mri demonstration of peritoneal implants

7
Abdominal Imaging Springer-Verlag New York Inc. 1994 MRI Demonstration of Peritoneal Implants C.-K. Chou, 1'* G.-C. Liu, 1 J.-H. Su, 2 L.-T. Chen, 3 R.-S. Sheu, 1 T.-S. Jaw s ]Department of Radiology, Kaohsiung Medical College, No. 100 Shih-Chuan I st Road, Kaohsiung, Taiwan, Republic of China 2Department of Obstetrics and Gynecology, Kaohsiung Medical College, No. 100 Shih-Chuan 1 st Road, Kaohsiung, Taiwan, Republic of China 3Department of Internal Medicine, Kaohsiung Medical College, No. 100 Shih-Chuan 1 st Road, Kaohsiung, Taiwan, Republic of China Received: 16 February 1993/Accepted: 22 March 1993 Abstract. The magnetic resonance imaging (MRI) find- ings of 12 proven cases of peritoneal implants, mainly carcinomatosis, were reviewed for evidence of perito- neal seedings. The seeded sites include the pouch of Douglas, the ileocecal and retrocecal regions, the fight and left paracolic gutters, Morison's pouch, the fight subdiaphragmatic parietal peritoneum, the greater and lesser omentum, the gastrocolic, gastrosplenic, and phrenicocolic ligaments, the small bowel mesentery, the sigmoid and transverse mesocolons, and the small and large bowel walls. Sizes varied from less than 1 cm to omental cake and bulky tumors. The findings include linear or tiny nodular infiltrations of the omentum and subperitoneal fat (ligamentous, mesenteric, and meso- colic), focal or segmental wall thickenings, loss of uni- lateral colonic haustration with sacculation on the con- tralateral side, and nodular soft tissue masses along dif- ferent locations of the peritoneal surfaces. Air was introduced via an antegrade or retrograde method to act as a gastrointestinal contrast agent and was found to be useful for delineating the seedings. As is true with com- puted tomography scan, miliary implants are also not detectable with MRI. The sensitivity and specificity of MRI in detecting peritoneal implants remain to be de- termined. Key words: Peritoneal implants--MR studies--Air-- Intraperitoneal metastases. Computed tomography (CT) is currently the choice im- aging modality in evaluating peritoneal implants. Bar- B * Present address: Department of Radiology, Chi Mei Foundation Hospital, No. 901 Chung-Wha Road, Tainan, Taiwan, Republic of China. Correspondence to: C.-K. Chou ium studies are also used on occasion. However, thus far, it is agreed that magnetic resonance imaging (MRI) is not satisfactory in detecting peritoneal seedings. The reasons include: (1) lack of an adequate gastrointestinal contrast agent; (2) motion artifacts due to respiration, peristalsis, and cardiac pulsation; and (3) the spatial res- olution of MRI is slightly inferior to CT. In an earlier report [1], we presented the MRI findings of peritoneal carcinomatosis in the colonic and gastric walls, small bowel mesentery, and greater omentum. In the follow- ing illustrations, we present the MRI findings of peri- 3) (i! LP( Q Abdom Imaging 19:95-101 (1994) Fig. 1. The seeding sites and number of implants (parentheses) are summarized in A and B. DP, pouch of Douglas: ICR, ileocecal region; SPR, subphrenic region; LPG, left paraeolic gutter; MP, Morison's pouch; RPG, right paracolic gutter; GSL gastrosplenic ligament; PCL, phrenicocolic ligament; SMC, sigmoid mesocolon; TMC, transverse mesocolon; GCL, gastrocolic ligament; GO, greater omentum; LO, lesser omentum; SBM, small bowel mesentery.

Upload: c-k-chou

Post on 06-Jul-2016

213 views

Category:

Documents


1 download

TRANSCRIPT

Page 1: MRI demonstration of peritoneal implants

Abdominal Imaging

�9 Springer-Verlag New York Inc. 1994

MRI Demonstration of Peritoneal Implants

C.-K. Chou, 1'* G.-C. Liu, 1 J.-H. Su, 2 L.-T. Chen, 3 R.-S. Sheu, 1 T.-S. Jaw s ]Department of Radiology, Kaohsiung Medical College, No. 100 Shih-Chuan I st Road, Kaohsiung, Taiwan, Republic of China 2Department of Obstetrics and Gynecology, Kaohsiung Medical College, No. 100 Shih-Chuan 1 st Road, Kaohsiung, Taiwan, Republic of China 3Department of Internal Medicine, Kaohsiung Medical College, No. 100 Shih-Chuan 1 st Road, Kaohsiung, Taiwan, Republic of China

Received: 16 February 1993/Accepted: 22 March 1993

Abstract. The magnetic resonance imaging (MRI) find- ings of 12 proven cases of peritoneal implants, mainly carcinomatosis, were reviewed for evidence of perito- neal seedings. The seeded sites include the pouch of Douglas, the ileocecal and retrocecal regions, the fight and left paracolic gutters, Morison's pouch, the fight subdiaphragmatic parietal peritoneum, the greater and lesser omentum, the gastrocolic, gastrosplenic, and phrenicocolic ligaments, the small bowel mesentery, the sigmoid and transverse mesocolons, and the small and large bowel walls. Sizes varied from less than 1 cm to omental cake and bulky tumors. The findings include linear or tiny nodular infiltrations of the omentum and subperitoneal fat (ligamentous, mesenteric, and meso- colic), focal or segmental wall thickenings, loss of uni- lateral colonic haustration with sacculation on the con- tralateral side, and nodular soft tissue masses along dif- ferent locations of the peritoneal surfaces. Air was introduced via an antegrade or retrograde method to act as a gastrointestinal contrast agent and was found to be useful for delineating the seedings. As is true with com- puted tomography scan, miliary implants are also not detectable with MRI. The sensitivity and specificity of MRI in detecting peritoneal implants remain to be de- termined.

Key words: Peritoneal implants--MR s tud ies - -Ai r - - Intraperitoneal metastases.

Computed tomography (CT) is currently the choice im- aging modality in evaluating peritoneal implants. Bar-

B

* Present address: Department of Radiology, Chi Mei Foundation Hospital, No. 901 Chung-Wha Road, Tainan, Taiwan, Republic of China.

Correspondence to: C.-K. Chou

ium studies are also used on occasion. However, thus far, it is agreed that magnetic resonance imaging (MRI) is not satisfactory in detecting peritoneal seedings. The reasons include: (1) lack of an adequate gastrointestinal contrast agent; (2) motion artifacts due to respiration, peristalsis, and cardiac pulsation; and (3) the spatial res- olution of MRI is slightly inferior to CT. In an earlier report [1], we presented the MRI findings of peritoneal carcinomatosis in the colonic and gastric walls, small bowel mesentery, and greater omentum. In the follow- ing illustrations, we present the MRI findings of peri-

3 ) �9

(i!

LP(

Q

Abdom Imaging 19:95-101 (1994)

Fig. 1. The seeding sites and number of implants (parentheses) are summarized in A and B. DP, pouch of Douglas: ICR, ileocecal region; SPR, subphrenic region; LPG, left paraeolic gutter; MP, Morison's pouch; RPG, right paracolic gutter; GSL gastrosplenic ligament; PCL, phrenicocolic ligament; SMC, sigmoid mesocolon; TMC, transverse mesocolon; GCL, gastrocolic ligament; GO, greater omentum; LO, lesser omentum; SBM, small bowel mesentery.

Page 2: MRI demonstration of peritoneal implants

96 C.-K. Chou et al.: MRI Demonstration of Peritoneal Implants

Fig. 2. Spin-echo Tl-weighted sagittal image of the case of gastric carcinoma. Tumor seedings are present at the pouch of Douglas (white arrow), mesentery (black arrows), and between the transverse colon (TC) and the jejunum (J).

Fig. 3. Spin-echo Tl-weighted coronal (A) and sagittal (B) images of a case of ovarian carcinoma. Tumors are noted at the ileocecal (white arrow) and retrocecal (black arrow) regions. The sigmoid mesocolon (white arrowhead) is displaced upward by a huge ovarian carcinoma. The small bowel mesentery (black arrowhead) is displaced by pa- raaortic lymphadenopathy (LAP).

toneal implants in various locations of the abdominal cavity in 12 proven cases. Familiarity with these find- ings may be helpful in the employment of MRI to eval- uate the intraabdominal spread of malignancy.

gastrointestinal contrast agent. The antegrade method uses a nasogas- tric tube to introduce about 2000 ml of air into the stomach. Metoclo- pramide is used to propel air through the entire gastrointestinal tract [3]. As regards the retrograde method, the first five patients received air insufflation in three different positions, changing from left lateral decubitus to prone, and finally to right lateral decubitus with about 400 ml of air introduced in each position [4]. The other five patients received MR examinations, although modified procedures were em- ployed. The patients' positions were changed from right lateral de- cubitus to supine, then left lateral decubitus, and finally prone position with about 330 ml of air introduced in each position. The total amount was about 1100-1500 ml. The patient received MR examinations in supine position. Scopolamine bntylbromide, 40 rag, was injected to suppress bowel peristalsis in both antegrade and retrograde methods.

Seven cases received CT examinations as well. However, due to the limitation of insurance payment, these CT examinations were lim- ited to the upper or lower half of the abdomen. Corresponding levels of scanning were not available in both CT and MRI examinations in most instances.

Materials and Methods

The subjects included 10 cases of surgically proven peritoneal carci- nomatosis, one case of surgically proven tuberculous peritonitis, and one case of peritoneal lymphomatosis, who did not receive laparot- omy. These included five men and seven women, who ranged from 21-70 years of age, with a mean age of 51 years. The underlying malignancies are ovarian carcinoma (four cases), colonic carcinoma (five cases, two of which were included in an earlier report [2]), and gastric carcinoma (one case, included in an earlier report [1]).

The MR examination was performed on a 0.5T superconductive scanner (GE, MR MAX plus system, Milwaukee, USA). The routine parameter was spin-echo Tl-weighted pulse sequence with TR/TE range of about 300-650/15-30 ms. Sagittal, coronal, and axial im- ages were obtained. The slice thickness was 7 - 1 0 mm, with a 1 -2 mm gap. The matrix size was 192 • 192 (coronal), 160 • 192 (sag- ittal), and 192 • 160 (axial). A T2-weighted pulse sequence was not used. Seven patients received a gadopentetate dimeglumine (Magne- vist, Schering AG, Germany) injection with a dose of 0.1 mmol/kg. Cardiac and respiratory gatings were not used. Air was introduced via the antegrade (2 cases) or retrograde (10 cases) method to act as a

Results

The seeding sites and number of implants are summa- rized in Fig. 1. These include the pouch of Douglas (Fig. 2), the ileocecal and retrocecal regions (Fig. 3), the right and left paracolic gutters (Fig. 4), Morison' s pouch (Fig. 5), the right subdiaphragmatic parietal peritoneum (Fig. 6), the greater and lesser omentum (Figs. 7 and 8), the gastrosplenic (Fig. 9), gastrocolic (Fig. 10), and phren- icocolic (Fig. 11) ligaments, the small bowel mesentery (Fig. 12), the sigmoid and transverse mesocolons (Fig. 13), and the bowel walls (Fig. 14). The sizes varied from less than 1 cm to omental cake and bulky tumors. The appearance of the seedings included linear and tiny nod- ular infiltrations within the greater omentum, various ligaments and mesocolons, a stellate pattern in the small bowel mesentery, focal or segmental wall thickenings of the air-distended small and large bowels, loss of uni-

Page 3: MRI demonstration of peritoneal implants

C.-K. Chou et al.: MRI Demonstration of Peritoneal Implants 97

Fig. 4. Two cases of ovarian carcinoma. Spin-echo Tl-weighted axial (A) and cor- onal (B) images show the small tumor nodules in the right (white arrow) and left (white arrowhead) paracolic gutters. The tumors are identified between the air-dis- tended colons and ascites. Paraaortic lym- phadenopathy (black arrows) is noted. Ovarian carcinoma (OC) with hemor- rhagic cystic components are surrounded by air-distended small bowels.

Fig. 5. Ovarian carcinoma. Spin-echo T1- weighted sagittal image shows a tumor mass in Morison's pouch (white arrow).

Fig. 6. Nodular thickening due to perito- neal seeding is noted along the right sub- diaphragmatic parietal peritoneum (white arrow) in a case of ovarian carcinoma.

lateral colonic haustration with sacculation on the con- tralateral side, tumor masses in Morison's pouch and the subdiaphargmatic parietal peritoneum, and small and large bowel obstructions (Fig. 15).

The case of peritoneal lymphomatosis did not re- ceive laparotomy. However, due to the similarity of omental infiltrations to those of peritoneal carcinoma- tosis and dramatic change after chemotherapy (Fig. 16), we have included it in this study. The case of tubercu- lous peritonitis had innumerable miliary implants, about 1 - 2 mm, scattered diffusely throughout the entire ab- dominal cavity. The subperitoneal fat infiltrations are similar to those of carcinomatosis and lymphomatosis. However, the infiltration along the right medial para- colic gutter is very thin and is only clearly visible on sagittal section (Fig. 17). The findings on axial and coro- nal sections are very subtle. The right subdiaphragmatic parietal peritoneum showed mild thickening with rela- tively smooth contour. This thickening is only identified after gadopentetate dimeglumine (Gd-DTPA) adminis- tration (Fig. 18). Gd-DTPA enhancement is useful in distinguishing the implants from the contiguous ascites/

soft tissue structures and excluding the possibility of edematous or fibrous components. However, this en- hancement made the implants less distinguishable from the surrounding fat (Fig. 19). The miliary implants on the mesentery and bowel wall, whether air-distended or not, were still not detectable by MRI as they were by CT. Ascites and paraaortic lymphadenopathy were pres- ent in six and three cases, respectively.

Discussion

The most common locations of peritoneal implantation are the pouch of Douglas, the ileocecal region, the right paracolic gutter, the sigmoid mesocolon, the greater omentum, and the right subdiaphragmatic parietal per- itoneum. This is related to the dependency of the pelvic cavity and the dynamic flow of intraperitoneal fluid and lymphatic drainage which have been widely studied and described [5-7]. In the present report, we show the var- ious appearances and locations of peritoneal implants in MRI. The factors which influence the detection of per-

Page 4: MRI demonstration of peritoneal implants

98 C.-K. Chou et al.: MRI Demonstration of Peritoneal Implants

Fig. 7. Gadolinium-enhanced spin-echo Tl-weighted axial image (A) and enhanced axial image of computed tomography (B) at the same level show similar infiltrations in the greater omentum (arrows).

Fig. 8. Ovarian carcinoma. A tumor mass with intermediate signal intensity is noted at the fissure for the ligamentum venosum, the site of the lesser omentum (white arrow). The sigmoid mesocolon (black arrows) is displaced upward by the pelvic-origin ovarian carcinoma.

Page 5: MRI demonstration of peritoneal implants

C.-K. Chou et al.: MRI Demonstration of Peritoneal Implants 99

Fig. 12. Spin-echo Tl-weighted sagittal im- age shows stellate infiltrations (black ar- row) in the small bowel mesentery. The in- ferior margin of the hepatic flexure is also thickened by tumor seeding (white arrow).

Fig. 13. Spin-echo T 1-weighted coronal (A) and sagittal (B) images demonstrate linear and nodular infiltrations in the sigmoid (white arrow) and transverse mesocolons (white arrowhead).

Fig. 14. Spin-echo T 1-weighted coronal im- ages in two cases of recurrent anastomotic rectosigmoid carcinoma with obstruction. A Small plaquelike seedings (arrows) are noted on the wall of the air-distended colon. B Tumor implants along the inferior margin of the transverse colon (arrow) cause loss of unilateral haustration with sacculation on the contralateral side (arrowhead).

"91 Fig. 9. Spin-echo Tl-weighted sagittal (A) and coronal (B) images show linear and tiny nodular infiltrations in the gastrosplenic ligament (white arrow) and greater omentum (white arrowhead). Stetlate pat- tern in the small bowel mesentery is noted (black arrow). Tumor seed- ing in the sigmoid mesocolon is also found (black arrowhead). SF, splenic flexure; S, spleen.

Fig. 10. Spin-echo Tl-weighted coronal (A) and sagittal (B) images show a smudged appearance of the gastrocolic ligament (white arrow). S, stomach; GPL, gastrophrenic ligament; TMC, transverse mesoco- Ion; TC, transverse colon.

Fig, 11. Ovarian carcinoma. Spin-echo Tl-weighted coronal image shows a linear infiltration (black arrow) in the phrenicocolic ligament. The right end of the transverse mesocolon also shows soft tissue per- meation (white arrow) similar to that of the greater omental infiltration commonly seen on CT.

itoneal implants on CT include size and location of the implants, and the presence of adjacent ascites. Some authors have stated that implants smaller than 1 cm are not detectable [8]. Others have emphasized the impor- tance of location of the implants and the presence of ascites, stating that implants located in the subphrenic region or profiled by ascites are readily detectable even when smaller than 5 mm [7]. In our experience with MRI examinations, the linear and tiny nodular infiltra- tions of various subperitoneal fat (omentum, ligaments, and mesocolons) are usually smaller than 1 cm and read- ily detected as they are in CT examinations. With regard to the locations of the implants, most of the favored sites of peritoneal seedings, such as the pouch of Douglas, the sigmoid mesocolon, the right paracolic gutter, and

Page 6: MRI demonstration of peritoneal implants

100 C.-K. Chou et al.: MRI Demonstration of Peritoneal Implants

Fig. 15. Spin-echo Tl-weighted axial image of a case with small bowel obstruction. The proximal dilated loops are filled with gas-fluid mix- ture (P). The distal bowel loops are distended by retrogradedly insuf- flated air (D). The tumor mass (white arrow) is delineated by the air- distended distal loops.

Fig. 16, Peritoneal lymphomatosis. Spin-echo T1-weighted sagittal images before (A) and after (B) chemotherapy show nearly complete disappear- ance of omental infiltrations (arrows). The paraaortic and mesenteric lym- phadenopathy is also markedly decreased in size (arrowheads).

Fig. 17. Tuberculous peritonitis. Spin-echo Tl-weighted, 7 mm thick, sagittal image shows a smudged infiltration along the right medial

paracolic gutter (arrow). The infiltration is present only in this slice. Coronal and axial sections did not show the infiltration as well.

Fig. 18. Tuberculous peritonitis. Spin-echo Tl-weighted axial images before (A) and after (B) Gd-DTPA administration. The thickened sub- diaphragmatic parietal peritoneum is identified only after Gd-DTPA enhancement (arrows).

Fig. 19. Spin-echo Tl-weighted sagittal images before (A) and after (B) Gd-DTPA administration. The tumor infiltrations (arrows) be- come less distinguishable from the surrounding fat after Gd-DTPA enhancement.

Page 7: MRI demonstration of peritoneal implants

C.-K. Chou et ai.: MRI Demonstration of Peritoneal Implants 101

the ileocecal region, are in contact with the colon and the ileum. In most of our cases, the colon, the ileum, and the bowel distal to the obstruction were very well distended by the antegradedly or retrogradedly intro- duced air. Many of the implants shown in our reports were indeed outlined between intraluminal air and ex- traluminal fat/ascites. The right subphrenic implants were also detectable by MRI. While most implants have a higher signal intensity than ascites, the differentiation of the implants from the ascites is easier if Gd-DTPA is used. Isolated involvement of the duodenum and the jejunum were not encountered in our series, and theo- retically are very rare. Consequently, the possibility of misinterpreting nondistended crowded jejunal loops as a tumor is relatively low if the other favored sites of peritoneal implants are not involved. The discrete mil- iary implants, about 1 -3 ram, on the air-distended bowel wall are still nondetectable. The possible reasons include truncation artifacts, motion artifacts, and partial volume averaging. The miliary seedings on the col- lapsed bowel are also not recognizable because both have similar signal intensities in pre- and postgadoli- nium images. The multiplanar capability of MRI can better demonstrate the seeding in different sections (Fig. 17). The lack of intraabdominal fat is a limiting factor for clear identification of peritoneal seedings.

With regard to the method of air introduction, we prefer the retrograde approach because the procedure and the amount of introduced air are similar to those of barium enema. One thousand to 1500 ml of air is suf- ficient to distend the colon and the ileum in most cases [4]. In this study, the retrograde method was employed in 10 cases while the antegrade method was used in only two.

The sensitivity and specificity of MRI in detecting peritoneal implants remain to be determined. So far, CT scanning is still the first choice of imaging modality in evaluating peritoneal implants. However, our results suggest that MRI also can show the various manifesta- tions of peritoneal seedings. Furthermore, the multi- planar images of MRI can sometimes better demonstrate the lesions. Air-distended bowels are helpful in deline- ating the implants if they are in the proximity of, or in contact with, the bowels. Both pre- and postgadolinium images should be obtained for better identification of the implants.

Acknowledgments. We thank Y. E. Chiu and F. O. Hsu for technical assistance and C. C. Lin for typing the manuscript.

References

1. Chou CK, Liu GC, Chen LT, Jaw TS. MRI manifestations of per- itoneal carcinomatosis. Gastrointest Radiol 1992;17:336-338

2. Chou CK, Liu GC, Chen LT, Jaw TS. The use of MRI in bowel obstruction. Abdom Imaging 1993; 18:131 - 135

3. Chou CK, Liu GC, Yang CW, Chen LT, Sheu RS, Jaw TS. Ab- dominal MR imaging following antegrade air introduction into the intestinal loops. Abdom Imaging 1993;18:205-210

4. Chou CK, Liu GC, Chen LT, Jaw TS. Retrograde air insufflation in MRI: a technical note. Abdom hnaging 1993;18:211-214

5. Meyers MA. Intraperitoneal spread of infections. In: Meyers MA, ed. Dynamic radiology of the abdomen." normal and pathologic anatomy (3rd ed). New York: Springer-Verlag, 1988:49-90

6. Meyers MA. Intraperitoneal spread of malignancies. In: Meyers MA, ed. Dynamic radiology of the abdomen: normal and patho- logic anatomy (3rd ed). New York: Springer-Verlag, 1988:91 - 178

7. Buy JN, Moss AA, Ghossain MA, et al. Peritoneal implants from ovarian tumors: CT findings. Radiology 1988;169:691-694

8. Levitt RG, Sagel SS, Stanley RI. Detection of neoplastic involve- ment of the mesentery and omentum by computed tomography. A JR 1978;I31:835-838