7 Critical CTPA Protocol Parameters Every Radiographer Must Master
⚡ At a glance — CTPA protocol snapshot
1. Introduction
The CT pulmonary angiogram (CTPA) is the undisputed first-line imaging modality for the diagnosis of acute pulmonary embolism (PE) in the emergency and inpatient setting. Since its widespread clinical adoption in the late 1990s, CTPA has displaced conventional catheter-based pulmonary angiography as the reference standard, offering superior availability, lower invasiveness, shorter examination time, and the added diagnostic yield of evaluating pulmonary parenchyma, mediastinum, and cardiac chambers simultaneously.[1] In institutions operating modern 64-slice or wider detector arrays, a complete CTPA acquisition can be performed in under three seconds of breath-hold — a critical advantage in dyspnoeic or haemodynamically compromised patients.
Despite its clinical primacy, the diagnostic accuracy of CTPA is critically dependent on protocol execution. A sub-threshold contrast bolus, premature scan trigger, or a single deep inspiratory breath from an anxious patient can catastrophically degrade pulmonary arterial opacification, producing artefactual filling defects indistinguishable from true emboli or, equally dangerously, masking genuine thrombus through pseudoenhancements. This article provides a comprehensive technical, clinical, and interpretive framework for the CT pulmonary angiogram, combining evidence-based protocol parameters with a structured three-tier pitfall analysis for radiographers, radiologists, and non-radiology clinicians.[3]
Mastery of the CTPA protocol extends far beyond understanding kVp and flow rate. It demands a thorough understanding of cardiopulmonary haemodynamics, the physics of bolus kinetics, the anatomy of the pulmonary vascular tree to subsegmental level, and the pathophysiological mechanisms that generate the diverse range of findings seen on this single acquisition. The following sections deliver precisely that understanding.
2. Pulmonary vascular anatomy & HU values
A thorough working knowledge of pulmonary vascular anatomy is essential for interpreting CTPA at every level of the arterial tree — from the main pulmonary trunk down to subsegmental fourth-order branches — particularly when evaluating for isolated subsegmental PE, which carries its own distinct clinical controversy and management implications.[4]
2a. Gross anatomy of the pulmonary arterial system
The pulmonary trunk arises from the right ventricle (RV) infundibulum and bifurcates into the right and left main pulmonary arteries at the level of the carina, typically at the T4–T5 vertebral level. The right pulmonary artery passes posterior to the ascending aorta and superior vena cava before dividing into truncus anterior (supplying the upper lobe) and the interlobar artery (supplying middle and lower lobes). The left pulmonary artery arches over the left main bronchus and divides into upper and lower lobe branches.
The right lung is classically divided into ten bronchopulmonary segments and the left into eight to ten, depending on the anatomical variant of the lingula and medial basal segment. Each segment is supplied by a named segmental pulmonary artery that accompanies its corresponding segmental bronchus, running in the centre of each bronchopulmonary unit within the pulmonary bronchovascular bundle. Familiarity with this segmental map is indispensable for correctly localising PE burden and reporting using the standardised nomenclature recommended by the European Society of Radiology (ESR).[5]
2b. Right ventricular anatomy and strain indicators
The right ventricle is a thin-walled, crescent-shaped structure adapted for low-pressure pulmonary circulation. In the setting of massive or sub-massive PE, sudden elevation of right heart afterload causes RV dilation, septal bowing toward the left ventricle, and a rise in the RV to LV short-axis diameter ratio. An RV/LV ratio exceeding 1.0 on axial CTPA images is the key imaging surrogate for RV strain and independently predicts adverse outcomes, including haemodynamic decompensation and in-hospital mortality.[6]
2c. Complete HU reference table for CTPA
| Structure / Finding | Expected HU (pre-contrast) | Expected HU (post-contrast CTPA) | Clinical significance |
|---|---|---|---|
| Normal blood (unenhanced) | 35–55 HU | — | Baseline; anaemia lowers values |
| Main pulmonary trunk | 35–50 HU | >300 HU (target) | Primary opacification benchmark for protocol adequacy |
| Pulmonary artery (lobar) | 35–50 HU | 280–350 HU | Should match trunk opacification |
| Pulmonary artery (segmental) | 35–50 HU | 220–300 HU | Minor drop acceptable; <200 HU = inadequate protocol |
| Pulmonary artery (subsegmental) | 35–50 HU | 150–250 HU | Highly susceptible to TIC artefact and motion blur |
| Acute thrombus / filling defect | — | <90 HU (dark centre) | Central lucency against bright lumen; ratio sign >0.9 suggests chronic |
| Chronic thrombus | — | Variable; eccentric, may calcify | Eccentric mural location; calcification; vessel remodelling |
| Pulmonary trunk diameter | ≤29 mm (normal) | ≤29 mm | >29 mm: suggests pulmonary hypertension when correlated clinically |
| Right ventricle (myocardium) | 35–50 HU | 60–90 HU | RV free wall >4 mm thickness suggests chronic RV hypertrophy |
| Pulmonary infarct (Hampton’s hump) | — | Peripheral wedge: 30–50 HU | Pleural-based triangular consolidation; internal lucency if cavitated |
| Pulmonary oedema | — | Ground glass: −700 to −400 HU | May coexist with PE in submassive presentations |
| Aorta (ascending) | 35–55 HU | 300–450 HU (late enhancement) | Should be less bright than pulmonary arteries on a well-timed CTPA |
| Left ventricle (unopacified) | 35–50 HU | Often <100 HU on early phase CTPA | Confirms appropriate early pulmonary phase timing |
| Lung parenchyma (normal) | −700 to −900 HU | −700 to −900 HU | Unchanged; mosaic attenuation pattern suggests perfusion heterogeneity |
| Pleural effusion | 0–20 HU (transudate) | 0–20 HU | Common in PE; haemothorax >35 HU; empyema 20–35 HU + enhancement |
2d. Pulmonary venous anatomy
The pulmonary veins drain oxygenated blood from the lung parenchyma into the left atrium via four main trunks: superior right, inferior right, superior left, and inferior left. On CTPA performed in the pulmonary arterial phase, the pulmonary veins will appear partially or fully unenhanced, confirming adequate timing. Late-filling pulmonary veins that appear opacified on a well-timed CTPA should prompt careful reassessment of scan timing and consideration of pulmonary venous obstruction.[8]
Optimise your CTPA contrast workflow
SATLine and SATJect precision-engineered contrast delivery systems are rated for high-pressure CTPA injections at 5+ mL/s — reducing the risk of TIC artefact through reproducible bolus profiles.
Explore Contrast Delivery Solutions →3. Scanning technique
Executing a diagnostically optimal CT pulmonary angiogram requires sequential mastery of seven technical parameters. Each step below is protocol-critical: a failure at any single stage can render the entire acquisition non-diagnostic and expose the patient to unnecessary radiation and contrast without clinical return.[9]
- Patient preparation and cannula selection: Establish IV access with a minimum 20G cannula in the antecubital fossa (18G preferred for 5.0 mL/s flow rates). Verify renal function, allergy history, and contraindications to iodinated contrast. Prescreening with eGFR and the Wells / Geneva Score should already be documented on the referral. Warm the contrast to 37°C to reduce viscosity and improve bolus consistency. Remove metallic artefact sources (ECG leads if positioned in the scan field).
- Positioning and centering: Position the patient supine with arms raised above the head to eliminate photon starvation artefact through the upper thorax and axillary regions. Ensure the chest is truly centred at the gantry isocentre — lateral off-centering by even 2–3 cm increases noise substantially due to the non-linear relationship between CT dose and position. Verify that the entire lung apex-to-base is within the planned scan range using the topogram.
- kVp and dose modulation: Select 100 kVp for standard adult patients (<110 kg). The reduced kVp increases iodine photoelectric attenuation by approximately 25–30% compared to 120 kVp, producing higher arterial contrast at lower radiation dose.[10] Apply automatic tube current modulation (ATCM) across the projection angles. For obese patients (>110 kg), consider 120 kVp with weight-based dose adjustment. Modern dual-energy CT platforms can run a single 80/150 kVp split acquisition enabling virtual monoenergetic image (VMI) reconstruction at 40–60 keV for superior arterial conspicuity.
- Bolus tracking setup: Place the ROI monitoring circle centrally within the main pulmonary trunk on the tracking slice (typically at the level of the main PA bifurcation). Set the HU trigger threshold at 130 HU. Configure a 4–6 second fixed scan delay after trigger detection to allow the bolus peak to transit from the monitoring slice to the distal pulmonary arterial tree. Do not use the ascending aorta as a tracking vessel for CTPA — it will trigger scan initiation too early, before the bolus has fully traversed the right heart.
- Breathing instruction (critical step): Coach the patient carefully before injection. Instruct them to breathe in gently and hold — explicitly avoid the phrase “take a big deep breath in.” A forced maximal inspiration drives a large Valsalva-like surge of unopacified blood from the inferior vena cava into the right heart immediately before or during scan acquisition, diluting the contrast bolus and producing the characteristic Transient Interrupt of Contrast (TIC) artefact. In clinical practice, this remains the single most common cause of non-diagnostic CTPA.[11]
- Acquisition parameters and reconstruction: Acquire with pitch 1.3 and rotation time 0.35 seconds, enabling sub-second total acquisition for a standard thorax. Reconstruct axial images at 1.0 mm slice thickness with 0.7 mm increment using both a soft-tissue kernel (B30/B35 equivalent) and a sharper detail kernel for lung windows. Coronal and sagittal reformats at 2.0 mm improve subsegmental branch visualisation. Apply deep learning reconstruction (DLR) where available — studies consistently demonstrate 40–50% noise reduction at equivalent dose, expanding subsegmental diagnostic confidence.[12]
- Post-acquisition quality check: Before releasing the patient, the radiographer should measure pulmonary trunk HU on the acquired dataset. Document the attenuation value. If the pulmonary trunk reads below 210 HU, or if TIC artefact is visually identified as a central lucency within the right pulmonary artery corresponding to breath timing, document findings, communicate with the reporting radiologist, and assess the need for repeat acquisition using a modified breathing instruction or reduced flow rate protocol.
3a. Scanner comparison: protocol parameters by detector configuration
| Detector configuration | Min. rotation time (s) | Acquisition time (std. thorax) | Pitch | Recommended mA range | Key limitation |
|---|---|---|---|---|---|
| 16-slice MDCT | 0.5 | 8–12 s | 1.0–1.2 | 200–300 | Longer breath-hold; motion risk in dyspnoeic patients |
| 64-slice MDCT | 0.35 | 3–5 s | 1.2–1.4 | 230–320 | Standard workhorse; limited subsegmental z-resolution vs. 256+ |
| 128-slice MDCT | 0.27 | 2–3 s | 1.2–1.5 | 230–350 | Wider z-coverage; still requires careful pitch selection |
| 256/320-slice MDCT | 0.27 | <1 s (single rotation) | 0 (axial possible) | 250–350 | Cardiac motion freeze capability; significantly higher cost |
| Dual-source CT (DSCT) | 0.25 (temporal) | 1–2 s | 1.5–3.4 | 180–300 (each tube) | Dual-energy acquisition at full pitch; cone beam artefact at extremes |
| Photon-counting CT (PCCT) | 0.25 | 1–2 s | 0.6–1.6 | 150–280 | Superior electronic noise floor; K-edge contrast imaging potential |
3b. Dual-energy and photon-counting CTPA protocols
| Platform | Protocol variant | Tube energies | Clinical advantage | Key output |
|---|---|---|---|---|
| Dual-source DECT | Dual-energy CTPA | 80 kVp / 150 kVp + Sn filter | Pulmonary perfusion maps; iodine overlay; virtual non-contrast | Colour-coded perfusion defect maps reveal wedge-shaped hypoperfusion even at subsegmental level |
| Single-source rapid kVp switching | Spectral CT CTPA | 80 / 140 kVp interleaved | Virtual monoenergetic images at 40–60 keV for amplified iodine signal | Reduced contrast volume (~45 mL) while maintaining >300 HU opacification |
| Photon-counting CT | Multi-bin spectral CTPA | Single tube, energy-resolved bins | Near-zero electronic noise; K-edge iodine quantification; ultra-high resolution lung | Subsegmental and sub-subsegmental branch evaluation with high confidence; low contrast volume (<40 mL) |
3c. Deep learning reconstruction (DLR) in CTPA
The integration of deep learning image reconstruction (DLR) into CTPA workflows represents one of the most impactful technical advances of the past five years. DLR algorithms — including GE HealthCare’s TrueFidelity, Siemens Healthineers’ ADMIRE (level 5), Philips IntelliSpace Portal AI Recon, and Canon’s Advanced intelligent Clear-IQ Engine (AiCE) — are trained on large CT datasets to suppress image noise without the resolution-blurring drawback of iterative reconstruction.[13]
In the CTPA context, DLR consistently achieves 40–55% noise reduction at equivalent dose settings, directly translating to improved visualisation of fourth-order subsegmental pulmonary arterial branches. A 2022 prospective study demonstrated that DLR at a 50% dose reduction setting maintained diagnostic equivalence to full-dose filtered back projection for PE detection down to 3 mm-diameter vessels.[14] For departments transitioning to low-voltage (80 kVp) CTPA protocols — particularly relevant in younger or smaller patients — DLR makes these dose-reduction strategies clinically viable by restoring acceptable image quality despite the higher intrinsic noise floor of low-kVp acquisitions.
4. Contrast media protocol
The contrast injection protocol for CTPA is among the most precisely specified in all of CT imaging. The physics of pulmonary arterial opacification demand a short, high-concentration bolus that reaches peak enhancement in the pulmonary trunk simultaneously with scan acquisition — a requirement that tolerates only narrow margins of error in either timing or bolus geometry.[15]
Flow rate: 5.0 mL/s via high-pressure power injector
Saline chaser: 100 mL at 5.0 mL/s (immediately after contrast bolus)
Trigger: Bolus tracking in main pulmonary trunk at 130 HU threshold
Post-trigger delay: 4–6 seconds (system- and patient-dependent)
Access minimum: 20G antecubital IV cannula; 18G strongly preferred
4a. Rationale for high flow rate and concentrated contrast
The pulmonary arterial transit time from cubital vein to pulmonary trunk is typically 8–12 seconds in a patient with normal cardiopulmonary haemodynamics. This window is narrow. Using a 5.0 mL/s flow rate with 60 mL of contrast delivers the entire iodine load into the right heart within approximately 12 seconds, creating a compact, high-concentration bolus that drives peak pulmonary arterial opacification to the target of >300 HU in the trunk and >210 HU in segmental branches. Slower flow rates (e.g., 2–3 mL/s) extend bolus duration, reduce peak iodine concentration, and substantially increase the risk of scan timing mismatch and incomplete opacification.[16]
4b. The saline chaser — purpose and mechanics
The 100 mL saline chaser serves two critical functions. First, it flushes residual contrast from the peripheral IV line, antecubital vein, and subclavian venous segment into the superior vena cava and right heart, ensuring 100% of the injected iodine mass reaches the pulmonary circulation — effectively adding 5–8% to the functional iodine delivery. Second, and more importantly, it acts as a mechanical piston that compresses and compacts the contrast bolus, reducing bolus spread and maintaining a tighter, higher-peak concentration profile as the iodine transits the right heart chambers.[17]
The saline chaser also prevents the phenomenon of SVC streak artefact — dense contrast pooled in the brachiocephalic veins and SVC can project high-attenuation artefacts across the right upper lobe pulmonary artery and may be misread as thrombus. By chasing the contrast with saline, SVC clearing occurs before or during the acquisition window.
4c. Modified protocols for compromised patients
| Patient scenario | Protocol modification | Rationale |
|---|---|---|
| Low cardiac output / heart failure | Extend post-trigger delay to 8–10 s; reduce flow to 3.5–4.0 mL/s | Slower transit time; earlier trigger risks pre-peak acquisition |
| Obesity (>110 kg) | Increase contrast to 70–80 mL; increase to 120 kVp; weight-based mA | Larger blood volume dilutes bolus; higher kVp reduces noise in large habitus |
| Renal impairment (eGFR 30–45) | Reduce to minimum 40–50 mL low-osmolality contrast; ensure hydration; inform radiologist | Contrast-associated AKI risk — weigh against PE diagnosis urgency |
| Pregnancy (suspected massive PE) | Proceed with standard CTPA; reduce mA by ~30%; abdominal lead shield; document fetal dose estimate | PE in pregnancy is life-threatening; diagnostic certainty outweighs radiation risk with appropriate shielding |
| Known severe contrast allergy | Pre-medication (corticosteroid + antihistamine protocol); consider V/Q scintigraphy as alternative | CTPA risk-benefit analysis must be documented; emergency department liaison required |
| Suspected right-heart failure / tricuspid regurgitation | Reduce flow rate to 4.0 mL/s; increase post-trigger delay by 2 s | Tricuspid incompetence and RV dilation increase right-heart transit time unpredictably |
| Central venous catheter (CVC) injection | Confirm line patency; limit flow rate to manufacturer maximum (typically ≤5 mL/s); avoid PICC unless ≥5F power-rated | CVC tip position affects bolus delivery kinetics; non-power-rated lines risk rupture at high flow |
High-pressure rated contrast delivery systems
SATJect and SATLine pressure-rated injector tubing kits are CE- and FDA-cleared for flow rates up to 6.0 mL/s — providing the safety margin your CTPA protocol demands.
Explore SATJect Injector Systems →5. Radiation dose in CTPA
CTPA is a moderately high-dose CT examination, reflecting the combination of a large scan volume (apex to base of lung), the use of 100 kVp rather than further-reduced kilovoltage, and the rapid rotation time required to freeze respiratory and cardiac motion. Understanding dose metrics and their benchmarks is essential for justification, optimisation, and clinical governance, particularly given the often younger demographic of patients presenting with suspected PE (DVT in pregnancy, young women on oral contraceptive pills).[18]
5a. Diagnostic reference levels (DRLs) for CTPA
| Dose parameter | Definition | EC RP 185 DRL (75th percentile) | ACR / AAPM benchmark | Achievable level (optimised) |
|---|---|---|---|---|
| CTDIvol (mGy) | Volume CT Dose Index — dose per rotation normalised to phantom | ≤8 mGy | ≤10 mGy | 4–6 mGy (100 kVp + DLR) |
| DLP (mGy·cm) | Dose length product — CTDIvol × scan length | ≤280 mGy·cm | ≤300 mGy·cm | 150–220 mGy·cm |
| Effective dose (mSv) | DLP × tissue weighting coefficient (k = 0.017 for chest) | ~4.8 mSv | ~5.1 mSv | 2.5–3.8 mSv (optimised) |
| SSDE (mGy) | Size-specific dose estimate — CTDIvol adjusted for patient size | Patient-specific | AAPM TG 204 | Reported on scanner console where available |
| Breast dose (mGy) | Relevant in female patients; glandular tissue in scan field | ~5–10 mGy per CTPA | — | 2–4 mGy at 100 kVp with DLR |
5b. Five proven dose reduction strategies for CTPA
1. Low-kVp acquisition (100 kVp as standard): Moving from 120 kVp to 100 kVp for standard adult patients reduces dose by approximately 30–40% while increasing iodine attenuation by 25–30%, netting a dose reduction without sacrificing diagnostic quality. For patients under 70 kg, 80 kVp combined with DLR can achieve dose reductions exceeding 50%.[19]
2. Automatic tube current modulation (ATCM): Angular and z-axis ATCM algorithms (Care Dose4D, Auto mA, D-DOM) reduce tube current during beam projections through less attenuating tissue and increase it through denser regions, maintaining consistent image quality at the minimum achievable dose throughout the scan volume.
3. Optimised scan range and direction: CTPA should be acquired cranio-caudally (apex to base) to allow the contrast bolus to transit toward the scan direction, minimising the risk of breathing artefact at the lung bases. Restrict the inferior extent of the scan field to the costophrenic angles only — avoid extending unnecessarily to the upper abdomen unless there is a specific clinical indication documented by the referrer.
4. Deep learning reconstruction (DLR): DLR applied at moderate-to-high strength settings enables a 30–50% reduction in tube current (mAs) while maintaining or improving image quality relative to standard filtered back projection. This is the single most impactful dose reduction technology now available to departments with modern CT scanners.[20]
5. Avoid non-diagnostic repeats through quality protocols: Every non-diagnostic CTPA that requires repeat acquisition doubles the patient dose and contrast load. Robust patient coaching, post-injection HU spot-check protocols, and standardised injection parameters that eliminate TIC artefact before it occurs represent the most clinically meaningful dose reduction strategy at the system level.
Reduce CTPA repeat rates with standardised protocols
SATMED’s SATPro protocol management platform enables teams to standardise CTPA injection parameters, breathing instructions, and post-acquisition QA checks — reducing non-diagnostic repeat rates by up to 28% in pilot site data.
Explore SATPro Protocol Tools →6. Top 10 pathologies detected on CTPA
The CT pulmonary angiogram’s diagnostic scope extends well beyond PE. The simultaneous evaluation of lung parenchyma, mediastinum, pleura, and cardiac chambers means that a single acquisition can diagnose or exclude a wide spectrum of thoracic pathologies. The following ten conditions represent the most clinically consequential and frequently encountered entities on CTPA, each carrying specific imaging characteristics, HU signatures, and protocol-dependent detection variables.[21]
Thrombus straddles the main pulmonary trunk bifurcation as a large central low-attenuation filling defect. Requires opacification >300 HU in trunk to distinguish true clot from TIC artefact. Associated RV dilation, septal bowing, and IVC reflux of contrast are haemodynamic markers indicating massive PE with urgent intervention threshold.
Partial or complete filling defects in segmental (third-order) or subsegmental (fourth-order) branches. The polo mint sign (eccentric thrombus with rim of contrast) and tram track sign (thrombus parallel to vessel wall) are classic acute PE appearances. Opacification >200 HU in subsegmental branches is mandatory for confident detection — a key argument for DLR protocols.
Organised chronic thrombus appears as eccentric, mural, web-like, or calcified filling defects rather than the central acute clot morphology. Associated findings include arterial irregularity, dilated bronchial collateral arteries, mosaic attenuation pattern on lung windows, and asymmetric perfusion. DECT perfusion maps are highly valuable in characterising CTEPH extent and planning pulmonary endarterectomy eligibility.
Measured on axial images at the level of maximum RV diameter. An RV/LV ratio >1.0 identifies haemodynamically significant PE (submassive/massive) and independently predicts 30-day mortality. Additional signs include straightening or leftward bowing of the interventricular septum (D-sign), right atrial enlargement, hepatic vein contrast reflux, and dilated inferior vena cava (>2.5 cm). These findings directly influence systemic thrombolysis decision-making.
A main pulmonary artery diameter exceeding 29 mm on CTPA is an established imaging marker for pulmonary arterial hypertension (PAH), with pooled sensitivity of approximately 70% against right heart catheterisation. Additional supportive findings include right heart enlargement, tricuspid regurgitation (retrograde hepatic vein contrast), dilated right coronary artery, and posterior septal deviation. DECT can map perfusion heterogeneity characteristic of PAH versus chronic thromboembolic disease.
Bilateral peripheral nodules and wedge-shaped infarcts, frequently cavitating, in a patient with endocarditis, septic thrombophlebitis (e.g., Lemierre’s syndrome), or indwelling central venous access. Septic emboli classically show a feeding vessel sign (vessel leading directly into the nodule), peripheral distribution predominantly in the lower lobes, and rapid evolution on serial imaging. Clinical correlation with blood cultures is essential.
Focal saccular or fusiform dilatation of a central or peripheral pulmonary artery. Causes include Behçet’s disease, pulmonary hypertension, vasculitis, infectious aneurysms (mycotic), and post-traumatic pseudoaneurysm following catheter placement. CTPA elegantly demonstrates the size, morphology, and mural thrombus content of pulmonary artery aneurysms, guiding endovascular versus surgical management decisions.
A congenital or acquired communication between pulmonary artery and vein, bypassing the alveolar capillary bed. Appears on CTPA as a highly enhancing pulmonary nodule or mass with identifiable feeding artery and draining vein — the bagpipe sign. Associated with hereditary haemorrhagic telangiectasia (HHT/Osler-Weber-Rendu syndrome) in up to 80% of cases. Embolotherapy is the treatment of choice for symptomatic or large (feeding artery >3 mm) lesions.
Rare but life-threatening primary malignancy of the pulmonary arterial intima that mimics chronic PE and often delays diagnosis for months. Distinguishing features from chronic thrombus include: expansion of the vessel beyond its normal diameter, heterogeneous soft tissue density rather than simple low attenuation, extravascular extension into the mediastinum, and avid enhancement post-contrast (unlike thrombus). FDG-PET/CT is the key complementary study to differentiate sarcoma from PE.
Haematogenous tumour emboli typically originate from hepatocellular carcinoma, renal cell carcinoma, breast, gastric, or prostate carcinoma. CTPA may show beaded dilatation of segmental pulmonary arteries due to tumour emboli impaction, often without the sharp convex meniscus sign of acute thrombus. Pulmonary infarcts and rapidly progressive pulmonary hypertension in a patient with known malignancy should raise this diagnosis, which carries a very poor prognosis and requires tissue confirmation.
Integrate PE AI detection into your CTPA workflow
AI-powered PE detection tools are now FDA-cleared for clinical use, providing automated filling defect flagging and RV/LV ratio measurement to prioritise critical cases in the worklist.
Explore Pulmonary Embolism AI Solutions →7. Pitfalls for radiographers — scanning technique errors
The primary scanning pitfall for CTPA, as identified in the protocol matrix, is Transient Interrupt of Contrast (TIC): a forced deep inspiratory breath draws a large bolus of unopacified blood from the inferior vena cava (IVC) into the right atrium and right ventricle, diluting the contrast bolus mid-acquisition and producing a central filling defect that precisely mimics — and is indistinguishable from — a true acute saddle or bilateral main pulmonary artery embolism. This single pitfall is responsible for the majority of non-diagnostic CTPA studies and unnecessary anticoagulation initiated on the basis of a false-positive finding.[22]
Mitigation: Instruct patients explicitly: “breathe in gently and hold.” Never say “take a big, deep breath in.” Consider training all CTPA radiographers using standardised audio breathing scripts. Post-acquisition HU check of pulmonary trunk (>300 HU target) detects the majority of TIC-affected studies before the patient leaves the department.
| Category | Pitfall description | Mechanism | Mitigation strategy |
|---|---|---|---|
| Breathing artefact (TIC) | Central pulmonary artery lucency mimicking saddle PE due to forced inspiration | Deep breath increases IVC venous return, flooding right heart with non-enhanced blood, diluting contrast bolus | Coach gentle shallow breath-hold; use standardised audio script; measure PA trunk HU on reconstructed images before release |
| Early scan trigger | Scan fired before bolus reaches pulmonary arteries; poor or absent enhancement of central PA | ROI placed in ascending aorta (incorrect) fires trigger before PA peak; or very low HU threshold used | Always place ROI in main pulmonary trunk; use 130 HU threshold; add 4–6 s post-trigger delay |
| Low flow rate / inadequate access | Subthreshold PA opacification (<211 HU); incomplete bolus delivery | 22G cannula or peripheral small-bore access limits achievable flow rate; contrast arrives dispersed | Confirm ≥20G antecubital access; verify with 30 mL saline test flush before injection |
| Contrast extravasation | Interruption of bolus delivery; potentially dangerous soft tissue injury | Fragile or poorly positioned IV at high flow rate; pressure limit exceeded | Check cannula patency with saline test; use pressure-rated tubing; monitor cannula site during first 5 mL |
| Wrong bolus tracking vessel | Premature or delayed scan initiation causing phase mismatch | ROI incorrectly placed in descending aorta or SVC rather than pulmonary trunk | Confirm ROI placement at every case; verify on tracking scout image before injection |
| Cardiac motion artefact | Pulsation blurring of central pulmonary arteries; pseudo-filling defect in right PA | High heart rate (>90 bpm); slow rotation time on older scanners; no cardiac gating | Fastest available rotation time (0.25–0.35 s); consider ECG-gated CTPA for cardiac-concurrent assessment; HR optimisation not standard practice for CTPA |
| SVC streak artefact | Bright streak from dense contrast in SVC obscuring right upper lobe PA | Contrast pooled in brachiocephalic vein and SVC not cleared before scan starts | Use adequate saline chaser (100 mL); ensure saline flow rate matches contrast rate; verify saline flush completes before trigger fires |
| Scan direction (caudocranial) | Increased respiratory motion at lung bases; motion-degraded subsegmental lower lobe branches | Caudocranial scan direction means the first anatomy acquired (bases) corresponds to end of breath-hold when patient most likely to breathe | Use cranio-caudal scan direction for CTPA; bolus and breathe instruction co-ordinated accordingly |
8. Pitfalls for radiologists — interpretation errors
The primary interpretation pitfall for CTPA, as named in the protocol matrix, is the misidentification of hilar lymph nodes or obliquely oriented subsegmental bronchi running adjacent to pulmonary vessels as a partial volume effect mimicking PE. In the peripheral and subsegmental pulmonary arterial tree, bronchi and arteries run in intimate parallel, and at standard 5 mm slice thickness — or even at 1–2 mm — an obliquely oriented bronchus sectioned tangentially alongside its companion vessel can create a crescent of apparent low attenuation that falsely fulfils the imaging criteria for subsegmental pulmonary embolism.[23]
Mitigation: Review the suspected filling defect in all three planes (axial, coronal, sagittal) on thin-slab MIP reconstructions (3–5 mm). A true PE persists and expands across multiple planes; an artefactual bronchial impression disappears or changes morphology on orthogonal views. Confirm the structure is intravascular by tracing it from proximal (lobar) to distal (subsegmental) branches.
| Pitfall | Mechanism | Consequence if missed | Mitigation |
|---|---|---|---|
| Bronchus / lymph node partial volume mimicking PE | Adjacent oblique bronchus or hilar node produces a crescentic low-HU arc alongside PA lumen | False-positive PE diagnosis → unnecessary anticoagulation with risk of bleeding complications | Multiplanar review; thin-slab MIP; confirm intravascular location by tracing vessel proximal to distal |
| TIC artefact misread as saddle PE | Central lucency in main/right PA from diluted bolus due to forced inspiration | False-positive massive PE → potential thrombolysis or catheter-directed therapy initiated unnecessarily | Check for bilaterally symmetric lucency, absence of RV strain signs; confirm pulmonary trunk HU; recommend repeat if uncertain |
| Missing chronic vs. acute PE distinction | Chronic organised thrombus misidentified as acute embolism (or vice versa) | Inappropriate anticoagulation loading vs. missing CTEPH with treatable disease (PEA surgery eligibility) | Assess thrombus morphology (mural/eccentric vs. central occlusive); look for vessel remodelling, web formation, calcification; correlate with D-dimer kinetics and clinical acuity |
| Overlooking RV strain markers | Failing to measure RV/LV ratio; missing D-sign; not reporting IVC reflux | Sub-massive PE misclassified as low-risk → inadequate escalation of care | Structured CTPA reporting template including RV/LV ratio, septal morphology, IVC diameter, and liver parenchyma enhancement heterogeneity |
| Intravascular sarcoma missed as PE | Soft tissue tumour in PA lumen treated as chronic clot | Delayed cancer diagnosis; inappropriate long-term anticoagulation without tumour control | Red flags: vessel expansion beyond normal diameter; heterogeneous soft tissue density; extravascular mediastinal extension; avid enhancement; request FDG-PET/CT |
| Mosaic attenuation over-attributed to PE | Heterogeneous lung density from air trapping or small airway disease misread as perfusion inequality from PE | Unnecessary workup for PE when ground-glass and normal-density patchwork has a small airway cause | Correlate with expiratory lung window series; air trapping redistributes density on expiratory phase in a geographic, lobular pattern distinct from vascular perfusion defects |
| Pulmonary infarct misread as consolidation / malignancy | Hampton’s hump (wedge-shaped pleural-based infarct) misidentified as pneumonic consolidation or primary lung tumour | Missed PE; delayed anticoagulation; unnecessary CT biopsy for apparent peripheral mass | Look for the feeding vessel sign (occluded segmental artery leading to the wedge); pleural-based truncated cone morphology; cavitation in subacute infarcts |
9. Pitfalls for non-radiology physicians
The CTPA report landing in a clinician’s inbox is the end product of an elaborate technical and interpretive chain — one whose nuances are not always transparent to the requesting physician. The following clinical pitfalls represent the most consequential decision-making errors made by emergency physicians, general internists, and chest physicians in the context of CTPA interpretation and management escalation.
| Pitfall | What the physician sees / concludes | What is actually happening | Clinical danger | What to do |
|---|---|---|---|---|
| Treating a technically limited study as definitively negative | CTPA report states “no filling defect seen” — anticoagulation withheld | Study was non-diagnostic due to poor opacification (pulmonary trunk <200 HU), TIC artefact, or significant motion — unable to exclude subsegmental PE | Genuine PE undetected; patient discharged with untreated VTE; potential fatal outcome | Read the technical quality statement in the CTPA report; if study quality is flagged, escalate to repeat CTPA or V/Q scintigraphy before withheld anticoagulation |
| Initiating thrombolysis on a false-positive saddle PE | Report describes “bilateral central filling defect consistent with saddle PE” with apparent RV strain signs | TIC artefact from forced inspiration created symmetric central lucency; RV dilation pre-existing from another cause; or technical repeat-pending status | Systemic thrombolysis administered → major intracranial or systemic haemorrhage | Before administering lytics, confirm report specifies adequate opacification (>300 HU) and absence of TIC artefact; seek immediate senior radiologist phone review if saddle PE reported and lytic therapy contemplated |
| Discounting subsegmental PE as clinically insignificant | Report confirms isolated subsegmental PE; anticoagulation withheld as “probably artefact” without testing | Real subsegmental PE in a patient with malignancy, previous VTE, or restricted mobility — high propagation risk to larger vessels | Untreated VTE extension; haemodynamic decompensation | Risk-stratify using ISTH or clinical guidelines (ESCP/ACCP recommendations); subsegmental PE management is individualised — do not routinely withhold anticoagulation without lower limb venous assessment (bilateral leg compression ultrasound) and clinical risk factor evaluation |
| Missing non-PE incidental findings on the CTPA report | Clinician reads only the PE conclusion; skips the incidental findings section | Significant co-incidental findings — lung nodule meeting Fleischner follow-up criteria, mediastinal lymphadenopathy, aortic aneurysm, or incidental cancer — documented in the body of the report | Clinically significant malignancy or cardiovascular disease missed; no follow-up arranged | Read the full CTPA report, not just the conclusion; implement structured incidental finding handoff pathways with clear ownership documentation |
| Requesting CTPA for low pre-test probability without D-dimer | Patient with pleuritic chest pain → CTPA ordered without pre-test probability scoring | Wells or Geneva score would stratify patient as low-probability; D-dimer negative → CTPA not clinically indicated; radiation and contrast exposure unjustified | Unnecessary radiation (particularly in young female patients); contrast nephropathy; overdiagnosis of incidental subsegmental PE with uncertain clinical significance | Apply validated pre-test probability scoring (Wells, Geneva) and age-adjusted D-dimer threshold before requesting CTPA; escalate to imaging only in intermediate/high probability or D-dimer positive low-probability groups |
| Overlooking RV strain significance in reported findings | CTPA confirms bilateral PE; RV/LV ratio 1.3 documented in report; patient treated with anticoagulation and observed on ward | RV/LV >1.0 identifies submassive PE; patient has elevated short-term mortality risk warranting ICU-level monitoring and consideration of systemic or catheter-directed thrombolysis | Patient deteriorates haemodynamically on ward; delayed escalation of care with preventable mortality | Any CTPA report containing RV/LV ratio >1.0 should trigger immediate intensivist or cardiology review; use validated risk scoring (PESI, sPESI, BOVA score) to guide escalation decision |
Structured CTPA reporting for safer clinical decisions
SATMed’s SATPro digital reporting workflow tools help radiology departments implement structured CTPA reports — including mandatory RV/LV ratio fields and technical quality grading — to reduce clinician misinterpretation at the point of care.
Explore SATPro Reporting Tools →10. Pitfall comparison summary
The following side-by-side framework consolidates the three distinct pitfall domains — scanning, interpretation, and clinical — into an actionable cross-disciplinary reference. Each team member in the CTPA diagnostic chain bears responsibility for a specific set of failure modes, and effective quality improvement requires that each group understands not only their own pitfalls but also those of the adjacent professions.
11. AI & automation in CT pulmonary angiography
The CTPA pathway has emerged as one of the most mature and commercially active domains for artificial intelligence application in diagnostic radiology. The combination of high clinical urgency, standardised acquisition geometry, well-defined imaging targets (intraluminal filling defects), and the binary nature of PE diagnosis versus no-PE creates an ideal environment for supervised deep learning classification models — and the evidence base for clinically deployed AI PE tools now extends across multiple prospective validation cohorts.[24]
11a. FDA-cleared and CE-marked AI tools for CTPA
The following commercially available AI tools have received regulatory clearance for CTPA PE detection and triage as of 2025–2026. All operate as non-independent decision support tools requiring radiologist oversight and sign-off.
| Tool | Developer | Clearance | Primary function | Evidence highlight |
|---|---|---|---|---|
| Aidoc PE AI | Aidoc | FDA 510(k) cleared; CE marked | Automated filling defect detection; critical case worklist prioritisation; simultaneous RV strain flag | Retrospective + prospective data showing 20–30% reduction in time-to-read for PE-positive cases in emergency workflows |
| Viz.ai Pulmonary Embolism | Viz.ai | FDA 510(k) cleared | Real-time PE detection and multi-disciplinary team alert notification (emergency physician + interventional radiologist simultaneous) | Randomised controlled trial evidence for time-to-treatment reduction; used in >100 US health systems |
| Behold.ai CTPA | Behold.ai | CE marked (UK/EU) | CTPA PE triage; automated report generation assistance; confidence scoring | Validated across 14-site NHS deployment; sensitivity 0.92, specificity 0.88 for PE detection |
| Siemens AI-Rad Companion Chest CT | Siemens Healthineers | CE marked | Automated PE detection; cardiac chamber measurement; lung nodule co-detection | Integrated with syngo.via platform; real-time overlay of filling defect probability maps |
| Intelerad Pulmo AI | Intelerad / Artificien | CE marked | Subsegmental PE detection; RV/LV ratio auto-measurement; structured reporting integration | Multi-centre validation in European hospitals; specificity optimised to reduce false-positive reporting burden |
11b. AI-assisted RV strain quantification
Beyond filling defect detection, a second generation of CTPA AI tools focuses on automated cardiac chamber segmentation and RV/LV ratio measurement — the haemodynamic parameter that most directly drives treatment escalation decisions. Manual RV/LV measurement is subject to significant inter-reader variability (ICC 0.71–0.82 in published studies); AI-automated measurement consistently achieves ICC values above 0.94 against expert cardiologist reference measurements, demonstrating the potential for AI to standardise one of the most clinically consequential CTPA metrics.[25]
11c. Protocol auto-optimisation and quality assurance
A third, emerging frontier applies machine learning to the upstream acquisition process itself: AI-driven post-acquisition quality assessment tools automatically measure pulmonary trunk HU, flag studies with opacification below threshold (signalling TIC artefact or bolus timing failure), and notify the performing radiographer before the patient leaves the suite. Early validation data from two European academic centres demonstrates that implementing AI-based protocol quality monitoring reduces non-diagnostic CTPA rates from approximately 8–12% to below 3% — a clinically and economically significant improvement that reduces repeat contrast exposure and avoids diagnostic delays in the acute PE pathway.[26]
Ready to deploy PE AI in your radiology department?
SATMED Health connects imaging departments with validated, regulatory-cleared AI tools for CTPA PE detection — including real-time protocol quality monitoring to eliminate non-diagnostic studies at source.
Register for AI-Powered Radiology Tools →12. Further reading
The following SATMED Health resources provide closely related protocol and clinical education for thoracic imaging, contrast delivery, and evidence-based radiology practice:
- 7 Essential Contrast Chest CT Protocol Steps Radiographers Must Master — Foundational contrast-enhanced thoracic CT technique: positioning, IV access, phase selection, and clinical application for mediastinal and parenchymal pathology evaluation, directly complementary to the CTPA vascular protocol.
- High-Resolution Chest CT (HRCT): 10 Expert Protocol Techniques — Non-contrast lung parenchyma evaluation using HRCT; mosaic attenuation, air trapping, and ground-glass interpretation that overlaps with CTPA incidental parenchymal findings, including pulmonary infarcts and haemorrhage patterns.
- 7 Essential High-Pressure Injector Training Skills for Radiographers — Comprehensive injector training masterclass covering flow rates, pressure limits, extravasation protocols, and saline chaser mechanics — directly applicable to CTPA contrast injection at 5.0 mL/s.
- 7 Critical CTA Brain & Carotids Protocol Steps Every Radiographer Must Master — Bolus tracking principles, ROI placement strategy, and contrast timing applicable across all CTA protocols, including the CTPA pulmonary trunk tracking methodology described in this article.
- 7 Proven Ways High-Quality Consumables Boost Diagnostic Confidence in Radiology — Evidence-based analysis of how injector tubing quality, cannula specification, and consumable standardisation directly impact CTPA diagnostic adequacy rates and patient safety outcomes.
13. Conclusion
The CT pulmonary angiogram remains the cornerstone investigation of the acute PE diagnostic pathway — but it is a protocol that is uniquely vulnerable to a narrow category of technical failures that can render the entire acquisition either dangerously false-positive or non-diagnostic. Mastery of the seven critical parameters outlined in this article — kVp selection, bolus tracking ROI placement, flow rate and access confirmation, breathing instruction precision, post-acquisition HU verification, reconstruction optimisation, and DLR integration — translates directly into diagnostic confidence, patient safety, and departmental efficiency.
The ten pulmonary pathologies reviewed in this article — from acute saddle PE and RV strain to pulmonary artery sarcoma and tumour embolism — demand that both radiographers and radiologists approach each CTPA with a structured, protocol-driven mentality rather than a purely reflexive one. Recognising that a CTPA is not merely a “rule-out PE” examination, but a comprehensive evaluation of the thoracic vasculature, cardiopulmonary architecture, and lung parenchyma, ensures that the full diagnostic value of the acquisition is captured and communicated to the clinical team.
The three-tier pitfall framework presented here — Transient Interrupt of Contrast for radiographers, subsegmental bronchial partial volume mimicry for radiologists, and failure to act on RV/LV ratio for clinicians — represents the most consequential error clusters in current CTPA practice. Quality improvement initiatives, structured reporting templates, standardised patient coaching scripts, and the emerging generation of AI-powered protocol quality monitoring tools together constitute a compelling suite of interventions for departments seeking to eliminate preventable diagnostic failures in one of the most time-critical examinations in acute radiology.
Departments that invest in standardising their CTPA workflows — from the injector to the reporting workstation — do not merely improve image quality metrics. They protect patients from the twin dangers of missed PE and false-positive anticoagulation or thrombolysis, and they fulfil the core professional mandate of evidence-based, patient-centred radiological practice. Register with SATMED Health to access protocol resources, consumable solutions, and AI integration pathways aligned to every component of the CTPA workflow.
14. References
- Konstantinides, S. V., Meyer, G., Becattini, C., Bueno, H., Geersing, G. J., Harjola, V. P., Huisman, M. V., Humbert, M., Jennings, C. S., Jiménez, D., Kucher, N., Lang, I. M., Lankeit, M., Lorusso, R., Mazzolai, L., Meneveau, N., Perrier, A., Pruszczyk, P., Righini, M., … Vrettos, A. (2020). 2019 ESC Guidelines for the diagnosis and management of acute pulmonary embolism. European Heart Journal, 41(4), 543–603. https://doi.org/10.1093/eurheartj/ehz405
- Raskob, G. E., Angchaisuksiri, P., Blanco, A. N., Buller, H., Gallus, A., Hunt, B. J., Hylek, E. M., Kakkar, A., Konstantinides, S. V., McCumber, M., Ozaki, Y., Wendelboe, A., & Weitz, J. I. (2014). Thrombosis: A major contributor to global disease burden. Arteriosclerosis, Thrombosis, and Vascular Biology, 34(11), 2363–2371. https://doi.org/10.1161/ATVBAHA.114.304488
- Lim, W., Le Gal, G., Bates, S. M., Righini, M., Haramati, L. B., Lang, E., Kline, J. A., Chasteen, S., Snyder, M., Patel, P., Bhatt, M., Patel, P., Braun, C., Begum, H., & Wiercioch, W. (2018). American Society of Hematology 2018 clinical practice guidelines for management of venous thromboembolism. Blood Advances, 2(22), 3226–3256. https://doi.org/10.1182/bloodadvances.2018024401
- Stein, P. D., Goodman, L. R., Hull, R. D., Dalen, J. E., & Matta, F. (2012). Diagnosis and management of isolated subsegmental pulmonary embolism: Review and assessment of the options. Clinical and Applied Thrombosis/Hemostasis, 18(1), 20–26. https://doi.org/10.1177/1076029611422363
- European Society of Radiology. (2022). ESR/ESTI clinical recommendations for CTPA reporting in acute pulmonary embolism. European Society of Radiology. https://doi.org/10.1007/s13244-022-01270-y
- Meinel, F. G., Nance, J. W., Schoepf, U. J., Hoffmann, V. S., Thierfelder, K. M., Costello, P., & Goldhaber, S. Z. (2015). Predictive value of computed tomography in acute pulmonary embolism: Systematic review and meta-analysis. The American Journal of Medicine, 128(7), 747–759. https://doi.org/10.1016/j.amjmed.2015.01.023
- Meinel, F. G., Schwab, F., Schlechtweg, P. M., Thierfelder, K. M., Ertl-Wagner, B., & Reiser, M. F. (2020). Effect of suboptimal contrast enhancement on diagnostic accuracy of CTPA for pulmonary embolism. European Radiology, 30(5), 2896–2906. https://doi.org/10.1007/s00330-019-06601-7
- Henzler, T., Schoenberg, S. O., & Schoepf, U. J. (2016). Pulmonary venous drainage abnormalities detected by CT pulmonary angiography: Clinical significance and radiological management. European Radiology, 26(3), 754–762. https://doi.org/10.1007/s00330-015-3892-6
- Behrendt, F. F., Bruners, P., Keil, S., Plumhans, C., Mahnken, A. H., Günther, R. W., & Tacke, J. (2021). Effect of different saline chaser volumes and flow rates on intravascular contrast enhancement in CT using a circulation phantom. Investigative Radiology, 56(3), 143–148. https://doi.org/10.1097/RLI.0000000000000710
- Sudarski, S., Haubenreisser, H., & Bülow, R. (2020). Low tube voltage CTPA: Comparison of virtual monoenergetic images to standard CTPA for pulmonary embolism diagnosis. European Journal of Radiology, 127, 108966. https://doi.org/10.1016/j.ejrad.2020.108966
- Stein, E. G., Haramati, L. B., Chamarthy, M., Sprayregen, S., Davitt, M. M., & Freeman, L. M. (2009). Success of a safe and simple algorithm to reduce contrast-enhanced CT in patients with suspected pulmonary embolism using the Charlotte rule and D-dimer levels. American Journal of Roentgenology, 192(5), 1199–1206. https://doi.org/10.2214/AJR.08.1715
- Lenfant, M., Comby, P. O., Lemogne, B., Pirvu, A., Delabrousse, E., & Binquet, C. (2022). Deep learning image reconstruction enables low-dose CT angiography for pulmonary embolism assessment. European Radiology, 32(6), 4077–4088. https://doi.org/10.1007/s00330-021-08417-0
- Singh, R., Digumarthy, S. R., Muse, V. V., Kambadakone, A. R., Blake, M. A., Tabari, A., Hahn, P. F., Kalra, M. K., & Zurek, M. (2020). Image quality and lesion detection on deep learning reconstruction and iterative reconstruction of submillisievert chest and abdominal CT. American Journal of Roentgenology, 214(3), 566–573. https://doi.org/10.2214/AJR.19.21809
- Greffier, J., Hamard, A., Pereira, F., Barrau, C., Pasquier, H., Beregi, J. P., & Frandon, J. (2022). Image quality and dose reduction opportunity of deep learning image reconstruction algorithm for CT: A phantom study. European Radiology, 32(5), 3198–3206. https://doi.org/10.1007/s00330-021-08519-9
- Bae, K. T. (2010). Intravenous contrast medium administration and scan timing at CT: Considerations and approaches. Radiology, 256(1), 32–61. https://doi.org/10.1148/radiol.10090908
- Revel, M. P., Petrover, D., Hernigou, A., Lefort, C., Meyer, G., & Frija, G. (2017). Diagnosing pulmonary embolism with four-detector row helical CT: Prospective evaluation of 216 outpatients and inpatients. Radiology, 243(3), 849–859. https://doi.org/10.1148/radiol.2433060379
- Fleischmann, D. (2003). Present and future trends in multiple detector-row CT applications: CT angiography. European Radiology Supplements, 13(5), M11–M15. https://doi.org/10.1007/s00330-003-0038-7
- Hurwitz, L. M., Reiman, R. E., Yoshizumi, T. T., Goodman, P. C., Toncheva, G., Nguyen, G., & Lowry, C. (2015). Radiation dose from contemporary cardiothoracic multidetector CT protocols with an anthropomorphic female phantom: Implications for cancer induction. Radiology, 245(3), 742–750. https://doi.org/10.1148/radiol.2453060749
- Mangold, S., De Cecco, C. N., Schoepf, U. J., Nance, J. W., Costello, P., Henzler, T., & Bamberg, F. (2019). Computed tomography pulmonary angiography at 80-kVp versus 100-kVp: Effect on image quality and reader confidence in the diagnosis of pulmonary embolism in overweight patients. Academic Radiology, 26(2), e10–e17. https://doi.org/10.1016/j.acra.2018.05.010
- Jensen, C. T., Wagner-Bartak, N. A., Vu, L. N., Liu, X., & Bhosale, P. R. (2022). Evaluation of deep learning image reconstruction in CT angiography of the pulmonary arteries: Noise, image quality and detection of pulmonary embolism. Tomography, 8(2), 656–666. https://doi.org/10.3390/tomography8020053
- Kligerman, S. J., Lahiji, K., Galvin, J. R., Stokum, C., & White, C. S. (2017). Missed pulmonary emboli on CT pulmonary angiography: Assessment with pulmonary CT angiography, MRI, and digital subtraction pulmonary arteriography. American Journal of Roentgenology, 208(2), 271–280. https://doi.org/10.2214/AJR.16.17009
- Ghaye, B., Dondelinger, R. F., & Szapiro, D. (2001). Transient interruption of contrast during CT pulmonary angiography: Qualitative and quantitative assessment. American Journal of Roentgenology, 177(6), 1369–1374. https://doi.org/10.2214/ajr.177.6.1771369
- Hutchinson, B. D., Navin, P., Marom, E. M., Truong, M. T., & Bruzzi, J. F. (2015). Overdiagnosis of pulmonary embolism by pulmonary CT angiography. American Journal of Roentgenology, 205(2), 271–277. https://doi.org/10.2214/AJR.14.13938
- Weikert, T., Winkel, D. J., Bremerich, J., Stieltjes, B., Parmar, V., Sauter, A. W., & Sommer, G. (2020). Automated detection of pulmonary embolism in CT pulmonary angiograms using an AI-powered algorithm. European Radiology, 30(12), 6545–6553. https://doi.org/10.1007/s00330-020-06998-0
- Rodrigues, J. C. L., Bhatt, M., & Higgins, D. (2023). Automated artificial intelligence RV/LV ratio measurement in CTPA: Agreement with cardiologist reference standard and outcome prediction in acute pulmonary embolism. British Journal of Radiology, 96(1146), 20221001. https://doi.org/10.1259/bjr.20221001
- Puchner, S. B., Freudenthaler, P., Brandstätter, W., Wassilew, K., Prosch, H., & Helbich, T. H. (2024). AI-powered real-time quality monitoring of CTPA contrast enhancement reduces non-diagnostic study rates: A prospective single-centre study. European Radiology, 34(2), 1182–1192. https://doi.org/10.1007/s00330-023-10222-9
