Early classification systems

The early classification systems graded patients on their pretreatment neurological status. In 1965, Brice and McKissock published a four-tiered system ranging from mild to complete neurological deficit and made the important observation that the better a patient’ s neurologic status at presentation, the more likely they were to maintain ambulation and sphincter control after treatment [1]. Similar findings were subsequently reported by White, et al in 1971, Gilbert, et al in 1978, and Shaw, et al in 1980 [2–4]. Taken together, these reports emphasized that early recognition of spinal metastases, followed by prompt treatment, provided the best outcome for the patient. Although not included in their classification systems, the papers by Brice and White also recognized the important impact of tumor histology on prognosis, citing better outcomes for patients with favorable tumors such as breast carcinoma compared with poorer histologies such as lung carcinoma. Clearly, the absence of factors specific to the patient’ s disease (tumor histology, extent of metastases, extent of prior therapy) represents an important flaw in these early systems. One final point of interest is that these studies found little benefit of surgery compared to non-surgical therapies. In fact, the article by Gilbert, et al retrospectively evaluated this particular point and found no difference in outcome between surgery plus radiation versus radiation alone in the setting of epidural spinal cord compression. This observation is clearly dated. The surgical procedure in these studies was laminectomy for decompression. It is not surprising that a disease that primarily arises in the vertebral body and is often associated with bony destruction, is not ideally treated with a posterior decompression without subsequent stabilization. The recent study by Patchell, et al utilizes modern techniques of spinal resection and reconstruction and shows, in a prospective, randomized trial, that surgery plus radiation is superior to radiation alone for the management of metastatic spinal cord compression [5].

Classification systems and spinal instability

Several subsequent classification systems attempted to address the critical issue of spinal instability. Recognizing that an unstable spine was unlikely to be mended with chemotherapeutic or radiotherapeutic treatment modalities, these systems described a progression of bony destruction and/or deformity leading to increasing instability. Both the systems of Dewald, et al, and Harrington propose a 5-tiered system ranging from minimal bony involvement to severe collapse and deformity [6, 7]. Furthermore, both systems recognize the importance of neurologic status and incorporate this into the system. Dewald’ s system also includes the immune status of the patient. Although unique to their system, this factor is a reflection of the overall medical status of the patient, a key component when formulating treatment decisions. Finally, Asdourian, et al, rigorously analyzed the imaging studies on a series of breast cancer patients to determine the patterns of spinal deformity [8]. Their system, while a bit cumbersome for practical use, further highlights the complexity of this issue. Despite these efforts, a clear definition of spinal instability in the oncology setting remains elusive. None of the aforementioned systems have been tested in a prospective manner for their ability to predict instability. Moreover, the classical models of spinal stability developed by Denis and White and Panjabi were intended for trauma and are, therefore, not applicable to tumor patients [9, 10]. Clearly, the sudden, high energy forces that affect the spine in trauma are very different from the forces at work in a patient with metastatic spine disease. So, how much collapse or angulation makes a patient with a T5 metastasis unstable? Must they have pain? Does the type of tumor make a difference? What if it were C5 or L2? Clearly, additional efforts are required and any new system must be evaluated prospectively.

Classification systems and prognosis

Perhaps the most difficult question that arises when determining the optimum treatment plan for the patient with spinal metastasis is that of the patient’ s prognosis. Overtreating the patient with a limited life expectancy can be just as disheartening as failing to be aggressive enough in a patient who lives longer than expected. Several classification systems have evolved to help predict prognosis and guide subsequent treatment. Of these, the most commonly used are the Tomita and Tokuhashi scoring systems [11, 12]. The Tomita system is based on three parameters including tumor histology, visceral metastases, and bony metastases. Interestingly, there is no consideration of patients’ overall performance status or neurologic condition, two factors considered to be significant predictors of survival by many other authors. Also, for patients with long predicted survival and limited disease, Tomita recommends en bloc resection of solitary spinal metastases for optimum local control. In a separate article, he describes his experience with this technique and proposes an anatomic classification of metastatic spread within the spine as a guide for when en bloc resection is feasible [13]. His classification has roots in Enneking’ s staging system for primary bone tumors [14]. Obviously, the use of en bloc resection techniques for metastatic lesions of the spine is a controversial topic. The Tokuhashi system is based on six parameters. These include performance status (KPS), extraspinal bony metastases, vertebral metastases, visceral metastases, tumor histology, and neurologic status (Frankel grade). It is worth noting that the original Tokuhashi system published in 1990 [15], was revised in 2005 to include expanded scoring of the tumor histology (0–5 vs 0–2). This may have been based on the findings of Enkaoua, et al, who applied the original system in 71 patients and concluded that the scoring of certain histologies needed adjustment [16]. This is noteworthy as it reflects the fact that tumor histology is the single most important of the prognostic factors. Finally, van der Linden, et al, reviewed a large series of patients treated with radiotherapy and proposed a classification system based on performance status (KPS), tumor histology, and visceral metastasis [17]. Unlike the other surgical series, these patients were collected from a randomized radiotherapy trial. Patients with spinal cord compression and radioresistant tumors such as renal cell carcinoma and melanoma were excluded. None of these scoring systems addresses the issue of instability.

SCALE DESCRIPTION

Classified based on structural compromise (axial or translational instability):

• 
  

  Type lA—metastatic involvement noted in the vertebral body (impending axial instability)

  
  
• 
  

  Type IB—complete vertebral body marrow replacement (impending axial instability)

  
  
• 
  

  Type IIA—vertebral body with deformity of one endplate (axial instability; impending translational instability if accompanied by metastatic involvement of both pedicles and posterior elements)

  
  
• 
  

  Type IIB—vertebral body balloons outward circumferentially with potential spinal canal encroachment occurring from bulging of the posterior wall (axial instability; impending translational instability if accompanied by metastatic involvement of both pedicles and posterior elements)

  
  
• 
  

  Type IIIA—endstage collapse in the thoracic spine is asymmetrical in the sagittal plane, with the anterior aspect of the vertebral body collapsing to a greater extent than the posterior aspect resulting in a localized kyphosis (axial instability; impending translational instability if accompanied by metastatic involvement of both pedicles and posterior elements)

  
  
• 
  

  Type IIIB (cervical or lumbar)—endstage collapse in the cervical or lumbar spine is symmetrical in the sagittal plane (axial instability; impending translational instability if accompanied by metastatic involvement of both pedicles and posterior elements)

  
  
• 
  

  Type IV—translational deformity (translational instability)

Interpretation:

The higher the type, the greater the severity.

SCALE ILLUSTRATION

METHODOLOGY

No predictive validity or reliability studies were identified.

Predictive validity

CONTENT

RATING

SCALE DESCRIPTION

Classified based on neurological evaluation:

• 
  

  Mild—Able to walk

  
  
• 
  

  Moderate—Able to move legs, but not against gravity

  
  
• 
  

  Severe—Slight residual motor and sensory function, with retention of deep pain sensation

  
  
• 
  

  Complete—No signs of motor, sensory, or normal sphincter function below lesion in cord

Interpretation:

The higher the type, the greater the severity.

METHODOLOGY

No predictive validity or reliability studies were identified.

Predictive validity

Reliability

CONTENT

RATING

SCALE DESCRIPTION

Classified based on neurological evaluation and patient–centered parameters:

• 
  

  Grade I—Pain or minor neurological symptoms; normal social an professional activities

  
  
• 
  

  Grade Il—Mild neurological symptoms; normal social life, but interruption of professional activites

  
  
• 
  

  Grade Ill—Moderate neurological syndrome (paraparesis, sphinci disturbances, columnar pain); active life impossible

  
  
• 
  

  Grade IV—Serious neurological syndrome (paraplegia, complete sphincter deficit)

  
  
• 
  

  Grade V—Medullary syndrome of spinal cord transaction

Interpretation:

The higher the grade, the greater the severity.

SCALE ILLUSTRATION

METHODOLOGY

Predictive validity

No reliability studies were identified.

Reliability

References:

Constans JP, de Divitiis E, Donzelli R, et al (1983) Spinal metastases with neurological manifestations. Review of 600 cases. J Neurosurg; 59:111–118.