12. Musculoskeletal Anatomy

The vertebral column is read on CT not as a stack of bones but as a serially repeated, biomechanically integrated organ in which the same five elements — the anteriorly load-bearing body, the paired pedicles, the laminae, the articular processes, and the posterior spinous and transverse processes — recur at every level with predictable morphologic drift from cervical lordosis to lumbar lordosis. On axial images the vertebral body presents as a near-circular block of low-attenuation trabecular bone (cancellous marrow-filled spongiosa, typically $50$–$300\,\mathrm{HU}$ depending on mineralization and age) sheathed in a thin rind of dense cortical bone ($>700\,\mathrm{HU}$, often exceeding $1000$), and the neural arch closes posteriorly to enclose the spinal canal. The single most consequential structure to localize is the neural foramen, bounded superiorly and inferiorly by the pedicles of adjacent vertebrae, anteriorly by the posterolateral vertebral body and intervertebral disc, and posteriorly by the facet (zygapophyseal) joint and ligamentum flavum; the exiting nerve root occupies its superior aspect, so disc material or osteophyte encroaching from below preferentially compresses the root. In the lumbar spine the conjoined concept of the lateral recess — the gutter bounded by the pedicle laterally, the vertebral body and disc anteriorly, and the superior articular facet posteriorly — is where the traversing (descending) root is trapped by facet hypertrophy and disc protrusion, the anatomic basis for the dissociation between the clinically symptomatic root and the disc level. The sagittal reformation, made routine by near-isotropic MDCT, is the plane of stability assessment: it displays vertebral body height, the anterior and posterior vertebral lines, the spinolaminar line, and the alignment of the facet joints, and it is on this plane that the three-column model of Denis is applied — the anterior column (anterior longitudinal ligament and anterior two-thirds of the body and anulus), the middle column (posterior third of the body, posterior anulus, and posterior longitudinal ligament), and the posterior column (the bony arch and posterior ligamentous complex). Mechanistically, the middle column is the fulcrum of stability: a burst fracture is defined by retropulsion of the posterosuperior body fragment into the canal precisely because the axial load has failed the middle column, and the percentage of canal compromise measured on the axial image carries prognostic weight for neurologic deficit. Normal variants are clinically load-bearing and routinely misread. The lumbosacral transitional vertebra (Castellvi sacralization or lumbarization, present in roughly $15$–$35\%$ of individuals) shifts the numbering frame and the level of accelerated degeneration (Bertolotti syndrome), and miscounting it is a recognized cause of wrong-level surgery; the expert counts from a fixed landmark (the lowest rib-bearing T12 or the iliolumbar relationship) on the localizer and the coronal reformat rather than trusting a foreshortened sagittal count. The ossiculum terminale and persistent dens synchondroses mimic odontoid fracture, while the limbus vertebra — a herniation of disc material through the ring apophysis — produces a corticated marginal fragment that must not be called acute. The expert search pattern traces each cortical line continuously across all three planes, interrogates the posterior elements (where laminar and facet fractures hide in the high-contrast bone window and are systematically missed once an obvious body fracture satisfies the search), and treats any paravertebral hematoma or asymmetric prevertebral soft-tissue thickening as a signpost to an adjacent osseous or ligamentous injury. The dominant cognitive trap is anchoring on a single conspicuous compression deformity and prematurely closing the search, when the biomechanics demand that a flexion-distraction (Chance) injury be excluded by inspecting the posterior ligamentous complex and the interspinous distance — an injury whose bony component may be trivial but whose instability is complete.

The bony pelvis is most usefully conceptualized on CT as a closed, weight-transmitting ring rather than a collection of three bones, and almost every error in pelvic trauma interpretation traces to forgetting that a ring cannot break in only one place without displacement. Each innominate bone is the fusion at the triradiate cartilage of the ilium, ischium, and pubis, and these meet the sacrum posteriorly at the paired sacroiliac (SI) joints and each other anteriorly at the pubic symphysis, forming a continuous osseoligamentous loop. The mechanically critical insight, visible on the coronal reformation, is that body weight is transmitted not through the bone uniformly but along defined trabecular trajectories — from the sacrum across the SI joints into the sciatic buttress and down the supra-acetabular bone to the femoral heads — so the posterior arch (the SI joints, sacral ala, and posterior ilium) is the true load-bearing structure, while the anterior arch (pubic rami and symphysis) functions largely as a strut that maintains ring geometry. This is why posterior arch disruption, not the often more conspicuous anterior rami fractures, dictates instability and hemorrhage risk: the principal arterial source of life-threatening pelvic bleeding, the internal iliac system and especially the superior gluteal and internal pudendal branches, runs immediately adjacent to the posterior ring and SI joints, so a widened SI joint or a sacral ala fracture on the axial image is simultaneously an instability marker and a vascular-injury predictor that should direct attention to the arterial-phase images for a contrast blush. The mechanistic classifications the CT reader is reporting toward — Young and Burgess (anteroposterior compression with symphyseal diastasis and the open-book hinge on intact posterior ligaments; lateral compression with sacral impaction and horizontal rami fractures; vertical shear with cephalad hemipelvic displacement) and Tile (rotational versus vertical stability) — are simply formalizations of where and how the ring has failed. Within the ring, the acetabulum is read through the two-column construct of Letournel: the anterior column (iliopubic, including the anterior wall) and the posterior column (ilioischial, including the posterior wall), with the iliopubic (iliopectineal) and ilioischial lines on the coronal image and the posterior wall on the axial image serving as the cortical landmarks whose disruption defines the fracture pattern and surgical approach. A posterior wall fragment is the structure most often under-appreciated on a single axial slice yet most predictive of hip instability after dislocation, and it is best confirmed on a 3D volume-rendered reconstruction. Normal variants matter: the os acetabuli (an unfused acetabular epiphysis or accessory ossicle) and the persistent triradiate cartilage in adolescents mimic fracture fragments, the sacral foraminal arcuate lines must remain unbroken on the AP-equivalent coronal view (their interruption is the cardinal sign of a subtle sacral fracture), and accessory SI joint articulations and a normally sclerotic osteitis condensans ilii must be distinguished from sacroiliitis. The expert search pattern is explicitly ring-based: having found one break, the reader is obligated to interrogate the entire loop for the mandated second lesion (a contralateral rami fracture, an SI diastasis, or a sacral fracture), to trace the sacral arcuate lines and the posterior acetabular wall deliberately, and to scan the arterial phase for extravasation. Satisfaction of search is the dominant failure mode — the conspicuous superior pubic ramus fracture draws the eye and closes the case while the genuinely destabilizing posterior sacral fracture, lying in a different plane and a less scrutinized region, is missed.

The synovial joints of the appendicular skeleton are interpreted on CT as congruent articulations defined by the geometry of their opposing subchondral plates, the disposition of their capsule and intracapsular fat, and the labral or meniscal fibrocartilage that deepens and stabilizes them — with the persistent caveat that CT images the consequences of cartilage and ligament far better than those structures themselves, which is why the relevant signs are frequently indirect. The hip is the archetypal ball-and-socket joint and the prototype for cross-sectional reasoning: on the axial and coronal planes the femoral head sits congruently within the acetabular fossa, the fibrocartilaginous acetabular labrum extends the socket peripherally, and the joint capsule attaches distally along the intertrochanteric line anteriorly, so that the entire femoral neck is intracapsular. This single anatomic fact governs the clinical behavior the reader must anticipate — a femoral neck fracture is intracapsular and therefore tamponades the retinacular vessels that supply the head, predisposing to avascular necrosis, whereas an intertrochanteric fracture is extracapsular and does not, a distinction that determines whether the surgical answer is arthroplasty or fixation and one the CT report must support by precisely localizing the fracture line relative to the capsular insertion. The shoulder, optimized for mobility at the expense of bony constraint, relies on the glenoid labrum and capsulolabral complex; CT (often as CT arthrography) demonstrates the bony Bankart fragment of the anteroinferior glenoid and the Hill-Sachs impaction of the posterolateral humeral head that together encode prior anterior dislocation, the engaging geometry of which predicts recurrence. The knee is read as paired femorotibial compartments plus the patellofemoral joint, with the menisci as low-attenuation fibrocartilaginous wedges and the cruciate and collateral ligaments inferred largely from their bony avulsions — the Segond fracture (a tiny lateral tibial rim avulsion) is anatomically trivial but is a near-pathognomonic marker of anterior cruciate ligament rupture, the canonical example of an osseous finding that shifts the differential to a soft-tissue injury CT cannot directly show. The ankle is a mortise whose congruence depends on the syndesmosis; widening of the medial clear space or the tibiofibular overlap on the coronal image betrays ligamentous disruption even when the malleoli appear intact. Across all of these, the named fat pads and recesses provide the indirect vocabulary of occult injury: displacement of the elbow's anterior and posterior fat pads (the sail sign and any visible posterior fat pad) implies an effusion and hence an occult radial head or supracondylar fracture; a lipohemarthrosis with a fat-fluid level in the suprapatellar recess of the knee indicates an intra-articular fracture releasing marrow fat even when the cortical break is not yet found. Physeal and apophyseal anatomy is a lifespan variable that the reader must calibrate to age — the open physis is a low-attenuation line that must not be called fracture, Salter-Harris geometry determines growth-arrest risk, and apophyses at tendon insertions (the anterior inferior iliac spine, the ischial tuberosity, the tibial tubercle) avulse under traction in adolescents and present as corticated fragments displaced along the line of muscle pull. The expert search pattern exploits the joint's expected congruence and symmetry: the subchondral plates of a normal joint are parallel and the spaces uniform, so any focal incongruity, step-off, or asymmetric widening is interrogated, and the comparison with the contralateral side (when in field) resolves variant from lesion. The dominant cognitive biases are anchoring on the named bony fracture while neglecting the indirect effusion or fat-pad sign that signals a second occult injury, and the framing effect of a request to evaluate one structure ("rule out hip fracture") that suppresses search of the SI joints, pubic rami, and proximal femoral shaft that lie within the same field.

Every long bone is, for the purposes of CT interpretation, a set of nested compartments whose layered organization both channels and contains disease, and reading bone pathology correctly is largely a matter of asking which compartment a process originates in, which boundary it has crossed, and how fast. From outside in these are the periosteum — a thin fibrovascular membrane, normally invisible but the source of the reactive new bone whose pattern grades aggressiveness; the cortex — dense, lamellar, weight-bearing compact bone of high attenuation ($>700$–$1000\,\mathrm{HU}$) whose continuity and thickness are the substrate of fracture detection and of cortical permeation by tumor; the endosteal surface lining the cortex internally; and the medullary cavity, filled in the appendicular skeleton by a mixture of hematopoietic (red) and fatty (yellow) marrow whose attenuation is correspondingly variable — fatty marrow approaching $-30$ to $-100\,\mathrm{HU}$ while cellular red marrow sits near soft-tissue density, a gradient that follows the predictable red-to-yellow conversion with age and that makes focal marrow replacement (the substitution of low-attenuation fat by soft-tissue-density tumor or the sclerosis of blastic deposit) a detectable, if subtle, CT finding best appreciated on a soft-tissue window rather than the bone window. Superimposed on this radial architecture is the longitudinal organization into epiphysis, physis (growth plate), metaphysis, and diaphysis, each with a distinct vascular supply and therefore a distinct disease predilection: the metaphysis, with its slow sinusoidal blood flow abutting the physeal barrier, is the seat of hematogenous osteomyelitis and of many primary bone tumors, while the epiphysis, vascularly isolated across the physis in the skeletally immature, is the site of chondroblastoma and of the avascular necrosis that follows interruption of its tenuous supply. The diagnostic power of the compartment model lies in margin and periosteal analysis. The Lodwick-Madewell grading of a lytic lesion reads the transition zone between lesion and host bone as a surrogate for biological rate: a sharply marginated, sclerotic-rimmed geographic lesion (grade IA) implies slow growth and a benign process the host has had time to wall off, whereas a permeative or moth-eaten margin with a wide, imperceptible transition zone (grades II–III) implies a rate of cortical destruction that outpaces host repair and reframes the differential toward aggressive infection or malignancy. The periosteal response is read in the same idiom: a solid, thick, undulating periosteal reaction signals an indolent insult the periosteum has had time to consolidate, while a lamellated (onion-skin), spiculated (sunburst), or interrupted Codman-triangle reaction signals a process growing faster than the periosteum can ossify behind it — the imaging signature of Ewing sarcoma, osteosarcoma, and aggressive infection. Beyond the bone itself, the fascial compartments of the limb — the osseofascial spaces bounded by deep fascia and interosseous membranes — are the relevant compartments for soft-tissue disease and for the surgical staging of sarcoma, since fascial boundaries channel hematoma, abscess, and tumor along predictable planes and define what an intracompartmental versus extracompartmental lesion means for resectability, and they are the anatomic basis of compartment syndrome when raised pressure within an inextensible fascial envelope compromises perfusion. The expert search pattern is compartment-by-compartment and window-aware: the bone window is interrogated for cortical continuity, periosteal reaction, and matrix mineralization (the rings-and-arcs of chondroid versus the cloud-like density of osteoid), and the soft-tissue window is then deliberately re-read for marrow attenuation, the soft-tissue mass that betrays cortical breakthrough, and the fascial planes — because the single most common and consequential error in bone CT is reading only the bone window, on which a marrow-replacing or soft-tissue-extending lesion may be nearly invisible. The cognitive hazard is the reverse anchoring of dismissing an aggressive permeative process as osteopenia or artifact precisely because its destruction is diffuse rather than focal, the diffuseness that defines its danger being mistaken for the diffuseness of a benign global change.