Predictive Scores Underestimate Survival of Patients With Metastatic Spine Disease

A Retrospective Study of 315 Patients in Sweden

Christian Carrwik, MD; Claes Olerud, MD, PhD; Yohan Robinson, MD, PhD, MBA


Spine. 2020;45(6):414-419. 

In This Article


This is one of the largest cohort studies investigating the accuracy of predictive survival scores for patients with spinal metastatic disease. We found that all scoring systems underestimated rather than overestimated survival.

The aim of surgery for metastatic spine disease is palliation rather than curing the disease. Health-related quality of life (HRQoL) is impaired in patients with spinal cord compression and there is increasing evidence that surgery increases HRQoL more than radiation alone and will benefit patients with survival estimated to at least 6 months.[13–15]

On the contrary, surgery for spinal metastatic disease is far from risk-free and the level of adverse events is high.[16] To justify the risk to the patient, the estimated survival should be long enough to give a positive net when evaluating HRQoL. There is also a need to find patients with long expected survival who could benefit from more extensive and radical surgery. These demands have driven the development of the scoring systems evaluated in this study.

Most predictive scores for spinal metastatic disease rely on retrospective clinical studies.[5,6,17] The main strengths of this study are the large cohort, the detailed data on each patient, and reliable survival data. By evaluating the predicted survival based on originally registered data, over- or underestimation were assessed with regard to the effect on treatment decision.

Our study is mainly retrospective and there is no control group with patients not undergoing surgery. The heterogeneity among the patients is large with several types of cancer, each with different outcome and prognosis. However, we believe that this cohort reflects the case mix at a regional hospital in northern Europe.

The definition of overestimation and underestimation of survival in this study might look arbitrary at a first glance, but is made from the clinician's point of view. Patients expected to live shorter than 3 months are often considered as inappropriate candidates for extensive spine surgery.[14] On the contrary, patients with estimated survival of longer than 6 months would benefit even from extensive surgery and should not be disqualified on the merits of an old scoring system.

It is no surprise that the scoring systems evaluated here underestimate rather than overestimate survival. The scores are developed on the basis of patient cohorts some of which were undergoing treatment for more than 25 years ago.[1,3,4] Although there have been significant advancements in spine surgery since then, the progress in the oncological field is even more extensive with new treatments and increased survival.[18]

Our findings are in line with a recent prospective multicenter study that evaluated six different scores among 1469 patients. The authors of that study conclude that no prognostic scoring system system had a good predictive value and suggested other methods of estimating survival.[19]

A step forward from the rigid point-based scoring systems is the NOMS (Neurologic-Oncologic-Mechanical-Systemic) framework,[20] where framework, where neurologic, oncologic, mechanical, and systemic considerations are taken before treatment decision. The NOMS framework divides the tumor types into "Radiosensitive/Radioresistant" without further specification, which makes the system more adaptable to oncological advancements.[20]

Future decision support will likely be based on real time data where multiple factors for survival and favorable outcome are analyzed by artificial intelligence routines to deliver an estimation of surgical results. This implies an ethical dilemma as it challenges the role of the physician.

Until then online tools in which the treating physician enters known parameters of the patients and gets a survival estimate based on similar cases in return could be used. These systems adapt quicker to oncological advancements and do not lag behind as the traditional scorings systems reviewed here.[21,22]

Based on our and other studies with similar conclusions, we believe that today's analogue scoring systems will have a very limited place in the future of surgical decision making. Further research efforts should focus on improving the online-based scoring systems and making them easily available.