Purpose: To integrate pathological information about margin status into the decision of treatment strategies for ductal carcinoma in situ (DCIS) of breast and to determine whether all patients undergoing breast conservation surgery (BCS) for DCIS should receive radiation therapy.
Method: We developed a decision-analytic Markov model to project outcomes for a hypothetical cohort of 55-year-old women with DCIS over a 20-year time horizon. We evaluated three treatment options: BCS alone, BCS plus radiation (RT), and mastectomy. We conducted a systematic literature review to estimate transition probabilities for local DCIS and invasive recurrences based on the margin status (free, close, or positive). Other probability estimates and utilities were collected from published literature. We also performed one-way and probabilistic sensitivity analyses.
Result: Using the conditions defined in this model, mastectomy yielded the greatest benefit regardless the margin status with expected 12.20 quality-adjusted life years (QALYs) over 20 years. However, the beneficial effect was small, especially in the free-margin group. Moreover, expected QALY after BCS alone was better than that after BCS plus RT under the free-margin scenario (12.12 vs. 12.10), and worse in the close margin (11.97 vs. 12.05) and positive margin patients (11.82 vs. 11.94). One way-sensitivity analysis showed that the optimal treatment was very sensitive to patients’ preferences and the side effects of RT. Probabilistic sensitivity analyses revealed that mastectomy would be the best strategy in 43.6% cases, BCS alone 29.6%, and BCS plus RT 26.8% under the free-margin scenario, whereas mastectomy would be the best strategy in 56.5% cases, BCS alone 17%, and BCS plus RT 26.5% in positive-margin scenario.
Conclusion: Although we found that the optimal treatment for women with DCIS is mastectomy, regardless of the pathological information of margin status, this finding was sensitive to patient preferences. Our analyses highlight the importance of patient preferences on the decision-making process. This study also illustrates that margin status is able to provide supplementary information on the decision of DCIS treatment. Our findings suggest that RT is not necessary for all patients with DCIS undergoing BCS.
Candidate for the Lee B. Lusted Student Prize Competition