Aims: Surveillance imaging guidelines for head and neck cancer lack consensus on imaging frequency. Excess imaging increases medical costs, patient time burden, and anxiety. We evaluated the cost‑effectiveness of a NI‑RADS–tailored surveillance imaging strategy compared with routine practice and quantified the value of additional research to reduce decision uncertainty. Methods: A decision tree classified patients into NI‑RADS categories, informing a state‑transition Markov model simulating costs and quality‑adjusted life years (QALYs) over five years. Health states included disease‑free, recurrence, re‑treatment, death. Transition probabilities and utilities were derived from published U.S. studies. Direct medical costs were estimated with Medicare and Healthcare Cost and Utilization Project data; indirect costs were estimated with Bureau of Labor Statistics data. Analyses were conducted from healthcare sector and societal perspectives with 3% annual discount rate. Uncertainty was assessed with one‑way and probabilistic sensitivity analyses(PSA), and value‑of‑information analysis. Results: NI-RADS strategy generated 2835 QALYs and cost 8.25 million(healthcare sector) and 1.30 billion USD(societal). NI-RADS strategy dominated routine practice in both perspectives with an incremental cost-effectiveness ratio(ICER) of 291,000 USD/QALY and 685,000 USD/QALY respectively. In PSA, NI-RADS strategy was dominant in 77.9% of simulations. Results were most sensitive to recurrence transition probabilities, particularly amongst patients with NI-RADS1 results. The population expected value of perfect information at a willingness-to-pay threshold of 100,000USD/QALY was 8.23 billion USD(healthcare sector) and 9.55 million USD(societal). Conclusions: NI‑RADS–tailored surveillance imaging is potentially cost‑saving compared with routine practice. Additional research to better characterize recurrence risk, especially among NI‑RADS 1 patients, may further reduce decision uncertainty.