Criteria for surgical intervention in patients with malignant cord compression
JC McKinley, JNA Gibson, PA Levack, JC Grahham, PFX Statham The Scottish Spinal Cord Compression Audit Office, Department of Neuro-oncology, the Western General Hospital, Edinburgh
There is still a wide diversity of opinion between clinicians about the relative value of different treatments in the management of malignant spinal cord compression. We have sought to determine whether a patient's true diagnosis, for a tumour of a given type, bears any relation to published prognostic criteria, and which factors are most likely to be predictive of a satisfactory and cost effective outcome from surgery.
During a nine month period, prospective data was collected from all patients with malignant cord or cauda equina compression presenting to hospital in Edinburgh, Glasgow and Aberdeen. The data from each patient was analysed for six key parameters; site of the primary tumour, neurological status, clinical performance (Karnofsky) index, vertebral and extraspinal bone metastases, and internal organ spread. From this we computed the Tokuhashi score. Excisional surgery has been considered beneficial if the score is greater or equal to 9, of uncertain benefit if between 6 and 8, and contraindicated if 5 or less.
Complete data were available from 280 of the 322 patients entered into the audit. The 42 with incomplete data, who had either chosen not to provide information or failed to attend for audit assessment, were a disparate group. 90% of patients presented to their GP in the first instance, mostly with incremental backache, and 40% could not walk by their date of referral to hospital. The three most common primary tumours were lung (20.7%), prostate (19.3%) and breast (17%).
A Tokuhashi score of six or more was associated with longer survival than a score of 5 or less (p=0.03). However, when we looked at the six components of the score, we found more significant associations. The Karnofsky score had a strong association with longevity (p<0.0001), and the neurological deficit was also significant (p=0,0084). Histology appeared important, especially when comparing lung cancer with breast and prostate cancer (p=0.0002).
The three other parameters of the Tokuhashi score (other vertebral, extraspinal and internal organ metastases) showed no significant association with longer survival.
We conclude that the Tokuhashi scoring system could be simplified to three rather than six parameters, concentrating on the Karnofsky score, neurological deficit and histology. The criteria for surgical decision-making could be improved by better weighting of these parameters