Background: Regorafenib was approved by the FDA in 2012 for the management of previously treated metastatic colorectal cancer (mCRC). It is now the standard of care in the third-line setting. Compared to placebo it improves median overall survival by 1.4 months but is associated with adverse effects and additional cost. The objective of this study was to examine the cost-effectiveness of regorafenib compared to best supportive care for patients receiving 3rd-line treatment for mCRC. Methods: We developed a Markov model to compare the cost and effectiveness of regorafenib compared to best supportive care in the 3rd-line treatment of mCRC based on randomized data from the CORRECT trial. Weibull models were fitted to the published overall and progression-free survival curves, and were used to extrapolate the cause-specific mortality and progression risks. Costs for administration and management of adverse events were based on Medicare reimbursement rates for hospital and physician services, and drug costs based on the Medicare average wholesale prices (all in 2014 US $). Health outcomes were measured in life years (LYs) and quality-adjusted life years (QALYs). Quality of life adjustments were calculated based on health utility values in the CORRECT trial and toxicity disutilities and durations were included for the most common toxicities: hand/foot syndrome, diarrhea, and hypertension. Model robustness was addressed by univariate and probabilistic sensitivity analyses (PSA). Results: In the model, regorafenib provided an additional 0.04 QALYs (0.13 LYs) at a cost of $39,391. The incremental cost-effectiveness ratio (ICER) was $897,411/QALY. In all one-way sensitivity analyses, the ICER of regorafenib was > $700,000/QALY. The ICER of regorafenib was greater than $200,000/QALY in > 99% of PSAs. Conclusions: This is the first US-based cost-effectiveness analysis of regorafenib in mCRC and our findings show that regorafenib provides minimal incremental benefit at high incremental cost per QALY. The ICER of regorafenib could be improved by use of an effective biomarker to select patients most likely to benefit, or by a lower price for payers.