Imaging for Upper GI Bleed: CT Angiography Without and With IV Contrast

Critical Distinction: CTA vs Standard CT with Contrast

• The American College of Radiology recommends CT angiography (CTA) of the abdomen and pelvis without and with IV contrast as the imaging modality of choice for upper GI bleeding when endoscopy is not performed or has failed, with a strength of evidence based on the 2024 ACR Appropriateness Criteria 1, 2, 3
• The 2024 ACR Appropriateness Criteria explicitly state that there is no significant literature supporting the use of standard CT abdomen/pelvis with IV contrast alone for upper GI bleeding, with standard CT with contrast alone rated as "usually not appropriate" (rating 2-3) across all clinical scenarios 1, 4, 5

Clinical Scenarios and Recommendations

• The American College of Radiology suggests that CTA abdomen and pelvis without and with IV contrast is usually appropriate (the preferred imaging) when patients present with large volume bleeding or clinical instability preventing initial endoscopy, with a rating of 8-9 for visceral arteriography as an equivalent alternative 1, 2, 3
• The American College of Radiology recommends CTA abdomen and pelvis without and with IV contrast when endoscopy confirms bleeding but the source is unclear, with visceral arteriography rated equally appropriate (rating 8-9) as it allows simultaneous diagnosis and therapeutic embolization 2, 3, 4, 5

Technical Performance Characteristics

• CTA can detect bleeding rates as slow as 0.3 mL/min, with a sensitivity of 79-85% for active GI bleeding, specificity of 92-95%, positive predictive value of 86%, and negative predictive value of 92% 6

Critical Protocol Requirements

• The ACR guidelines emphasize specific technical requirements for CTA in GI bleeding, including multiphasic acquisition (non-contrast phase followed by arterial and venous phases), no positive oral contrast, and thin collimation (≤1 mm) to enable high-quality multiplanar reformations 1, 6

Subsequent Management

• If CTA identifies active bleeding, catheter angiography with intent to embolize should be performed, with technical success rates up to 95% and clinical success around 67%, and the CTA provides a roadmap that allows faster selective catheterization and more efficient embolization 6

CT Angiography of the Abdomen

Protocol Components

• The American College of Radiology recommends CTA without and with IV contrast as the preferred protocol for abdominal CT angiography, providing optimal visualization of vascular structures while detecting active bleeding as slow as 0.3 mL/min 7
• Multiphasic imaging is essential for abdominal CTA, typically including non-contrast phase to identify baseline high-attenuation materials that could mimic bleeding or contrast extravasation 8, 9
• Thin collimation (≤1 mm) enables high-quality multiplanar reformations 9, 7

Clinical Applications

• CTA can detect bleeding rates as slow as 0.3 mL/min, which is more sensitive than conventional angiography, and is particularly valuable when endoscopy is not feasible due to large volume bleeding or clinical instability 9, 7, 8, 10
• The Society of Abdominal Radiology Gastrointestinal Bleeding Disease-Focused Panel recommends non-contrast followed by contrast-enhanced phases for gastrointestinal bleeding 9

Technical Parameters

• 3D reconstructions are required elements of CTA, providing better vascular detail and differentiation from standard CT with contrast 11

Advantages Over Other Modalities

• Compared to conventional angiography, CTA is less invasive, has better sensitivity for detecting slow bleeding, and provides comprehensive evaluation of non-vascular structures 7
• Compared to standard CT with IV contrast, CTA provides better vascular detail, required 3D reconstructions, and optimized contrast timing 11
• Compared to MRA, CTA has faster acquisition time, making it more suitable for acute bleeding scenarios 7

Common Pitfalls and Limitations

• Intermittent bleeding may result in false-negative studies 7
• Radiation exposure is a consideration, particularly in younger patients 12

Special Considerations

• Dual-energy CT can generate virtual non-contrast images from contrast-enhanced data, potentially reducing radiation exposure, but this remains user/site-specific 9
• Timing of acquisition is critical, with early performance (within first 5 hours of presentation) associated with significantly higher visualization of extravasation in bleeding cases 8

CT-Angiography for Identifying Specific Sites of Intrahepatic Tumor Bleeding

Diagnostic Performance for Bleeding Detection

• CT-angiography can identify the specific site of intrahepatic tumor bleeding with high accuracy, demonstrating sensitivity of 79-85% and specificity of 92-95% for detecting and localizing active gastrointestinal bleeding, including hepatic sources 13
• CTA can detect bleeding rates as slow as 0.3 mL/min, which is significantly more sensitive than conventional angiography and comparable to nuclear medicine studies 13, 14
• Overall accuracy for detecting and localizing gastrointestinal bleeding is 94-95%, with positive predictive value of 86% and negative predictive value of 92% 13, 14
• The primary goals of CTA are to determine the location and assess the intensity of the bleed, with a secondary goal of identifying the underlying cause 13

Critical Technical Requirements

• Multiphasic acquisition is essential: noncontrast phase, late arterial phase, and portal venous phase 13
• Positive oral contrast must be avoided as it obscures active hemorrhage and can render the examination nondiagnostic 13, 14, 15
• Thin collimation (≤1 mm) enables high-quality multiplanar reformations necessary for precise localization 13

Tumor-Specific Capabilities

• Contrast extravasation volume can be quantified, with larger volumes associated with higher transfusion requirements and need for hemostatic therapy 13

Important Caveats

• Intermittent bleeding may result in false-negative studies if the patient is not actively bleeding during the scan acquisition 14
• The cause of bleeding may not be ascertained if there is variant vascular anatomy or inadequate contrast resolution, though CTA still provides anatomic information superior to conventional angiography 14

Advantages Over Alternative Imaging

• Noninvasive nature, short examination time, and widespread availability make CTA well-suited for hemodynamically unstable patients 13
• CTA is superior to nuclear medicine scintigraphy, which has poor positive predictive value for determining the precise site of bleeding and is rarely available emergently 14, 15
• CTA is more comprehensive than conventional angiography alone, as it evaluates both vascular and non-vascular structures and can detect etiologies outside the GI tract 13

CT Angiography Protocol for Acute Gastrointestinal Bleeding When Endoscopy Is Not Feasible

Imaging Modality Selection

• Order CT angiography (CTA) of the abdomen and pelvis with and without IV contrast using a multiphasic protocol; do not order a standard “CT abdomen with contrast.” The American College of Radiology (ACR) specifies this as the appropriate study when endoscopy cannot be performed. 16
• Standard contrast‑enhanced CT is not appropriate for evaluating acute GI bleeding; the ACR notes a lack of supporting literature and rates it “usually not appropriate” (rating 2‑3) across clinical scenarios. 17

CTA Protocol Requirements

• Three‑phase acquisition (unenhanced, arterial, portal‑venous) yields the highest diagnostic sensitivity (≈ 92%). This multiphasic approach is recommended by the ACR with strong consensus. 16
• Single‑phase protocols (arterial‑only or portal‑venous‑only) provide lower sensitivity (≈ 83%) and are discouraged. 16
• Non‑contrast images are required; the ACR reaches 100 % consensus for including a true unenhanced phase in single‑energy CT. 18

Rationale for Each Phase

• Unenhanced phase – Detects baseline high‑attenuation material (e.g., sentinel clot) that could mimic active bleeding on contrast‑enhanced images. 18
• Arterial phase – Captures active arterial extravasation, capable of detecting bleeding rates as low as 0.3 mL/min. [no citation needed for rate]
• Portal‑venous/delayed phase (≈ 70‑90 s post‑bolus) – Identifies slower or intermittent bleeding not seen on the arterial phase. 16

Anatomic Coverage

• Image both the abdomen and pelvis because the bleeding source is often indeterminate without prior endoscopy. 16
• Abdomen‑only imaging lacks supporting evidence and should be avoided. 19
• Chest imaging is unnecessary unless there is a specific concern for esophageal bleeding. 16

Technical Considerations

• Avoid positive oral contrast agents; they obscure intraluminal hemorrhage and render the study nondiagnostic. 16
• Oral water should also be omitted because it can dilute intraluminal blood and reduce detection. 19

Diagnostic Performance

• Meta‑analysis of 22 studies (672 patients) shows CTA sensitivity ≈ 85 % and specificity ≈ 92 % for active GI bleeding. This high‑quality evidence supports CTA as a reliable diagnostic tool. 16
• CTA accelerates triage to definitive therapy compared with endoscopy or nuclear‑medicine studies, improving time to intervention. 16
• Larger volumes of contrast extravasation on CTA correlate with the need for hemostatic therapy, intra‑procedural active bleeding, and massive transfusion. 19

Ordering Pitfalls

• Do not order “CT abdomen and pelvis with IV contrast” (a standard CT); it is not a CTA. 17
• Do not limit CTA to the abdomen alone; include the pelvis to ensure complete coverage. 19
• Do not use single‑phase protocols; the ACR mandates a three‑phase approach. 16

Alternative Imaging Modalities

• Visceral arteriography is rated equally appropriate (rating 8‑9) when endoscopy confirms bleeding but treatment is not feasible or bleeding persists after endoscopic therapy. It allows simultaneous diagnosis and therapeutic embolization, with diagnostic success rates of 88‑100 %. 19
• Tc‑99m‑labeled red‑blood‑cell scintigraphy lacks significant literature support as an initial imaging modality for suspected GI bleeding without endoscopy. 19
• CTA is superior to nuclear‑medicine studies for rapid triage to definitive treatment. 16
• Upper‑GI fluoroscopic series (barium or iodine) and MR enterography have no role in the acute evaluation of overt GI bleeding. 19

Advanced Imaging Options

• Virtual non‑contrast images derived from dual‑energy CT can substitute true non‑contrast scans, but this technique is not universally accepted and remains site‑specific. 18

CT Angiography Guidelines for Detecting Active Bleeding

Diagnostic Performance

• Contrast‑enhanced CT angiography (CTA) demonstrates a sensitivity of 85–90 %, specificity of 92 %, and overall accuracy of 94–95 % for identifying active hemorrhage in patients with suspected bleeding. High‑level evidence (meta‑analysis). 20
• In gastrointestinal bleeding, a pooled analysis of 22 studies (672 patients) reported CTA sensitivity of 85 % and specificity of 92 %. High‑level evidence (meta‑analysis). 21
• CTA can detect bleeding rates as low as 0.3 mL/min, outperforming conventional catheter angiography (0.5–1.0 mL/min) and nuclear‑medicine scintigraphy (0.2 mL/min). Radiology evidence. 22
• Larger volumes of contrast extravasation on CTA are associated with greater transfusion requirements, increased need for hemostatic therapy (endoscopic, angiographic, or surgical), and confirmation of active bleeding on subsequent procedures. Observational data. 21
• A negative CTA predicts a favorable prognosis, with lower rates of re‑bleeding and reduced necessity for therapeutic intervention. Observational data. 20

Imaging Protocol Recommendations (American College of Radiology)

• Multiphasic acquisition is mandatory for optimal detection of active bleeding. Guideline recommendation. 22
• Three‑phase protocol (non‑contrast + arterial + portal‑venous) yields 92 % sensitivity, compared with 83 % for single‑phase protocols. Guideline recommendation. 22
• Non‑contrast phase identifies baseline high‑attenuation material (e.g., ingested substances, sentinel clot) that could mimic active bleeding. Guideline recommendation. 22
• Late arterial phase (≈35 s after contrast injection) captures active arterial extravasation and provides arterial opacification. Guideline recommendation. 20
• Portal‑venous phase (≈60–70 s) detects slower venous bleeding and improves solid‑organ evaluation. Guideline recommendation. 20
• Delayed phase (≥90 s, optional) further enhances detection of slow or intermittent bleeding. Guideline recommendation. 20
• Oral positive contrast agents (including oral water) must be avoided because they obscure intraluminal hemorrhage and can render the study nondiagnostic. Guideline recommendation. 22

Clinical Implications

• CTA‑detected contrast extravasation correlates with higher transfusion needs, necessity for hemostatic interventions, and confirmation of ongoing bleeding on later procedures, guiding urgency of therapeutic decisions. Observational data. 21
• A negative CTA provides reassurance, allowing clinicians to consider conservative management with a lower risk of re‑bleeding. Observational data. 20

Comparative Imaging Modalities

• CTA vs. nuclear‑medicine scintigraphy: Both detect active bleeding in ~38 % of cases, but CTA localizes the bleeding source in 53 % of studies versus 30 % for scintigraphy (p = 0.008). CTA also offers faster diagnosis, supporting rapid triage to definitive therapy. Guideline recommendation. 22
• CTA vs. catheter angiography: When CTA confirms bleeding but endoscopic or surgical options are unavailable, catheter angiography receives an equally high appropriateness rating (8–9), providing both diagnostic information and the ability for therapeutic embolization. Guideline recommendation. 22

Anatomic Coverage

• The ACR recommends imaging both abdomen and pelvis when the bleeding source is indeterminate; abdomen‑only scans lack supporting evidence. Guideline recommendation. 22
• Chest imaging is unnecessary unless esophageal bleeding is specifically suspected. Guideline recommendation. 22

Common Diagnostic Pitfalls

• Intermittent bleeding can lead to false‑negative CTA if the patient is not actively bleeding during image acquisition. Radiology evidence. 20
• High‑attenuation ingested material may mimic active bleeding; inclusion of a non‑contrast phase helps differentiate true extravasation. Radiology evidence. 20