CORR Insights®: What Is the Diagnostic Performance of Conventional Radiographs and Clinical Reassessment Compared With HR-pQCT Scaphoid Fracture Diagnosis?

Where Are We Now?

Scaphoid fractures are problematic because they are hard to diagnose and difficult to treat. Despite surgeons’ best efforts in those areas, complications—the most dreaded one being osteonecrosis further complicated by painful wrist arthritis—are dispiritingly common. For those reasons, discussions about diagnosis and treatment of these injuries continue to pack rooms at meetings and fill the pages of journals. The widespread use of scintigraphy, CT, and MRI has generally increased our ability to detect scaphoid fractures, but these modalities have some limitations. With “normal” CT, it is possible to overlook fractures depending on the resolution and cross-sections; with MRI, an increased signal is sometimes interpreted as a fracture that is not confirmed with the use of thin-slice CT. In practice, most of these imaging studies are only used when there is some clinical suspicion of a scaphoid fracture being present.

An article in this month’s Clinical Orthopaedics and Related Research® by Daniels et al. [3] deals with the question of whether a repeated clinical examination within 7 to 14 days combined with conventional radiography is sufficient to identify patients who will benefit from additional imaging. One message of this study is clear: The combination of conventional radiographs and two clinical examinations does not provide adequate diagnostic certainty, because a true fracture was identified in only about 40% of patients. This is important because a fracture that is not diagnosed will not be treated. Unfortunately, this study does not provide a solution for this problem. In contrast, it reveals that a commonly used clinical algorithm to detect these fractures is often misleading. Nevertheless, this is a very important result because it might affect clinical standards and behavior. Because scaphoid fractures are unforgiving, it is crucial to develop effective diagnostic tools and algorithms that are also economical to identify these injuries.

Where Do We Need To Go?

The question remains: Which method (or combination of methods) gives us the highest possible likelihood of detecting scaphoid fractures? According to this study [3], our diagnostic problems could be solved with high-resolution CT. But because this method is a research tool and not yet available in a hospital setting, it does not answer the question for now. Furthermore, because even high-quality CT and MRI are not available in all locations globally, and regular reassessment might be problematic in certain healthcare systems, any diagnostic strategy must deal with the fact that care in many parts of the world is limited by resource constraints. For example, in Germany, hospitals are intended to provide emergency diagnostics and treatment only. After that, patients must be sent to practitioners (or physicians in private practice) for further treatment. Therefore, it is not possible to perform regular reassessments. To put everyone with hand or wrist trauma in a CT or MRI machine is not practicable. Because patient selection seems to be the most problematic issue, further studies should focus on identifying robust diagnostic algorithms using commonly available tools to select patients who should undergo further diagnostic procedures.

How Do We Get There?

Diagnosing and treating scaphoid fractures can be divided into two parts: first, identifying patients for further extended diagnostic procedures with easily accessible clinical and radiologic means, and second, performing these extended diagnostics. The current study by Daniels et al. [3] might be a blueprint for testing the validity of existing clinical pathways. Other pathways (such as different conventional radiographs, clinical assessments, and time intervals between assessments) must also be checked. If we don’t succeed in identifying feasible clinical algorithms (and the current study does not encourage me to be very optimistic, because a lot of investigation has already been done on this problem) we can only hope for the development of new diagnostic tools or further improvement of existing ones, or increasing and easier availability of cross-sectional techniques such as CT and MRI.

In the meantime, we should thoroughly explore available existing diagnostic methods. To do this, we need a trustworthy gold standard against which we can test those alternatives. Current studies [1, 2, 4] have shown the superiority of high-resolution peripheral quantitative CT over conventional CT; therefore, this method might serve us well as a gold standard when evaluating other diagnostic methods.

Although most surgeons regard clinical examination as indisputable, we may need to accept that in identifying and treating scaphoid fractures, serious technical support is necessary. The old strategy of “if you are not sure, apply a cast, then wait a week and reassess” is, as we learned in this article [3], very likely to let us down. Until a robust and effective diagnostic algorithm can be developed, I suggest that if there is any doubt based on a patient’s clinical presentation, the clinician needs to obtain advanced cross-sectional imaging (such as a CT or MRI), if at all possible.

Therefore, the protocol to detect a high number of scaphoid fractures might be: If wrist trauma seems to be adequate to produce a scaphoid fracture, don't trust a first impression, don't trust conventional radiography, and don't trust even a repetitive clinical examination. Instead, perform CT or MRI for the patient.