br DISCUSSION br A justifiable focus of
A justifiable focus of patients with cancer and their treating physicians is mortality. For cancers such as stage I breast cancer, in which the vast majority of patients are cured, quality of life and risks vs benefits of treatment recommen-dations are becoming more relevant than an exclusive focus on mortality. Effective decision making in these patients requires tradeoff calculations by the physician and buy-in from the patient. To better inform such discussions, we quantify competing risks of death in elderly patients with breast cancer. In our study, we show that Mitomycin C age, race, comor-bidity grouping, and ER status of the patient have signifi-cant influence on the probability of death from breast cancer, other cancers, and non-cancer causes for any given patient. We believe these data are important because breast cancer and the impact on a patient’s life does not occur in a vacuum, but in the context of age, race, and other comor-bidity. Traditional cancer survival statistics often do not ac-count for these competing risks of death.
Similar to our study, Mell and colleagues12 used institu-tional data to show that competing mortality in patients with stage I and II breast cancer is associated with increasing age, black race, and comorbid disease. The 10-year cumulative incidence of competing mortality was 7.2% in the low risk vs 30.6% in the high-risk group
34 Wasif et al Competing Risk of Death in Breast Cancer J Am Coll Surg
Table 1. Five- and 8-Year Probability of Death
Breast cancer Other cancer
(p < 0.001). With regard to comorbid conditions, Bayliss and colleagues13 looked at patients with cancer and multi-morbidity to show that the influence of cancer prognosis was greatest in year 1, and the effect of comorbidities increased long term, especially in patients with good prog-nosis cancers. In an older study using just SEER data, Schairer and colleagues14 showed that non-cancer mortal-ity exceeds breast cancer mortality for patients older than 50 years with localized breast cancer. An important impli-cation of these data are that the 5-year overall survival sta-tistic used as a “benchmark” for reporting prognosis in cancer can be misleading, especially in older and sicker patients.15 For example, Spheroplast statistic can be improved, not due to any underlying improvements in cancer care but rather better treatment of comorbid conditions.
Studies looking at other cancer types report similar re-sults. In particular, prostate cancer is the archetype of a cancer with low long-term mortality and treatment op-tions with the potential to adversely impact quality of life. In a study by Daskivivh and colleagues,16 older men with low- or intermediate-risk prostate cancer with major comorbidity were at high risk for other-cause mor-tality within 10 years of diagnosis. The authors suggest that these patients and their treating physicians take this information into consideration when deciding between conservative management and aggressive treatment. For localized renal carcinoma, a nomogram was developed incorporating commonly available clinical information to help make tradeoff calculations about treatment.17 Another approach was to use competing mortality to r> Table 2. Adjusted Cause Specific Hazard Ratios
HR, hazard ratio. *Referent.
risk stratify patients with head and neck cancers into cat-egories most likely to benefit from treatment intensifica-tion.18 Finally, for patients with early-stage endometrial cancer an increasing competing mortality risk score was associated with a diminishing likelihood of benefit from treatment intensification.19