Page 108 - Predicting survival in patients with spinal bone metastasesL

CHAPTER VII
significantly associated with the cumulative incidence of an adverse event within the studied population.
The levels of agreement found in this study were obtained by calculating Fleiss’ kappa based on the results of four highly experienced observers and ranged from fair to moderate. The results were comparable to those obtained in previous studies6-8. Especially the SINS component bone quality, describing the lesion as being lytic, blastic, or mixed, continues to be a challenge for observers to agree upon, ranging from only 0.21 to 0.306-8.
Apparently, although seemingly objective, measures such as displacement and kyphosis, are still difficult to interpret, even for highly experienced clinicians. Previous research comparing the interobserver agreement between groups of similar specialist (e.g. orthopedic surgeons vs. no orthopedic surgeons and surgeons vs. radiation oncologists), has shown that the clinicians’ background seems to affect their agreement on the total SINS score as well as different components of the SINS8,9. Accordingly, a mixed group of clinicians agrees less than, for example, a group of experienced orthopedic surgeons. It has also been shown that experience affects the degree of agreement. Highly experienced spine surgeons agree more than less experienced spine surgeons8.
The highest levels of sensitivity and specificity found in this study (69% and 48%) differed considerably when compared to the levels found in the study performed by Fourney et al. (96% and 80%)6. This might be attributed to the fact that Fourney et al. used consensus opinion to a priori define a metastatic spine as either stable or unstable, whereas in this study actual clinical endpoints were used to define stable and unstable spines. Nonetheless, a sensitivity of 69% suggests a good capacity to identify patients with a low risk of complication due to spinal instability. In clinical practice, this could translate to fewer surgical consultations for patients who are classified as stable according to the SINS. On the other hand, the rather low specificity found in this study indicates that no treatment decisions should be made based on the classifications potentially unstable and unstable. In daily practice, surgical consultation is still warranted for these categories.
In line with previously reported studies10,11, no predictive value of the total SINS score was observed for the occurrence of an adverse event. Total radiotherapy
106