Postgrad Med J: first published as 10.1136/pgmj.62.729.615 on 1 July 1986. Downloaded from http://pmj.bmj.com/ on 23 April 2019 by guest. Protected by copyright.

Postgraduate Medical Journal (1986) 62, 615-620

Review Article

Lymphoma of the colon and rectum

M.A. Richards
ICRFDepartment ofMedical Oncology, St. Bartholomew's Hospital, London ECIA 7BE, UK.

Introduction
Lymphomatous involvement of the colon and rectum the following grounds: (1) When the patient was first
may occur either as a localized entity or as a manifesta- seen there was no palpable supetficial lymphaden-
tion of generalized lymphoma. It is important to opathy. (2) Chest radiographs showed no obvious
distinguish between primary and secondary colorectal enlargement of the mediastinal nodes. (3) The white
lymphoma as the natural history and management of blood cell counts, total and differential, were within
the two conditions differs significantly. normal limits. (4) At laparotomy the bowel lesion
Colorectal lymphoma is a rare condition although predominated, the only lymph nodes obviously affec-
several major series have been published as well as ted being those in its immediate neighbourhood. (5)
many case reports. A review of the literature is, The liver and spleen appeared free of tumour in every
however, complicated by a number of factors: (1) The case.
distinction between primary and secondary in- Normal bone marrow biopsy examination and
volvement is not always clearly made and several absence of lymphadenopathy on computed tomogra-
different staging classifications have been used. (2) phic (CT) scan of the mediastinun should probably
Colorectal lymphoma is sometimes not separated now be added to these criteria. Although there has
from other gastrointestinal lymphomas. (3) Compar- been wide acceptance of these criteria, some authors
ison between different histological classifications is (Lewin et al., 1978; Herrman et al., 1980) have used a
difficult, particularly as some of the major series were somewhat broader definition for primary gastrointes-
published over twenty years ago. (4) Childhood cases tinal lymphoma. They include all patients who present
are included by some authors and excluded by others. with obvious predominant alimentary tract lesions or
(5) The exact site of origin of ileocaecal masses is who present with gastrointestinal involvement with
difficult to determine. (6) Referral patterns differ lymphoma. In practice one of the major differences
between institutions. (7) Geographical factors may be between these two definitions lies in the inclusion or
important. (8) All the series reported are retrospective exclusion of patients with para-aortic node (as
making the contribution of surgery, radiotherapy or opposed to mesenteric node) involvement. As long as a
chemotherapy difficult to assess. (9) Some individual staging system which differentiates between these sites
cases have been reported more than once. of nodal involvement such as that proposed by
This review therefore concentrates on the major Musshof & Schmidt-Vollmer (1975), is used, it seems
series reported since 1960 in which sufficient informa- reasonable to include such patients in an analysis of
tion regarding these factors has been given to allow primary gastrointestinal lymphoma.
conclusions to be drawn.
Incidence
Primary colorectal lymphoma Primary colonic and rectal lymphomas are rare disor-
ders. Lymphomas made up only 0.2% (3/1822) of
The standard criteria for the diagnosis of primary cases of colonic malignancy seen at the Methodist
intestinal lymphoma were established by Dawson et Hospital, Memphis, Tennessee between 1966 and 1979
al. (1961). Tumours were considered to be primary on (Fleming et al., 1982). In a large survey from Japan
(Jinnai et al., 1983), 130 cases of large intestinal
lymphoma were reported; 19,850 cases of large bowel
Correspondence: M.A. Richards M.R.C.P. cancer were treated during the same period. Lym-
Received: 25 November 1985 phoma therefore accounted for 0.65% of cases oflarge
© The Fellowship of Postgraduate Medicine, 1986
 Postgrad Med J: first published as 10.1136/pgmj.62.729.615 on 1 July 1986. Downloaded from http://pmj.bmj.com/ on 23 April 2019 by guest. Protected by copyright.
616 M.A. RICHARDS

bowel malignancy. The incidence of rectal lymphoma found no apparent association between the macros-
compared with rectal carcinoma is similarly low- copic and the microscopic appearance of intestinal
0.2% at St Mark's Hospital (Perry et al., 1972). lymphomas. However, Blackshaw (1980) found a very
In two large studies of non-Hodgkin's lymphoma high degree of correlation between the 'lymphomatous
(NHL) (Rosenberg et al., 1961; Jones et al., 1973) polyposis' macroscopic pattern of infiltration and one
about 5% of patients presented with lymphoma particular histological type - namely centrocytic lym-
confined to the gastrointestinal tract. Gastrointestinal phoma.
involvement with Hodgkin's disease is even rarer - 2% The relative incidence of histological types of lym-
at presentation in the study reported by Peters et al. phoma involving the colon and rectum differs marked-
(1968). ly from that of primary nodal lymphomas. Hodgkin's
The colon and rectum are uncommon sites for disease and 'follicular' non-Hodgkin's lymphoma are
lymphoma even compared with other gastrointestinal both very rarely found amongst cases of primary
sites, where stomach and small bowel predominate. colorectal lymphoma. The large majority ofsuch cases
Colorectal involvement accounts for between 10% have a 'diffuse' histological pattern - although the
and 20% of cases in most studies of gastrointestinal ratio of 'low grade' to 'high grade' non-Hodgkin's
lymphoma (Lewin et al., 1978; Bush & Ash, 1969; lymphoma differs considerably between series. This is
Blackledge et al., 1979; Contreary et al., 1980; Loehr et again affected by inclusion of children in whom the
al., 1969; Freeman et al., 1972; Dragosics et al., 1985), lymphomas are normally high grade (immunoblastic
but may be as low as 3% if ileocaecal cases are or lymphoblastic). Lymphoplasmacytoid differentia-
excluded (Isaacson et al., 1979). The proportion of tion of gastrointestinal lymphomas, reflecting B cell
colorectal cases may be higher in India, with figures up origin, is also reported to a varying extent with some
to 45% (Nirmala et al., 1981). probable overlap with true plasmacytomas.
Within the large bowel the caecum is the commonest
site (Jinnai et al., 1983; Wychulis et al., 1966), Possible associated disorders
particularly in children (Lewin et al., 1978), followed
by rectum and then the remainder of the colon. Colonic lymphoma may develop in patients with
Primary lymphoma is occasionally found in the longstanding ulcerative colitis or may present simulat-
appendix. Jinnai et al. (1983) suggested that the ing ulcerative colitis. Lymphoma is a rare complica-
frequency of caecal lesions might reflect the greater tion of ulcerative colitis compared to carcinoma, with
extent of normal lymphoid tissue in this region. They only about 20 recorded cases (Bartolo et al., 1982;
also considered that the frequency of involvement of Emanuel & Isbister, 1979). The duration of symptoms
the ampullary portion of the rectum might be due to of colitis prior to the diagnosis of lymphoma
retention of intestinal contents at this site. varies between 5 and 30 years (Dawson et al., 1961;
The age of reported cases ranged from 3 to 83 years. Wychulis et al., 1966; Renton & Blackshaw, 1976;
However, the maximum incidence is in the 50 to 70 Wagonfeld et al., 1977) with an average of 12 years
year age group (Naqvi et al., 1969) with a mean age (Renton & Blackshaw, 1976). It usually develops in the
between 50 and 55 years (Jinnai et al., 1983; Perry et context of total colitis. A short period of increased
al., 1972; Loehr et al., 1969). The number of men and pain, diarrhoea and rectal bleeding before the diag-
women was equal in Perry's study (1972) of rectal nosis of lymphoma is made has been noted (Renton &
lymphoma, but most studies show a male predomin- Blackshaw, 1976). In at least 7 cases multicentric
ance for lymphoma ofthe large bowel ofapproximate- tumours have been found (Wagonfeld et al., 1977).
ly 2:1 or higher if children are included. Colonic lymphoma occasionally simulates inflam-
matory colitis (Wagonfeld et al., 1977; Weir et al.,
Macroscopic and microscopic appearance 1980; Friedman et al., 1968) and differentiation bet-
ween the two conditions may be difficult clinically,
Discrete lesions are found in about 90% of primary radiologically and histologically. The paper by Sagar
colonic lymphomas (Dawson et al., 1961; Wychulis et et al. (1986) shows that modern immunohistochemical
al., 1966). Normally the lesion is single, but occasion- methods allow the diagnosis of lymphoma to be made
ally two or more may occur. Macroscopically the with much greater objectivity.
tumours may be protuberant intraluminal growths, A few patients have been described with both
infiltrative intramural thickenings or extramural primary large bowel lymphoma and large bowel
growths (Dawson et al., 1961; Jinnai et al., 1983). In adenocarcinoma. Two of the three cases reported by
the remaining 10% of cases the tumours present as Cornes (1960) fall into this category. In one case the
diffuse involvement of the colon with or without two lesions were diagnosed synchronously and in the
polyps. Perry et al. (1972) found an indurated mass in other adenocarcinoma of the rectum and sigmoid
13/22 patients with rectal lymphoma, an ulcer in 8 presented 9 months after the excision of an ascending
patients and a polyp in one case. Dawson et al. (1961) colonic lymphoma. It is therefore important not to
 Postgrad Med J: first published as 10.1136/pgmj.62.729.615 on 1 July 1986. Downloaded from http://pmj.bmj.com/ on 23 April 2019 by guest. Protected by copyright.
LYMPHOMA OF THE COLON AND RECTUM 617

assume that a second growth is always a recurrence or Treatment and survival

metastasis from the original primary tumour. More
recently, a case of adenocarcinoma developing within Surgical excision of the tumour, if technically feasible,
an area of lymphomatous bowel has been reported is the main-stay of treatment for colonic lymphoma.
(Kalisman et al., 1979). A further interesting case of Patients who have received a 'curative' resection (i.e.
coexisting adenocarcinoma and primary malignant all visible tumour excised) undoubtedly have a better
lymphoma of the large intestine in an IgA deficient 14 prognosis than those who have received a palliative
year old boy has been reported from Tehran (Mir- resection (Jinnai et al., 1983; Blackledge et al., 1979;
Madjlessi et al., 1984). Contreary et al., 1980). This may, of course, reflect
differences in the initial extent of disease rather than
Presentation/diagnosis the importance of surgery itself. Although surgery
alone may be curative, recurrence may occur following
Patients present with symptoms that cannot be dif- 'complete excision' either within the abdomen or at a
ferentiated from carcinoma (Perry et al., 1972; distant site. This is of grave prognostic significance
Wychulis et al., 1966). In the Mayo clinic series (Bush & Ash, 1969).
(Wychulis et al., 1966), 90% of patients had Adjuvant radiotherapy is frequently given after
abdominal pain at presentation, 80% weight loss and surgical excision, but no prospective study has been
76% change in bowel habit. Weakness, nausea and performed to assess its efficacy. Retrospective studies
vomiting, anorexia, fever and bleeding per rectum all are difficult to assess as the reason radiotherapy was
occurred in descending order of frequency. A mass given is frequently unclear. Dawson et al. (1961),
was palpable in 88% of cases. Bleeding and diarrhoea reviewing the literature prior to 1961, noted that too
are the commonest symptoms of rectal lymphoma few cases had been recorded to assess the influence of
(Perry et al., 1972). Tenesmus and weight loss are also radiotherapy. Although a remarkable clinical im-
important presenting features. Occasionally patients provement was observed in individual cases, no case
present with complications such as intussusception treated with radiotherapy alone had survived 10 years.
(especially children), perforation or obstruction. A significant improvement in the 2 year survival rate
Typically the time from onset of symptoms to diag- with postoperative radiotherapy has been observed
nosis is 4 to 6 months (Wychulis et al., 1966; Naqvi et however for gastrointestinal lymphoma as a whole by
al., 1969). Bush & Ash (1969). Sixty-four per cent of their
Barium enema examination was abnormal in 41 out patients treated with surgery alone died within 2 years.
of 44 cases in Wychulis' series (1966), though the The recurrence rate at 2 years for those who had
appearances are not necessarily pathognomonic for received radiation therapy was 44%. The difference
lymphoma. Five different patterns have been des- remained significant for those evaluable at 5 years.
cribed (O'Connell & Thompson, 1978) which reflect Perry et al. (1972) recommended surgical excision
the varied macroscopic appearances of the disease: (1) with or without radiotherapy as the treatment of
Mucosal nodularity - nodules may range from 2 mm choice for rectal lymphoma. Alternatively, radioth-
to 2.5 cm, the pattern resembles the pseudopolyposis erapy can be given following left iliac fossa colostomy.
of ulcerative colitis, but the haustral pattern is main- The value of chemotherapy either alone or as
tained and ulceration is uncommon. The appearances adjuvant therapy for colorectal lymphoma cannot be
may be similar to those found in Crohn's disease, assessed from the current literature.
amoebiasis or pseudomembranous colitis (Bruneton The five year survival rate for all patients is about
et al., 1983). (2) Endo-exocentric mass - an extensive 35% (Jinnai et al., 1983; Loehr et al., 1969; Naqvi et
mass with gross mucosal destruction due to mural al., 1969). A high percentage of deaths occurs in the
infiltration. Occasionally contrast may track ex- first 2 years (Dawson et al., 1961; Jinnai et al., 1983).
traluminally into a necrotic mass. (3) Intraluminal Deaths due to lymphoma may still occur after 5 years
mass - these are often lobulated and may be up to (Dawson et al., 1961) although the risk appears small
20 cm in size, without causing obstruction unless (Jinnai et al., 1983). At the Mayo clinic the 10 year
intussusception occurs. (4) Infiltrative forms - the survival rate for those who had received a curative
lesion may appear either as a rigid anhaustral segment resection (with or without adjuvant radiotherapy) was
or as an annular stricture resembling a carcinoma. (5) 50%. Dawson et al. (1961) projected an overall 10 year
Mesenteric invasion - extraluminal tumours appear as survival rate of not more than 25%. For inoperable
a soft tissue mass causing extrinsic compression cases the 5 year survival rate in 2 series was zero (Jinnai
without mucosal involvement. et al., 1983; Contreary et al., 1980).
Histology is, of course, required to establish a firm These survival rates are worse than those for gastric
diagnosis. This may be obtained by endoscopic lymphoma (Dawson et al., 1961; Lewin et al., 1978;
biopsy, but more frequently is only available following Dragosics et al., 1985) perhaps because the disease is
laparotomy. more advanced at presentation (Contreary et al.,
 Postgrad Med J: first published as 10.1136/pgmj.62.729.615 on 1 July 1986. Downloaded from http://pmj.bmj.com/ on 23 April 2019 by guest. Protected by copyright.
618 M.A. RICHARDS

1980), in terms of spread to lymph nodes and adjacent of gastrointestinal involvement. Thus colonic in-
viscera. The prognosis is also worse than that of volvement in the context of widespread (stage IV)
carcinoma of the colon (Jinnai et al., 1983; Freeman et disease at presentation was only observed in 8 (2%) of
al., 1972). the patients. Six of these had histologically 'diffuse'
lymphoma and 2 had nodular histology.
Prognostic factors Rosenberg et al. (1961) reviewed 1269 cases of
lymphosarcoma, of whom only 7 presented with
Although there is a general agreement regarding the evidence of lymphoma involving the large intestine
importance of operability as a prognostic factor, the (primary or secondary). However, a total of 36
importance of other factors is less clear cut. Lymph patients (2.8%) developed clinical evidence ofcolorec-
node involvement was a significantly adverse factor in tal lymphoma subsequently. These figures contrast
three studies (Jinnai et al., 1973; Contreary et al., 1980; markedly with results from post-mortemn examina-
Freeman et al., 1972), and was probably significant in tions. In the same review, Rosenberg found colorectal
another (Wychulis et al., 1966). However, Dawson et lymphoma in 68 out of 277 (24.5%) autopsy cases.
al. (1961) found that regional lymph node involvement Fifty of these were in the colon and 18 in the rectum. In
did not have any effect on prognosis. The different only 20 of these patients had the colorectal in-
histological classifications used make it very difficult volvement been demonstrated clinically. Over 50% of
to assess the impact of different histological subtypes. the autopsy series had evidence of lymphoma in some
The Japanese study, comprising by far the largest part of the gastrointestinal tract.
single survey, was also able to identify size and Peters et al. (1968) in a study of 1406 cases of
macroscopic type of the initial lesion as important Hodgkin's disease and NHL, also found a much
factors. A great difference was seen between patients higher incidence of gastrointestinal involvement at
with tumours of 5 cm or less in diameter and those with post-mortem (45%) than at presentation (11%),
tumours larger than this. Intramural tumours had the though the exact location within the gastrointestinal
worst prognosis and intraluminal tumours the best. tract is not recorded. Furthermore, these figures
represented gross involvement at autopsy not micros-
copic lesions. The incidence at presentation was 16%
Secondary lymphoma for NHL (reticulum cell sarcoma and lymphosar-
coma) and only 2.4% for Hodgkin's disease. At post-
Secondary involvement of the gastrointestinal tract mortem the figures were 70% and 34% respectively.
and particularly the large bowel has received less Ehrlich et al. (1968) who studied 323 autopsy cases of
attention than primary gastrointestinal lymphoma malignant lymphoma, also found approximately 55%
despite the fact that the incidence is higher and the had tumours in the gastrointestinal tract. However,
prognostic implications are grave. only 7% had tumour in the colon (10% for NHL, 3%
The incidence of gastrointestinal involvement in for Hodgkin's disease). Prolla & Kirsner (1964) found
systemic lymphoma increases during the course of the macroscopic lesions in 3 out of 18 patients with
disease, with relatively low rates at presentation, a chronic lymphocytic leukaemia at autopsy. A further 9
higher number of patients developing clinical evidence patients had microscopic lesions giving a total of 60%
of gastrointestinal involvement subsequently and a with either macroscopic or microscopic evidence of
very high incidence in post-mortem studies. colorectal involvement. A further case of large bowel
In the study of 405 patients with NHL reported by infiltration by chronic lymphocytic leukaemia was
Jones et al. (1973), 64 patients had gastrointestinal published recently in this journal (Tucker & Cachia,
involvement before therapy. Nineteen of these had 1986).
localized (i.e. primary) gastrointestinal lymphoma. Do these striking differences between clinical and
The remaining 45 patients (11% of the series) had post-mortem findings mean that gastrointestinal in-
disseminated disease. This is a minimum figure as less volvement is underdiagnosed at presentation? The
than half of the patients in the study underwent study by Goffinet et al. (1973) suggests that this is not
laparotomy or radiological assessment ofthe gastroin- the case. Staging laparotomies performed on 69
testinal tract. Involvement was significantly higher in previously untreated patients revealed no cases of
patients with diffuse lymphoma than those with colonic lymphoma. Furthermore. 37 of these patients
nodular histology. The true incidence of gastrointes- had barium enema examinations as part ofthe staging
tinal involvement was calculated to be between 7% procedure and all were normal. By contrast O'Connell
and 12% for nodular lymphoma and between 22% & Thompson (1978) demonstrated colonic pathology
and 39% for diffuse lymphoma. Only 11/405 had on barium enema examination in 4 patients with
colonic involvement (3 localized, 8 stage IV), the systemic lymphoma who had no clinical symptoms
stomach and small intestine being the commonest sites referrable to the colon.
 Postgrad Med J: first published as 10.1136/pgmj.62.729.615 on 1 July 1986. Downloaded from http://pmj.bmj.com/ on 23 April 2019 by guest. Protected by copyright.
LYMPHOMA OF THE COLON AND RECTUM 619

Significance of gastrointestinal symptoms Development of gastrointestinal symptoms during

the course of the patients' illness is also of adverse
In order to account for the discrepancy between prognostic significance. Erhlich et al. (1968) noted that
clinical and autopsy findings, it must be assumed that the onset of abdominal pain or the destruction of an
most patients whose disease spreads to the colon and abdominal mass preceded death by 6 months in
rectum do not develop symptoms referrable to this. patients with 'reticulum cell sarcoma', as against a
Conversely, gastrointestinal symptoms do not neces- mean survival of 24 months from the onset of clinical
sarily imply that tumour is present. Other lesions may symptoms. The trend was similar for other non-
develop as a result of the patients' immunosuppressed Hodgkin's lymphomas and for Hodgkin's disease,
state and as a result of radiotherapy or chemotherapy. although the time spans were longer. Bleeding, per-
Ehrlich et al. (1968) found non-tumourous ulcerations foration or obstruction usually heralded death within
of the colon in 13 out of their 323 autopsy cases, 4 of 3 months in all histological groups.
which had bled. They also noted fungal and bacterial Specific data for survival of patients with secondary
invasion. With increasingly intensive therapy the colonic or rectal lymphoma is not available, but
incidence of these non-tumourous lesions is likely to probably reflects that of the gastrointestinal tract as a
rise and complications may result. Indeed Ehrlich et al. whole.
(1968) observed that bleeding was more frequently due
to non-tumourous causes than to tumour. In 7 cases of
bleeding arising from the intestine, 2 were due to Conclusions
tumour and 5 due to other causes. However, obstruc-
tion and perforation were more usually due to tumour. Primary colorectal lymphoma is a rare condition
which carries a poor prognosis compared with either
Pathological features primary gastric lymphoma or carcinoma of the colon.
Optimal management is still uncertain, particularly
Secondary colonic involvement is often multicentric regarding the role of radiotherapy and chemotherapy.
and may have a different distribution from primary No single institution can hope to answer these ques-
lymphoma. The rectosigmoid region and the left tions and it is therefore important that cases of
hemicolon are more frequently involved than the right colorectal lymphoma should continue to be reported.
hemicolon (O'Connell et al., 1978). Radiologically the Special attention should be given to histological
'mucosal nodularity' pattern is seen more frequently in subtype and the specific sites of nodal involvement, in
secondary than in primary colonic lymphoma as is the order that valid comparisons can be made between
'endo-exoenteric mass' (O'Connell et al., 1978). institutions. Cooperative prospective studies should
be strongly considered.
Survival Secondary colorectal lymphoma has received very
little attention, but probably carries an even worse
The identification of gastrointestinal involvement by a prognosis. However, the incidence and prognostic
diffuse lymphoma may imply a very poor prognosis significance of both tumourous and non-tumourous
with the exception of 'diffuse well differentiated' colonic lesions complicating systemic lymphoma
lymphoma. Jones et al. (1973) reported a median should be reassessed in view of changes in treatment
survival of 6 months for patients with gastrointestinal which have been introduced in the past 10 years,
involvement at presentation. particularly for high grade NHL.

References
BARTOLO, D., GOEPEL, J.R. & PARSONS, M.A. (1982). Rectal Fortschritte auf dem Gebiete der Rontgenstrahlen, 138, 283.
malignant lymphoma in chronic ulcerative colitis. Gut, 23, BUSH, R.S. & ASH, C.L. (1969). Primary lymphoma of the
164. gastrointestinal tract. Radiology, 92, 1349.
BLACKLEDGE, G., BUSH, H., DODGE, O.G. & CROWTHER, CONTREARY, K., NANCE, F.C. & BECKER, W.F. (1980).
D. (1979). A study of gastrointestinal lymphoma. Clinical Primary lymphoma of the gastrointestinal tract. Annals of
Oncology, 5, 209. Surgery, 191, 593.
BLACKSHAW, A.J. (1980). Non-Hodgkin's lymphomas of the CORNES, J.S. (1960). Multiple primary cancers: Primary
gut. In Recent Advances in Gastrointestinal Pathology, malignant lymphomas and carcinomas of the intestinal
Wright, R. (ed.) W.B. Saunders: Eastbourne. tract in the same patient. Journal of Clinical Pathology, 13,
BRUNETON, J.N., THYSS, A., BOURRY, J., BIDOLI, R. & 483.
SCHNEIDER, M. (1983). Colonic and rectal lymphomas. DAWSON, I.M.P., CORNES, J.S. & MORSON, B.C. (1961).
 Postgrad Med J: first published as 10.1136/pgmj.62.729.615 on 1 July 1986. Downloaded from http://pmj.bmj.com/ on 23 April 2019 by guest. Protected by copyright.
620 M.A. RICHARDS

Primary malignant lymphoid tumours of the intestinal lymphoma and adenocarcinoma of the large intestine in an
tract. British Journal of Surgery, 49, 80. IgA-deficient boy. Diseases of Colon and Rectum, 27, 822.
DRAGOSICS, B., BAUER, P. & RADASZKIEWICZ, T. (1985). MUSSHOFF, K. & SCHMIDT-VOLLMER, H. (1975). Prognosis
Primary gastrointestinal non-Hodgkin's lymphomas. Can- of non-Hodgkin's lymphomas with special emphasis on
cer, 55, 1060. the staging classification. Zeitschrift fur Krebsforschung,
EHRLICH, A.N., STALDER, G., GELLER, W. & SHERLOCK, P. 83, 323.
(1968). Gastrointestinal manifestations of malignant lym- NAQVI, M.S., BURROWS, L. & KARK, A.E. (1969). Lym-
phoma. Gastroenterology, 54, 1115. phoma ofthe gastrointestinal tract. Annals ofSurgery, 170,
EMANUEL, J.C.M. & ISBISTER, W.H. (1979). Chronic 221.
mucosal ulcerative colitis with malignant lymphoma: NIRMALA, V., THOMAS, J.A. & ANTHONY, A.J. (1981).
report of a case. Australia and New Zealand Journal of Primary malignant lymphoma of colon. Indian Journal of
Surgery, 49, 95. Cancer, 18, 47.
FLEMING, I.D., MITCHELL, S. & DILAWARI, R.A. (1982). O'CONNELL, D.J. & THOMPSON, A.J. (1978). Lymphoma of
The role of surgery in the management of gastric lym- the colon: the spectrum of radiologic changes. Gastrointes-
phoma. Cancer, 49, 1135. tinal Radiology, 2, 377.
FREEMAN, C., BERG, J.W. & CUTLER, S.J. (1972). Occurrence PERRY, P.M., CROSS, R.M. & MORSON, B.C. (1972). Primary
and prognosis of extranodal lymphomas. Cancer, 29, 252. malignant lymphoma of the rectum (22 cases). Proceedings
FRIEDMAN, H.B., SILVER, G.M. & BROWN, C.H. (1968). of the Royal Society of Medicine, 65, 8.
Lymphoma of the colon simulating ulcerative colitis. PETERS, M.V., HASSELBACK, R. & BROWN, T.C. (1968). The
American Journal of Digestive Diseases, 13, 910. natural history of the lymphomas related to the clinical
GOFFINET, D.R., CASTELLINO, R.A., KIM, H., DORFMAN, classification. Proceedings of the International Conference
R.F., FUKS, Z., ROSENBERG, S.A., NELSON, T. & KAPLAN, on Leukaemia-Lymphoma. 357.
H.S. (1973). Staging laparotomies in unselected previously PROLLA, J.C. & KIRSNER, J.B. (1964). The gastrointestinal
untreated patients with non-Hodgkin's lymphomas. Can- lesions and complications of the leukemias. Annals of
cer, 32, 672. Internal Medicine, 61, 1084.
HERRMANN, R., PANAHON, A.M., BARCOS, M.P., WALSH, RENTON, P. & BLACKSHAW, A.J. (1976). Colonic lymphoma
D. & STUTZMAN, L. (1980). Gastrointestinal involvement complicating ulcerative colitis. British Journal of Surgery,
in non-Hodgkin's lymphoma. Cancer, 46, 215. 63, 542.
ISAACSON, P., WRIGHT, D.H., JUDD, M.A. & MEPHAM, B.L. ROSENBERG, S.A., DIAMOND, H.D., JASLOWITZ, B. &
(1979). Primary gastrointestinal lymphomas. Cancer, 43, CRAVER, L.F. (1961). Lymphosarcoma: A review of 1269
1805. cases. Medicine, 40, 31.
JINNAI, D., IWASA, Z. & WATANUKI, T. (1983). Malignant SAGAR, S., SELBY, P., SLOANE, J. & McELWAIN, T.J. (1986).
lymphoma of the large intestine - operative results in Colorectal lymphoma simulating inflammatory colitis and
Japan. Japanese Journal of Surgery, 13, 331. diagnosed by immunohistochemistry. Postgraduate
JONES, S.E., FUKS, Z., BULL, M., KADIN, M.E., DORFMAN, Medical Journal, 62, 51.
R.F., KAPLAN, H.S., ROSENBERG, S.A. & KIM, H. (1973). TUCKER, J. & CACHIA, P.G. (1986). Gastrointestinal bleeding
Non-Hodgkin's lymphomas IV. Clinicopathological due to large bowel infiltration by chronic lymphocytic
correlation in 405 cases. Cancer, 31, 806. leukaemia. Postgraduate Medical Journal, 62, 45.
KALISMAN, M., DEBEER, R. & PFEFFER, H. (1979). Aden- WAGONFELD, J.B., PLATZ, C.E., FISHMAN, F.L., SIBLEY,
ocarcinoma arising from a lymphoma - case report. R.K. & KIRSNER, J.B. (1977). Multicentric colonic lym-
Clinical Oncology, 5, 85. phoma complicating ulcerative colitis. Digestive Diseases
LEWIN, K.J., RANCHOD, M. & DORFMAN, R.F. (1978). and Sciences, 22, 502.
Lymphomas of the gastrointestinal tract. Cancer, 42, 693. WEIR, A.B., POON, M-C., GROARKE, J.F. & WILKERSON, J.A.
LOEHR, W.J., MUJAHED, Z., ZAHN, D., GRAY, G.F. & (1980). Lymphoma simulating Crohn's colitis. Digestive
THORBJARNARSON, B. (1969). Primary lymphoma of the Diseases and Sciences, 25, 69.
gastrointestinal tract: A review of 100 cases. Annals of WYCHULIS, A.R., BEAHRS, O.H. & WOOLNER, L.B. (1966).
Surgery, 170, 232. Malignant lymphoma of the colon. Archives of Surgery,
MIR-MADJLESSI, S.H., VAFAI, M., KHADEMI, J. & 93, 215.
KAMALIAN, N. (1984). Coexisting primary malignant