Early identification and treatment of epidural neoplasm, before the development of significant neurologic deficits, provides the best opportunity for a favorable outcome. Among the many patients with symptoms, signs, or scintigraphic or radiographic findings suggesting possible epidural disease, a small proportion will have the lesion. The selection of patients for definitive imaging of the epidural space should be based on a determination of the risk of this complication. In this study, the medical records, plain spinal radiographs, bone scintigraphs and myelograms of 43 patients were analyzed retrospectively to assess the risk of epidural disease associated with specific clinical, radiographic, and scintigraphic findings. Cervical, thoracic, and lumbosacral spinal segments were evaluated independently. Symptomatic segments (SS) (N = 41), defined by focal pain or neurologic dysfunction, were distinguished from asymptomatic segments (AS). At SS, epidural disease was found at 86% and 8% of abnormal and normal spinal radiographs, respectively (P < 0.001), and at 69% and 0% of abnormal and normal scintigrams, respectively (P < 0.001), whereas at AS epidural disease occurred in 43% and 3% of abnormal and normal spinal radiographs, respectively (P < 0.001), and 14% and 7% of abnormal and normal scintigrams, respectively (P = NS). Vertebral collapse was highly predictive of an epidural lesion. Epidural disease occurred in 12% of SS and 0% of AS with an abnormal scintigram and normal radiograph, 86% of SS and 45% of AS with abnormalities on both scintigram and radiograph, and at two AS when both were normal. Decision analysis applied to these data yielded a specific conditional probability of epidural disease for each combination of clinical, scintigraphic, and radiographic findings. These data provide a basis for the selection of patients for additional evaluation of the epidural space before neurologic deficits develop.
|Original language||English (US)|
|Number of pages||7|
|State||Published - Dec 1 1989|
ASJC Scopus subject areas
- Cancer Research