Reading Musculoskeletal X-Rays with Confidence: A Practical Framework for PTs and ATs

A patient comes in with an obvious shoulder dislocation. The humeral head clearly isn't where it belongs. If you only had one x-ray view, would you know which way it went?

Most people guess anterior, since that's the direction dislocations usually go. But “usually” is not good enough when you're the one deciding what happens next. Add a lateral view to that same shoulder and the picture changes. A second view, a systematic look, and a little humility about what you're qualified to call go a long way.

I'm a physical therapist, not an orthopedic surgeon or a radiologist. But I've ordered and reviewed a lot of imaging, going back to my time working direct access in the military, and I've picked up a framework that holds up whether you're a physical therapist with ordering privileges or an athletic trainer trying to decide if an athlete needs to go to imaging tonight or can wait until morning.

Even the specialists miss things

Studies auditing radiology reports have found that a meaningful share of statements made by radiologists and residents were inaccurate, some with real clinical consequences.^1-3 These fall into two buckets: errors of observation, where the reader doesn't fully search the image, and errors of interpretation, where an abnormal finding doesn't get linked to what's actually going on with the patient.

That second category is where we come in. The radiologist reading the film didn't evaluate your patient. You did. You know the history, the mechanism, and the exam findings. That's why almost every imaging report ends with some version of “correlate clinically.” They're handing that job back to you.

That handoff matters because a radiology report isn't always the final word by the time it reaches you. Research on preliminary reads has shown that early interpretations sometimes get revised once a more experienced reader looks again,⁴ and a clinical history shared with the reader up front changes how a film gets interpreted in the first place.⁵

In other words, the report you're holding was shaped by how much the radiologist knew about your patient, and it may still change. Correlating clinically is how you catch the gap either way.

I still order plenty of imaging and trust the people reading it, but the report doesn't get the last word. That belongs to whatever I already know about the patient in front of me.

A repeatable way to scan a film: the ABCS

If you don't know anatomy, you won't know what you're looking at. That's step one. Step two is resisting the pull toward the obvious finding, because tunnel vision on the fracture everyone already saw is exactly how the second fracture gets missed.

I use four categories, in order, every time:

• Alignment. Are all the bones present and accounted for? Are they positioned normally in relation to the joint? For a cervical spine film, that means literally counting to seven.

• Bone. How does the density look? Weight-bearing bone should be denser than non-weight-bearing bone. Does the texture match what you'd expect for this patient's age and this body part?

• Cartilage. You can't see cartilage directly on a plain film, but you can see the space it occupies. Watch that space, especially around growth plates in younger patients.

• Soft tissue. You won't get a detailed read on muscle or tendon, but you can see outlines and swelling, and that's often enough to flag something worth a closer look.

Run through all four, every time, on every film. Not just the body part someone told you hurts.

One view is no view

We're three-dimensional people. A single x-ray view is a two-dimensional shadow of that, and shadows can lie about depth. That's why the standard is two views at minimum, typically an anteroposterior and a lateral, with a tangent view whenever the anatomy calls for it.

When you order imaging, also think about what's proximal and distal to the area of concern. A calcaneal fracture from a fall carries enough compressive force that it's worth asking about back pain and considering the lumbar spine, even if the foot is the obvious complaint. Chest x-rays can pick up findings unrelated to why you ordered them. The joint above and below the injury deserves a look—not necessarily a separate x-ray order every time, but at minimum a careful clinical check.

With trauma, expect more than one finding

Try this at home: take a hard pretzel and see if you can break it in exactly one place. It's nearly impossible. Bone rings behave the same way. Structures like the pelvis and the cervical spine tend to fail in more than one location when enough force is involved, which is why a Jefferson fracture or a pubic rami fracture should prompt you to keep looking rather than stop at the first thing you find.

I've had patients where I picked up an obvious malleolar fracture, kept palpating, and found a second fracture in the fibula that would have been easy to miss if I'd stopped at the first one. Trauma doesn't play by tidy rules. Do the full scan every time.

After that full scan, how you describe what you found matters too, especially when you're calling it in to an orthopedic surgeon or documenting for the next clinician who sees this patient. Fractures are typically described by location, pattern (transverse, oblique, spiral, comminuted), displacement, and whether they're open or closed. Spiral fractures deserve extra attention in pediatric patients, specifically, since a twisting-pattern fracture and a story about falling down the stairs don't always add up. If the story doesn't match the pattern on film, that's worth a second, more careful conversation with the family.

When it's not about a fracture at all

Bone changes that aren't fractures still matter, and you don't need to name the pathology to recognize that something's off. Osteoarthritis has four classic radiographic hallmarks: joint space narrowing, bone spurs, sclerotic borders, and subchondral bone cysts. Seeing all four on a hip or knee film tells you something clinically useful, even though imaging findings and reported pain don't always line up the way you'd expect.

Other findings—lytic lesions, unusual growth patterns, bone that looks like it's been eaten away—won't always come with a name you recognize. That's okay. Working out exactly what's causing those changes belongs to someone with more specialized training. Your job is to recognize that something doesn't look like the hundreds of normal films you've seen before and to route that patient to someone who can sort out what it is.

The bottom line

Get good quality images, know your anatomy, get more than one view, and don't stop looking once you've found the obvious thing. Then correlate everything against the patient sitting in front of you, because that link between the image and the exam is the piece no one else can supply for you.

And when something looks wrong and you can't say exactly what it is? Send it on, ask the question, or walk it down the hall if you have to. Knowing your limits is part of doing this well.

If you're interested in going deeper on musculoskeletal imaging, I cover it at greater length in my two-part course series:

• Musculoskeletal Radiology and Imaging: Fundamentals, Part 1—Radiodensity, image quality, the ABCS method, fracture patterns, pediatric considerations, and recognizing arthritis, infection, and neoplasms on plain film.

• Musculoskeletal Radiology and Imaging: Fundamentals, Part 2—When to reach for fluoroscopy, bone scan, ultrasound, CT, or MRI, and how to match the right modality to the patient in front of you.

Whether you order imaging, review it, or just want to hold your own in a conversation with a radiologist, both courses are built for you.

References

1. Christensen, E. E., Murry, R. C., Holland, K., Reynolds, J., Landay, M. J., & Moore, J. G. (1981). The effect of search time on perception. Radiology, 138(2), 361–365. https://pubmed.ncbi.nlm.nih.gov/7455116/

2. Rhea, J. T., Potsaid, M. S., & DeLuca, S. A. (1979). Errors of interpretation as elicited by a quality audit of an emergency radiology facility. Radiology, 132(2), 277–280. https://pubmed.ncbi.nlm.nih.gov/461779/

3. Swensson, R. G., Hessel, S. J., & Herman, P. G. (1977). Omissions in radiology: Faulty search or stringent reporting criteria? Radiology, 123(3), 563–567. https://pubmed.ncbi.nlm.nih.gov/870932/

4. Holman, B. L., Aliabadi, P., Silverman, S. G., Weissman, B. N., Rudolph, L. E., & Fener, E. F. (1994). Medical impact of unedited preliminary radiology reports. Radiology, 191(2), 519–521. https://pubmed.ncbi.nlm.nih.gov/8153332/

5. Doubilet, P., & Herman, P. G. (1981). Interpretation of radiographs: Effect of clinical history. AJR. American Journal of Roentgenology, 137(5), 1055–1058. https://pubmed.ncbi.nlm.nih.gov/6975000/