Predictive Analytics Determines Throat Cancer Outcomes

August 12, 2021 - University of Michigan researchers are developing a model that uses predictive analytics to determine an individual’s survival and recurrence outcomes for patients recently diagnosed with oropharyngeal cancer.

With increasing rates of human papillomavirus (HPV)-associated disease, researchers’ understanding of biologic characteristics and treatment of oropharyngeal squamous cell carcinoma (OPSCC) has developed. Currently, there are multiple studies underway to evaluate treatment for patients with HPV-positive OPSCC. However, data suggests that patterns of recurrence differ depending on HPV status.

Individual prognostic calculators are a critical tool in evaluating potential risks. Unlike previous staging systems that place patients in broad risk categories, prognostic calculators use predictive analytics to determine cancer recurrence or death by examining individual characteristics.

To quickly determine and diagnosis OPSCC, researchers created a prognostic calculator to predict survival outcomes by evaluating individual characteristics and biomarkers.

“We hypothesized that this multistate model-based prognostic calculator combining clinical, oncologic, and imaging data can provide clinicians and patients with robust, individualized predictions that may facilitate cancer treatment decision making and counseling,” the study researchers wrote.

Researchers tested the model in a prognostic study using a data set comprised of 840 patients with newly diagnosed oropharyngeal cancer. Patients received treatment at a National Cancer Institute-designated center between January 2003 and August 2016.

With the data, researchers created a Bayesian multistate model to generate individualized predictions of patient’s survival and recurrence outcomes. The prognostic performance of the model was validated with data from 447 patients treated for oropharyngeal cancer at Erasmus Medical Center in the Netherlands.

Of the 840 patients, 715 were men and 268 were current smokers. The Erasmus Medical Center group was made up of 300 men and 350 current smokers.  

“Model predictions for 5-year overall survival demonstrated good discrimination, with area under the curve values of 0.81 for the model with and 0.78 for the model without imaging variables. Application of the model without imaging data in the independent Dutch validation cohort resulted in an area under the curve of 0.75,” the study stated.

Researchers concluded that the prognostic study appeared to estimate and discriminate locoregional recurrence from distant metastases. The model’s ability to provide personalized predictions will serve as an important tool for physicians, especially when making decisions regarding treatment options.

“In this study, we developed a prognostic calculator informed by clinical, oncologic, and imaging data, resulting in good discrimination for overall survival in training and validation cohorts. Interaction with our web application may aid in communicating prognosis to patients and informing medical decisions for practitioners,” the researchers concluded.