| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Gastrointestinal Imaging |
1 From the Departments of Radiology (J.H.B., H.K.H., A.Y.K., T.K.K., E.Y.K., J.K.L.), Pathology (E.S.Y.), and Internal Medicine (S.J.M., S.K.Y., H.Y.J., J.H.K.), Asan Medical Center, University of Ulsan College of Medicine, 388-1 Poongnap-dong, Songpa-gu, Seoul 138-736, Korea. Received January 4, 2002; revision requested March 4; revision received June 3; accepted July 25. Address correspondence to H.K.H. (e-mail: hkha@amc.seoul.kr).
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSIONREFERENCES |
|---|
MATERIALS AND METHODS: CT scans were retrospectively reviewed in 14 patients with pathologically proved PTCLs of the gastrointestinal tract for the following considerations: sites, patterns of involvement (ie, morphologic features, bowel wall thickness or mass size, and contrast enhancement pattern), and ancillary findings at other sites (ie, lymphadenopathy, bowel perforation, and involvement of other organs).
RESULTS: PTCL involved the stomach in three patients, the small intestine in eight, both the stomach and the small intestine in one, and the sigmoid colon in two; multifocal involvement was seen in three (21%) patients. CT failed to demonstrate the bowel lesions in three of 14 patients. At CT, 11 patients had gastric or bowel wall thickening (n = 10) and a polypoid mass (n = 1). In 10 patients, the gastric or bowel wall thickening was mild (<1 cm) in six, moderate (1–2 cm) in three, and severe (>2 cm) in one. Lymphadenopathy was noted in nine (64%) patients, with the nonbulky type in eight and the bulky type in one. Bowel perforation occurred in four (29%) patients. Other organs were involved in eight (57%) patients.
CONCLUSION: CT can depict PTCL involving the gastrointestinal tract if it is not confined to the mucosa. There is a tendency toward preferential jejunal or duodenal involvement, as well as bowel perforation.
© RSNA, 2003
Index terms: Gastrointestinal tract, CT, 70.12112 • Gastrointestinal tract, neoplasms, 70.343 • Lymphoma, CT, 70.343 • Lymphoma, diagnosis
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSIONREFERENCES |
|---|
In general, PTCLs that develop primarily in the gastrointestinal tract are commonly associated with malabsorption, but many of the patients with this disease have no history of celiac disease or no biopsy evidence of gluten sensitivity and have normal-appearing jejunal mucosa (4). In contrast to the most common intestinal B-cell lymphomas, different clinicopathologic behaviors have also been noted (5). Although findings in previous reports (6,7) have suggested that the prognosis of PTCL is poor, those in other reports (1,8) indicate that there is no significant difference in outcome between patients with PTCL and those with B-cell lymphoma.
Computed tomography (CT) is increasingly used for evaluation of the patient who has or is suspected of having gastrointestinal lymphoma, because it offers the benefit of obtaining information about both the mural and extramural components of the disease. Thus, it allows for accurate staging of the disease and follow-up of the therapeutic response. A wide variety of CT and radiologic manifestations about gastrointestinal lymphoma have been described in the literature; however, most of them have focused on B-cell lymphoma. To our knowledge, little information has been published until recently about the radiologic findings in PTCL (5,9). The purpose of our study was to evaluate whether CT accurately depicted gastrointestinal tract involvement in PTCL.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSIONREFERENCES |
|---|
Histopathologic proof of the existence of PTCL was obtained at endoscopic biopsy in six patients (ie, with the duodenofiberscope in five and with the colonoscope in one) and at surgery in eight. The surgical indications in eight patients included acute abdominal emergency caused by bowel perforation in four and a bulky palpable mass in four. Segmental resection of the small intestine was performed in six patients; segmental resection of the sigmoid colon, in one patient; and subtotal gastrectomy, in one patient. Clinical symptoms in the 14 patients included weight loss (mean, 8.1 kg) in 10, abdominal pain or hunger pain in eight, acute severe abdominal pain in four, diarrhea in four, vomiting in three, fever in two, melena in one, and paraplegia with low back pain in one. Four patients had histories of malabsorption ranging from 20 days to 6 months, and the episodes of malabsorption were clinically characterized by chronic diarrhea and steatorrhea, weight loss, and low serum albumin levels (<3 g/dL [<30 g/L]).
Immunohistochemical studies (monoclonal antibodies to T-cell antigen, clonal rearrangement of the ß chain of the T-cell receptor gene) for the specimens obtained at biopsy or surgery revealed a peripheral T-cell origin in 10 patients and a natural killer cell origin in the remaining four. Epstein-Barr virus association was found in two patients. One had evidence of pulmonary pneumoconiosis, but the others had no underlying disease, such as celiac disease, systemic lupus erythematosus, Crohn disease, autoimmune disease, or acquired immunodeficiency syndrome. According to a recent classification system devised by the International Lymphoma Study Group (10), six patients had PTCL unspecified, four had angiocentric lymphoma, three had intestinal T-cell lymphoma with enteropathy, and one had intestinal T-cell lymphoma without enteropathy.
Imaging
Contrast material–enhanced abdominal CT had been performed in all 14 patients. CT was performed with one of three scanners (Somatom Plus-S or Somatom Plus-IV, Siemens, Erlangen, Germany; or HiSpeed Advantage, GE Medical Systems, Milwaukee, Wis) with 7- or 10-mm collimation and a pitch of 1.0–1.5 from the dome of the diaphragm to the symphysis pubis. The patients routinely received intravenously administered contrast material. In six patients, at least 900 mL of a 2% barium suspension (500 mL 1 hour before the scan and 400 mL immediately before the scan) was administered orally, and in the other eight patients, oral contrast material was not administered. The intravenously administered contrast medium iopamidol (Iopamiro 300; Bracco, Milan, Italy) was administered as a bolus through an 18-gauge angiographic catheter inserted in an antecubital vein with a digital injection system (LF CT 9000; Liebel-Flarsheim, Cincinnati, Ohio) at 3 mL/sec for a total dose of 120 mL. Contrast-enhanced CT scans were obtained at approximately 72 seconds after initiation of contrast agent injection.
Image Review
CT images were retrospectively reviewed in consensus by two radiologists (J.H.B., H.K.H.) who were aware that every case was PTCL involving the gastrointestinal tract but who were blinded to the location of involvement and to the endoscopic, surgical, and biopsy findings. The CT findings were evaluated with regard to the primary site involved, extent of involvement, and bowel wall involvement patterns of the lesions (ie, morphologic features, bowel wall thickness or mass size, and contrast enhancement pattern). Also evaluated were the presence or absence of lymphadenopathy, bowel obstruction, bowel perforation, intussusception, ascites, and other abdominal organ involvement.
The patterns of gastric lymphoma were classified into the following categories (11): diffuse infiltration involving greater than 50% of the length of the stomach, segmental infiltration, and localized polypoid form. The patterns of small and large intestinal lymphoma were classified as having a circumferential form, a cavitary form, a localized polypoid form, or a mesenteric nodal form (11–13).
In analyzing the gastrointestinal tract involvement patterns, we considered that bowel wall thickening was present when the thickness of the bowel wall exceeded 5 mm in the stomach and 3 mm in the small intestine and colon, with the lumen well distended (14,15). The degree of gastric or intestinal wall thickening was categorized as follows: mild, when the thickening was less than 1 cm; moderate, when the thickening was 1–2 cm; and severe, when the thickening exceeded 2 cm. The contrast enhancement of the gastrointestinal lesions was compared with that of the normal-appearing gastric or intestinal wall.
Lymphadenopathy was classified into four types, namely, bulky versus nonbulky and localized versus diffuse. Bulky lymphadenopathy was considered to be present when it was massive and conglomerate, and nonbulky lymphadenopathy was considered to be present when there were minimally enlarged and separable lymph nodes. Lymphadenopathy was considered localized when the lymph nodes were confined to the region surrounding the gastrointestinal lesions and diffuse when the lymph nodes extended to distant sites far from the gastrointestinal lesions. Hepatomegaly was defined when the liver was greater than 13 cm in longitudinal diameter at the midclavicular line, and splenomegaly was defined when the spleen was greater than 12 cm in the greatest diameter (16,17).
Histologic and Pathologic Findings
The histologic and pathologic findings were collected with a review of pathologic reports by one radiologist (J.H.B.), and these findings had been reported by a gastrointestinal pathologist (E.S.Y.). The pathologist classified PTCL according to a recent classification system delineated by the International Lymphoma Study Group (10).
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSIONREFERENCES |
|---|
|
At CT in 11 of 14 patients, the gastrointestinal lesions of PTCL were well demonstrated as bowel or gastric wall thickening in 10 and as a polypoid mass in one; there was no false-positive diagnosis of bowel involvement. In 10 patients, the degree of gastric or bowel wall thickening was mild (<1 cm) in six, moderate (1–2 cm) in three, and severe (>2 cm) in one. However, in the remaining three patients, CT failed to depict gastrointestinal involvement. Pneumoperitoneum was noted in four of the 14 patients.
Lymphadenopathy was noted in nine (64%) patients; it was bulky and localized in one, nonbulky and localized in four, and nonbulky and diffuse in four. The affected sites included mesenteric lymph nodes in five patients, paraaortic lymph nodes in three patients, superior mesenteric artery root lymph nodes in two patients, portacaval and celiac axis lymph nodes in one patient, and both iliac and inguinal lymph nodes in one patient.
Stomach
In four patients, the sites of gastric lymphoma were in the prepyloric antrum in two, the cardia in one, and the lower body in one; one patient had lesions in both the stomach and the small intestine. CT failed to depict the gastric lesion in one patient; however, this patient had hepatosplenomegaly, multiple hypoattenuating masses in the liver, and multiple osteolytic bone lesions. In the remaining three patients, PTCL of the stomach appeared on a CT scan as an infiltrative lesion (Fig 1) in two patients and as a polypoid mass (Fig 2) in one patient. In the former two patients, the gastric wall thickening was mild (8 mm) in one and moderate (13 mm) in the other (Fig 1), with a length of 65 mm and 83 mm, respectively. In the last patient, the size of the polypoid mass was measured as 32 x 23 mm. Mild gastric wall thickening was depicted as homogeneous isoattenuation, and moderate gastric wall thickening and the polypoid lesion were depicted as homogeneous hypoattenuation at CT (Figs 1, 2).
|
|
|
|
Endoscopic (n = 3) or gross pathologic (n = 1) findings in the four patients included a 1.8-cm oval ulcer, a linear ulcer with a surrounding nodular mucosal surface, huge ulcerations with a central nodular ulcer base, and a large mass with ulcer ("ulcerofungating mass") (Table).
Small Intestine
If one patient with involvement of both the stomach and jejunum is included, the sites of PTCL were as follows in nine patients: the duodenum, in three patients; the jejunum, in four patients; and the ileum, in two patients. Multifocal lesions were noted in two of nine patients. CT failed to depict gastrointestinal involvement in one patient who had only diffuse bowel dilatation, with excess fluid in the entire small intestine. Seven of the remaining eight patients in whom CT was successful in depicting the gastrointestinal lesions had circumferential bowel wall thickening. The involved intestinal wall ranged in maximum thickness from 7 to 28 mm (mean, 14 mm), and the length was 10–200 mm (mean, 94 mm). The intestinal wall showed mild thickening (7, 8, 9, and 9 mm) in four patients, moderate thickening (17 and 18 mm) in two patients, and severe thickening (28 mm) in one (Fig 3) patient.
|
Bowel perforation occurred in three patients with CT evidence of pneumoperitoneum, a large amount of ascites, diffuse omental and mesenteric infiltration, and peritoneal thickening (Fig 4); intramural and portal venous gas was demonstrated in one patient (Fig 4). None of the patients had bowel obstruction. Hepatomegaly was seen in two patients, and splenomegaly was noted in one patient.
|
|
|
|
Large Intestine
The sites of PTCL in the large intestine were in the sigmoid colon in two patients. CT failed to depict the lesions in one patient. CT in the remaining patient depicted mild circumferential wall thickening of the sigmoid colon along with diffuse bowel dilatation, with fluid excess and fecal material in the entire large intestine; the involved bowel wall was about 4.5 mm in thickness and enhanced homogeneously with isoattenuation (Fig 5).
|
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSIONREFERENCES |
|---|
Most lymphomas involving the gastrointestinal tract are non-Hodgkin lymphomas of B-cell origin. The gastrointestinal tract is not commonly affected in primary extranodal PTCL, with an incidence of only 4%–6% of the reported cases (1,3). Despite their rarity, PTCLs developing primarily in the gastrointestinal tract have been reported to have some distinguishing features from those of B-cell lymphomas (4,19–21).
In contrast to lymphomas of B-cell origin in which the stomach is the most frequent site and the distal ileum is a more common site in the small intestine, the preferential site of PTCL is the small intestine and the upper part of the jejunum (19). Findings in our study also indicated similar results in that the most common site of involvement is the small intestine (64%), especially at both the duodenum and jejunum. In fact, the jejunum has been shown to be the most common site for PTCL–complicating celiac disease (4,19), as well as for the "Mediterranean" or "Middle Eastern" lymphomas that occur most often in young patients of poor socioeconomic status in Middle Eastern and Mediterranean areas (22).
PTCL has a high incidence of multifocal bowel involvement, although this was noted in only three (21%) of our 14 patients. The reported incidence of this finding in PTCLs ranged between 50% and 72% (4,19), compared with 10%–25% in B-cell lymphomas (19,21).
The incidence of bowel perforation is higher in PTCLs (41%–50%) than in B-cell lymphomas (<30%) (4,19,20,23–25). According to some reports (23,26), tumor perforation is extremely rare in gastric and colorectal lymphomas but is higher in small-bowel lymphoma. Almost half of the PTCLs and one-third of the B-cell lymphomas were perforated at the time of surgery (19). In our study, bowel perforation occurred in four (29%) of the 14 patients; none of these patients had a history of having received chemotherapy, and three of them had angiocentric lymphoma. In fact, bowel perforation is a common and critical problem in this type of lymphoma (10). Its characteristic morphologic features included angiocentric and angioinvasive infiltration of the vascular walls by lymphoid cells with varying degrees of cytologic atypia, which caused vascular occlusion with subsequent development of ischemic necrosis of both the tumor and the normal tissues (10) and, subsequently, bowel perforation.
In our study, PTCL involving the gastrointestinal tract was successfully detected at CT in 11 of our 14 patients. At CT, the incidence of gastric or bowel wall thickening (n = 10) was much higher than that of a polypoid mass (n = 1). Despite optimum use of the oral and intravenous contrast agents, CT failed to depict the lesions in the remaining three patients. The involved gastric and intestinal walls in most patients were either mildly (<1 cm, n = 6) or moderately (1–2 cm, n = 3) thickened. Therefore, although the number of our study cases was small, gastric or bowel wall thickening was not considered to be as severe (>2 cm) as that in B-cell lymphoma (11,27). This result is supported by findings in several studies (4,9,19). Chott el al (4) demonstrated that the majority of PTCLs in the small intestine were derived from mucosal T-cell lymphocytes at immunomorphologic analysis.
In a study by Domizio et al (19), the macroscopic features of PTCLs in the small intestine included predominantly thickened plaques, ulcers, or strictures, whereas most B-cell lymphomas were exophytic or annular tumor masses. Similar findings were also observed in the colon. According to findings in several studies (28–30), the most distinguishing characteristic in colonic PTCLs was the extensive mucosal ulcerations. Accordingly, when the lesions are confined to the mucosa with an absence of, or only mild, mural involvement, CT may not have a critical role in the detection of primary lesions in the gastrointestinal tract. In addition, the lesions of gastrointestinal PTCL showed homogeneous isoattenuation or hypoattenuation at CT and, therefore, were not different from the contrast enhancement patterns seen in B-cell lymphomas (31).
The bulky and diffuse regional or mesenteric lymphadenopathy in gastrointestinal lymphoma is a reliable CT feature that helps to distinguish it from other gastrointestinal lesions (11,31). However, our study findings showed that the lymphadenopathy seen in eight (57%) of our 14 patients was nonbulky. Thus, nonbulky lymphadenopathy is also considered strikingly different from B-cell lymphomas. When abdominal solid organs were involved, the main CT finding included hepatosplenomegaly or multiple hypoattenuating masses; this did not differ from those characteristics in B-cell lymphoma.
When CT reveals minimal bowel wall thickening along with nonbulky lymphadenopathy or pneumoperitoneum, the diagnosis may be very difficult or may be confused with various other gastrointestinal diseases, such as inflammatory bowel diseases and vascular disorders. The rarity of PTCL among the general population may contribute to this diagnostic difficulty. However, PTCL should be included in the list of differential diagnoses at CT when the duodenum or the jejunum is a preferential site of intestinal involvement, when multifocal bowel involvement is seen, and when there is evidence of bowel perforation or diffuse peritonitis. Observation of other organs is also very helpful in establishing the diagnosis, because eight of our 14 patients had involvement in other organs. It is also noteworthy that most patients with clinical histories of malabsorption received a diagnosis of enteropathy-associated PTCL (19).
In conclusion, CT can depict PTCL involving the gastrointestinal tract if it is not confined to the mucosa. In PTCL involving the gastrointestinal tract, there is a tendency toward preferential jejunal or duodenal involvement, as well as bowel perforation with both minimal bowel wall thickening and lymphadenopathy, and thus this condition mimics inflammatory or vascular gastrointestinal disorders.
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
Author contributions: Guarantor of integrity of entire study, H.K.H.; study concepts, J.H.B.; study design, A.Y.K., H.K.H.; literature research, E.Y.K., J.K.L.; clinical studies, S.J.M., S.K.Y., H.Y.J., J.H.K.; data acquisition, E.Y.K., J.K.L.; data analysis/interpretation, J.H.B., H.K.H., E.S.Y.; manuscript preparation, J.H.B., H.K.H.; manuscript definition of intellectual content, H.K.H.; manuscript editing, J.H.B., H.K.H.; manuscript revision/review, A.Y.K., T.K.K.; manuscript final version approval, H.K.H.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  N. P. Leite, N. Kased, R. F. Hanna, M. A. Brown, J. M. Pereira, R. Cunha, and C. B. Sirlin Cross-sectional Imaging of Extranodal Involvement in Abdominopelvic Lymphoproliferative Malignancies RadioGraphics, November 1, 2007; 27(6): 1613 - 1634. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. Ghai, J. Pattison, S. Ghai, M. E. O'Malley, K. Khalili, and M. Stephens Primary Gastrointestinal Lymphoma: Spectrum of Imaging Findings with Pathologic Correlation RadioGraphics, September 1, 2007; 27(5): 1371 - 1388. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| RADIOLOGY | RADIOGRAPHICS | RSNA JOURNALS ONLINE |
