Tumour Imaging: ABCDS Approach

Musculoskeletal tumours run from trivial "do-not-touch" lesions to lethal high-grade sarcomas. The candidate's task is not to memorise every entity but to apply a reproducible systematic framework to any radiograph and answer one question: is the lesion non-aggressive, which can be observed or treated by a generalist, or aggressive, which must be referred to a specialist sarcoma centre before any intervention? Age, zone of transition, cortical integrity, periosteal reaction and a soft-tissue mass answer that question faster than trying to name the tumour.

Why the radiograph comes first. Matrix mineralisation, periosteal reaction and the zone of transition are best appreciated on plain film, and they frequently allow a confident diagnosis or a tight differential. Cross-sectional imaging then answers specific questions: MRI for marrow and soft-tissue extent and local staging, CT for cortical detail and pulmonary metastases, and FDG-PET/CT for whole-body staging and response.

Do-not-touch lesions. Non-ossifying fibroma, bone island, fibrous dysplasia, osteochondroma and simple bone cyst have such typical features that biopsy is unnecessary and may cause harm. Recognising them avoids over-investigation.

The danger of acting before staging. A poorly planned biopsy or an unwary excision of an unsuspected sarcoma can convert a limb-salvageable tumour into an amputation or compromise survival. The candidate who states early that any suspicious lesion is referred intact to a sarcoma multidisciplinary team, fully staged, and biopsied along a planned track demonstrates safe practice.

What Lodwick grades. The classification grades the pattern of bone destruction, and with it the growth rate and likely behaviour, from a geographic lesion with a sclerotic rim to permeative destruction. Type I lesions (A, B, C) can often be observed or treated conservatively; Type II and III lesions require urgent investigation and often indicate malignancy or aggressive infection.

View larger

The modified scheme. The Lodwick-Madewell modification in the figure below numbers the patterns differently from the table above: ill-defined geographic destruction becomes grade II, a margin that changes on serial radiographs is IIIA, and moth-eaten and permeative destruction are IIIB and IIIC. Serial change and the zone of transition are biological observations, not merely labels.

View larger

View larger

Reading the periosteum. A solid, continuous reaction means a slow, benign process; an interrupted one means a rapid, aggressive one. Lamellated onion-skin layers record intermittent growth, the Codman triangle is the periosteum lifted at the edge of the lesion, and sunburst spiculation is the tumour outgrowing its containment. Infection and healing can mimic the periosteal reaction of an aggressive tumour, so always correlate with the clinical features.

View larger

What the tumour makes. Osteoid matrix is dense, cloud-like and amorphous, and means bone formation by the tumour; seen with aggressive features it is osteosarcoma until proven otherwise. Chondroid matrix, the rings-and-arcs or popcorn calcification, is pathognomonic of cartilage tumours, but it is seen in both the benign enchondroma and the malignant chondrosarcoma, so the matrix must be interpreted with margins, cortical integrity and clinical symptoms. Fibrous matrix is hazy ground-glass, and mixed patterns belong to dedifferentiated tumours.

View larger

Benign lesions are staged by behaviour: latent lesions are inactive and self-limiting and stay within their capsule, active lesions grow and deform but remain contained, and aggressive lesions breach the capsule and the cortex.

Malignant lesions are staged by combining histological grade (G1 low, G2 high), anatomic site (T1 intracompartmental, T2 extracompartmental) and metastasis (M0 / M1). The T is an imaging decision: MRI shows whether the tumour has breached its compartment, which, with grade and metastasis, sets the stage and the required margin. The Enneking system is increasingly supplemented or superseded by AJCC/UICC TNM, which adds tumour size and depth, for soft-tissue sarcoma staging.

Location, age and the pattern of matrix and margin identify most of the common tumours, and the two tables set out the features that make each diagnosis. Osteosarcoma is the prototype of the aggressive lesion: a metaphyseal tumour in a young patient with osteoid matrix, a sunburst periosteal reaction and a Codman triangle.

View larger

View larger

View larger

Why imaging defines the margin. Imaging does not just diagnose; it defines whether a tumour can be removed with an oncologically adequate margin, which is why MRI assessment of compartmental boundaries is mandatory. MRI-defined tumour extent, compartmental boundaries and neurovascular relations determine whether a wide margin (limb salvage) is achievable or a more radical procedure or amputation is needed. The planned biopsy track must lie within the tissue that will be excised, or it contaminates the margin and the resection must be widened.

Enneking described four margins by the plane of dissection relative to the tumour and its reactive zone.

The decisive radiographic discriminator is the zone of transition combined with periosteal reaction and cortical status, integrated with patient age. The table collapses the most common pattern-based decisions a candidate must make at the viva.

Primary bone sarcomas are rare worldwide (osteosarcoma incidence approximately 3-4 per million per year, peaking in adolescence, with a second peak in older adults often related to Paget disease or prior radiation). Ewing sarcoma is rarer still and shows marked geographic and ancestral variation, being notably less common in populations of African ancestry. Soft-tissue sarcomas are commoner than bone sarcomas (roughly 4-5 per 100,000 per year) but still account for under 1% of adult malignancies. Bone metastases vastly outnumber primary bone tumours overall and dominate the differential in patients over 40.

Reliability of radiographic classification. Despite their teaching value, the Lodwick and Enneking radiographic classifications show only poor-to-moderate interobserver agreement (kappa 0.23-0.28) and fair-to-moderate intra-observer reproducibility (0.42-0.45), and training level made no measurable difference, so this is not a gap experience closes. The accuracy figure is the one to carry: across the three systems, lesions were correctly called malignant in only 73%, 59% and 62% of cases. A framework that is right about two-thirds of the time guides thinking; it cannot be the sole basis for a benign-versus-malignant decision, which must integrate clinical data and cross-sectional imaging.

Enchondroma versus low-grade chondrosarcoma. No single imaging feature is definitive. Deep endosteal scalloping (over two-thirds of the cortical thickness), cortical destruction, a soft-tissue mass, pain and an axial or proximal location favour chondrosarcoma, but the overlap is substantial, and biopsy of cartilage tumours is itself notoriously unreliable for grade.

Whole-body MRI versus PET-CT. Both detect skip and distant disease; the optimal staging algorithm, and the place of whole-body diffusion-weighted MRI, is still evolving and varies between centres. CT chest remains mandatory for pulmonary metastases regardless, as small lung nodules may be PET-negative.

Surveillance imaging after sarcoma resection. The value of routine post-treatment MRI for detecting local recurrence is debated, and the numbers are sharper than the debate suggests. Among 366 patients, 17% developed a local recurrence; 66% were found by clinical signs and symptoms and only 34% by surveillance imaging, and the tumours found by imaging were not smaller than those found clinically (4.5 cm versus 3.9 cm), which removes the main rationale for scanning, since earlier detection is only worth having if it means smaller disease. No patient subgroup has been reliably shown to benefit preferentially from imaging surveillance over examination. The case for continuing surveillance MRI within current guidelines rests on the third of recurrences that were imaging-only, not on earlier detection.