MRI Soft Tissue Interpretation

Know normal first. Soft tissue MRI interpretation rests on a single principle: knowing what normal looks like on each sequence, so that any deviation can be recognised as pathology. Muscle, tendon, ligament, nerve, fat and the fibrous capsular structures each have a characteristic, predictable signal that reflects their water and collagen content. MRI is the dominant modality for these structures because of its unmatched soft tissue contrast, multiplanar capability and absence of ionising radiation.

Three questions answer most soft tissue MRI problems:

1. Is the lesion fluid, fat, fibrous tissue or something more complex, the four basic signal building blocks?
2. Where does it sit relative to the deep fascia, and how large is it? Depth and size are the two strongest red flags for a sarcoma.
3. Is an apparent abnormality real, or an artefact such as the magic angle phenomenon that mimics tendon degeneration on short-echo sequences?

The four building blocks. Almost every soft tissue lesion can be characterised by comparing it with four reference tissues:

• Fluid - low T1, very high T2/STIR (cysts, oedema, fresh haematoma)
• Fat - high T1, intermediate-high T2, suppresses on fat-saturated and STIR sequences (lipoma, fatty infiltration)
• Fibrous or collagenous tissue - low on all sequences (tendon, ligament, scar, fibroma)
• Cellular or solid tissue - intermediate T1, intermediate-high T2, and it enhances (most tumours, granulation tissue)

Anything that does not fit a single block, heterogeneous, partly necrotic or partly haemorrhagic, should raise concern.

What changes management. A small number of findings do: the gap and retraction that determine whether a tendon is reparable, the Goutallier grade that predicts rotator cuff repair failure, and the depth-and-size combination that mandates referral to a sarcoma unit before any biopsy.

View larger

Why tendons are dark. Tendons contain highly organised collagen fibres with few mobile protons and very short T2 relaxation times, so they give low signal on every pulse sequence. Any increased signal within a tendon is abnormal and suggests degeneration, a tear, or the magic angle artefact.

Tendinopathy. Four findings, each with its own meaning:

Tears. The classification turns on which surface is breached and whether the tendon is still in continuity:

Reading the full-thickness tear. The fibres are completely discontinuous and fluid signal, bright on T2, fills the gap; the stump may retract. Chronic tears carry atrophy and fatty infiltration of the muscle. Measure the gap for surgical planning.

View larger

Magic angle artefact. A tendon oriented at 55° to the main magnetic field B0 shows artifactually increased signal on short-TE sequences, T1 and PD, that is not present on true T2-weighted images. The artefact is intermediate at about 45 and 65 degrees and disappears at 0 and 90; the common sites are the rotator cuff and the ankle tendons. If increased tendon signal is seen only on T1 or PD and not on T2, suspect magic angle artefact.

Why "check it on T2" works, and the number behind it. Magic angle signal decays exponentially as echo time lengthens. Once TE exceeds roughly 37ms the signal falls back to what the same tendon shows when it lies parallel to B0: the artefact is abolished while anatomical detail is preserved. That critical TE was independent of the coil used and of field strength, which is what makes the rule portable between scanners rather than a local quirk. So the instruction is not merely to look at another sequence: any adequately T2-weighted sequence is long enough to settle the question, and if increased signal persists on it, the lesion is real.

Describing a mass. Measure it in three planes and describe it against the seven features below. Heterogeneous signal from haemorrhage or necrosis, irregular or infiltrative margins, infiltrative growth, surrounding oedema and rapid growth on serial imaging all count against it.

View larger

Referral before biopsy. Any deep mass greater than 5cm, or any heterogeneous or enhancing lesion, should be referred to a sarcoma unit before biopsy or excision. Local staging is MRI of the entire compartment, with CT chest for pulmonary metastases. The biopsy is planned in consultation with the tumour surgeon so that the track can be excised at definitive surgery.

Common benign masses. The signal signature and the key feature of each:

Lipoma versus well-differentiated liposarcoma. Both follow fat signal. The features that concern for liposarcoma are size greater than 10cm, thick septae (greater than 2mm), nodular non-fat components and deep location. If any is present, biopsy or excision is recommended.

View larger

Myxoid tumours. Myxoid matrix is characteristically very bright on fluid-sensitive sequences and can appear deceptively cyst-like.

View larger

The haematoma that mimics a tumour. A subacute haematoma can be intrinsically bright on T1 (methaemoglobin) and may have an enhancing wall, so knowing how blood signal evolves over time is essential before calling such a lesion a tumour. The signal changes predictably with the oxidation state of haemoglobin:

Two techniques beyond the conventional sequences increasingly support soft-tissue mass characterisation, biopsy targeting and treatment-response assessment.

Adjuncts, not arbiters. Neither technique reliably separates benign from malignant on its own because of major overlap, so they supplement, and do not replace, conventional multiplanar MRI and sarcoma-MDT review. Their highest-value uses are directing the biopsy to viable tissue and monitoring treatment response, where a rising ADC and falling enhancement suggest necrosis or response.

View larger

View larger

A mass on a nerve declares itself. A mass arising from a peripheral nerve has a characteristic constellation of signs, and recognising it both makes the diagnosis and warns against an ill-judged biopsy through a functioning nerve. The combination of an entering/exiting nerve, split-fat sign, target sign and fascicular sign identifies a benign nerve sheath tumour.

View larger

View larger

Schwannoma, neurofibroma and the malignant one. A schwannoma typically sits eccentric to the nerve, so it can sometimes be shelled out, whereas a neurofibroma is fusiform with the nerve passing through it and is harder to separate. The major concern is the malignant peripheral nerve sheath tumour (MPNST): suspect it when a nerve lesion is large (over about 5 cm), rapidly enlarging, heterogeneous with necrosis, ill-defined, or has peritumoral oedema, especially in a patient with neurofibromatosis type 1. Loss of the target sign with rapid growth, large size and heterogeneity in an NF1 patient should raise the alarm for malignant transformation and prompt sarcoma referral rather than local excision. Imaging cannot always separate an atypical neurofibroma from an MPNST, so FDG-PET and image-guided biopsy planned with the sarcoma team are used.

View larger

A fixed search pattern. A reproducible search pattern prevents the two classic errors: satisfaction of search, stopping after the obvious finding, and over-calling artefact as disease. The seven steps below work for any soft tissue study and are the order in which to present a film.

Start from the pattern, not the suspicion. Soft tissue lesions are best approached by their dominant MRI signal pattern rather than by clinical suspicion alone. The table groups the common differentials by the pattern that first catches the eye, with the discriminating feature that separates them.

Soft tissue MRI interpretation is broadly consistent worldwide because it rests on physics-based signal characteristics, but referral pathways, access and protocol emphasis vary by health system and resource setting.

The 5cm and depth thresholds. Size over 5cm and depth to fascia are the classic sarcoma red flags, but they are sensitive rather than specific: most deep lesions over 5cm are still benign (a lipoma, a haematoma). They define who needs specialist review, not who has cancer. Khan et al. found that even established imaging red flags performed modestly for lipomatous tumours, which reinforces MDT review over any single sign.

Meniscal grade 2 versus grade 3. The distinction between intrasubstance signal and signal reaching the articular surface underpins MRI meniscal reporting, but interobserver agreement at the surface is imperfect, and signal can persist for years after repair without re-tear. MR arthrography is more reliable in the post-operative knee.

Does MRI grade predict return to play? Higher BAMIC grade broadly predicts longer recovery, but considerable overlap exists, and intratendinous ('c') injuries within a grade carry a worse, more variable prognosis. MRI informs but does not replace serial clinical assessment when timing an athlete's return.

The interpretive frameworks used in soft tissue MRI are underpinned by classification and validation studies. The following are the landmark and supporting references.