Sex- and Site-Related Significance in Cutaneous Head and Neck Melanoma

Predicting cutaneous malignant melanoma patients' survival using deep learning: a retrospective cohort study

BACKGROUND

Cutaneous malignant melanoma (CMM) has the worst prognosis among skin cancers, especially metastatic CMM. Predicting its prognosis accurately could direct clinical decisions.

METHODS

The Surveillance, Epidemiology, and End Results database was screened to collect CMM patients' data. According to diagnosed time, patients were subdivided into three cohorts, train cohort (diagnosed between 2010 and 2013), validation cohort (diagnosed in 2014), and test cohort (diagnosed in 2015). Train cohort was used to train deep learning survival model for cutaneous malignant melanoma (DeepCMM). DeepCMM was then evaluated in train cohort and validation cohort internally, and validated in test cohort externally.

RESULTS

DeepCMM showed 0.8270 (95% CI, confidence interval, CI 0.8260-0.8280) as area under the receiver operating characteristic curve (AUC) in train cohort, 0.8274 (95% CI 0.8286-0.8298) AUC in validation cohort, and 0.8303 (95% CI 0.8289-0.8316) AUC in test cohort. Then DeepCMM was packaged into a Windows 64-bit software for doctors to use.

CONCLUSION

Deep learning survival model for cutaneous malignant melanoma (DeepCMM) can offer a reliable prediction on cutaneous malignant melanoma patients' overall survival.

Extirpative Considerations of Melanoma of the Head and Neck

The incidence of melanoma is continuing to rise in the United States, and head and neck melanomas account for 25% of all cutaneous melanomas. The National Comprehensive Cancer Network guideline recommendations for surgical margins and sentinel lymph node biopsy in head and neck melanomas are the same as cutaneous melanoma located in other regions, but require special considerations when performing wide local excision, sentinel lymph node biopsy, and completion lymph node dissection and reconstruction taking into account the location of the melanoma and structures involved in and around the suggested margins.

Novel Prognostic Models Predicting the Cancer-Specific Survival in Patients with Cutaneous Melanoma Based on Metastatic Lymph Node Status

BACKGROUND

In cutaneous melanoma (CM), the present methods of lymph node (LN) staging have not sufficiently utilized the prognostic information of metastatic LNs. In this study, we aimed to construct prognostic nomograms based on the number of positive LNs (PLNs) and other clinicopathologic characteristics of CM patients.

METHODS

Two prognostic models were constructed in the none/single PLN (PLN_none/single) and multiple PLN (PLN_multiple) cohorts, respectively. Independent prognostic predictors associated with cancer-specific survival (CSS) in the above two cohorts were integrated to construct two nomograms for predicting the probability of 2-, 4-, and 6-year CSS in the PLN_none/single and PLN_multiple cohorts. The nomograms were evaluated by the area under the receiver operating characteristic curves (AUC), the calibration plots, and the decision curve analyses (DCAs).

RESULTS

A total of 31,065 CM cases were included in this study. Factors included in the prognostic nomogram for patients in the PLN_none/single cohort were age, sex, race, marital status, insurance, primary tumor site, T stage, and number of PLNs, while factors included in the nomogram for cases in the PLN_multiple cohort included age, sex, marital status, insurance, primary tumor site, T stage, and number of PLNs. The AUC values for 2-, 4-, and 6-year CSS in the validation group of the PLN_none/single cohort were 0.833, 0.811, and 0.818, respectively, while in the validation group of the PLN_multiple cohort, the AUC values for 2-, 4,- and 6-year CSS were 0.720, 0.723, and 0.745, respectively. Compared with the American Joint Committee on Cancer 7th edition staging system, our two nomograms showed better predictive values. Additionally, the calibration plots and DCA curves for 2-, 4-, and 6-year CSS prediction demonstrated good coordination and net benefit in both the PLN_none/single and PLN_multiple cohorts.

CONCLUSION

Our nomograms, based on the number of PLNs and other clinicopathologic characteristics, showed good predictive ability for predicting the survival of CM patients.