The Lateral Pillar Classification of Herring et al for Legg-Calve-Perthes Disease

 

The femoral head is divided into 3 sections based on 2 lines drawn perpendicular lines to the epiphyseal line.

Pillars of the Femoral Head	Width of the Femoral Head
lateral	lateral 15-30%
central	central 50% (more or less)
medial	medial 20-35%

 

The precise location for the pillars depends on the demarcation between the central sequestrum and the remainder of the epiphysis.

• At a early stage this would be fairly clear-cut. The degeneration would be limited to the central pillar and the point where the concavity reaches the spherical surface would be the point to draw the line.

• This may be somewhat subjective in more advanced disease.

 

Rules for reading radiographs:

(1) Radiographs must be true AP projections. Any that are not true AP views should not be used. External rotation may cause result in measurements that cause misclassification.

(2) It is best to examine several radiographs taken at the time of early fragmentation if possible. Classification is based on the one showing greatest involvement of the lateral pillar.

(3) The measurements for both lateral pillars must be taken exactly the same.

 

relative height of lateral pillar =

= (height of lateral pillar of affected side in mm) / (height of lateral pillar of normal side in mm)

 

Lateral Pillar	Class
lateral pillar radiographically normal	A
lateral pillar affected but height >= 50% of normal	B
lateral pillar affected and height < 50% of normal	C

 

where:

• When measuring the height of a severely affected lateral pillar, I have the following questions:

(a) Do I use the maximal height within the pillar, or measure along the line separating the lateral and central pillars? (The edge with the central pillar might be more severely affected than a more lateral portion.)

(b) Do I measure the maximal point on the affected and unaffected side? or do I measure the height at the same position within the pillar on both sides?

• I am not sure the process to use if there is the unaffected side is not normal.

• On page 144 of Herring et al, the Group B section title has "> 50%" but text implies ">= 50%".

• I wonder if measuring the area or volume remaining as a percent of normal might be more accurate.

 

Interpretation:

• Group A: uniformly good outcome. This corresponds to Catterall groups I and II.

• Group B: good outcome if age of onset < 9 years; less favorable if >= 9 years old at the time of onset. This includes patients who would be classified as Catterall groups II and III; this represents the group most difficult to classify using the Catterall method.

• Group C: Most femoral heads become aspherical. The fragmentation and reossification stages have longer durations. This corresponds to Catterall groups III and IV.

 

Performance:

• The interobserver reliability is 0.78 with this method, while the Catterall and other methods may be < 0.50.

• The intraobserver agreement was 0.79.

• The kappa statistic was 0.52.

• The chance that 2 random observers would rate a random radiograph the same was 0.68.