Emergency neoplastic spinal cord compression without known histology: should we operate?

Patients presenting emergently with neoplastic spinal cord compression without known histology pose a complex decision for spine surgeons. In this patient population, the authors sought to 1) describe presentation/histology, 2) determine if emergency surgery was indicated, and 3) compare survival to that of patients with known histology prior to surgery.

A retrospective, cohort study of patients who underwent extradural spine tumor surgery between 2010 and 2021 was undertaken. The inclusion criteria were neoplastic spinal cord compression and emergency department or emergency clinical presentation without known tumor histology or cancer history. Histology was classified as radiosensitive (lymphoma, leukemia, small cell lung cancer, germ cell cancer), radio-intermediate (breast, prostate), and radioresistant (all others). Stability was determined with the Spine Instability Neoplastic Score (SINS). Descriptive/bivariate and multivariable Cox regression statistics were performed.

Of 371 patients who underwent surgery for extradural spine tumors, 80 (21.6%) had emergency presentation without a known histological diagnosis. The mean ± SD age was 60.5 ± 12.4 years, 65% of patients were male, and the mean follow-up was 530.5 ± 738.0 days. Although spinal cord compression was seen in all (100%) patients, a neurological deficit was seen in only 43 (53.7%). Final pathology revealed that 59 (73.7%) patients had radioresistant tumors, 12 (15.0%) radio-intermediate, and 9 (11.3%) radiosensitive; of note, 2/80 (2.5%) had primary bone tumors. Accepting that those patients with radioresistant pathology would need separation surgery regardless, 11 of the 21 (52.4%) patients with radio-intermediate or radiosensitive pathology had a neurological deficit, and thus surgery was indicated. In the remaining 10 patients without a neurological deficit, SINS was stable/indeterminate in 8 patients (10% of patients with unknown diagnosis), for whom surgery may have been avoided. Multivariable Cox regression showed that patients without known histology at the time of surgery had decreased overall survival compared to those with known histology preoperatively (HR 1.64, 95% CI 1.17-2.30, p = 0.004).

Among 80 patients with emergency presentation of neoplastic spinal cord compression without known histology, 90% ultimately had histology, neurological status, or instability favorable for surgery, thus affirming the surgeons' choice to operate. However, 1 in 10 patients could have potentially been treated with radiation and/or systemic treatment alone. Two patients (3%) had primary bone tumors, foregoing the chance of a negative margin resection. Not knowing histology preoperatively was independently associated with decreased overall survival. These results provide empirical data for a challenging yet not uncommon situation for spine surgeons and reinforce the importance of knowing histology prior to spine tumor surgery when safe and feasible.