In the early phase of HO, triple phase bone scanning demonstrates increased uptake of osteotropic radionucleotides. Bone scanning has proven to be more sensitive than plain radiography in detecting early HO. Neurogenic HO becomes evident on plain radiography approximately two to six weeks after diagnosis using the triple phase bone scan (Orzel et al. 1985; Freed et al. 1982). However, bone scans have lower specificity than radiography (Freed et al. 1982). Computed tomography (CT) or magnetic resonance imaging (MRI) scanning may be a useful tool when considering surgery as it allows for better visualization of the heterotopic bone (Amendola et al. 1983). Some studies have examined diagnosis of HO through elevations in biochemical markers such as alkaline phosphatase (Singh et al. 2003; Tibone et al. 1978) and creatine phosphokinase (Singh et al. 2003; Welch et al. 1973; Rossier et al. 1973). The predictive value of alkaline phosphatase has not been validated (Singh et al. 2003; Welch et al. 1973; Rossier et al. 1973), although there is conflicting evidence of an association with HO and increased serum creatine phosphokinase levels (Singh et al. 2003; Welch et al. 1973). Schurch et al. (1997) studied individuals with acute SCI and found increases in the 24-hour prostaglandin E2 (PGE2) urinary excretion a valid indicator of early HO formation.
The Brooker Classification Scheme is typically used to diagnose HO in the pelvic region (Zychowicz 2013). The system is based on an anteroposterior radiograph of the pelvis which classifies HO into one of five classes. The classes are based on the progression of ossification: Class 0 – no presence of ossification, Class 1 – islands of bone within soft tissue of any size, Class 2 – bone spurs from pelvis or femur with at least one cm between opposing bone surfaces, Class 3 – bone spurs from pelvis and femur reducing space between opposing bone surfaces to less than 1 cm, and Class 4 – complete ankylosis of hip (Zychowicz 2013).
The Brooker Classification Scheme has been criticized by some clinicians and adjustments to the traditional classification system have been proposed. Mavrogenis et al (2012) have suggest focusing on the location of the HO formation around the hip joint using the following scheme: Type 1–anterior, Type 2–posterior, Type 3–anteromedial, Type 4–circumferential HO. The adjustments are based on locating the anatomical position of HO which permits an estimation of the prognosis regarding blood loss, transfusion requirements, and recurrence. The new classification system improves ease of use and provides the opportunity for more rapid post-operative planning for surgical approach, evaluation, and prognosis (Mavrogenis et al. 2012). However, Citak et al. (2012) suggested the use of ultrasound, CT or MRI rather than radiograph in order to improve diagnosis and reduce the use of methods with less sensitivity for early diagnosis.
A new classification has been proposed by Arduini et al. (2015) based on the use of a 3D-CT reconstruction scan. The authors propose that this method allows the surgeon to observe the bone bridge in its entirety, can assist in assessing the relationship between neurovascular structures and the HO, and can identify the site of discontinuity therefore allowing the surgeon to plan the best surgical approach. Arduini et al. (2015) contend that their approach is more advantageous in comparison to the Brooker Classification, which they believe is inadequate at classifying the anatomy of the HO, and Mavrogenis et al. (2012)’s proposal, which the authors argue cannot give precise details regarding the continuity of the bone bridge or the extent to which surrounding muscles and soft-tissues are being affected. Genêt et al. (2015) also highlight that the Brooker Classification cannot be used to determine the potential drawbacks of surgery and gives pessimistic results of hip range of motion with some hips classified as III or IV despite no clinical ankylosis found. Arduini et al. (2015) state that their classification system has reproducible high intra- and inter-observer reliability and that previous methods including Garland’s radiological classification and Mavrogenis et al. (2012) CT approach have not been validated. However, it is also important to note that Arduini et al. (2015) used a mixed population of SCI and TBI with SCI accounting for 29% of the sample. Although their approach is of interest, further research is required with an all-SCI, or at least SCI-majority, group of patients.