Rockwood, Green, and Wilkins' Fractures, 10e Package

119

CHAPTER 5 • Classification of Fracture

A

B

Figure 5-3. AP radiograph ( A ), coronal CT scan ( B ), and 3D CT rendering ( C ) show ing a proximal humerus fracture and the challenge of the Neer classification. Note the differences in appearances of the frac ture lines. How and from where does an observer measure the magnitude of dis placement and/or angulation necessary to be characterized as a part?

C

Observer agreement is low for the Neer classification. Frac tures of the proximal humerus often have poor radiographs because of difficulty in obtaining a good lateral, especially while acutely injured. Without good radiographs, reliable classifica tion is difficult. Although the definitions that separate the cat egories described above seem clear, they may be inadequately applied or misunderstood by observers. The definitions of 1 cm of displacement or 45 degrees of angulation are clear and com mendable but present difficulties with adequately measuring these variables that are critical to assigning categories (Fig. 5-3). Finally, the proximal humerus often has complex fracture lines that do not fit a consistent pattern or definition. Despite these limitations, the Neer classification continues to be widely uti lized. These observations indicate that the proximal humerus is a deeply entrenched classification and fracture description lan guage that works well in some ways. However, poor observer reliability and resulting poor validity are clear issues. Similarly, proximal humerus fractures have among the widest variation and disagreement on natural history and optimal treatment tech niques for various patterns. While there may be other reasons for this lack of knowledge and disagreement on optimal treatment, the difficulties with classification are almost certainly a contrib uting factor. This is an area in which AI systems have the poten tial to propagate an already well-known and easily understood system, while increasing the reliability of the classification. Calcaneus Fractures Classification of calcaneus fractures illustrates how advances in imaging technology have and will undoubtedly continue to add

new information and how that new information can be used to expand and develop new classifications. The calcaneus also illustrates the complexity of fracture classification since despite many different important classifications there are important aspects of this fracture that are not included in any of the major classifications. The most popular early classification of fractures of the cal caneus was described by Essex-Lopresti in 1952. 12 This classi fication, which was based upon plain radiographs, accounted for two major fracture patterns—a joint-depression type frac ture where the articular fragment is separate from the calcaneal tuberosity and a tongue-type fracture where the articular sur face remains attached to the tuberosity. If the force is directed more posteriorly toward and into the posterior facet, a joint depression-type fracture occurs; whereas if the force is directed more axially, a tongue-type fracture occurs (Fig. 5-4). 12 With the rise of CT imaging, other authors described injury patterns based upon the new information that was obtained with this more advanced imaging technique. This included informa tion on joint alignment and displacement of fragments, and surgeons used this increased information to determine whether or not surgical reconstruction may be indicated. 55 These early descriptions added useful information to our knowledge on the mechanism of injury and treatment of calcaneus fractures. Several authors began to base systems of classification on data obtained from CT. 9,50 The most well-known calcaneus fracture classification system based upon CT scans was described by Sanders et al. The Sanders classification focuses upon the number and location of articular fracture fragments at the widest part of the posterior facet viewed