Localization of Gastrointestinal Bleeding by Cinematic 99m Tc Labeled Red Blood Cell Scan Chia-Shang Wu 1, Chang-Chung Lin 1, Nan-Jing Peng 1, 1 Department of Nuclear Medicine, Kaohsiung Veterans General Hospital, Kaohsiung, Taiwan National Yang-Ming University, Taipei, Taiwan Purpose: To evaluate the accuracy of correct localization of gastrointestinal (GI) bleeding by 99m Tc-red blood cell (RBC) scans. Methods: We reviewed our records with 30 patients and 34 positive cinematic RBC scans for GI bleeding from Aug. 1999 to Mar. 004, and compared with the records from other examinations including panendoscopy, abdominal computed tomography (CT), angiography and colonic scope, and prior surgical procedure were also reviewed, to evaluate the accuracy of our records. Results: Thirty patients with 34 positive cinematic RBC scans were enrolled. Twenty-one of the 30 patients (70%) and 4 of the 34 scans (71%) were successful to localize the bleeding sites by cinematic RBC scans, the accuracy of cinematic RBC scans compared with other examinations or operation was 88% for gastric bleeding, 56% for small intestinal bleeding and 100% for colonic bleeding. Among 30 patients, 5 of 6 patients with recent operation history for GI bleeding were proven that their recurrent bleeding was happened at the formal operation site. Conclusion: The cinematic RBC scans have high accuracy of correct localization of GI bleeding and the operation history of patients is important clue to determine location of bleeding sites. Received 7/14/006; revised 8/17/006; accepted 8/18/006. For correspondence or reprints contact: Nan-Jing Peng, M.D., Department of Nuclear Medicine, Kaohsiung Veterans General Hospital, 386 Ta-Chung 1st Rosd, Kaohsiung 813, Taiwan, ROC. Tel: (886)7-3411 ext. 6418, Fax: (886)7-346834, E-mail: njpeng@isca.vghks.gov.tw Key words: gastrointestinal bleeding, stomach, 99m Tc- RBC scans Ann Nucl Med Sci 007;0:1-5 Introduction Gastrointestinal (GI) bleeding was occurred very often in clinical condition; the incidence has been reported around 40 to 150 episodes per 100,000 persons per year for acute upper GI bleeding and 0 to 7 episodes per 100,000 persons for lower GI bleeding [1]. Several methods were used to locate bleeding sites such as computed tomography (CT), endoscopy, angiography and nuclear scintigraphy, and each of them has its own advantages and disadvantages. For example, angiography could be used not only as a diagnostic tool but also as a therapeutic tool; but it is an invasive procedure. The disadvantage of 99m Tc-red blood cell (RBC) scans was weak in anatomic accuracy, but strong in detecting a small amount of bleeding. The purpose of this study was to evaluate the accuracy of correct localization of GI bleeding by cinematic RBC scans. Materials and Methods Patients We reviewed the results of 3 RBC scans in 193 patients for GI bleeding at Kaohsiung Veterans General Hospital from Aug. 1999 to Mar. 004. Patients were excluded if bleeding sites were not found by the RBC scans or a positive RBC scan but a negative operation, endoscopy, angiography or abdominal CT scan. Thirty four scans in 30
Wu CS et al patients were confirmed with a positive bleeding by other studies such as endoscopy, abdominal/pelvic CT, angiography or by operation. Therefore, there were totally 30 patients enrolled to this study, 1 were males and 18 females, aged from 34 to 86 years. Protocol An in-vivo method of labeling RBC was used. First, an intravenous injection of mg stannous pyrophosphate (cold PYP) was given, and 15 min later, another 0 mci (740 MBq) 99m Tc-pertechnetate injection would be following to give and those two injections should be given in the different site. Scintigraphy The patients were supine imaged with a single head gamma camera (Siemens Orbitor 75; Siemens Corp., Hoffman Estates, Illinois, USA) equipped with a low-energy, parallel-hole, all-purpose collimator and interfaced with a computer (Siemens MicroDelta/MaxDelta or ICON AP). Images of anterior abdominal were acquired at frame rate of one frame per minute in a 18 18 matrix with 15% windows centered on the 140-keV. A frame-mode of data, 1 frame/per min was used throughout the study, and images started after the second injection of 0 mci (740 MBq) 99m Tcpertechnetate and stopped when got enough information to define the bleeding sites. Image Interpretation Then, the nuclear physician checked those images and according to its result showed the active bleeding to localize the bleeding site. The bleeding sites were traced according to the location of abnormal RBC accumulation when it first appeared and the time it took to pass through the GI tract. A quadrant division of abdomen (right upper quadrant; RUQ, left upper quadrant; LUQ, right lower quadrant; RLQ and left lower quadrant; LLQ) was used for localizing the bleeding site for small intestine. The bleeding sites in a small intestine with initial presentation at upper quadrants were considered in proximal small intestine and the bleeding sites at lower quadrants were in distal small intestine. In addition, the time it drained to colon was also considered as a useful parameter if the time of scan was long enough. Results Among the 30 patients enrolled in this study, 13 bleeding sites were defined by operation, 11 by endoscopies, 3 by angiography and 3 by abdominal CT. In the identification of bleeding sites, the accuracy of cinematic RBC scans for gastric, small intestinal and colonic bleeding is 88%, 56% and 100% by patient and 89%, 57% and 100% by scan (Table 1). We noticed that the history of operation before GI bleeding is important for localizing the bleeding site. Among 6 patients who had the history of operation in GI tract, 5 were proven that its recurrent bleeding was at the formal operation site. Only one of them the bleeding site did not match with the formal operation site, it might be due to the formal operation was performed more than 0 years before this GI bleeding (Table ). A mistake happened in both RBC scan and abdominal CT about bleeding location, it happened in one patient who had a gastrointestinal stromal tumor (GIST) of jejunum with 7 cm in diameter in RLQ of abdomen. And the nuclear medicine physician and radiologist were misled by this tumor during they defined the bleeding location according to the data showed by RBC scan and abdominal CT, and the final diagnosis was done by operation (Figure 1). Discussion Among the 30 patients with 34 scans enrolled to this study, 1 of 30 patients and 4 of 34 scans had correct results and the correct rate is 70% (1/30) by patient and 71% (4/34) by scan. Nine patients with incorrect bleeding localization were occurred with 1 gastric ulcer bleeding and 8 bleeding sites in small bowel. Among the 18 patients of small intestinal bleeding, ten were correctly localized and eight were incorrect. 6 of the 8 incorrect localized patients Table 1. The rate of correct localization of the bleeding sites of 99m Tc-RBC scans by patients and scans Bleeding sites Patients (n = 30) Scans (n = 34) Stomach 7/8 (88 %) 8/9 (89%) Small intestine 10/18 (56%) 1/1 (57%) Colon 4/4 (100%) 4/4 (100%) Ann Nucl Med Sci 007;0:1-5 Vol. 0 No. 1 March 007
-99m 99m Tc-RBC scans for GI bleeding Table. The relationship among the previous gastrointestinal surgical sites, the time interval of surgery, 99m Tc-RBC scans and GI bleeding sites of six patients Patient Surgical site Procedure Time interval * GI bleeding site 1 GE junction Esophegoenterostomy 11 years GE junction Stomach Subtotal gastrectomy 0 years Subtotal gastrectomy wound 3 Stomach Subtotal gastrectomy weeks Subtotal gastrectomy wound 4 Small bowel Segmental resection 4 days Ileojejunal small bowel 5 Small bowel Segmental resection to 4 days Terminal ileum 6 ** Stomach For peptic ulcer 0 years Small bowel * Time interval: The interval between previous GI surgery and this GI bleeding Gastroesophageal The patient received three times of surgery for GI bleeding and 3 99m Tc-RBC scans for initial and post-operative GI bleeding localization within one week. ** The bleeding site was different from his previous operation site. Figure 1. A 64-year-old female with gastrointestinal stromal tumor at jejunum was incorrectly reported the bleeding site in ileum by 99m Tc-RBC scan due to the lesion presented in RLQ of abdomen. Only the duration of the tracer (84 to 1 min) circling in small intestine provided a clue to suppose the bleeding site above ileum. The arrows indicated the blood activity in small intestine. were due to underestimate the bleeding site to colon. After reviewing the data, we found that the bleeding sites in upper quadrants were mostly defined as jejunal or duodenal bleeding; even the bleeding sites in the lower quadrant were sometimes confirmed as jejunal bleeding. For example, a patient with small intestinal bleeding was incorrect localized of a bleeding site at downward displaced regional jejunum (GIST), because of the initial presentation of both RBC scans and CT studies showed it was at RLQ of abdomen, thus the bleeding site was defined as at terminal ileum. However, after reviewed the original reading data, according to the unreasonable 4 h circling time of blood in bowel loop, the bleeding site at proximal small intestine might be suggested. The rate of correct localization of GI bleeding had been reported in a wide range. The lowest rate in other studies we found were only 48% which was report- 007;0:1-5 007 3 0 1 3
Wu CS et al ed by Olds et al []. In his study, he found a higher positive rate in patients received more than units of blood transfusion within 4 h before RBC scans, but a higher positive rate didn t mean the localization being more accurate. Orellana et al reported a high rate of correct localization of active GI bleeding up to 93% by RBC scans [3], he had collected 59 patients with 40 positive scans and 57% of positive scans were positive during the first hour. So an active bleeding during scans had the great effect for the results. Other studies we had reviewed reported the rate of correct localization was similar to this study, such as in Dolezal et al study, he reviewed 31 patients with in vivo labeled RBC scans with 1 positive scans, 16 were confirmed bleeding and 11 of the 16 were proved correct localization with 69% of correct rate [4]. Until now we do not know the exact rate of re-bleeding after GI tract operation, and it might change according to many factors of the patients own condition. O Neill et al had reviewed 6 patients with GI bleeding, and after RBC scan, it was proven that twenty-five patients required an operation to control the bleeding and of them (88%) were correctly localized the bleeding site [5], and among them the 8% re-bleeding rate was also reported. Anthony et al had reviewed a total of 119 patients with acute lower GI bleeding and the re-bleeding rate was documented at years in 14 of 10 patients who survived more than weeks after lower GI bleeding [6]. According to our data, we suppose that the formal operation history of GI tracts is a helpful clue to define the bleeding sites. Alan et al had evaluated 5 scans to determine whether cinematic acquisition and display localized bleeding sites would be more accurate than traditional static imaging or not. They had showed that cinematic acquisition and display were not only helpful for nuclear physicians to localize the bleeding sites more accurate, but also increased the sensitivity of the studies to show the bleeding sites [7]. O Neill et al thought that cinematic RBC scan could show a real time imaging and avoid the false localization at interval imaging due to the retrograde peristalsis or rapid transit [5]. In this study, the advantage of cinematic RBC scan for correct localization of bleeding site was found. The accuracy of this study compared to the study of O Neill et al (70% vs. 88%) is similar, although the labeling method of RBC and case distribution were different. Conclusion RBC scans for GI bleedings could correctly locate the bleeding sites in 70% of patients. RBC scan has a non-invasive advantage and carefully defining initial presentation, monitoring blood drainage in GI tract and performing scans during active bleeding make the RBC scans in localizing the bleeding sites more accurate. In addition, it would be more accurate in reading the data of RBC scan, if the operation history of the patient is considered. References 1. Manning-Dimmitt LL, Dimmitt SG, Wilson GR. Diagnosis of gastrointestinal bleeding in adults. Am Fam Physican 005;71:1339-1346.. Olds GD, Cooper GS, Chak A, Sivak MV Jr, Chitale AA, Wong RC. The yield of bleeding scans in acute lower gastrointestinal hemorrhage. J Clin Gastroenterol 005;39:73-77. 3. Orellana P, Vial I, Prieto C, et al. 99m Tc red blood cell scintigraphy for the assessment of active gastrointestinal bleeding. Rev Med Chil 1998;16:413-418. [In Spanish; English abstract]. 4. Dolezal J, Vizd a J, Bures J. Detection of acute gastrointestinal bleeding by means of technetium-99m in vivo labelled red blood cells. Nucl Med Rev Cent East Eur 00;5:151-154. 5. O Neill BB, Gosnell JE, Lull RJ, et al. Cinematic nuclear scintigraphy reliably directs surgical intervention for patients with gastrointestinal bleeding. Arch Surg 000;135:1076-1081. 6. Anthony T, Penta P, Todd RD, Sarosi GA, Nwariaku F, Rege RV. Rebleeding and survival after acute lower gastrointestinal bleeding. Am J Surg 004;188:485-490. 7. Maurer AH, Rodman MS, Vitti RA, Revez G, Krevsky B. Gastrointestinal bleeding: improved localization with cine scintigraphy. Radiology 199;185:187-19. Ann Nucl Med Sci 007;0:1-5 Vol. 0 No. 1 March 007 4
-99m 99m Tc-RBC scans for GI bleeding -99m 1 1 1, 1-99m 1999 8 004 3-99m -99m 30 1 (70%) 34 4 (71%) 88% 56% 100% -99m -99m 007;0:1-5 95 7 14 95 8 17 95 8 18 813 386 (07)- 3411 6418 (07)-346834 njpeng@isca.vghks.gov.tw 007;0:1-5 007 3 0 1 5