Five-year survival in spinal metastatic disease, and what it means for the construct you build

Design

retrospective healthcare system registry cohort with external death linkage, multivariable Cox proportional hazards regression; Level III evidence

Population

926 patients presenting with spinal metastatic disease 2017-2025 across two tertiary and two community hospitals; 60% treated surgically, median Charlson index 8 in both groups

Primary outcome

five-year survival stratified by tumour type, and performance of the SORG nomogram and NESMS beyond one year

Effect

five-year survival 8.9% (95% CI 7.2-10.9); SORG HR 0.98 per point (0.98-0.99); NESMS HR 0.61 per point (0.56-0.66)

Spinal metastatic surgery has long been planned around a short survival horizon, with older literature suggesting few patients live long enough for construct durability to matter. This registry study revisits that in the immunotherapy era, drawing on 926 patients presenting with spinal metastatic disease across two tertiary and two community hospitals between 2017 and 2025, 60% of whom had surgery. Vital status was captured through records plus state and federal linkage, and missingness was under 1% for every variable.

Kaplan-Meier five-year survival was 8.9% (95% CI 7.2 to 10.9). Patients with lung primaries did worst at every time point. Both prognostic tools held up beyond the first year: after adjustment for age, sex, race, comorbidity, immunotherapy and treatment, each point of the SORG nomogram carried a 2% lower mortality (HR 0.98, 95% CI 0.98 to 0.99) and each point of the New England Spinal Metastasis Score a 40% reduction (HR 0.61, 95% CI 0.56 to 0.66).

The authors draw two conclusions and they pull in different directions, which is honest. About one patient in ten will still be alive at five years, so a construct that fails at three years is a real problem for a real group of patients — and conversely, nine in ten will not, which argues for minimising surgical morbidity with shorter, simpler operations wherever the goal can be met. What the study genuinely adds is that the scores can be used to tell those groups apart at the point of decision, rather than only for the first year. It is observational and from one health system, and surgical versus non-operative survival should not be compared directly: those groups were chosen, not randomised.

• Use SORG or NESMS to inform the durability the construct actually needs, not just the first-year decision
• Counsel that about one in ten patients survives five years — and that this varies sharply by primary tumour
• Expect the poorest survival in lung primaries at every time point
• Do not read the operative versus non-operative comparison as an effect of surgery; allocation was by indication
• Single health system, Level III evidence — apply the framework, verify the numbers against your own case mix

The statistics, in plain English

A hazard ratio of 0.98 per SORG point looks trivially small until you remember the score runs from 0 to 100 — across its full range that is a large difference, which is why per-point hazard ratios must always be read against the scale. The NESMS ratio of 0.61 per point is far larger per unit because the score runs only 0 to 3. The 8.9% survival interval (7.2 to 10.9) is reasonably tight, but Kaplan-Meier estimates assume that patients lost to follow-up would have fared like those still observed, and 1% here had unknown vital status.

Read the rest in the app

You have read your two free briefings this month. The app carries all 27 specialties, every morning, free — and this finding is waiting in it.

Get Daily Doctor — free

Scan to keep reading on your phone. No account needed to start.