a b s t r a c t

Purpose: The indications for requesting a diagnostic test are important for the selection and timing of imaging protocols. We sought to evaluate the diagnostic yield and impact on patient disposition when evaluating com- puted tomography (CT) of the abdomen and pelvis in adult patients presenting with gastrointestinal bleeding to the Emergency Department (ED).

Methods: This study was an observational cohort study of consecutive adult ED patients with ICD10 codes related to GIB between 5/5/2018 and 6/1/2020. CT reports were reviewed for indications, exam type and findings. Re- ports were classified as positive (active bleeding, recent bleeding or suspected etiology for GIB), negative or other significant findings. Methodological guidelines for reporting observational studies were followed (STROBE).

Results: Among 943 patients with GIB during the study period, 33% (n = 312) had an abdominopelvic CT ordered. Most CTs included contrast, 64.1% (n = 200) used a single portal venous pHase and 28.9% (n = 90) were multi- phase. CT identified active bleeding in 4.2% (n = 13/312) and intraluminal blood in 2.9% (n = 9/312) patients. Patients that had GIB indications on the CT order (n = 142) were more likely to receive a multiphase study com- pared to those without GIB indication (n = 94) (43.0% vs. 8.5%, difference 34.5%, 95% CI 23.7% to 43.7%, p < 0.0001). Patients that received multiphase studies were more likely to have a source of GIB identified compared to single-phase (18.9% vs 1.5%, OR 15.3, 95% CI 4.4 to 53.7, p < 0.0001). In 40.3% (n = 117/290) of patients with- out bleeding, an intra-abdominal cause for their symptoms was identified. Those with GIB or with an identified cause were more likely to be admitted.

Conclusions: One-third of patients evaluated in the ED with GIB had a CT ordered. Active GIB was detected more often when multiphase exams were performed. Multiphase exams are done more often if GIB is listed in the CT indication. When a CT is positive, patients are more likely to be admitted or observed. Accurate indications are critical to optimize Exam performance.

(C) 2022

1. Introduction

Gastrointestinal (GI) bleeding is a common presentation to the emergency department (ED). The incidence of upper GI bleeding is ap- proximately 50-150 per 100,000 [1]. Lower GI bleeding, including distal small bowel (distal to the ligament of Treitz) and colonic bleeding is seen in half as many patients with an incidence of 25-27 per 100,000 [1,10]. Common etiologies for upper GI bleeding include Peptic ulcer disease, gastritis/esophagitis as well as varices and Mallory Weiss tears [1], while lower GI bleeds are more commonly caused by divertic- ular disease, angioectasia and colitis [1].

* Corresponding author at: Mayo Clinic, 200 First Street SW, Rochester, MN 55905, USA.

E-mail address: Flicek.kristina@mayo.edu (K.T. Flicek).

Despite advances in therapy, GI bleeding still carries up to a 10% mortality and significant morbidity [1,2]. Accurate diagnosis and local- ization of GI bleeding is critical for proper management of patients [3]. While endoscopy remains the standard of care for detection and treat- ment of GI bleeding, this is not always feasible to perform emergently. CT Angiography (CTA) provides a critical role for assessment of an overt GI bleed, as it is almost universally available and can be performed rapidly [2]. CTA can also be useful in scenarios in which upper endos- copy reveals non variceal upper GI bleeding, in patients with negative endoscopy, postsurgical, and Traumatic causes of GI bleeding, or when endoscopy is contraindicated [4]. Similarly, in the case of lower GI bleeding, diagnostic and therapeutic colonoscopy is typically the first line therapy. CTA is indicated in hemodynamically unstable patients with active bleeding (including both melena and hematochezia) and

https://doi.org/10.1016/j.ajem.2022.12.003

0735-6757/(C) 2022

can also be utilized to evaluate intermittent or obscure non-localized re- current bleeding with negative prior endoscopy (obscure GI bleed) [5]. Despite current guidelines recommending CTA for unstable patients, CT is being more widely utilized in stable patients presenting to the ED given the widespread 24/7 availability, noninvasiveness, speed, and ability to screen the entire abdomen and pelvis for other etiologies [4]. CT can be protocolled in a variety of ways and the specific protocol is tailored to the clinical indication and optimized to answer the clin- ical question. For those patients presenting with overt GI bleeding, multiphasic CTA is usually performed. This technique includes non- contrast images followed by a multiphasic contrast-enhanced exam- ination. Non-contrast images are helpful to identify high attenuation ingested material which can mimic bleeding and lead to false posi- tive interpretations. One of the contrast-enhanced phases is ac- quired during the arterial pHase and the second during the portal venous or delayed phase [5]. The Arterial phase images are helpful to demonstrate high flow vascular lesions which may be invisible on the second phase [6], demonstrate the severity and briskness of bleed which may lead to more emergent intervention, and provide a roadmap for subsequent angiography. The second phase is helpful at detecting slower venous bleeds and showing the changing ap- pearance of intraluminal high attenuation which is more convincing for active extravasation. Therefore, these specific phases allow for better detection of active GI bleeding (IV contrast extravasation), se- verity of the bleed, and detection and characterization of the poten- tial etiology of the bleed. Despite these theoretical advantages of multi-phase CTA, there is a paucity of data comparing the perfor- mance of multi-phase versus single-phase CTA in the same patient. CTA is especially helpful when the suspected cause of bleeding is di- verticular or vascular which may present with more brisk bleeding. Some other causes of GI bleeding can also be detected on routine single-phase examinations (colitis, some masses, etc.); however, they may be missed or misinterpreted in only single-phase imaging. If bleeding is suspected, without other symptoms to suggest inflam- mation (i.e., diarrhea or abdominal pain) experts suggest a multi- phase exam is the most appropriate exam [4,5]. Therefore, thorough clinical history and accurate indications are needed to optimize the

utilization of CT and CTA.

The goals of this study were to describe the utilization and yield of CT for patients presenting to the ED with GI bleeding. Secondary aims were to evaluate diagnostic detection rate by clinical indication of the exam, and to evaluate the impact of CT findings on patient disposition.

1. Methods
   1. Study design and setting

We adhere to the STROBE guidelines for reporting observational studies [7]. This study was approved by the institutional review board. Only patients that provided authorization for medical record review were included. This was a single center retrospective cohort study con- ducted at an academic institution with an annual ED census of 77,000 patients per year. Radiology coverage and CT availability is provided for the ED 24 h a day, 7 days a week.

• 1. Study population and identification

The study population consisted of consecutive patients presenting to the ED with a final diagnosis of GI bleeding including ICD10 codes K92.1 (melena), K62.5 (hemorrhage of anus and rectum), K92.2 (gastrointes- tinal hemorrhage, unspecified) and K92.0 (hematemesis) between May 5, 2018 and June 1, 2020. Patients who underwent abdominopelvic CT while in the ED were included. Patient demographic information and disposition was recorded.

• 1. Imaging review

When CTs are ordered for an ED patient at our academic center, a ra- diologist reviews the indications and determines the protocol to be per- formed, including unenhanced, single portal venous phase, dual phase (arterial and portal venous phase), dual phase (arterial and delayed phase), or triple phase (arterial, enteric and delayed). All CT reports were reviewed by a single radiologist during the clinical encounter at the time of patient presentation to the ED.

For this study, a second radiologist retrospectively reviewed all the CT reports. Imaging findings and reports were classified as 1) active bleeding, 2) probable/definite intraluminal blood without active bleed- ing, 3) no evidence of active or recent bleeding but suspected etiology for GI bleeding visualized, 4) no active or recent bleeding or suspected etiology, 5) other significant findings potentially leading to admission, or 6) no findings of GI bleed or other etiology that may require admis- sion. If GI bleed was identified, the site and potential cause for bleeding was noted.

Indications for the CT exam were reviewed to determine its associa- tion with the type of exam performed and detection rate of active bleed- ing. Indications were classified according to whether GI bleeding was mentioned, and the number of indications listed was recorded.

• 1. Patient outcomes

Patient disposition (admit/observation or discharge) was docu- mented and reported according to CT findings and exam type. In patients who had a negative initial CT or CTA, the medical record was reviewed to determine if there was readmission for GI bleeding or if additional imaging (CTA, CT enterography, tagged RBC scan, or capsule endoscopy) for GI bleeding was performed in follow-up.

• 1. Statistical analysis

Descriptive statistics are reported as number and precents for dis- crete variables and as mean or median for continuous variables depend- ing on the distribution. Odds ratio (OR) and 95% confidence interval (CI) were calculated. Mean differences were calculated when applicable. The alpha-level was set at 0.05 for statistical significance.

1. Results
   1. Patient cohort

Nine hundred and forty-three patients met inclusion criteria and presented to the ED with a chief complaint and ICD 10 code related to GI bleeding during the study period. Among the 943 patients, 33% (312/943) had an abdominopelvic CT ordered, 52% (163/312) were male with a mean age of 59 years (range 18 to 96 years). Of the patients in which GI bleeding was detected (active bleeding or intraluminal blood), the average age was 67 (range 20 to 96).

• 1. CT technique

Most (64.1%, 200/312) CTs were performed using a single portal ve- nous phase and 28.9% (90/312) were multiphase CT studies. Most of the multiphase exams were performed with an arterial and portal venous phase (51/89, 57.3%). Twenty-two (7.1%) unenhanced CTs were per- formed because of contraindications to IV contrast. Appendix A summa- rizes the different protocol parameters used for each type of exam performed in this cohort.

• 1. Findings

Thirteen patients 13/312 (4.2%) had an active GI bleed detected, 12/ 13 (92%) of the patients with active bleed had a multiphase CT scan. One

Table 1 Number of patients without GI bleeding but with an additional significant CT find- ing that would likely lead to admission. Findings are categorized by organ system most responsible for emergent/significant finding.

Organ system Number of patients

Bowel 26 (42%)

Hepatobiliary 8 (13%)

Pancreatic 7 (11%)

Genitourinary 8 (13%)

Cardiovascular 3 (5%)

Peritoneal/retroperitoneal 3 (5%)

Thorax 5 (8%)

Musculoskeletal 1 (2%)

Other 1 (2%)

patient received a single portal venous phase exam (1/13, 7.7%). There were 9/312 (2.9%) patients with findings of intraluminal high attenua- tion blood without active bleeding; of these, 5 of 9 had multiphase, 2 had single-phase, and 2 had unenhanced CT. Table 1 summarizes the number of patients without evidence of GI bleeding but with other sig- nificant CT findings (categorized by organ system) that would likely lead to admission. Sixty-two patients without GI bleed had other signif- icant CT findings that would likely lead to admission. 42% had a non-GI bleed bowel etiology (e.i Small bowel obstruction) identified. Other common etiologies included hepatobiliary (n = 8, 13%), genitourinary (n = 8, 13%) and pancreatic (n = 7, 11%) causes. Of these 62 patients, 52/62 (84%) were admitted or placed under observation. Only 10 patients from this group were discharged.

• 1. Diagnostic yield based on exam type

Patients that received multiphase studies (n = 90) were more likely to have a source of GIB identified when compared to single-phase (n = 200) (18.9% vs 1.5%, OR 15.3, 95% CI 4.4 to 53.7, p < 0.0001) (Table 2). In

40.3% (n = 117/290) of patients without bleeding, an intra-abdominal cause for their symptoms was identified.

• 1. CT protocols and findings based on chief complaints

Table 2 lists the most common primary ED chief complaints in this cohort, the type of CT performed, and whether bleeding (active bleeding or intraluminal blood) was identified at CT. Primary chief complaints re- lated to GI bleeding were present in 61.9% (193/312) of the patients, in- cluding Rectal bleeding (31.1%, 97/312), black/bloody stools (16.7%, 52/ 312) and GI bleed, not otherwise specified (5.1%, n = 16/312). Abdom- inal pain was the most common non-GI bleeding primary chief com- plaint (13.8%, n = 43/312). Of patients in which GI bleeding was

Table 2

Detection of GI bleeding on different CT protocols based on primary chief complaint.

detected, rectal bleeding was the most common primary chief com- plaint (31.8%, n = 7/22).

Table 3 describes the most common primary chief complaints in this cohort (rectal bleeding, hematemesis/hemoptysis, black bloody stools, abdominal pain and GI bleed) and if GI bleeding was listed in the CT in- dication (the primary area where radiologists determine the appropri- ate CT protocol) and what type of CT exams were performed (multi- phase versus single portal venous phase or unenhanced). Unenhanced exams were completed due to contraindications to IV contrast including renal function or contrast allergy.

A total of 236 patients made up this group with primary chief com- plaints, including rectal bleeding, hematemesis/hemoptysis, black bloody stools, abdominal pain and GI bleed. Of those, 60.2% (142/236) had bleeding listed in the CT indication. Of those patients, 61/142 (43.0%) had a multiphase exam, and 71/142 (50.0%) had a portal venous phase exam. Patients that had GIB indications on the CT order (n = 142) were more likely to receive a multiphase study compared to those with- out GIB indication (n = 94) (43.0% vs 8.5%, difference 34.5%, 95% CI

23.7% to 43.7%, OR 8.1, 95% CI 3.7 to 18.0, p < 0.0001).

• 1. Patient disposition based on CT results

Table 4 illustrates the disposition of patients (admit/observation or discharge) and its association with CT findings. No patient with active bleeding (n = 13) identified at CT was discharged. Most patients (n = 8, 89%) in which intraluminal hemorrhage was identified were admit- ted. Of patients with other significant CT findings identified, 84% (52/ 62) were admitted/observed. A majority of these patients had a single portal venous phase exam (34/52, 65%).

In total there were 239 patients who had no active bleeding identi- fied at CT. The average follow-up for this group of patients was 672 days or 1.84 years. Only 14 (6%) patients had no follow up after their ini- tial emergency room visit.

Twenty-two patients initially scanned (7 multiphase; 13 single phase and 2 unenhanced) in the emergency department for GI bleed had at least one additional follow up CTA or triple phase CT Enterography (CTE) to as- sess for GI bleed (range = 1-5 CT scans, mean = 1.36.

CT scans). These exams were not necessarily completed during their initial admission but may have been performed on subsequent outpa- tient, emergency room visits or hospital admissions. 9% (2/22) of pa- tients with repeat CTA/CTE had evidence of active small bowel bleeding at CT. No patients who had a single portal venous phase exam in the Emergency department were admitted and had a subse- quent CTA/CTE performed during that admission.

1. Discussion

GI bleeding is a common complaint for ED visits [8,9]. While endoscopy remains the gold standard for diagnosis and treatment of GI bleeding, en- doscopic evaluation, particularly colonoscopy, can be challenging due to the need for bowel preparation and often intermittent or short active

Chief Complaint Multiphase CT (n = 90)

Single Portal Venous Phase CT (n = 200)

non-contrast CT (n = 22)

bleeding times [9,10]. As such, in an urgent setting like the emergency de- partment, CT is a valuable and rapid tool for evaluating GI bleeding as well as other Potential causes of Patient symptoms [10-13]. Our study found that when GI bleeding indications are part of the indication for ordering a CT of the abdomen and pelvis (melena, hematochezia, etc), radiologists are more likely to select multiphase protocols, and the likelihood of finding the source of GI bleeding is higher. CT can help identify causes of bleeding and abdominal symptoms in nearly one half of the patients.

In our study, approximately 33% of patients presenting with a com- plaint of GI bleeding or those found to have GI bleeding while in the ED were ordered a CT in the ED. Decisions for who will benefit from imag- ing is complex and related to many factors which were not assessed as part of this study, including clinician gestalt, severity and duration of

Bleed + = either active extravasation or intraluminal hemorrhage. Bleed – = no active extravasation or intraluminal hemorrhage.

	Bleed	Bleed		Bleed	Bleed		Bleed	Bleed
	+	–		+	–		+	–
GI Bleed (n = 16)	1	7		0	8		0	0
Abdominal Pain (n = 43)	1	5		0	33		0	4
Diarrhea (n = 2)	0	0		0	2		0	0
Rectal Bleeding (n = 97)	6	23		1	62		0	5
Black/bloody stools (n = 52)	3	14		0	29		1	5
Hematemesis/hemoptysis	2	7		0	17		0	2
(n = 28)
Shortness of Breath (n = 6)	1	2	0		3	0		0
Chest Pain (n = 3)	0	0	1		2	0		0
Other (n = 65)	3	15	1		41	1		4

bleeding, hemodynamic instability and co-existing signs and symptoms which may be more concerning.

Table 3

Correlation of CT indication compared to primary ED chief complaint and the type of CT exam performed (n = 236).

Bleeding was in the CT indication (n = 142) Bleeding was not in the CT indication (n = 94)

Primary Chief	Multi-phase	Single PV Phase	Unenhanced		Multi-phase	Single PV Phase	Unenhanced
Complaint	(n = 61)	(n = 71)	(n = 10)		(n = 8)	(n = 81)	(n = 5)
Rectal Bleeding (n = 97)	27	37	2		2	26	3
Hematemesis/ Hemoptysis (n = 28) Black blood stools	8 17	8 14	1 4		2 0	8 16	1 1
(n = 52) Abdominal Pain	2	8	3		3	27	0
(n = 43)
GI Bleed (n = 16)	7	4	0		1	4	0

In our practice, the performing radiologist determination of the spe- cific CT protocol is primarily based on the clinical indication that is pro- vided at the time the study is ordered. A surprising finding in our study was that most of these patients were scanned using a single-portal ve- nous phase protocol. If GI bleeding was not listed in the indication, pa- tients were more likely to have a single-phase exam. However, to detect active bleeding, a multi-phasic protocol using higher iodine concentration and faster injection rates is recommended [5]. In our study only 7% (1/13) of cases with active bleeding were detected using the single-phase pro- tocol. It is possible that the use of the single phase based on incomplete clinical indication may have led to an overall decreased performance of CT. Not all patients require multi-phasic examination. For example, if there is concern for an infectious or inflammatory colitis or if there is ab- dominal pain with minimal GI bleeding, a single-phase examination may suffice to identify a potential etiology. However, many other sce- narios would benefit from the use of CTA, including concerns for ische- mia, mass, vascular lesion or even ulcer.

Kim et al. evaluated 386 patients with upper GI bleed and reported accuracy of 32.9% for patients with active bleeding and 27.4% for those with recent bleeding. CT’s accuracy was 60.9% for esophageal bleeding, 60.6% for stomach sources, and 50.9% for duodenal bleeding [14]. A study by Lee et al. evaluated 382 patients who presented to the ED with hematochezia at a single center using colonoscopy and/or sig- moidoscopy as gold standard. CT had a sensitivity of 85.7%, specificity of 100%, and negative predictive value of 96.9% for detecting active bleeding sites, and a sensitivity of 87.4%, specificity of 40.0%, and nega- tive predictive value of 47.6% for identifying the etiology of lower GI bleed [15]. A study by Nagata et al. included 126 patients who under- went colonoscopy following CT and found high agreement between im- aging and colonoscopy findings (kappa 0.83). The authors report that CT added value for detecting vascular lesion compared to colonoscopy alone in 15% of the cases [16]. Kim et al. included 111 consecutive pat- ents who presented to the ED with overt GI bleeding and used clinical diagnosis as gold standard (instead of colonoscopy). The authors re- ported a diagnostic yield of CTA of 61.3% (n = 68/111), sensitivity of 84.8% and specificity of 96.9% [13].

Tse et al., in a study of 50 patients with overt GI bleed, reported that

larger extravasation volumes correlated with higher bleeding rates, with current CTAs detecting bleeding rates of 0.1 mL/min [17].

Table 4

ED disposition of patients by CT results.

Regarding the timing for obtaining CTA, Obana et al. recommended CT to be performed within 2 h of hematochezia to increase its diagnostic yield in cases of suspected diverticular bleeding [18]. A prospective study by Umezawa et al. included 202 patients with diverticular bleed- ing and reported higher sensitivity among those who had CT within 4 h of hematochezia (n = 33/51, 64.7%), and high specificity of the findings when compared to colonoscopy [19].

Most patients in whom GI bleeding was detected at CT were at least 47 years old. Only two patients under the age of 25 had GI bleeding identified, both cases attributable to post procedural complications. The more advanced age of most patients presenting with GI bleed symptoms/signs may help alleviate some concern regarding the in- creased radiation dose required from multiphase CT.

CT played an important role in patient disposition from the emer- gency department. No patients with active bleeding were discharged and only one with intraluminal blood from a recent GI bleed. Addition- ally, when a likely etiology for GI bleeding was identified at CT, patients were more likely to be admitted.

There are several limitations to this study. First, this was a retrospec- tive study. Second, multiple different types of multiphase CT protocols were used in this study. Given the variability of the multiphase exams, this could have impacted GI bleed detection rates. While multi-phase CTA is recommended for GI bleeding, we did not perform a retrospective blinded review to compare the performance of single-phase images to multi-phases in the same patient. Further studies are needed to evaluate the added benefit of additional phases. Our study assumed that CT find- ings drove the decision to admit, observe or discharge patients, however, there may have been other reasons that impacted that decision. Similarly, in some of the cases using an ICD code for GI bleeding, the GI bleeding may have been a minor presenting sign or symptom. The CT exams for this cohort were not retrospectively reviewed, but rather the prospective clinical reports were reviewed for presence or absence of GI bleeding. In our practice, these exams are interpreted by several different specialties (abdominal, chest, musculoskeletal, cardiovascular) covering a busy emergency department. A retrospective review by specialized Abdominal imaging experts in a more controlled reading environment may have led to different detection rates. Lastly, endoscopy reports were not reviewed as most patients did not undergo endoscopy (the gold standard), so we did not calculate Diagnostic test accuracy of CTA in this study.

In conclusion, CT plays an important role in the diagnosis and man- agement of patients presenting with GI bleeding in the emergency department. The diagnostic yield for significant findings was approxi-

CT results Admit/Observation

n = 213

Discharge

N = 99

mately 25% including active bleeding (4%), evidence for recent bleeding (3%), suspected bowel etiologies for bleeding (8%), and other significant

Active bleed (n = 13) 13 0

Intraluminal blood (n = 9) 8 1

Other significant CT findings (n = 62) 52 10

extraluminal findings (10%). The detection or exclusion of these impor-

tant findings led to significant differences in patient disposition. In order to optimize CT technique and enhance diagnostic yield it is crucial for

No bleeding or other significant extra abdominal CT findings (n = 228)

140 88

emergency physicians and other health care providers to provide con- cern for GI bleeding at the time of CT ordering.

types of CT exams performed “>Funding

None.

IRB approval

Approved and exempt.

Availability to data/material

Not applicable.

Code availability

Not applicable.

Declaration of Competing Interest

The authors have no conflicts of interest to declare.

Acknowledgements

The authors acknowledge the assistance of Lucy Bahn, PhD, in editing the manuscript.

Appendix A. Types of CT exams performed and parameters

Protocol Phases Iohexol Concentration (mgl/mL) and Rate Injection

Volume Iohexol 647 (mL) by weight 0.9% NaCl Flush

(mL)

			<64 kg	64-109 kg	110-136 kg	>136 kg
GI Bleed Single or Dual Energy	Arterial: Bolus tracking Portal: 70 s after start of contrast	Iohexol 350 4 mL/s	60	85	120	135		50
Single Phase	injection 60-70 s delay	Iohexol 350	100	140	200 @ 4	200 @ 5	50
Triple Phase Enterography without	Arterial Phase: Bolus tracking	3 mL/s Iohexol 300	100	140	mL/s 200	mL/s 200 @ 5		50
oral contrast		4 mL/s				mL/s
Mesenteric ischemia	Enteric: 50 s after beginning injection Delayed: 90 s after beginning of injection Arterial Phase: Bolus tracking	Iohexol 300	100	140	200	200 @ 5	50
	Delay Phase: 35 s after end of	4 mL/s				mL/s
Dual Phase	arterial phase Late arterial at 45 s	Iohexol 300	100	140	200	200	50
	Portal venous phase at 70 s	4 mL/s

References

1. Kumar R, Mills AM. Gastrointestinal bleeding. Emerg Med Clin North Am. 2011;29 (2):239-52. viii. https://doi.org/10.1016/j.emc.2011.01.003.
2. Wortman JR, Landman W, Fulwadhva UP, Viscomi SG, Sodickson AD. CT angiography for Acute gastrointestinal bleeding: what the radiologist needs to know. Br J Radiol. 2017;90(1075):20170076. https://doi.org/10.1259/bjr.20170076.
3. Almadi MA, Barkun AN. Patient presentation, risk stratification, and initial Manage- ment in Acute lower gastrointestinal bleeding. Gastrointest Endosc Clin N Am. 2018;28(3):363-77. https://doi.org/10.1016/j.giec.2018.02.008.
4. Fidler JL, Guglielmo FF, Brook OR, Strate LL, Bruining DH, Gupta A, et al. Management of gastrointestinal bleeding: Society of Abdominal Radiology (SAR) institutional sur- vey. Abdom Radiol (NY). 2022;47(1):2-12. https://doi.org/10.1007/s00261-021- 03232-3.
5. Fidler JL, Gunn ML, Soto JA, Brook OR, Allen BC, Guglielmo FF, et al. Society of abdom- inal radiology gastrointestinal bleeding disease-focused panel consensus recom- mendations for CTA technical parameters in the evaluation of acute overt gastrointestinal bleeding. Abdom Radiol (NY). 2019;44(9):2957-62. https://doi. org/10.1007/s00261-019-02131-y.
6. Huprich JE, Barlow JM, Hansel SL, Alexander JA, Fidler JL. Multiphase CT enterography evaluation of small-bowel vascular lesions. AJR Am J Roentgenol. 2013;201(1):65-72. https://doi.org/10.2214/AJR.12.10414.
7. von Elm E, Altman DG, Egger M, Pocock SJ, Gotzsche PC, Vandenbroucke JP, et al. The strengthening the reporting of observational studies in epidemiology (STROBE) statement: guidelines for reporting observational studies. Ann Intern Med. 2007; 147(8):573-7. https://doi.org/10.7326/0003-4819-147-8-200710160-00010.
8. Khamaysi I, Gralnek IM. Acute upper gastrointestinal bleeding (UGIB) – initial eval- uation and management. Best Pract Res Clin Gastroenterol. 2013;27(5):633-8. https://doi.org/10.1016/j.bpg.2013.09.002.
9. Moss AJ, Tuffaha H, Malik A. Lower GI bleeding: a review of current management, controversies and advances. Int J Color Dis. 2016;31(2):175-88. https://doi.org/10. 1007/s00384-015-2400-x.
10. Copland A, Munroe CA, Friedland S, Triadafilopoulos G. Integrating urgent multide- tector CT scanning in the diagnostic algorithm of active lower GI bleeding. Gastrointest Endosc. 2010;72(2):402-5. https://doi.org/10.1016/j.gie.2010.04.014.
11. Geffroy Y, Rodallec MH, Boulay-Coletta I, Julles MC, Ridereau-Zins C, Zins M. Multi- detector CT angiography in acute gastrointestinal bleeding: why, when, and how. Radiographics. 2011;31(3):E35-46. https://doi.org/10.1148/rg.313105206.
12. Artigas JM, Marti M, Soto JA, Esteban H, Pinilla I, Guillen E. Multidetector CT angiog- raphy for acute gastrointestinal bleeding: technique and findings. Radiographics. 2013;33(5):1453-70. https://doi.org/10.1148/rg.335125072.
13. Kim J, Kim YH, Lee KH, Lee YJ, Park JH. Diagnostic performance of CT angiography in patients visiting emergency department with overt gastrointestinal bleeding. Ko- rean J Radiol. 2015;16(3):541-9. https://doi.org/10.3348/kjr.2015.16.3.541.
14. Kim D, Kim JH, Ko DR, Min IK, Choi A, Beom JH. Usefulness of contrast-enhanced multi-detector computed tomography in identifying upper gastrointestinal bleed- ing: a retrospective study of patients admitted to the emergency department. PLoS One. 2022;17(4):e0266622. https://doi.org/10.1371/journal.pone.0266622.
15. Lee HS, Kang SH, Rou WS, Eun HS, Joo JS, Kim JS, et al. Computed tomography versus lower endoscopy as initial diagnostic method for evaluating patients with hematochezia at emergency room. Medicine (Baltimore). 2020;99(22):e20311. https://doi.org/10.1097/MD.0000000000020311.
16. Nagata N, Niikura R, Aoki T, Moriyasu S, Sakurai T, Shimbo T, et al. Role of urgent contrast-enhanced Multidetector computed tomography for acute lower gastroin- testinal bleeding in patients undergoing early colonoscopy. J Gastroenterol. 2015; 50(12):1162-72. https://doi.org/10.1007/s00535-015-1069-9.
17. Tse JR, Shen J, Shah R, Fleischmann D, Kamaya A. Extravasation volume at computed tomography angiography correlates with bleeding rate and prognosis in patients with overt gastrointestinal bleeding. Investig Radiol. 2021;56(6):394-400. https:// doi.org/10.1097/RLI.0000000000000753.
18. Obana T, Fujita N, Sugita R, Hirasawa D, Sugawara T, Harada Y, et al. Prospective eval- uation of contrast-enhanced computed tomography for the detection of colonic di- verticular bleeding. Dig Dis Sci. 2013;58(7):1985-90. https://doi.org/10.1007/ s10620-013-2629-6.
19. Umezawa S, Nagata N, Arimoto J, Uchiyama S, Higurashi T, Nakano K, et al. Contrast- enhanced CT for colonic diverticular bleeding before colonoscopy: a prospective multicenter study. Radiology. 2018;288(3):755-61. https://doi.org/10.1148/radiol. 2018172910.