According to current National Comprehensive Cancer Network (NCCN) treatment guidelines, the specificity of sentinel lymph node biopsy (SLNB) for the detection of lymph node metastasis in head and neck melanoma (HNM) is low. Therefore, Oliver et al. developed multiple machine learning (ML) algorithms to identify HNM patients at very low risk of occult nodal metastasis using National Cancer Database (NCDB) data for 8466 clinically node negative HNM patients who underwent SLNB.
The researchers compared SLNB performance under NCCN guidelines and ML algorithm recommendations on independent test data from the NCDB and an academic medical center.
The results showed that the top-performing ML algorithm recommendations obtained significantly higher specificity when compared to the NCCN guidelines in both internal and external test populations. The sensitivity achieved was >97%.
In conclusion, the authors noted that clinically node negative HNM patients at very low risk of nodal metastasis and who may not benefit from SLNB can be identified with the use of machine learning.
Oliver JR, Karadaghy OA, Fassas SN, Arambula Z, Bur AM. Machine learning directed sentinel lymph node biopsy in cutaneous head and neck melanoma [published online ahead of print, 2022 Feb 6]. Head Neck. 2022;10.1002/hed.26993. doi:10.1002/hed.26993