Bleeding in the
gastrointestinal (GI) system can be conveniently divided into two general
categories:
- upper GI bleeding
occurs proximal to the Ligament of Trietz and includes bleeding in the
esophagus,
stomach
or duodenum.
Common causes include:
- lower GI bleeding
occurs distal to the Ligament of Trietz and includes bleeding in the jejunum,
ileum,
colon
or rectum. Common causes include:
The prognosis for a
patient presenting with acute bleeding depends upon a number of factors
including the:
- rate
of blood loss
- total volume of blood
lost
- patient's
age
- presence of related
disease in the patient
- effectiveness of
treatment of the bleeding.
Because gastrointestinal
bleeding can be life-threatening, timely identification and localization of the
bleeding site is an important aspect of the management of acute GI
bleeding.
Gastrointestinal bleeding
is often intermittent or slow. Nuclear medicine imaging (especially using
Tc-99m red blood cells) is very effective and is up to ten times more sensitive
than radiographic angiography for the detection of lower GI bleeds. The
sensitivity and specificity are not as good for upper GI bleeds.
 Return to Top of
Page
Gastrointestinal bleeding
studies are based on detection of the extravasation of an injected
radiopharmaceutical into the lumen of the GI tract at a bleeding site. Since
blood is an irritant within the GI tract, it is rapidly moved distally (usually
- but movement can be bidirectional) along the tract and away from the bleeding
site by peristaltic activity. The movement also results in dispersal of the
activity making it less evident on images. For these reasons,
continuous imaging / data collection is
recommended during performance of a GI bleeding study.
Two radiopharmaceuticals
have been used successfully for detecting GI bleeds:
- Tc-99m sulfur
colloid
- Tc-99m red blood
cells
Each has its own inherent
advantages and disadvantages.
Protocols for each of the
radiopharmaceuticals are presented separately.
 Return to Top of
Page
Gastrointestinal bleeding
studies are indicated in patients with known or suspected GI bleeding to detect
the presence of bleeding and to localize the site of bleeding.
It may also assist in
identifying patients requiring angiography.
 Return to Top of
Page
Patient
Preparation
No specific patient
preparation is required.
Explain the test procedure
to the patient and obtain relevant patient history.
Pharmacological Interventions
No specific
pharmacological interventional techniques are commonly used with this
study.
Radiopharmaceutical(s)
- Tc-99m sulfur
colloid
- Tc-99m tin
colloid
About 300 to 600 MBq of
Tc-99m labelled colloid is injected as a bolus. Significantly higher doses are
used compared to liver / spleen imaging due to the expectation that only small
amounts of the radiopharmaceutical will be extravasated.
Equipment
A LFOV gamma camera with a
low-energy all purpose or high resolution parallel hole collimator interfaced
to a nuclear medicine computer is preferred. The camera's analyzer is set at
140 keV with a 15 to 20% window. The computer acquisition uses a 128 x 128
matrix.
Image /
Data Acquisition Parameters
The patient should be
placed supine with the field of view adjusted to include only the lower portion
of the liver and spleen and the entire abdomen and pelvic region. If images are
to be recorded on film, the intensity should be set high in order to optimize
visualization of the potentially small amounts of extravasated
radiopharmaceutical.
An anterior flow is
obtained for 2 to 3 min at a rate of 2 to 5 sec/frame. This is followed by 1
min acquisitions per image for 20 to 30 min. Delayed images obtained at about
45 to 60 min may allow better visualization of colloid extravasated at the
hepatic or splenic flexures. Continuous imaging improves the sensitivity of the
procedure.
More than one dose (up to
~3) may be administered if bleeding is not detected with the previous dose but
is strongly suspected.
The protocol posted by
The
Crump Institute for Molecular Imaging at UCLA provides a clinical example
of a GI Bleeding protocol using Tc-99m labelled colloid.
Data
Analysis
Images are viewed in cine
mode. Images may be combined together for production of hard copy
images.
Procedure
using Tc-99m Red Blood Cells
Patient
Preparation
No specific patient
preparation is required.
Explain the test procedure
to the patient and obtain relevant patient history. An informed concent form
should be completed.
Pharmacological Interventions
No specific
pharmacological interventional techniques are commonly used with this
study.
Radiopharmaceutical(s)
In
vitro labelling of the red blood cells is preferred to obtain a high
labelling yield (>~97%). The higher amounts of free pertechnetate resulting
from
in
vivo labelling (~75 to 85% labelling) and less so from
in
vivtro labelling (~90 to 95% labelling) can produce false positive results
due to free pertechnetate secreted into the lumen of the GI tract by the
gastric mucosa. The activity will eventually arrive in the bowel and may
present interpretation difficulties. For a movie presentation on invivtro
labelling,
click
here. Note: The file is 2.551 Megabytes in size and takes some time
to download.
About 700 to 1000 MBq of
Tc-99m labelled red blood cells is injected. Pediatric doses are adjusted
appropriately.
Equipment
A LFOV gamma camera with a
low-energy, high resolution parallel hole collimator interfaced to a nuclear
medicine computer is preferred. The camera's analyzer is set at 140 keV with a
15 to 20% window. The computer acquisition uses a 128 x 128 x byte (or word)
mode matrix (64 x 64 x byte mode if storage space is limited).
Image /
Data Acquisition Parameters
The patient should be
placed
supine
with the field of view and adjusted to include only the lower portion of
the liver and spleen and the entire abdomen and pelvic region.
An anterior flow is
obtained at a rate of 2 to 5 sec/frame for 1 to 2 min. This is followed by 10
to 60 second acquisitions per image for 60 to 90 min (i.e. cinescintigraphy).
If computer acquired cinescintigraphy is not possible, acquire 750k to 1000k
count static images every 2 to 5 minutes for 60 to 90 minutes. Shorter imaging
intervals increase the likelihood of identifying the origin of a
bleed.
If bleeding is not
detected, delayed images may be obtained at appropriate intervals between 2 to
6 hours and 18 to 24 hours. The more frequently images are taken, the higher
the probability of detecting an intermittent bleed.
The protocols posted by
The Society of Nuclear
Medicine (in Adobe *.pdf format) and
The
Crump Institute for Biological Imaging at UCLA provide clinical examples of
GI Bleeding protocols using Tc-99m labelled red blood cells.
Data
Analysis
Images are viewed in cine
mode. Images may be combined together for production of hard copy
images.
 Return to Top of
Page
Because of its rapid
clearance from the circulatory system, the Tc-99m sulfur colloid method
requires that the bleeding site be actively hemorrhaging during the study. This
being the case, the bleeding site will be observed as a focal accumulation of
the radiopharmaceutical that increases in intensity with time and moves
distally (usually, although retrograde movement is possible) through the
intestine during the study.
Structures
normally visible include the liver, spleen and bone marrow of the vertebrae
and pelvic girdle.
Failure to observe a
bleeding site may only indicate that the patient was not actively bleeding
during the study. This is the primary disadvantage of the sulfur colloid method
(as it is with radiographic angiography). If bleeding is strongly suspected, a
second injection may be indicated.
This method is very
sensitive due to the
very
low background and bleeding rates in the order of 0.05 to 0.10 mL/min may
be detected.
False positive
interpretations may result from:
- transplanted kidneys
which will take up colloid during rejection
- splenic tissue such as
ectopic or accessory spleens
- modified, asymmetric
marrow uptake caused by myelofibrosis, tumor (primary or metastatic) or
avascular necrosis
- male
genitalia
- arterial grafts
False negative
interpretations may result from:
- patient not actively
bleeding at time of injection
Interpretation of a Tc-99m Red Blood Cell
Study
Because they remain within
the circulatory system, the Tc-99m red blood cell method does not require that
the bleeding site be actively hemorrhaging at the time of injection. The flow
study normally demonstrates the major vessels and organs of the abdomen
including the aorta, inferior vena cava, iliac vessels, liver, spleen and
kidneys. Genitals may also be visualized, especially the uterus.
A
positive
bleeding site appears as an abnormal accumulation of activity that persists
or increases with time and conforms to the shape of the bowel. The activity
moves distally (usually) through the intestine during the study.
Although not as sensitive
(~0.10 to 0.50 mL/min) as the sulfur colloid method due to the higher vascular
background, a major advantage of the red cell method is its ability to detect
intermittent bleeding. If bleeding is not detected within the first 60 to 90
minutes of the study, the patient can be periodically returned to the Nuclear
Medicine department for additional imaging. The probability of identifying the
location of the bleeding site is proportional to the frequency of
imaging.
False positive
interpretations may result from:
- free Tc-99m
pertechnetate secreted by the gastric mucosa
- renal tissue such as
pelvic or ectopic kidneys
- renal pelvic retention
of activity
- hepatic
hemangioma
- varices and aneurysms
of abdominal veins and arteries.
 Return to Top of
Page
The primary technical
concern when using the red blood cell method is ensuring that an adequate cell
labelling efficiency has been achieved. Several medications and therapeutic
medications can interfere with the efficiency of cell labelling. Using an in
vitro labelling procedure is the best way to ensure a high quality yield is
achieved.
Artifacts leading to false
positive interpretations have been discussed above.
 Return to Top of
Page
Radiographic
Angiography
Endoscopy
An endoscope may be
inserted into both the
upper
and lower gastrointestinal tract in an attempt to visually locate and examine
bleeding sites.
 Return to Top of
Page
Case #1 is located at
the Mallinckrodt Institute of Radiology at Washington University Medical
Center. A 53-year-old presents with dizziness and maroon stools. Anterior
images at 5 minute intervals using Tc-99m RBCs are provided. Note the option to
view the study in cine format. Try it.
Case #2 is located at
the Mallinckrodt Institute of Radiology at Washington University Medical
Center. The patient presented with a history of blood per rectum with the
latest episode appearing to have stopped the day following admission. Anterior
images at 2 minute intervals using Tc-99m RBCs are provided.
Case
#3 is hosted by The Joint Program in Nuclear Medicine (JPNM) based at
the Harvard Medical School. A 62-year-old man with known hypertension suffers a
pulmonary embolism post surgery for tumor resection. He is placed on
anticoagulant therapy after which he demonstrates hematochezia x 2.
Case
#4 is hosted by The Crump Institute for Biological Imaging at UCLA. A
71-year-old woman presents with diverticular bleeding.
Case
#5 is hosted by The Crump Institute for Biological Imaging at UCLA. A
30 year old with hematemesis and hematochezia is evaluated for possible lower
GI bleeding.
Case
#6 is hosted by The Crump Institute for Biological Imaging at UCLA. A
63 year old female with history of G.I. Bleeding requiring 1 unit packed RBC's
every week for approx. 6 months. The patient also had a negative exploratory
laparotomy approximately 1 week ago and now presents with black-tarry
stools.
Case
#7 is from the Virtual Hospital at the University of Iowa. The document
is actually a complete lesson on GI bleeding and is very good. Some links on
this page were not functioning properly at the time this document was
posted.
 Return to Top of
Page
Pathology
Texts
Damjanov I. Pathology for the
Health - Related Professions. Philadelphia: W.B. Saunders Company, 1996:
257.
Kumar V, Cotran RS, Robbins SL.
Basic Pathology. 5th ed. Philadelphia: W.B. Saunders Company, 1992:
473.
McCance KL, Huether SE.
Pathophysiology. 2nd ed. St. Louis: Mosby - Year Book, 1994: 1324 -
1326.
Price SA, Wilson LM.
Pathophysiology: Clinical Concepts of Disease Processes. 4th ed. St.
Louis: Mosby - Year Book, 1992: 356.
Procedure
Texts
Datz FL. Handbook of Nuclear
Medicine. 2nd ed. St. Louis: Mosby - Year Book, 1993: 133.
Bernier DR, Christian PE, Langan JK.
Nuclear Medicine: Technology and Techniques. 3rd ed. St. Louis: Mosby -
Year Book, 1994: 329 - 331.
Early PJ, Sodee DB. Principles
and Practice of Nuclear Medicine. 2nd ed. St. Louis: Mosby - Year Book,
1995: 516 - 520.
Henkin
Thrall JH, Ziessman HA. Nuclear
Medicine: The Requisites. St. Louis: Mosby - Year Book, 1995: 241 -
248.
Wagner H
Journals
Videos
Internet
URL's
GI
Bleeding Scintigraphy in
Nuclear
Medicine Review Manual by Dr. Scott C. Williams
 Return to Top of
Page
|