Artificial intelligence (AI) is quickly expanding within the sphere of health care, offering the potential to enhance the efficiency of care delivery, diminish costs, and reduce diagnostic and therapeutic errors. As the field of otolaryngology also explores use of AI technology in patient care, a number of ethical questions warrant attention prior to widespread implementation of AI. This commentary poses many of these ethical questions for consideration by the otolaryngologist specifically, using the 4 pillars of medical ethics—autonomy, beneficence, nonmaleficence, and justice—as a framework and advocating both for the assistive role of AI in health care and for the shared decision‐making, empathic approach to patient care.
Objective: Characterize the sensitivity and specificity of TransImpedance Matrix (TIM) recordings compared with x-rays in detecting cochlear implant tip foldovers. Study Design: Retrospective. Setting: Multi-institutional, academic and private surgery centers. Patients: Patients 6 months and older undergoing cochlear implantation. Interventions: After cochlear implant array insertion, intraoperative TIM recordings and x-rays were conducted to confirm appropriate placement. If a foldover was identified, repeat insertion, followed by TIM and x-ray, was performed. Main Outcome Measures: Presence of tip foldovers as determined by TIM recordings and x-rays. Results: There were 13 tip foldovers out of 484 insertions, yielding a 2.7% tip foldover rate. Using x-rays as the “gold standard,” TIM recordings showed 100% sensitivity, 99.6% specificity, 84.6% positive predictive value (PPV), and 100% negative predicative value in detecting tip foldovers. The x-ray images for two TIM-identified tip foldovers were interpreted as normal intraoperatively (apparent false positives), but then were interpreted as foldovers when presented to the operating surgeon postoperatively in a blinded fashion. If these false positives had been reclassified as true positives, TIM specificity and PPV would both improve to 100%. Conclusions: These results provide further support regarding the utility of TIM recordings to evaluate appropriate cochlear implant electrode array position. Despite the low tip foldover rate reported across the literature and within this study, this multicenter patient group provided a large sample size to calculate the sensitivity and specificity of TIM recordings in identifying tip foldovers. Compared with the reference-standard, x-ray imaging, TIM recordings show equivalent, and in some cases superior, performance in identification of tip foldovers. Thus, use of TIM can potentially preclude the need for routine intraoperative x-rays without compromising patient care.