What Is a Romanus Lesion in Spondyloarthritis?

A Romanus lesion is an inflammatory change at the corner of a vertebral body, closely associated with axial spondyloarthritis and, in particular, ankylosing spondylitis. It appears on MRI as bone marrow edema where the outer fibers of the spinal disc attach to the vertebra, and over time it can evolve into a fatty deposit visible on different MRI sequences. The term carries weight in rheumatology and radiology because these small corner signals help clinicians catch a disease that can otherwise go undiagnosed for years, and the lesions themselves offer clues about whether the spine is headed toward permanent stiffening.

What Actually Happens at the Vertebral Corner

The spine’s intervertebral discs are anchored to each vertebra by a ring of fibrous tissue. In spondyloarthritis, the immune system targets these anchoring points, called entheses. When inflammation settles at the front corners of the vertebral endplates, the bone marrow in that small triangular zone becomes swollen with fluid and inflammatory cells. Radiologists have historically called this “anterior spondylitis,” and the lesion itself is named after the Swedish radiologist Carl Romanus, who described the characteristic erosive changes on plain X-rays decades ago.

On MRI, these acute lesions show up as dark areas on T1-weighted images and bright areas on T2-weighted or fluid-sensitive sequences, reflecting the water content of inflamed bone marrow.¹ The shape is distinctive: a sharply marginated triangle hugging the corner of the vertebral body, without the irregular contour you would see with a degenerative disc problem or a Schmorl’s node (a small herniation of disc material into the bone). MRI picks up these lesions far more reliably than standard X-rays, which tend to show changes only after significant erosion or new bone has already formed.²

Where Romanus Lesions Tend to Appear

Although Romanus lesions can show up anywhere along the spine, they favor certain regions. The thoracic spine is the most common location. In one study of patients with spondyloarthritis, roughly 60 percent of fatty Romanus lesions were found in the thoracic vertebrae.³ This distribution makes sense when you consider the biomechanics of the mid-back, where the rib cage limits spinal flexibility and the entheseal attachment points take on different mechanical stresses. Because inflammation tends to concentrate in the thoracic region, the fatty changes that follow inflammation track to the same area.⁴

This thoracic predilection matters clinically. Many radiologists focus heavily on the sacroiliac joints and the lumbar spine when evaluating suspected spondyloarthritis. If the thoracic spine is not included in the imaging protocol, a meaningful proportion of Romanus lesions will be missed entirely. Research into whole-spine MRI protocols has tried to address this gap by showing that thoracic-specific inflammatory lesion counts can match the diagnostic specificity of sacroiliac MRI when appropriate cutoffs are used.⁵

How Well Romanus Lesions Identify Spondyloarthritis

The diagnostic value of these corner lesions depends on how you define and count them. A single Romanus lesion at one vertebral corner is a weak signal on its own: inflammatory changes at a vertebral corner can occur in people with ordinary disc degeneration or even in people with no spinal disease at all. The strength of the finding comes from specificity and from the number of lesions present.

A study comparing patients with ankylosing spondylitis to controls found that the sharply marginated MRI corner sign had a specificity of 96 percent and a positive predictive value of 92 percent. Sensitivity was lower, at 44 percent, meaning that more than half of people with the disease did not have the classic sign.⁶ In practical terms, when a radiologist sees a well-defined Romanus lesion on MRI, the chance it belongs to someone with spondyloarthritis is high. But the absence of such lesions does not rule the disease out.

Counting lesions improves diagnostic accuracy. An international expert consensus defined a positive spinal MRI for inflammation in axial spondyloarthritis as requiring anterior or posterior spondylitis at three or more vertebral sites.⁷ For corner inflammatory lesions specifically, a threshold of five or more across the whole spine, or three or more in the thoracic spine alone, has been proposed to match the high specificity (around 97 percent) that sacroiliac MRI provides.⁸

Fatty Romanus lesions, the chronic remnants of earlier inflammation, have their own diagnostic power. Among patients with spondyloarthritis, about a third showed fatty Romanus lesions on MRI, compared with much smaller proportions in patients with degenerative arthritis, spinal cancer, or other conditions. When more than five fatty Romanus lesions were present, the likelihood ratio for spondyloarthritis jumped to about 12.6, making it a strong diagnostic indicator. Notably, a quarter of patients who had these fatty lesions showed no active bone-edema lesions on fluid-sensitive MRI at all, suggesting that fatty Romanus lesions can flag the disease even during periods of apparent quiescence.⁹

How a Romanus Lesion Evolves Over Time

A Romanus lesion is not a static finding. It reflects a snapshot of a biological process that changes over months and years. In its acute phase, the vertebral corner is inflamed and swollen with fluid, visible as bone marrow edema on MRI. If the inflammation resolves, which happens in the majority of cases, the bone marrow in that corner undergoes fatty replacement. On T1-weighted MRI sequences, the corner then appears bright rather than dark, reflecting fat content instead of water. This bright corner on T1 is what radiologists call a fatty Romanus lesion or fatty corner lesion.

This transition from edema to fat is the most common outcome. In one long-term study of patients on anti-TNF biologic therapy, more than 70 percent of inflamed vertebral edges resolved completely within two years, while about 29 percent converted to fatty deposits.¹⁰ A much smaller fraction progressed directly to new bone. The signal intensity pattern that shows up most often in these corner lesions resembles what spine radiologists call a Modic type II change, a fatty marrow signal seen in about three-quarters of MRI corner lesions in ankylosing spondylitis patients.¹¹

Understanding this evolution has practical importance. A fatty corner lesion on a single MRI scan does not tell you when the inflammation happened. It could represent a burned-out Romanus lesion from years ago or a relatively recent conversion. Repeat imaging, ideally compared against a prior study, gives a much richer picture of disease trajectory.

Do Romanus Lesions Lead to New Bone Growth?

One of the defining features of ankylosing spondylitis is the gradual formation of syndesmophytes, bony bridges that grow between adjacent vertebrae and can eventually fuse the spine into a rigid column. A natural question is whether Romanus lesions are the starting point of this process. The relationship turns out to be real but less straightforward than the simple story of “inflammation causes bone” would suggest.

Vertebral corners that show fatty deposits on MRI do go on to develop new syndesmophytes more often than corners that look normal. In a clinical trial cohort, about 10 percent of corners with baseline fat developed a new syndesmophyte, compared with about 3 percent of corners without any MRI abnormality. An observational cohort showed a similar pattern, with roughly 6.5 percent of fatty corners developing new bone versus 1.4 percent of normal corners.¹² Corners where both inflammation and fatty deposition were present at the same time carried the highest risk, with about a three-fold increase in the relative chance of syndesmophyte formation over five years.¹³

But the neat sequential model, where an inflammatory Romanus lesion converts to fat and then to bone, is actually quite rare when researchers track individual corners over time. In two separate cohorts, the complete sequence of corner inflammation followed by fatty change followed by new syndesmophyte occurred at only about 0.2 percent of all vertebral corners studied.¹⁴ Many syndesmophytes appear at corners that never showed detectable inflammation on MRI, and many inflamed corners heal without ever producing new bone. The pathway from Romanus lesion to spinal fusion clearly exists, but it is not the dominant route. Other factors, possibly mechanical stress, subclinical inflammation below MRI detection thresholds, or genetic predisposition, are driving much of the new bone formation independently.

This has fueled an ongoing debate about whether suppressing inflammation with biologic drugs can slow structural progression. The early hope was that if you put out the inflammatory fire at the vertebral corners, you would prevent the downstream bone formation. But the rarity of the clean inflammation-to-fat-to-bone sequence, even in patients on anti-TNF therapy for years, has led researchers to question whether the relationship is that direct.¹⁵ Whether earlier or more aggressive suppression of inflammation changes the long-term bone-formation picture remains an open question.

Telling Romanus Lesions Apart from Similar Findings

Not every bright or dark signal at a vertebral corner means spondyloarthritis. Several other conditions produce corner changes that can overlap with Romanus lesions on MRI, and distinguishing them is one of the trickier tasks in spinal imaging.

Degenerative Modic changes are the most common mimics. These are signal changes in the vertebral endplates caused by ordinary disc degeneration, and they become increasingly common with age. Modic type I changes (bone marrow edema near a degenerating disc) can look very similar to an acute Romanus lesion. Modic type II changes (fatty marrow replacement) can resemble a fatty Romanus lesion. The key differences are location and shape. Modic changes tend to involve the central or lateral portions of the endplate and often appear adjacent to a clearly degenerating disc, while Romanus lesions sit at the very corner of the vertebral body and have that characteristic triangular outline. In clinical practice, though, the overlap is real enough that fatty Romanus lesions have been observed in a small percentage of patients with degenerative arthritis and even in patients with spinal malignancy.¹⁶

Andersson lesions are another finding in the spondyloarthritis family but involve a different structure. While Romanus lesions target the vertebral corner, Andersson lesions, sometimes called discovertebral lesions, affect the disc space and adjacent endplates more centrally. They can represent either an inflammatory process (destructive spondylodiscitis) or a stress fracture through an already fused spinal segment. Both Romanus and Andersson lesions fall under the broader umbrella of spinal changes in spondyloarthritis, but they carry different clinical implications: Andersson lesions in a fused spine may indicate an unstable fracture requiring surgical consideration, while Romanus lesions typically do not.¹⁷

Schmorl’s nodes, small herniations of disc material into the vertebral body, can also cause focal signal changes near the endplate. These usually have a rounded or lobulated morphology quite different from the sharp triangular Romanus pattern. Infectious spondylitis, spinal metastases, and traumatic compression fractures can all produce corner edema as well, but these typically involve larger areas of the vertebral body, affect adjacent discs, or have clinical features (fever, history of cancer, acute injury) that point away from spondyloarthritis.

What the Lesion Count Tells You About Disease Burden

Beyond their role in initial diagnosis, Romanus lesions serve as a rough yardstick for how active and advanced the disease is. Whole-spine MRI studies have shown that the number of both acute and chronic Romanus lesions correlates with measures of spinal mobility restriction.¹⁸ Patients with more widespread lesions tend to have stiffer spines as measured by clinical mobility assessments, and the total count of spinal MRI lesions, including both Romanus and Andersson types, tracks with overall disease severity.

This correlation has practical value for monitoring treatment response. If a patient starts biologic therapy and repeat MRI shows a reduction in the number of actively inflamed corners, that is a tangible sign the drug is working at the tissue level, even if symptoms have not fully improved yet. Conversely, an increase in active lesions on follow-up imaging can prompt a change in treatment before irreversible structural damage accumulates. The international ASAS/OMERACT consensus recognized both inflammatory and structural corner lesions as the most characteristic spinal findings in axial spondyloarthritis, reinforcing their role as benchmark features on imaging.¹⁹

When Romanus Lesions Appear Without a Clear Diagnosis

One situation that trips up both clinicians and patients is the incidental finding of a vertebral corner lesion on an MRI done for unrelated reasons, such as evaluating back pain thought to be mechanical. A single fatty or edematous corner can appear on MRI in people who do not have spondyloarthritis. Degenerative changes, minor trauma, and even normal aging can produce corner signals that resemble Romanus lesions on a superficial read.

The specificity data discussed earlier shows why a one-off corner lesion should not trigger alarm on its own. It is the pattern that matters: multiple corners involved, the sharp triangular morphology, a thoracic-dominant distribution, and the clinical context of inflammatory back pain (gradual onset before age 40, worse at rest, improvement with activity). If only one or two corners are involved and the patient profile does not fit spondyloarthritis, the finding is more likely degenerative and does not warrant a rheumatology referral. On the other hand, when five or more fatty corner lesions cluster across the spine, particularly in a younger person with chronic back stiffness, the diagnostic odds shift dramatically in favor of spondyloarthritis.²⁰ In that scenario, even if the sacroiliac joints look unremarkable on imaging, the spinal findings alone can tip the diagnostic scales and prompt further workup including blood tests for the HLA-B27 gene and clinical evaluation for inflammatory features.

The quarter of spondyloarthritis patients whose fatty Romanus lesions were present without any active bone-edema signal on fluid-sensitive MRI underscores an underappreciated point: you can have diagnostically meaningful spinal disease that looks quiet on the sequences designed to detect active inflammation. A T1-weighted scan showing fatty corners may be the only imaging clue in someone whose inflammatory phase happened years earlier and has since burned out. Relying exclusively on edema-sensitive sequences risks missing these patients entirely.