Vetting a request, answered from the clinical question.

Confirmation and mapping of a suspected intra-abdominal collection, usually to answer whether there is a drainable target. The vetting decision is almost always about contrast, because the diagnosis depends on it.

The archetypal age- and pregnancy-forked pathway: contrast-enhanced CT for most adults, ultrasound first in children, and ultrasound then MRI in pregnancy. The same request deserves three different answers.

Whole-body CT is justified by MECHANISM and PHYSIOLOGY, not by the fact that someone has been injured. The patient with a high-energy transfer, multi-region signs or an examination that cannot be trusted — intubated, obtunded, intoxicated, or distracted by one dominant painful injury — is the patient the pan-scan exists for. The awake, stable, examinable patient who is tender in exactly one place should have that place imaged, and irradiating the rest of them is a real cost for no benefit. And in the patient who does not respond to volume resuscitation the correct answer is not a different protocol: it is theatre or interventional radiology, because a CT that delays haemostasis is the harm.

Left lower quadrant pain where the question is not only whether there is diverticulitis but whether it is complicated. Contrast-enhanced CT answers both in one pass; the vetting failure is accepting an unenhanced or oral-only study that cannot grade it.

The question is almost never "is there an obstruction" alone. It is level, cause, and whether the bowel is compromised — and only the last of those three changes what happens in the next hour.

Free gas is easy; the site of perforation is the useful answer. CT provides both, and the common vetting errors are settling for an erect chest radiograph and forgetting that a suspected anastomotic leak needs luminal contrast from the correct end.

The most time-critical abdominal CT there is, and the one most often ordered as the wrong protocol. A routine single portal-venous abdomen is not a mesenteric CT angiogram, and the difference decides whether the occlusion is seen.

The whole study rests on catching contrast leaving the vessel while it is still leaving. That needs an arterial acquisition, an unenhanced series to prove the density is new, and a later venous acquisition to show it accumulating and moving — which is why a routine single portal-venous abdomen is the wrong study, not merely a suboptimal one.

A card mostly about timing. The diagnosis is biochemical, so imaging is asked to grade necrosis and find collections — and both take days to declare themselves, which makes a day-one CT the commonest avoidable study in this disease.

The clearest ultrasound-first indication in the abdomen, and the request most often submitted as a CT. Ultrasound answers the question with no radiation, and it is the only modality that combines the imaging findings with a sonographic Murphy sign obtained at the point of tenderness.

A two-question pathway: is the jaundice mechanical, and if so at what level and from what cause. Ultrasound answers the first cheaply; the second usually needs cholangiography, and the choice between MRCP and a pancreas-protocol CT depends on whether a mass is suspected.

The indication where imaging alone makes the cancer diagnosis, and therefore the indication where the protocol is not negotiable. Arterial hyperenhancement followed by washout is the diagnosis, so a scan without those phases cannot make it, unmake it, or be salvaged by reporting.

Most of these lesions are benign and the job is to say so confidently and stop. The pre-test probability, and therefore the pathway, is set by whether the patient has a known malignancy or chronic liver disease — which is why this card is separate from the cirrhosis card even though the modalities overlap.

The indication that most clearly proves the design point: the right body part imaged in the wrong phase is a wrong study. A routine portal-venous abdomen and a pancreas-protocol CT are both "CT abdomen with contrast", and only one of them shows the tumour.

A tumour defined by its fibrous stroma, which is why its enhancement runs backwards compared with most liver lesions and why a study that stops at the portal venous phase can miss or mischaracterise it entirely.

A surveillance question rather than a diagnostic one, and one where the modality choice is driven by what the follow-up costs over a decade. Two features decide almost everything — communication with the duct and the presence of an enhancing mural nodule — and MRI shows both better than CT.

A lifelong disease of young people, which makes cumulative radiation the dominant vetting consideration and MR enterography the default. The recurring protocol failure is not the modality but the preparation: without luminal distension the study answers nothing.

A small-field-of-view, high-resolution question that a general pelvic MRI cannot answer. The surgeon needs the track classified against the sphincter complex, and that requires slices angled to the anal canal, not to the patient.

Imaging rarely names the cause of a colitis, but it does two things that change management on the day: it shows the distribution, which narrows the differential, and it finds the complications that make this an urgent surgical problem.

Two studies answer this question — CT pulmonary angiography and perfusion scintigraphy — and both are rated appropriate. Which one is right depends on the chest radiograph, on renal function and on whether the patient is pregnant. This is the card where withholding imaging does the most harm.

A nodule found by accident on a scan done for another reason. Almost every vetting decision here is about technique and interval rather than about whether to scan: an unenhanced thin-section acquisition reproducing the previous technique, at an interval set by size, morphology and risk.

Two different questions wear the same words. Acute suspected pneumonia is a radiograph question; consolidation that has not cleared after adequate treatment is a CT question, because the point is no longer the infection but what is obstructing or mimicking it.

The question is whether the pleural fluid is infected and organised, and that is answered by pleural enhancement — so this is a contrast decision above all. The corollary is that an unenhanced CT reported as "effusion, no empyema" is not an answer.

Three studies, and the vetting decision is almost never which one — it is the order. Ultrasound characterises the fluid and marks a safe needle site, which no CT can do. CT looks for the cause, and it can only do that properly while the fluid is still there: pleural nodularity is obvious outlined by effusion and nearly invisible once the pleura are drained together. So the sequencing rule that matters is CT before complete drainage, not after it.

A pattern-recognition study, not a detection study. The diagnosis is made from the distribution of reticulation, the presence of honeycombing and the behaviour of the lung in expiration — which means the supine, prone and expiratory series together are the examination, and contrast has no part in it.

A morphological diagnosis made by comparing the calibre of a bronchus with the artery running beside it. Thin-section CT is definitive; the chest radiograph is not, and a normal radiograph in a patient with a productive cough for months does not close the question.

Imaging has to answer two questions at once: what is bleeding, and where is the artery that supplies it. Both need intravenous contrast, and neither is answered by a normal chest radiograph — which is why a normal radiograph in a smoker with haemoptysis is a reason to scan, not a reason to stop.

A radiograph question in nearly every case. CT is for the situations a radiograph genuinely cannot resolve — surgical emphysema, complex bullous lung, a supine trauma patient, or planning for surgery — and not for confirming what an erect film already showed.

Two things must come out of one acquisition: the level and length of the venous obstruction, and the cause. That makes it a venous-phase study with a deliberately chosen injection strategy — the arm you inject is part of the protocol, not a nursing detail.

Echocardiography is the screening test that raises or lowers the probability; once it has raised it, imaging exists to answer one question first: is this chronic thromboembolic disease, the form that is potentially curable by surgery. Perfusion scintigraphy answers that better than CT, which is why it comes first here and nowhere else in the chest.

A staged hyperacute sequence rather than a single study: unenhanced CT to exclude haemorrhage and permit thrombolysis, angiography to find a treatable occlusion, and perfusion only where it will decide thrombectomy in the extended window.

Unenhanced CT is definitive for acute blood and needs nothing added to it. The vetting value lies downstream: deciding which patients need angiography for an underlying vascular cause, and which need delayed MRI for an underlying lesion.

Unenhanced CT first, and then a pathway whose next step is decided by the clock: within roughly six hours of ictus a negative modern CT reported by a radiologist is close to definitive, and beyond that window a negative CT must be completed by lumbar puncture or angiography.

A diagnosis that a routine unenhanced head CT misses in a substantial minority of cases, so the vetting job is to convert a plain CT request into a dedicated venographic study — by MR or by CT, which perform comparably.

The structural survey after a first unprovoked seizure is an MRI, not a CT. CT earns its place only in the emergency department, where the question is whether something needs treating tonight.

The study is the same anatomical region as a routine brain MRI and a completely different examination. What is being asked for here is a dedicated epilepsy protocol, and requesting or accepting a general brain MRI is the commonest way this pathway fails.

Contrast-enhanced MRI is the study, and the reason is countability: the number and location of lesions determines whether the patient receives stereotactic radiosurgery, whole-brain radiotherapy or neither.

Contrast-enhanced MRI with a standardised tumour protocol, because the request is simultaneously a diagnostic question and the first step of a surgical plan, and a non-standard baseline compromises every follow-up study that comes after it.

A diagnostic study whose whole purpose is to satisfy formal criteria, which makes the protocol non-negotiable: specific sequences, specific coverage, and a contrast decision that answers dissemination in time rather than reflexive practice.

Two questions arrive in one request. Contrast-enhanced MRI answers the diagnostic one; a pre-lumbar-puncture CT answers a procedural one in a defined minority — and neither may be allowed to delay antimicrobial and antiviral treatment.

The counterpart to the uncomplicated headache card. Where a red flag is present the pre-test probability of structural disease rises enough to justify imaging, and MRI is the study because it answers far more of the differential than CT.

A stable primary headache disorder with a normal neurological examination has a yield from imaging close to the background rate of incidental findings, so the scan is at least as likely to create a new problem as to solve the presenting one.

A whole-brain MRI is the wrong study for a gland that is roughly a centimetre across. The dedicated dynamic protocol exists because a microadenoma is defined by a difference in enhancement timing, not by a difference in signal.

A targeted internal auditory meatus study, not a brain scan and not a temporal bone CT. The lesion being sought is often only a few millimetres and sits inside a canal that routine brain sequences do not resolve.

The imaging question is not simply whether the ventricles are large — it is whether they are large out of proportion to atrophy, which is a pattern judgement that MRI supports and a report of "ventriculomegaly" does not.

Imaging here does two jobs: exclude the small number of structural causes that are treatable, and characterise the pattern of atrophy and vascular burden that supports a specific diagnosis. Neither needs gadolinium.

Whether to scan is decided by a validated clinical decision rule, not by the request form. Once the rule is satisfied the study is an immediate unenhanced CT, and MRI has no place in the acute assessment.

The decision rule matters more here than anywhere else, because the balance between a missed injury and a lifetime radiation risk is at its tightest. Where the rule indicates imaging the study is an unenhanced CT with paediatric parameters.

A card with a genuine paediatric fork: in adults who fail a validated clinical decision rule the first study is CT, while in children the same failure usually leads to radiographs first, because the injury pattern and the radiation stakes are both different. The obtunded adult asks a third question — not "does this patient need a scan" but "can the collar come off after a normal one" — and EAST answers it with the CT alone.

Whole-spine MRI within 24 hours, and the word that carries the clinical weight is whole. Multilevel and skip metastatic disease is common, so imaging only the symptomatic level treats the level that hurts and misses the one that will paralyse.

An emergency MRI question with one specific trap: CT of the lumbar spine cannot exclude cauda equina compression, so a CT request for this indication is a redirect rather than a protocol adjustment.

Contrast-enhanced MRI, with a whole-spine survey rather than a single level, because non-contiguous involvement is characteristic and because the finding that changes management fastest — an epidural abscess — is defined by enhancement.

Uncomplicated low back pain, with or without radicular symptoms, is self-limiting in most people, and imaging it early neither improves outcomes nor changes early management. The degenerative findings it produces are near-universal and poorly correlated with symptoms.

An acute cord syndrome is a compressive question until proven otherwise. The first study must therefore be capable of excluding compression as well as demonstrating intrinsic cord signal change, which means contrast-enhanced MRI of the whole cord rather than of the segment the sensory level suggests.

CT does not diagnose sinusitis; it maps it. Acute uncomplicated rhinosinusitis is a clinical diagnosis and imaging it changes nothing, so the two questions worth scanning for are chronic disease that has failed maximal medical therapy and is heading for endoscopic surgery, where the study doubles as the navigation roadmap, and suspected orbital or intracranial complication, where contrast, wider coverage and an MRI question all enter at once.

Two things make this different from low back pain. Thoracic pain carries a materially higher pre-test probability of malignancy, infection and fracture — degenerative thoracic disease is comparatively uncommon, so the symptom is closer to a red flag in its own right — and if there are myelopathic features the lesion is frequently not at the level the patient points to, which is an argument for imaging the whole cord rather than the painful segment.

The bony question, and the one place in the lumbar spine where CT is the right first study rather than a substitute for MRI. Adults with a significant mechanism or clinical findings go straight to thin-section CT with reformats; children go to radiographs first, because their injury pattern and their dose stakes are both different; and MRI enters afterwards for the cord, the disc and the posterior ligamentous complex, which decide stability and which CT cannot see.

A well-standardised pathway in which the vetting value lies in the details: that the chest is included, that a portal venous phase is not optional, and that an indeterminate liver lesion in this population is a resection-planning question rather than an incidental one.

A geometry problem more than a contrast problem. The answer depends on slices angled perpendicular to the tumour at high in-plane resolution, and gadolinium adds nothing — which makes "MRI pelvis with contrast" the classic mis-specified request here.

Staging is a coverage-and-phase problem. A contrast-enhanced chest CT extended to include the liver, adrenals and lower neck comes first, before any biopsy; FDG PET-CT follows for everyone who might be treated with curative intent; brain imaging is added by histology and stage rather than by symptoms.

The staging request is rarely wrong about the STUDY and frequently wrong about the PHASE, and which phase is right is a property of the primary. A single portal-venous acquisition stages most solid tumours and adding an arterial phase to those is dose and cost for no yield. A late arterial acquisition is not optional for a hypervascular primary, and renal and pancreatic primaries each need their own timing. Saying that no arterial phase is required is as much of an answer as adding one.

The question is the LIVER in a primary whose hepatic deposits are hypervascular — neuroendocrine above all, and renal cell. A portal-venous-only CT is the wrong study for that question and reports as a false negative, because the window in which the lesion is visible closes before the standard staging phase opens. The card exists to teach why the late arterial acquisition is not optional where it is needed, and equally why it is not a licence to add two abdominal passes to every staging CT.

A measurement study, not a detection study. The protocol is dictated by the baseline: same modality, same phase, same coverage, comparable slice thickness. A technically excellent scan that differs from the baseline produces numbers that cannot be compared with it.

For the FDG-avid histologies, PET-CT is the staging study and CT is the fallback. The vetting decisions are getting the sequence right relative to biopsy and treatment, and not accepting a request that expects imaging to make a diagnosis imaging cannot make.

The radiographic skeletal survey has been superseded. Modern criteria define myeloma-defining bone disease by lytic destruction on cross-sectional imaging or by focal marrow lesions on whole-body MRI, and both detect disease in patients whose radiographs are normal. Vetting a skeletal survey request usually means replacing it.

Two questions are being asked at once: where is the primary, and what is the most accessible site to biopsy. The second is usually the more useful, because histology and immunohistochemistry identify the origin more often than imaging does.

A survey question: is there skeletal disease anywhere. Bone scintigraphy remains the standard whole-skeleton survey for most solid tumours, with anatomical imaging used to characterise what it finds. The trap is a request that is really about cord compression, which is urgent and needs MRI.

Two related questions with different answers. Staging needs high-resolution local imaging matched to the subsite plus nodal and thoracic assessment. Surveillance needs a baseline at the right interval after treatment — too early and treatment change masquerades as residual tumour.

Ultrasound is the whole of the initial answer: it characterises the nodule, stratifies risk against a structured system, and directs fine-needle aspiration. Cross-sectional imaging has two narrow roles, and iodinated contrast in a patient heading for radioiodine is a sequencing decision the vetter owns.

Ultrasound characterises the mass, MRI resolves what ultrasound leaves indeterminate, and CT stages disease that is already presumed malignant. Sending an adnexal mass straight to CT inverts that order and produces a study that cannot characterise what it finds.

The measurement that drives management is depth of myometrial invasion, and it can only be made in a plane perpendicular to the endometrial cavity. Straight axial imaging systematically distorts it, so the vetting decision is as much about geometry as about contrast.

The decision imaging supports is surgery versus chemoradiotherapy, and it turns on parametrial invasion — an interruption of the low-signal cervical stromal ring. That ring is only visible on thin, small-field-of-view T2 acquired in planes referenced to the cervical canal, not to the patient.

The question is whether the lesion enhances, and enhancement is a subtraction. Any protocol without an unenhanced acquisition in matched geometry cannot answer it, which is why a routine single-phase portal-venous CT — the study most often requested — is non-diagnostic here.

The unenhanced attenuation does most of the work and comes first. A homogeneous nodule at or below about 10 HU is a lipid-rich adenoma and the study can stop there; only above that threshold does the washout protocol have anything to add — and washout cannot be calculated at all without the unenhanced number.

Imaging follows biochemistry, never precedes it. Once catecholamine excess is confirmed, the task is locating a tumour that is adrenal in most cases but can lie anywhere along the sympathetic chain — so coverage, not adrenal-specific protocol, is the vetting decision.

MRI comes before biopsy, not after it. Doing it first lets a proportion of men avoid biopsy altogether and directs the needle in those who still need one; doing it after leaves post-biopsy haemorrhage obscuring the peripheral zone for weeks.

Once cancer is proven, the questions are whether disease has left the prostate and, after treatment, where a rising PSA is coming from. PSMA PET-CT answers both more accurately than the conventional bone scan and CT combination, and does so at PSA levels where conventional imaging is blank.

Urothelial cancer is a field disease, so staging a bladder tumour means imaging the entire urothelium, not just the pelvis. A routine portal-venous CT abdomen and pelvis leaves the upper tracts unassessed and is explicitly not the staging study.

Painless visible haematuria in an adult is urothelial cancer until proven otherwise, and the study that excludes it is a CT urogram. The classic vetting failure is the request arriving as a stone protocol: an unenhanced CT KUB looks superficially like the right scan and cannot see a urothelial tumour at all.

The rare indication where contrast is not merely unnecessary but actively destroys the diagnosis: opacified urine has the same density as a stone. Low-dose unenhanced CT is the protocol, and the forks are pregnancy and childhood, where ultrasound comes first.

Uncomplicated pyelonephritis is a clinical diagnosis and needs no imaging at all. Imaging is for the complicated patient: not responding, immunocompromised, diabetic, stone disease, or a single or transplanted kidney — and there the question is drainable collection or obstruction, which requires contrast.

Ultrasound with colour Doppler is definitive for both presentations and should not be delayed or substituted. The crucial vetting point is the one the scan cannot make: clinically convincing torsion goes to theatre, and imaging must never become the reason for delay.

Two questions, in this order. First, does this patient warrant imaging at all — which means genuinely resistant hypertension, recurrent flash pulmonary oedema, deterioration in renal function after starting an ACE inhibitor or angiotensin receptor blocker, unexplained asymmetric kidney size, or young-onset hypertension where fibromuscular dysplasia is in play. Second, which test — and here the awkward fact drives the answer: the people most likely to have atherosclerotic renal artery stenosis are the people whose kidneys make iodinated contrast and gadolinium least attractive, which is what makes an operator-dependent duplex genuinely the right first study rather than a compromise.

Localisation of a pregnancy in a woman with a positive test and pain or bleeding. Transvaginal ultrasound is the whole first line; the vetting work is protecting its urgency and making sure a non-diagnostic scan is not read as reassurance.

The general entry point for acute pelvic pain in a premenopausal woman. Transabdominal plus transvaginal ultrasound with Doppler is the initial study for almost every version of this question; contrast-enhanced CT earns its place when the differential is genuinely broad or a tubo-ovarian abscess needs mapping for drainage.

A time-critical question answered by pelvic ultrasound with Doppler, where the decisive findings are morphological — an enlarged oedematous ovary with peripherally displaced follicles — and normal Doppler flow does NOT exclude the diagnosis.

Risk-stratifying an adnexal lesion so that benign disease is left alone and malignant disease reaches a gynaecological oncology service. Ultrasound assigns the risk; MRI is the problem-solver for the genuinely indeterminate lesion; CT stages, and only once malignancy is likely.

Detection and mapping of endometriosis. Specialist transvaginal ultrasound is the first-line test; MRI answers the negative or inconclusive scan in a symptomatic woman and maps disease before surgery, and it must be protocolled for endometriosis rather than as a general pelvic MRI.

Transvaginal ultrasound is the initial test for abnormal uterine bleeding at any age. In postmenopausal bleeding it functions as a triage test for endometrial sampling; in premenopausal bleeding it is a structural assessment, and endometrial thickness is not interpretable in the same way.

A palpable lump is assessed by clinical examination, imaging and — where indicated — needle biopsy, together, not sequentially. Which imaging test leads depends on age: targeted ultrasound in the dense young breast under 30, either study between 30 and 39 where the guidance genuinely does not choose, and diagnostic mammography with targeted ultrasound from about 40 onwards.

A screen-detected finding is an incomplete assessment, not a diagnosis. The pathway is targeted problem-solving mammographic views with tomosynthesis and targeted ultrasound, proceeding to image-guided biopsy where the finding persists.

Ultrasound is the first-line test for placenta accreta spectrum and, in expert hands, performs comparably to MRI. MRI complements it where the placenta is posterior, where invasion depth or lateral extension is uncertain, or where parametrial involvement is suspected.

Obstetric ultrasound with biometry, liquor volume and umbilical artery Doppler is the whole assessment. The vetting question is rarely which study, and almost always whether the interval and the Doppler components requested match the degree of concern.

Clinical scaphoid tenderness with normal scaphoid-series radiographs. MRI is the definitive next test: it shows the fracture line and the marrow oedema that radiographs cannot, and it either confirms the injury or releases the patient from immobilisation immediately.

Persistent hip or groin pain with normal or equivocal radiographs. MRI of both hips with a large field of view is the answer to both halves of the question — the undisplaced proximal femoral fracture in the elderly faller, and early osteonecrosis in the patient with risk factors.

Radiographs first — cheap, quick and occasionally diagnostic — then MRI of the symptomatic region when they are normal and the clinical suspicion persists. The site does more work than the modality: high-risk locations change management even when the fracture is barely visible.

A joint infection is diagnosed by aspirating the joint and treated by washing it out, not by imaging it. Imaging exists to find the effusion, to guide the needle and to answer the next question — is there osteomyelitis. The vetting priority is that nothing in the imaging pathway delays aspiration, antibiotics or surgical lavage: cartilage is destroyed in hours, and an MRI booked in front of the theatre list is the mechanism by which that happens.

Radiographs first, MRI of the symptomatic region when they are normal or equivocal and infection is still suspected. MRI is the test because osteomyelitis begins in the marrow, and marrow is the one compartment radiographs and CT describe poorly.

Two studies are both correct starting points, and which one leads is set by the lump rather than by a ranking. Ultrasound triages the superficial palpable lump: confidently benign, or not. A lump that is already deep to fascia, above about 5 cm, growing or painful has passed the referral threshold before anyone scans it, and the right first study is a dedicated soft-tissue mass MRI of the compartment — not a joint MRI, which has the wrong coil, the wrong planes and a field of view built around an articulation. Anything not confidently benign escalates to that MRI and a sarcoma service, not to reassurance.

Radiographs first, then ultrasound or MRI depending on the question. For the cuff itself the two are close to equivalent and ultrasound is dynamic and cheap; MRI wins when the labrum, the marrow or the deeper soft tissues are in question.

Radiographs first for the bony morphology, then a direct arthrogram for the labrum itself. MR arthrography is the reference standard; CT arthrography is its genuine substitute — and only its substitute — when the patient cannot go in the magnet or when metalwork makes the MR study non-diagnostic.

Radiographs first — they answer the fracture question and grade the arthritis that often makes MRI unnecessary — and then MRI without contrast for the meniscal, ligamentous and cartilage question in a knee whose management would change.

A time-critical study with two non-negotiable technical elements: an unenhanced series before the contrast, and coverage that runs from the thoracic inlet to the femoral arteries. The first is what makes intramural haematoma visible; the second is what defines the extent and the access.

Two different surveillance problems in one clinical thread, and one emergency that must never be answered as either. An untreated aneurysm is a diameter measurement and belongs to ultrasound; a repaired one is a leak-detection problem and needs a multiphase CT with a delayed series, because the commonest endoleak is the slowest. A painful, tender or newly symptomatic aneurysm is neither: it is a rupture until proved otherwise, and it goes to a contrast CT within the hour — or, if the patient is shocked, to theatre with the scanner not allowed to become the delay.

Anatomical imaging of the coronary arteries has become the first test for new stable chest pain in most systems. Almost every vetting decision here is about whether the patient can be prepared to give a diagnostic study: heart rate, rhythm, breath-hold and nitrate tolerance.

An unenhanced, gated, low-dose acquisition whose only job is to move a preventive treatment decision that is currently uncertain. Requested for a symptomatic patient it answers the wrong question; requested for someone already committed to treatment it answers no question at all.

Cardiac MRI is the only test that can see myocardial inflammation directly, and it does so by measuring water and by showing where gadolinium lingers. Both signals decay with time from the acute episode, which makes when the scan happens as much a part of the decision as whether it happens.

Echocardiography measures how the heart moves; cardiac MRI says what the muscle is made of. The pattern and distribution of late gadolinium enhancement is what separates ischaemic from non-ischaemic disease and points to a specific phenotype, and it also carries the arrhythmic risk information that changes device decisions.

This pathway is routed by a blood test, not by imaging. NICE NG106 measures NT-proBNP and, if it is raised, buys a transthoracic echocardiogram — the level only sets how quickly. The echocardiogram is unambiguously the first-line imaging study. The chest radiograph is not a cheaper version of it: it answers a different question, which is whether there is pulmonary congestion and whether the breathlessness is something else entirely.

Echocardiography is both the diagnosis and the grading, and the ESC/EACTS pathway assumes it in every patient. Everything beyond it exists to answer one of two narrow questions: CT supplies the annular dimensions and the access route once a transcatheter valve is genuinely being planned, and cardiac MRI supplies a regurgitant volume when the echocardiographic measurements disagree with each other or the window is poor. Neither is a better echocardiogram.

Compression ultrasound answers this question completely, immediately and with nothing injected. Almost every vetting decision is either a redirection away from cross-sectional imaging, or a timing question about how quickly the scan is needed given the Wells score and whether interim anticoagulation has been given.

Imaging is for planning an intervention, not for making the diagnosis — which is clinical and ankle-pressure based. NICE CG147 makes duplex ultrasound the first imaging test; cross-sectional angiography comes after it, once duplex has established that revascularisation is realistic or has failed to answer the question. The characteristic failure of the cross-sectional study is technical: the table outruns the bolus and normal calf vessels are reported as occluded.

Imaging exists to answer one question — is there a surgically treatable stenosis on the symptomatic side — and to answer it inside days, because the benefit of endarterectomy falls away rapidly after the index event. That deadline is what makes the most accessible adequate test the right one.

Bilious vomiting in a neonate is midgut volvulus until proven otherwise, and it is a surgical emergency: the whole small bowel hangs on a narrow mesenteric pedicle and infarcts within hours of twisting. The correct vetting output is usually "yes, now, and phone the paediatric surgeons" rather than a protocol tweak — imaging is arranged alongside the surgical review, never in front of it. The upper GI contrast study is the reference standard for the position of the duodenojejunal flexure; ultrasound is a legitimate and increasingly used first study where the expertise exists; a normal abdominal radiograph excludes nothing.

Two studies, two jobs. Ultrasound makes the diagnosis and is close to definitive in experienced hands; the air or contrast enema that follows is the definitive treatment, not a confirmatory investigation. Vetting this request therefore means booking a procedure — a child who is resuscitated, a paediatric surgeon who knows, and a room that can decompress a tension pneumoperitoneum on the spot. CT has no place in the initial assessment.

Ultrasound of the pylorus is the diagnostic test in an infant with non-bilious projectile vomiting. The vetting points are that this is an ultrasound question rather than a fluoroscopic or CT one, and that bilious vomiting is a different and more urgent problem.

A prescribed radiographic series performed as part of a multi-agency child protection process, not an ad-hoc set of films. The imaging is standardised, reported to a defined standard, and followed by a repeat survey after around two weeks because healing changes reveal injuries invisible at presentation.

Scrotal ultrasound with colour Doppler is the imaging test, performed immediately or not at all: where clinical suspicion is high the patient goes to theatre, because testicular salvage falls sharply with time and a scan that delays exploration causes the harm it was meant to prevent.