Volume 69, Issue 3, Supplement , Page S92, 1 November 2007
Partial-Breast Irradiation Versus Whole-Breast Irradiation for Early-Stage Breast Cancer: A Decision Analysis
Article Outline
Purpose/Objective(s)
To compare the quality-adjusted life expectancy (QALE) between women treated with whole-breast radiation therapy (WBRT) and partial-breast irradiation (PBI) for early-stage breast cancer.
Materials/Method
We developed a Markov model to describe health states in the 15 years following radiotherapy (RT) for early-stage breast cancer. After treatment, patients entered the NED-PBI or NED-WBRT health state, from which they could either remain well or transition to ipsilateral breast tumor recurrence (IBTR), distant metastases (DM), or death. Breast recurrences were separated into local recurrences (LR) and elsewhere failures (EF), and all IBTRs were salvaged with mastectomy. IBTR risk was extracted from the Oxford overview, and DM rates were taken from NSABP B-14 and calibrated for consistency with current survival data. Utilities were adapted from the literature; the baseline NED-WBRT utility was set to 0.92, and the NED-PBI utility for the first 5 years after PBI was adjusted in sensitivity analyses. The NED-PBI utility was set equal to NED-WBRT for years 6–15. We studied two cohorts of women (one aged 40 years and one aged 55 years) and evaluated the influence of estrogen-receptor (ER) status. ER-positive patients received adjuvant tamoxifen. We assumed patients did not receive chemotherapy.
Results
Assuming a NED-PBI utility of 0.93, QALE after PBI (and WBRT) was 12.61 (12.57) and 12.10 (12.06) QALYs for 40-year-old and 55-year-old women, respectively. The NED-PBI utility thresholds for preferring PBI over WBRT were 0.923 and 0.921 for 40 year-old and 55 year-old ER-positive women, respectively, both slightly greater than the NED-WBRT utility. For ER-negative tumors, the respective thresholds were 0.926 and 0.923. Model outcomes were sensitive to the utility of NED-PBI, the PBI hazard ratio (HR) for local recurrence, and the fraction of IBTRs that were true local failures. Overall, the degree of superiority of PBI over WBRT was greater for 55-year-old women than for 40-year-old women, given younger women's increased risk for IBTR. Three-way sensitivity analysis (Fig.) revealed that the optimal strategy is dependent upon age, PBI HR, and the percentage of IBTRs that are local.
Conclusions
For most utility values of the NED-PBI health state, PBI is the preferred treatment modality. This result is highly sensitive to patient preferences and is also dependent upon patient age, PBI efficacy and the fraction of IBTRs that are local.
Author Disclosure: D.J. Sher, None; E. Wittenberg, None; R. Punglia, None.
PII: S0360-3016(07)01350-8
doi:10.1016/j.ijrobp.2007.07.168
© 2007 Elsevier Inc. All rights reserved.
Volume 69, Issue 3, Supplement , Page S92, 1 November 2007

