Clustering of Largely Right-Censored Oropharyngeal Head and Neck Cancer Patients for Discriminative Groupings to Improve Outcome Prediction

Abstract

clustering is the task of identifying groups of similar subjects according to certain criteria. the AJcc staging system can be thought as a clustering mechanism that groups patients based on their disease stage. This grouping drives prognosis and influences treatment. The goal of this work is to evaluate the efficacy of machine learning algorithms to cluster the patients into discriminative groups to improve prognosis for overall survival (oS) and relapse free survival (RfS) outcomes. We apply clustering over a retrospectively collected data from 644 head and neck cancer patients including both clinical and radiomic features. in order to incorporate outcome information into the clustering process and deal with the large proportion of censored samples, the feature space was scaled using the regression coefficients fitted using a proxy dependent variable, martingale residuals, instead of follow-up time. Two clusters were identified and evaluated using cross validation. The Kaplan Meier (KM) curves between the two clusters differ significantly for OS and RFS (p-value < 0.0001). Moreover, there was a relative predictive improvement when using the cluster label in addition to the clinical features compared to using only clinical features where AUC increased by 5.7% and 13.0% for OS and RFS, respectively. Every year over 50,000 new cases of head and neck cancers are diagnosed in the United States. This number is projected to rise in the future, especially for oropharyngeal cancers, recently been associated with the incidence of HPV16 genotype infections 1. The American Joint Committee on Cancer (AJCC) and the Union for International Cancer Control, maintains an internationally used standardized TNM Staging System. This system serves as a way to systematically assess the severity of the cancer on individual subjects 2. The vast majority of risk stratification of head neck cancer patients uses staging systems that sub classify patients into four or less groups, based primarily on committee derived treatment standards and approaches using existing data sets. These consider physical examinations, imaging and laboratory tests, pathology and surgical reports, etc. Establishing the AJCC stage for a patient considers various important anatomic classifications and other risk factors that contribute to the overall assessment such as T, N and M Categories. T Category relates to the extent of the primary tumor, N Category relates to the spread to lymph nodes, and M Category indicates the spread outside the T and N related areas. These classifications play a critical role in the ultimate diagnosis and prognosis. The ability to more accurately assess the underlying condition such that it improves the prediction on various outcomes is a long-standing clinical goal. In the era of personalized cancer medicine, innovative sources of meaningful data are critically needed. For head and neck cancer, radiomics is one such "big data" approach that applies advanced image refining/data characterization algorithms to generate imaging features t...

Utilizing Artificial Intelligence for Head and Neck Cancer Outcomes Prediction From Imaging

“…The AUCs of both survival variables increased in the models containing both the cluster labels and clinical features. 35 The 7 studies predicting OS had good accuracies, with most models achieving an AUC or c-index above 0.7. 10,[36][37][38][39][40] Most models achieved the highest AUC when evaluating on a combination model, encompassing both clinical and radiomic features.…”

“…Radiomic signatures have demonstrated promise for predicting tumor characteristics associated with OS in multiple cohorts of patients with HNSCC. 10,[35][36][37][38][39][40][41] Chen et al 36 Most constructed models have performed better in the training cohort than the validation cohort, suggesting that overfitting may be occurring. Furthermore, most studies have used data sets from a single institution, which may poorly represent the larger patient population.…”

“…This approach simplifies the ethical challenges of data sharing between hospitals, when performing validation studies using an external data set. 37 Tosado et al 35 developed a model to predict OS and PFS in a CT data set of patients with oropharyngeal cancer. A baseline model comprising clinical information based on patient demographics was established and compared to models comprising staging features, radiomic features, cluster labels, and combinations of these features.…”

“…Many studies attempt to minimize bias in their models by validating on independent data sets, 10,[36][37][38][39][40][42][43][44][45] but others rely on cross-validation. 35,46,47 Future efforts should target larger, prospective studies to better understand how machine learning models can predict patient survival.…”

Oropharyngeal Cancer Patient Stratification Using Random Forest Based-learning Over High-dimensional Radiomic Features

“…Nevertheless, as can be seen in the model evaluation, the addition of the RFS clusters to other predictive clinical covariates including N-staging, HPV status, and Therapeutic combination improves model performance for both training and testing. Prior work has also effectively leveraged clustering to improve outcome prediction for OPC patients [39][40][41]49], however, none of these works have attempted to use the entire set of radiomic features or Random Survival Forest learning as we have done in this work.…”