A nomogram for predicting survival in patients with upper limb melanoma: A study based on the Surveillance, Epidemiology, and End Results database

Abstract

Abstract Introduction: In order to predict cancer-specific survival (CSS) in patients with upper limb melanoma (ULM) at three-, five-, and eight-year intervals following diagnosis, a nomogram was developed and validated in this study. Methods: Patient data about the upper limb melanoma (ULM) cases were collected from the comprehensive Surveillance, Epidemiology, and End Results (SEER) database. A training cohort consisting of 70% of the patient population and a validation cohort comprising 30% of the patients were established through a random selection process. A stepwise backward Cox regression model was employed to identify independent prognostic variables. The nomogram was then updated to integrate these factors, making it possible to estimate the rates of cancer-specific survival (CSS) after 3, 5 and 8 years after diagnosis. A number of metrics were used to assess the nomogram's performance, including the calibration curves, decision-curve analysis (DCA), net reclassification improvement (NRI), concordance index (C-index), integrated discrimination improvement (IDI), and the area under the time-dependent receiver operating characteristic curve (AUC). Results: This study involved a comprehensive cohort comprising 36,621 patients diagnosed with upper limb melanoma (ULM). Through an analysis of the Cox regression model within the training cohort, a total of 13 prognostic factors were identified, namely age, RNP (regional node positive), sex, race, marriage, AJCC (American Joint Committee on Cancer) stage, surgical status, radiation status, chemotherapy status, income status, survival time and current status. An extensive set of internal and external validation processes were then applied to the development of a nomogram. The nomogram demonstrated excellent discriminatory abilities, as reflected by significantly high C-index and AUC values. Calibration curves provided further confirmation of the nomogram's reliability. Notably, the nomogram exhibited superior performance compared to the AJCC model, as evidenced by improved NRI and IDI values. The decision-curve analysis (DCA) curves further validated the clinical utility of the nomogram, underscoring its practical relevance in prognostication for patients with ULM. Conclusions: The current investigation has effectively developed and validated an initial nomogram for prognosticating outcomes in patients with upper limb melanoma (ULM). The nomogram's impressive performance and practical applicability highlight its potential usefulness within clinical settings. However, it is important to note that additional external validation is necessary to further substantiate its reliability and generalizability.