 (4) 244–52.

6. Clarke, L. E. et al. Cancer 2017. 123 (4) 617–28.

7. Ko, J. S. et al. Cancer Epidem Biomar Prev 2017. 26 (7) 1107–13.​

8. Ko, J. S. et al. Human Pathology 2019. 86 213–21.​

9. Clarke, L. E. et al. Personalized Medicine 2020. 17 (5) 361–71.​

Similar Articles

You may also start an advanced similarity search for this article.

Most read articles by the same author(s)