Predicting Hearing Loss for Cancer Patients Before Treatment

Abstract

Most of us will have a loved one hear the devastating words, “You have cancer.” As he or she processes this news, quick treatment decisions must be made. Cancer treatment side effects, such as hearing loss, often come with uncertainties. Our team sought to provide clarity to patients with head and neck cancer on their expected hearing loss before treatment.shutterstock/Iurii Motov, Health care, hearing loss, cancer.Figure 1: The standard model (post-treatment data) is as accurate as the predictive model (pretreatment data). Health care, hearing loss, cancer.Figure 2: Screenshot of the web-based app. Health care, hearing loss, cancer.Each year, more than half a million people develop head and neck cancer, and more than a quarter million die from it.1 Treatment options include surgery, radiation therapy, and chemotherapy (cisplatin). These therapies can cause permanent hearing loss and other side effects, including kidney damage (nephropathy), extreme dry mouth (xerostomia), and loss of healthy tissue (radiation necrosis).2-8 Despite the negative impact of hearing loss on patients’ quality of life, it can be overlooked.9 In 2014, our division of adult audiology at Washington University School of Medicine in St. Louis began providing care for head and neck cancer patients in collaboration with otolaryngology physicians. One of our team members attended a weekly multidisciplinary tumor board to better understand a patient's treatment journey and their challenges in streamlining the process for our patients. We were able to provide audiological care within existing structures to avoid disruptions by reaching out to our multidisciplinary partners. Working with the radiation oncology department, we built standard referral patterns to increase patient follow-up while reducing patient burden through same-day appointments. Baseline audiological testing was incorporated into our department's Head & Neck Cancer Multi-Disciplinary Clinic launched in 2016. The clinic allowed patients to see multiple medical professionals—a radiation oncologist, a medical oncologist, a head and neck cancer surgeon, a speech therapist, and an audiologist—all on the same day. This was appreciated by patients since they could complete their pre-treatment cancer appointments in one morning instead of a month. At these appointments, many patients often asked about what to expect of their hearing after treatment, which made us realize a gap: We could only provide general information about the effects of ototoxic medications, but couldn't address our patients’ needs. As a result, our team decided to explore options for how we could predict post-treatment hearing loss before treatment to provide better counseling. PREDICTING HEARING LOSS A literature review only revealed options that could predict hearing loss after treatment because they relied on final treatment doses. After identifying this gap in information, our team decided to create a new resource for patients. We studied 242 patients using an internal database and data from Siteman Cancer Center at Barnes-Jewish Hospital and Washington University School of Medicine in St. Louis, the Washington University Oncology Data Services Cancer Registry, Varian (a radiation oncology treatment planning software), and AudBase (a clinical audiometric database). We collected demographic information, tobacco and alcohol use, tumor characteristics (stage, histology, and proximity to the cochlea), and treatment methods (surgery, surgery with clean margins, cisplatin treatment regimen, total cumulative cisplatin dose, and mean cochlear radiotherapy dose) for each patient. Next, we analyzed the data to determine if they could be used to predict post-treatment hearing (pure-tone average [PTA] of 1, 2, 4 kHz) and help counsel patients on their hearing before treatment. We built two key models: one that used proxy measures available before treatment (location of the tumor and planned chemotherapy treatment regimen) and another that relied on final treatment doses. The results were identical: Using pre-treatment measures was as accurate as using post-treatment measures (Fig. 1). The pretreatment factors that we found predicted hearing loss were age, pre-treatment hearing, location of the tumor, and planned treatment chemotherapy dose. Using these factors, our model explained 80 percent of a patient's post-treatment hearing (1, 2, 4 kHz PTA), had a sensitivity of 80 percent and a specificity of 75 percent for predicting post-treatment hearing loss greater or less than 35 dB HL. To our knowledge, we created the first model capable of accurately predicting post-treatment hearing loss before treatment. Our team converted the model into a user-friendly internet-based tool. To use the tool, an audiologist inputs the predictive factors into a web-based application at the baseline hearing test (Fig. 2). The audiologist can then use this information to counsel his or her patient on realistic expectations regarding the patient's post-treatment hearing; however, the model is not yet ready for widespread implementation. We are working to validate it in a new cohort of patients with head and neck cancer.10 Once completed, we plan to publish the data and share our findings to support patients across the country. Each patient we see is someone's loved one. We hope our tool makes their journey a little easier.