Evaluation of Prognosis in Nasopharyngeal Cancer Using Machine Learning

Abstract

Background and Aim: Although the prognosis of nasopharyngeal cancer largely depends on a classification based on the tumor-lymph node metastasis staging system, patients at the same stage may have different clinical outcomes. This study aimed to evaluate the survival prognosis of nasopharyngeal cancer using machine learning. Settings and Design: Original, retrospective. Materials and Methods: A total of 72 patients with a diagnosis of nasopharyngeal cancer who received radiotherapy ± chemotherapy were included in the study. The contribution of patient, tumor, and treatment characteristics to the survival prognosis was evaluated by machine learning using the following techniques: logistic regression, artificial neural network, XGBoost, support-vector clustering, random forest, and Gaussian Naive Bayes. Results: In the analysis of the data set, correlation analysis, and binary logistic regression analyses were applied. Of the 18 independent variables, 10 were found to be effective in predicting nasopharyngeal cancer-related mortality: age, weight loss, initial neutrophil/lymphocyte ratio, initial lactate dehydrogenase, initial hemoglobin, radiotherapy duration, tumor diameter, number of concurrent chemotherapy cycles, and T and N stages. Gaussian Naive Bayes was determined as the best algorithm to evaluate the prognosis of machine learning techniques (accuracy rate: 88%, area under the curve score: 0.91, confidence interval: 0.68-1, sensitivity: 75%, specificity: 100%). Conclusion: Many factors affect prognosis in cancer, and machine learning algorithms can be used to determine which factors have a greater effect on survival prognosis, which then allows further research into these factors. In the current study, Gaussian Naive Bayes was identified as the best algorithm for the evaluation of prognosis of nasopharyngeal cancer.

Progression-Free Survival Prediction in Patients with Nasopharyngeal Carcinoma after Intensity-Modulated Radiotherapy: Machine Learning vs. Traditional Statistics

“…Right-censored data are quite common in cancer survival analysis, which means that follow-up ends before subjects experience a specific event, such as disease progression. As mentioned above, there are several studies that applied ML techniques for cancer prognosis prediction [ 16 , 17 , 18 , 19 , 20 ]. However, many ML approaches have an assumption that all patient outcomes are known (disease progression or no disease progression).…”

“…Furthermore, we also compared the performances of the three models using the log-rank test by stratifying the testing set into high- and low-risk groups. Many similar studies in the past only adopted conventional metrics, such as accuracy and AUROC [ 16 , 17 , 18 , 19 , 20 ], which are not suitable for censored data. Moreover, to the best of our knowledge, this study is the first one to compare CPH, CSF, and DeepSurv using three different evaluation metrics in NPC.…”

“…Recently, ML has been applied for prognosis prediction in some types of cancer [ 9 , 10 , 11 , 12 , 13 , 14 , 15 ]. Notably, series of study applying ML for survival prediction of NPC have reported good performance [ 16 , 17 , 18 , 19 , 20 ]. Du et al [ 18 ] built a ML model to predict 3-year disease progression based on 525 radiomics features extracted from magnetic resonance imaging (MRI) scan images and five clinical features.…”

Severity Grade Recognition for Nasal Cavity Tumours Using D閏or CNN

“…The overall performance of the DenseNet model was noted with the average sensitivity of 95%. In [10] Melek akcay et al, proposed an automated machine learning model to evaluate the prognosis of nasopharyngeal cancer. The study was conducted with 72 patients who are diagnosed with nasopharynx cancer and undergone chemotherapy with proper follow ups.…”

Emerging landscape of circHIPK3 and its role in cancer and other diseases (Review)

“…circHIPK3 and nasopharyngeal carcinoma (NPC). NPC is a common malignant tumor in humans, which occurs in the head region (74). In 2018, Ke et al (75) revealed that circHIPK3 expression was higher in NPC tissues and cell lines compared with in normal tissues and cells.…”