Purpose: Magnetic resonance lymphangiography (MRL) is a recently developed, non-invasive, high-resolution imaging technique that enables detection of malignancy in lymph nodes. Our aim was to determine the cost-effectiveness of MRL alone, and MRL in combination with sentinel lymph node biopsy (SLNB) relative to SLNB alone for axillary staging in patients with operable, early breast cancer.

Methods: We developed a decision-analytic Markov model to simulate outcomes for a hypothetical cohort of 60-year-old women with operable, clinically node-negative early breast cancer (1-2-cm, estrogen-receptor positive). In the two single-test strategies examined (MRL alone, SLNB alone), positive test results were followed by axillary dissection while negative results ended axillary work-up. In the combined MRL-SLNB strategy, positive MRL results were followed by axillary dissection while negative results were followed by SLNB. Markov models implementing yearly transition rates were used to estimate quality-adjusted life expectancy and lifetime costs associated with each possible final test result for axillary node status. Patients with positive test results received an anthracycline-based chemotherapy regimen (6 cycles) and tamoxifen (5 years), while patients with negative test results received tamoxifen only. Transition rates to advanced cancer states were determined by true node status, test performance, and consequent choice of chemotherapy course. Sensitivity analyses were performed to evaluate the impact of key model assumptions and input parameters on results.

Results: Quality-adjusted life expectancy (discounted) was similar for all diagnostic strategies: MRL alone, 14.54 life years (LY); SNLB alone, 14.57 LY; and combined MRL-SLNB, 14.58 LY. The cost (discounted) of the single-test MR strategy was the lowest ($46,950), followed by SNLB ($48,730) and MRL-SLNB ($49,260). In an incremental cost-effectiveness analysis, the single-test SLNB strategy was eliminated through extended dominance; the incremental cost-effectiveness ratio for the combined MRL-SLNB strategy compared with MRL alone was $40,730/QALY. Elimination of SLNB by extended dominance was sensitive to MRL and MRL-SLNB test performance, in particular. Strategy rankings were otherwise robust through a range of sensitivity analyses.

Conclusions: MRL-based diagnostic strategies may be more cost-effective than a single-test SLNB strategy in the axillary work-up of patients with operable, early breast cancer. However, further studies of MRL test performance are necessary prior to routine clinical implementation of MRL-based diagnostic strategies.