Page 79 - Predicting survival in patients with spinal bone metastasesL
P. 79

                                Even though in general patients with symptomatic SBM from breast cancer have a relatively long survival, the limited availability of systemic treatment options in patients with a triple negative phenotype results in a shorter survival. Predictive models based on previous studies overestimated survival in these patients, leading to overtreatment. The effect of correcting for molecular phenotypes was confirmed when a previously published survival prediction model was adjusted for the current findings, resulting in an improvement of the accuracy (table 3 and figure 2). A precise model can be a useful tool for clinicians. Extensive surgical treatment and long-course radiotherapy regimens could be avoided in patients with a short expected survival and in patients with a longer expected survival extent of treatment could be increased, if needed.
The most important limitation of this study is its retrospective design. However,
since the endpoint was death and the only clinical information gathered was
derived from pathology reports, the design is not expected to have a large impact
on the quality of our results. Also, receptor status was assessed in different V laboratories and by different pathologists. HER2 immunohistochemistry can be
difficult to interpret, however, in all three institutes equivocal cases were subjected to molecular testing using in situ hybridization. The distribution of patients over the four categories was slightly different when compared to larger studies16,17. The luminal B category was underrepresented and the luminal A category was overrepresented in this study. However, since the luminal B and HER2 categories contained few patients and were joined with the luminal A category the overall effect on the survival data was negligible. This means that a statement on the exact survival times in the luminal B and HER2 categories cannot be made based on small sample size. From a clinical perspective however, we believe it is more relevant to identify patients with a short survival in order to avoid overtreatment, than to give an exact estimation of survival for patients that have an expected survival of more than a year.
A previously published study on SBM from breast cancer investigated the role of ER and PR status on survival18 and found that patients negative for ER had a decreased survival when compared to patients with a positive ER status. As this study was conducted before the onset of HER2 testing, the effects of this receptor were not considered. A study conducted by Rades et al.19 examined the effects of several other risk factors, such as performance status and the presence of visceral
BREAST CANCER
 77


























































































   77   78   79   80   81