Gastrointestinal non-Hodgkin's lymphoma

ClinicalRadiology (1987)38, 609-614

Gastrointestinal Non-Hodgkin's Lymphoma A. R. MAKEPEACE, D. C. FERMONT* and M. H. BENNETT* The Meyerstein Institute of Radiotherapy and Oncology, The Middlesex Hospital, Mortimer Street, London W1 and * The Mount Vernon Hospital, Northwood, Middlesex Seventy-two patients with gastrointestinal nonHodgkin's lymphoma treated between 1952 and 1980 are reviewed. The small intestine was involved in 49% of cases and the stomach in 29%. Surgical resection of the tumour was performed whenever feasible. Radiotherapy was used either adjuvantly or for incompletely excised tumours and chemotherapy was more often reserved for advanced, unresected disease. The overall 5 year survival was 36% and the 5 year relapse free survival was 22%. Forty-one (57%) patients relapsed of w h o m 33 (80%) subsequently died of non-Hodgkin's lymphoma. The histology in each case was reviewed using the British National L y m p h o m a Investigation criteria and 94% of cases were reclassified as Grade 2 non-Hodgkin's lymphoma.

Non-Hodgkin's lymphoma may arise in the gastrointestinal tract as a primary event or involve this site incidentally in Stage IV disease. As a result, considerable variability exists between the patients included in different studies. Whereas some authors maintain that tumours classified as primary gastrointestinal lymphoma should be confined to the gastrointestinal tract and adjacent lymph nodes (Dawson et al., 1961; Shimm et al. 1983; Brooks and Enterline, 1983; Maor et al., 1984; Skudder and Schwartz, 1985) others also include patients with widespread disease at presentation (Bush and Ash, 1969; Hande etal., 1978; Rosenfelt and Rosenberg, 1980; Economopoulos et al., 1985). The gastrointestinal tract is the most common site of primary extranodal non-Hodgkin's lymphoma and accounts for 4.6% of all patients (Rosenberg et al., 1961; Freeman etal., 1972; Lewin etal., 1978; Rao etal., 1984; Skudder and Schwartz, 1985). However, 50 to 70% of patients with disseminated lymphoma will incidentally have gastrointestinal involvement (Rosenberg et al., 1961; Peters et al., 1968). In 1961, Dawson et al. proposed criteria in an attempt to define those patients with primary disease. These included: 1 No palable superficial lymphadenopathy at presentation; 2 No obvious enlargement of mediastinal nodes on the chest radiograph; 3 The white blood cell, total and differential count to be within normal limits; 4 At laparotomy the bowel lesion to predominate and the only lymph nodes obviously involved being those in the immediate neighbourhood; 5 The liver and spleen to be grossly free of tumour. The most appropriate staging system for gastrointestinal lymphoma also remains controversial. The Ann Arbor classification (Carbone et al., 1971) is used most Correspondence to: A. R. Makepeace.

frequently (Hande et al, 1978; Lewin et al., 1978; Rosenfelt & Rosenberg, 1980; Brooks & Enterline, 1983) but fails to distinguish between locoregional and para-aortic lymph node involvement. A variety of other classifications have also been used (Bush and Ash, 1969; Lira et al., 1977; Crowther and Rankin, 1982) and further complicate the comparison of different studies. As the prognosis of these patients is related to the local extent of the tumour, local lymph node involvement and to whether the viscus has perforated, Blackledge et al. (1979) proposed a new staging classification incorporating these factors (Table 1).

PATIENTS AND M E T H O D S This study reviewed 72 patients treated during the 28 year period from 1952 to 1980 with primary nonHodgkin's lymphoma of the gastrointestinal tract. Patients in Stage I and If fulfilled the criteria proposed by Dawson et al. (1961) and those in Stages III and IV had concomitant involvement of the para-aortic lymph nodes, spleen, liver or bone marrow. The mean age of the 44 male patients was 45 years (range 6 to 80 years) and of the 28 females, 57 years (range 6 to 77 years). The age distribution of the patients is illustrated in Fig. 1. The small bowel was involved in 35 cases, the stomach in 21 cases, the ileocaecal region in 10 cases, the colon in five cases and the rectum in one case. Presenting Symptoms

Patients with involvement of the small intestine were symptomatic for an average period of 8 weeks (range 2 days to 10 months); abdominal pain was reported in 72% of patients, vomiting or diarrhoea in 28% and weight loss in 22%. A mass was palable in the right iliac fossa in eight patients and twice diagnosed as acute appendicitis. Ten patients (31%) had multiple areas of bowel involvement. Fifteen patients presented as surgical emergencies of whom five had intestinal obstruction, five had a perforation of the small intestine, three had severe melaena and two presented with an intusussception. Patients with gastric lymphoma had symptoms for an average of 10 months (range 6 weeks to 4 years). Epigastric pain was present in 48% of patients, weight loss in 39% and vomiting or diarrhoea in 23%. Four patients had a palable epigastric mass and the barium meal was suggestive of gastric carcinoma in seven cases. Whilst four patients presented with gastric haemorrhage none presented with gastric perforation. Each of the 10 patients with ileocaecal lymphoma complained of pain in the right iliac fossa. Five patients presented with a palpable mass which in three cases was caused by a colonic intusussception.

610

CLINICAL RADIOLOGY 18-

Table 1 - Staging classification of Blackledge et al. (1979)

IA IB IIA fib IIC III IV

Single tumour confined to the gut Multiple tumours confined to the gut Tumour with local lymph node involvement Tumour with local extension to adjacent tissues Tumour with perforation and peritonitis Tumour with widespread lymphadenopathy Tumour with disseminated disease in non-lymphoid tissue

K\\N\'~

16

x\xxx~

12 ~x\xx~ ~\xxxs

~x~xx~ ~xxxx,

~xxxx~

......

I0-

This study used the staging classification proposed by Blackledge et al. (1979) although insufficient information was available to reliably subdivide cases into Stages IIA, liB and IIC. In the 71 staged patients, 18 (25%) presented in Stage I, 42 (59%) in Stage II, 3 (4%) in Stage III and 8 (12%) in Stage IV (Table 2). All the patients in Stage III had para-aortic lymph node involvement and of the eight patients in Stage IV, five had hepatic involvement and three bone marrow infiltration.

xl

22222=

8--

. . . . . . ,X.\\\'q ,"..\\\N .N\\\'q

z

b.\\\N .x.\\\'q .',.\\\'q ,N.N\\N . . . . . .

6-

N® (I0

10

19

20

29

Histology

30

39"40

49150

59

Age of patients (years)

The original histological specimens were reviewed by one histopathologist (M.H.B.) using the British National Lymphoma Investigation (BNLI) criteria (Bennett et al., 1974; Henry et al., 1978; Farrer-Brown, 1981). The National Cancer Institute (NCI) International 'Working Formulation' terms are also given (Henry, 1981; NCI, 1982). Sixty-eight patients (94%) presented with Grade 2 non-Hodgkin's lymphoma (Table 3); diffuse large cell undifferentiated lymphoma (34 patients) and diffuse poorly differentiated lymphocytic (lymphoblastic) lymphoma (17 patients) occurred most frequently. Only four patients were classified as Grade 1 non-Hodgkin's lymphoma.

I

Fig. 1 - Distribution of patients by age.

palliative by-pass procedure and nine biopsies of inoperable tumours. Complete resection of the tumour was possible in 19 patients, the turnout was incompletely excised in 28 patients and the adequacy of surgery could not be defined in a further 13 patients. After a complete resection of tumour, 13 patients (68%) received adjuvant radiotherapy, five patients Table 2 - Stage at presentation by each primary site of tumour

Site of disease

Treatment Details of surgery were available in 69 cases; the procedures included 29 small bowel resections, 13 partial gastrectomies, one total gastrectomy, 16 right hemicolectomies (four in patients with ileal lymphoma), one

G

l 60-69 I 70-79 I > 8 0

Stomach Small bowel Caecum Colon Rectum

Stage 1

I1

1H

IV

Unclassified

4 10 2 1 1

13 20 6 3 0

2 1 0 0 0

2 3 2 1 0

0 1 0 0 0

Table 3 - Distribution of patients by histological subtype

Working formulation Malignant lymphoma, follicular Malignant lymphoma, small lymphocytic Malignant lymphoma, diffuse small cleaved cell Malignant lymphoma, lymphoblastic Malignant lymphoma, diffuse mixed small and large cell Malignant lymphoma, diffuse large cell Malignant lymphoma, diffuse histiocytic Extramedullary plasmacytoma Unclassifiable

BNLI

o)

No. of patients

Architecture

Description

Male

Follicular

Follicular lymphoma

o

Diffuse

Lymphocytic, well differentiated Lymphocytic, intermediately differentiated (small follicle lymphocyte) Lymphocytic, poorly differentiated (lymphoblast) Lymphocytic, mixed small and large cell (mixed follicle cell)

3

0

1

0

12

5

0

1

Undifferentiated large cell (large lymphoid cell) Histiocytic (mononuclear phagocytic cell origin) Plasma cell (extramedullary) Malignant lymphoma unclassifiable

20

Female

Grade 1

14 ,Grade 2

4

4

3 1

2 2

NON-HODGKIN'S LYMPHOMA 100

(26%) had no further treatment and one patient was treated with both radiotherapy and COP chemotherapy (cyclophosphamide, vincristine, prednisone). Following an incomplete resection, 17 patients (61%) were irradiated, eight (28%) received chemotherapy using either single agents (cyclophosphamide, chlorambucil, nitrogen mustard) or combination regimes, COP or CHOP (cyclophosphamide, doxorubicin, vincristine, prednisone), one patient received both radiotherapy and chemotherapy (vinblastine) and two had no further treatment. Nine of the 13 patients in whom the adequacy of resection was uncertain were irradiated. Following a biopsy alone, six patients were irradiated and three treated with chemotherapy (COP). The radiotherapy given during this study was not standardised but the patients could be divided into two broad categories: 1 Small intestinal lymphoma was treated using large abdominal fields to encompass both the ileum and jejunum. The average field size was 26 cmx25 cm (range 19 cmx20 cm to 30 cmx30 cm) and the average tumour dose was 3440 cGy in 22 fractions over 35 days (range 1095 c G y / 7 / l l to 5000 cgy/32/47). Five patients received whole abdominal radiotherapy for an average dose of 2230 cGy in 16 fractions over 23 days. 2 For gastric lymphoma the average field size was 19 cmx18 cm (range 10 c m x l 8 cm to 30 cmx23 cm) which encompassed the gastric bed or the remaining stomach with an adequate margin. The average tumour dose was 3750 cGy in 19 fractions over 30 days (range 578 cgy/9/14 to 5600 cgy/22/31). One patient received whole abdominal radiotherapy to a dose of 1245 cGy in 15 fractions over 36 days. A similar field size was used to treat ileocaecal lymphoma which averaged 21 c m x l 8 cm (range 19 c m x l 8 cm to 25 cmx26 cm) with a dose of 2740 cGy being given in 20 fractions over 30 days (range 2000 cGy/19/25 to 3550 cGy/18/32).

611

80

60

m]•v

40

[18]

>

\

[9]

20

w

r

lO-

I

I

I

I

I

0

5

10

15

20

Time since presentation

(years)

Fig. 2 - Survival, all patients.

ing, weight loss and depression of her peripheral blood count despite a normal bone marrow; death occurred within 3 weeks of treatment. One surgical death was reported due to haemorrhage from a splenic laceration. Only one patient has died from a second malignancy, a 100

Abdominal irradiation is frequently associated with a significant morbidity. In this study 21 patients (46%) experienced severe nausea, 14 patients (30%) complained of vomiting and 12 (26%) of diarrhoea. Radiotherapy was given to all the patients in Stage III and four of the Stage IV patients were also irradiated, one with palliative intent and the other three using large abdominal fields.

80

60

~o

RESULTS

The 5 year actuarial survival achieved in this study was 37% (Fig. 2) with a median survival of 10 months (mean 42 months; range 1 to 242 months). The survival curves for Stages I, II and IV are shown in Fig. 3. Sixty patients have died, five are in remission and seven have been lost to follow-up. Forty-seven patients (65%) died from non-Hodgkin's lymphoma. Two further deaths resulted from radiotherapy; one patient developed a perforation in an area of bowel necrosis 28 months after whole abdominal irradiation to a dose of 4300 cGy. The second patient, a 38 year old woman, was treated for a gastric lymphoma to a dose of 3695 cGy. During the radiotherapy she developed severe vomit-

q0

20

I

I

I

I

I

0

5

10

15

20

Years since presentation

Fig. 3 - Survival, by stage. ~ = S t a g e A - - A =Stage IV.

I; ~ = S t a g e

II;

612

CLINICAL RADIOLOGY

bronchial carcinoma. In addition one patient committed suicide, three died from cardiovascular disease and the causes of death in the remaining five patients are unknown. Forty-one patients (57%) relapsed following initial treatment. The median time to relapse was 8 months (mean 17 months: range 1 to 128 months) and 68% of relapses occurred within i year of presentation and 81% within 2 years. Relapse occurred in 71% of gastric lymphomas, 60% of ileocaecal lymphomas, 51% of small intestinal lymphomas and 40% of colonic lymphomas. The initial site of relapse was known in 39 cases; the recurrence was confined below the diaphragm in 24 cases, was extra-abdominal in 11 cases and involved both regions in four cases. Of the 11 relapses which occurred within 3 months of treatment, eight involved sites distant from the primary tumour. The extent of surgery appeared not to influence the incidence of relapse. After complete resection of the tumour 12 patients relapsed (60%) of whom four had received adjuvant radiotherapy and one chemotherapy (COP). Following an incomplete resection, 19 patients relapsed (70%), of whom 11 had been irradiated and six treated with chemotherapy using CHOP, COP or cyclophosphamide. In the 10 remaining patients, in whom the adequacy of excision was not known, only one obtained a lasting remission despite seven being irradiated postoperatively and one being treated with chemotherapy. At relapse 31 patients (76%) were actively treated. Radiotherapy was used in 16 cases, single agent chemotherapy in nine cases (nitrogen mustard, chlorambucil, cyclophosphamide or vinblastine), combination chemotherapy (CHOP) in two cases and the combination of both modalities in four patients.

100

80

). > L

D.

60

40

c~o

20

F

[131

I 0 -

[81

I

1

1

I

I

5

I0

15

20

Time since presentation (years) F i g . 4 - R e l a p s e - f r e e s u r v i v a l , all p a t i e n t s .

In the 41 relapsed patients, 33 died from lymphoma, three died in remission, three died with active disease present but from unrelated causes and only two are alive (94 and 89 months). The 5 year relapse free survival was 24% (Fig. 4) and the median survival following relapse was 4 months (mean 15 months; range 1 to 138 months). Ten patients received no treatment at relapse; eight died within 2 months from lymphoma, one died from a postoperative haemorrhage and one from complications or radiotherapy.

DISCUSSION Non-Hodgkin's lymphoma accounts for only 1 to 4% of gastrointestinal malignancies (Loehr et al., 1969; Blackledge et al., 1979) and, in the absence of any pathognomonic symptoms or unique clinical signs, the tumour is rarely suspected at presentation. Treatment is frequently initiated with a presumptive diagnosis of carcinoma or gastric ulceration by a surgeon whose experience in the management of this malignancy may be limited. Consequently during surgery attention is frequently focused on the primary lesion and the opportunity to stage the disease pathologically is not taken; whilst a formal staging laparotomy would not be advocated in this disease, valuable additional information may be obtained incidentally during surgery which might influence subsequent treatment. Primary gastrointestinal lymphoma is normally considered to have a poor prognosis (Lim and Hartnell, 1977; Fleming et al., 1982; Brooks and Enterline, 1983; Paulson et al., 1983) and the 5 year survival of 36% in this study compares with values of 39% to 51% reported previously (Lim et al., 1977; Lewin et al., 1978; Skudder and Schwartz, 1985). In comparison with nodal lymphoma, high grade histological subtypes predominate in gastrointestinal disease and undoubtedly adversely affect the prognosis (Fisher and DeVita, 1977; Lewin et al., 1978; Herrmann et al., 1980; Fisher and Hubbard, 1981; Fleming et al., 1982; Brooks and Enterline, 1983; Paulson et al., 1983; Rao et al. , 1984; Economopoulos et al., 1985). Other relevant factors include poor nutrition, general debility, gastrointestinal obstruction, perforation and haemorrhage. The depth of turnout penetration may also influence prognosis; Shimm et al. (1983) reported a 5 year survival of 91% when the serosa was intact compared with 32% when the turnout extended to the peritoneal surface or involved adjacent structures. Most relapses occur within 2 years of presentation (Bush and Ash, 1969; Maor et al., 1984; Skudder and Schwartz, 1985). However, unlike Hodgkin's disease where second line treatment will often achieve a complete remission, relapse in gastrointestinal nonHodgkin's lymphoma is frequently fatal (Bush and Ash, 1969); in this study 80% of the patients who relapsed subsequently died from lymphoma. Obtaining an initial complete remission is therefore of paramount importance. Although gastric lymphoma accounted for only 29% of cases reported here, an incidence of approximately 50% is more frequently reported (Bush and Ash, 1969; Lewin et al., 1978; Green et al., 1979; Rosenfelt and Rosenberg, 1980; Gary et al., 1982; Skudder and Schwartz, 1985). Turnouts arising at this site may be

NON-ItODGKIN'S LYMPHOMA

assessed and b.iopsied endoscopically. However, only small fragments of tissue will be obtained whose histological interpretation may prove difficult, especially when differentiating between large cell lymphoma and anaplastic carcinoma (Loehr et al., 1969; Saraga et al., 1981; Fleming et al., 1982; Shimm et al., 1983; Brooks and Enterline, 1983). The application of this technique to gastric lymphoma therefore remains controversial; both Economopoulos et al. (1985) and Fleming et al. (1982) found little benefit and in several patients gastroscopic biopsy resulted in the incorrect diagnosis of gastric carcinoma. However, other studies have reported success rates of 15% to 18% (Rosenfelt and Rosenberg, 1980; Skudder and Schwartz, 1985) and one author reported a positive diagnosis in 88% of cases (Maor et al., 1984). Although surgery is rarely used in the routine treatment of non-Hodgkin's tymphoma, it is frequently considered an integral part of the management of primary gastrointestinal disease (Dawson et al., 1961; Peckham and McElwain, 1977). Surgical resection alone may be curative for localised disease (Azzopardi and Menzies, 1960; Dawson et al., 1961; Weaver and Batsakis, 1964). However, major surgery carries a significant risk in the elderly; Skudder & Schwartz (1985) reported that 37% of their patients were unfit for surgery and other studies have reported an operative mortality of 9% to 19% (Dawson et al., 1961; Naqvi et al., 1969; Maor et al., 1984). Between 37% and 88% of patients have their tumours completely resected (Connors and Wise, 1974; Lim etal., 1977; Hande e t a l . , 1978; Fleming etal., 1982; Maor et al., 1984 Skudder and Schwartz, 1985) and this variability no doubt reflects differences in the extent of pre-operative assessment. The primary use of either radiotherapy or chemotherapy to treat unresected tumours is frequently considered to be disappointing (Hande et al., 1978; Blackledge et al., 1979; Rosenfelt and Rosenberg, 1980; Economopoulos et al., 1985) and may entail an increased risk of gastrointestinal haemorrhage or perforation due to the rapid lysis of full-thickness tumour deposits (Hande et al., 1978; Herrmann et al., 1980; Weingard et al., 1982; Sheridan et al., 1985). In a study of 18 patients, Hande et al. (1978) reported eight episodes of perforation or massive gastrointestinal haemorrhage following treatment with C-MOPP (cyclophosphamide, vincristine, procarbazine, prednisone) or BACOP (cyclophosphamide, doxorubicin, vincristine, bleomycin, prednisone). Four of the six patients who perforated had an incompletely resected turnout and two deaths were caused by massive haemorrhage from necrotic tumour; these complications interrupted or curtailed the treatment in five responding patients and represented a major cause of treatment failure. Rosenfelt and Rosenberg (1980) reported similar complications in 25% of patients treated with chemotherapy or radiotherapy. However, Economopoulos et al. (1985), reported no similar complications in their series of gastric lymphomas treated with COP (cyclophosphamide, vincristine, prednisone) or CHOP (cyclophosphamide, doxorubicin, vincristine, prednisone) despite the tumour being unresectable in 31% of the cases studied. The role of radiotherapy in the management of gastrointestinal lymphoma remains poorly defined because treatment, even within individual studies, is

613

rarely given in a uniform way. Radiotherapy used alone produces poor results (Dawson etal., 1961 ; Connors and Wise, 1974; Lira et al., 1977; Rosenfelt and Rosenberg, 1980; Brooks and Enterline, 1983; Shimm et al., 1983) and more often follows surgery (Faulkner and Docherty, 1952; Bush and Ash, 1969; Fu and Perzin, 1972; Gary et al., 1982; Rao et al., 1984; Skudder and Schwartz, 1985). Lira et al. (1977) reported no local recurrence in 15 gastric lymphomas treated by curative resection and radiotherapy compared with a recurrence in 2 of 20 patients treated with resection alone. Bush and Ash (1969) used large abdominal radiotherapy fields to encompass the area of the gastrointestinal tract in which the turnout had arisen and gave 2500 cGy over 4 to 6 weeks, in early stage disease, an improved 2 year disease free survival from 44% to 64% was seen which was reflected in a significant survival advantage. In gastric lymphoma the survival of Stage IIE patients (55%) compares unfavourably with Stage IE (86%) and consequently adjuvant radiotherapy has been proposed (Brooks and Enterline, 1983). Rosenfelt and Rosenberg (1980) used adjuvant whole abdominal radiotherapy (2500 cGy) for completely resected Stage IE tumours and obtained a complete remission rate of 85%. However in Stage IIE, 55% of those patients who relapsed had also received post-operative radiotherapy. Connors and Wise (1974) found no benefit from low dose radiotherapy and both Blackledge et al. (1979) and Shimm et al. (1983) also reported no significant survival advantage with similar adjuvant treatment. Only 50% of patients who relapse do so locally (Lewin et al., 1978; Maor et al., 1984) and disease frequently recurs at extra-abdominal sites. The use of adjuvant chemotherapy therefore merits consideration. However, gastrointestinal lymphoma is usually of an aggressive histological type requiring the use of combination chemotherapy regimes. Sheridan et al. (1985) used post-operative chemotherapy in 23 patients with gastric lymphoma; patients in Stages IE and liE, with no residual macroscopic disease, were treated with CVP (cyclophosphamide, vincristine, prednisolone) and the remaining patients, including those in Stages IIIE and IV, were treated using CHOP. All 17 patients who completed treatement were alive and disease free at mean follow-up of 44 months and two of the three patients with inoperable tumours achieved a complete remission with CHOP. Hande et al. (1978) treated 18 patients (14 in Stage IV) with C-MOPP or BACOP; six failed to respond to treatment but five, all treated with BACOP, obtained a complete remission, In their study, the addition of radiotherapy (2500 to 4500 cGy) did not appear to influence survival and morbidity was least when the tumour had been resected. Economopoulos et al. (1985) reported a series of 29 cases of gastric lymphoma; 26 patients underwent laparotomy of whom 22 were found to have involvement of the perigastric, mesenteric or regional lymph nodes. A gastrectomy was possible in 20 cases and all the patients in the study were also treated with either CHOP or COP. After six cycles of chemotherapy, 62% were in complete remission and only two relapses were reported with a median follow-up of 31.5 months. However, only three of the nine patients with unresected tumours achieved a complete remission. The recurrence rate in gastrointestinal non-Hodgkin's

614

CLINICAL RADIOLOGY

l y m p h o m a is largely s y n o n y m o u s with m o r t a l i t y rate a n d p r i m a r y t r e a t m e n t t h e r e f o r e r e q u i r e s careful cons i d e r a t i o n . W h i l s t c o m p l e t e surgical r e s e c t i o n m a y be c u r a t i v e in early stage d i s e a s e , t h e s e p a t i e n t s m u s t be i d e n t i f i e d r e l i a b l y using b o t h i m a g i n g t e c h n i q u e s and p a t h o l o g i c a l staging at t h e t i m e of the s u r g e r y . T h e role of r a d i o t h e r a p y r e m a i n s i l l - d e f i n e d a n d whilst substantial a r e a s of t h e a b d o m e n m a y b e i r r a d i a t e d , the t o l e r a n c e of n o r m a l s t r u c t u r e s within t h e t r e a t m e n t field m a y c o m p r o m i s e such t r e a t m e n t . C h e m o t h e r a p y is o p t i m a l l y given w h e n t h e p r i m a r y t u m o u r has b e e n r e s e c t e d , e v e n if s u b t o t a l l y , a n d is t h e only effective t r e a t m e n t for a d v a n c e d disease. In an a d j u v a n t role c h e m o t h e r a p y also offers t h e p o t e n t i a l to e r a d i c a t e r e s i d u a l local d i s e a s e a n d occult m e t a s t a s e s with a possible r e d u c t i o n in the i n c i d e n c e of r e l a p s e , for which t r e a t m e n t has p r o v e d so difficult. H o w e v e r , gastrointestinal l y m p h o m a is n o r m a l l y o f an aggressive histological s u b t y p e r e q u i r i n g t r e a t m e n t with c o m b i n a t i o n c h e m o t h e r a p y r e g i m e s such as C H O P which a r e associa t e d with a significant m o r b i d i t y . Such a d j u v a n t t r e a t ment therefore requires careful consideration, e s p e c i a l l y in the e l d e r l y , and s h o u l d be e x p l o r e d within t h e a u s p i c e s of a c o n t r o l l e d m u l t i c e n t r e clinical trial to e n a b l e sufficient cases to be a c c r u e d for m e a n i n g f u l c o n c l u s i o n s to b e d r a w n . Acknowledgements. The authors wish to thank the consultant staff at Mount Vernon Hospital for allowing their patients to be included in this study. REFERENCES

Azzopardi, JG & Menzies, T (1960). Primary malignant lymphoma of the alimentary tract. British Journal of Surgery, 47, 358-366. Bennett, MH, Farrer-Brown, G, Henry, K & Jelliffe, AM (1974). Classification of non-Hodgkin's lymphomas. Lancet, ii, 404-406. Blackledge,G, Bush, H, Dodge, OG & Crowther, D (1979). A study of gastro-intestinal lymphoma. Clinical Oncology, 5, 209-219. Brooks, JJ & Enterline, HT (1983). Primary gastric lymphomas. A clinicopathologic study of 58 cases with long term follow-up and literature review. Cancer, Sl, 701-711. Bush, RS & Ash, CL (1969). Primary lymphoma of the gastrointestinal tract. Radiology, 92, 1349-1354. Carbone, PP, Kaplan, HS, Musshoff, K, Smithers, DW & Tubiana, M. (1971). Report of the committee on Hodgkin's disease staging classification. Cancer Research, 31, 1860-1861. Connors, J & Wise, L (1974). Management of gastric lymphomas. American Journal of Surgery, 27, 102-108. Crowther, D & Rankin, EM (1982). Staging patients with nonHodgkin's lymphoma. British Journal of Haematology, 52, 357367. Dawson, IMP, Cornes, JS & Marson, BC (1961). Primary malignant lymphoid tumours of the intestinal tract. British Journal of Surgery, 49, 80-89. Economopoulos, T, Alexopoulos, C, Stathakis, N, Styloyannis, J, Dervenoulas, J, Tsouis Set al. (1985). Primary gastric lymphoma The experience of a general hosptial. Britis'hJournal of Cancer, 52, 391-397. Farrer-Brown, G (198l). Prognosis and survival in the non-Hodgkin's lymphomas: analysis of the different histopathological types (BNLI Report No. 8). Clinical Radiology, 32, 501-504. Faulkner, JW & Docherty, MB (1952). Lymphosarcoma of the small intestine. Surgery, Gynecology and Obstetrics, 95, 76-84. Fisher, RI & DeVita, VT (1977). Prognostic factors for advanced diffuse histiocytic lymphoma following treatment with combination chemotherapy. American Journal of Medicine, 63, 177-182. Fisher, RI & Hubbard, SM (198l). Factors predicting long term survival in diffuse mixed histioeytic or undifferentiated lymphoma. Blood, 58, 45-51.

Fleming, ID, Mitchell, S & Dilawari, RA (1982). The role of surgery in the management of gastric lymphoma. Cancer, 49, 1135-1141. Freeman, C, Berg, JW & Cutler, SJ (1972). Occurrence and prognosis of extranodal lymphomas. Cancer, 29, 252-260. Fu, YS & Perzin, KH (1972). Lymphosarcoma of the small intestine: a clinicopathologic study. Cancer, 29, 645-659. Gary, GM, Rosenberg, SA, Cooper, AD, Gregory, AD,Stein, DT & Herzenberg, H (1982). Lymphomas involving the gastrointestinal tract. Gastroenterology, 82, 143-152. Green, JA, Dawson, AA, Jones, PF & Brunt, PW (1979). The presentation of gastrointestinal lymphoma: study of a population. British Journal of Surgery, 66, 798-801. Hande, KR, Fisher, RI, DeVita, VT, Chabner, BA & Young, RC (1978). Diffuse histiocyticlymphoma involving the gastrointestinal tract. Cancer, 41, 1984-1989. Henry, K, Bennett, MH & Farrer-Brown, G (1978). Classification of non-Hodgkin's lymphoma. In Recent Advances in Histopathology, ed. Anthony, PP & Woolf, N. Churchill Livingstone, Edinburgh. Henry, K (1981). National Cancer Institute Retrospective Study of the Non-Hodgkin's Lymphomas: Abstract (BNLI Report No. 7). Clinical Radiology, 32, 499-500. Herrmann, R, Panahan, A, Barcos, MP, Walsh, D & Stutzman, L (1980). Gastrointestinal involvement in non-Hodgkin's lymphoma. Cancer, 46, 215-222. Lewin, KJ, Ranchod, M & Dorfman, RF (1978). Lymphomas of the gastrointestinal tract. A study of 117 cases presenting with gastrointestinal disease. Cancer, 42, 693-707. Lim, FE, Hartman, AS, Tan, EGC, Cady, B & Meissner, WA (1977). Factors in the prognosis of gastric lymphoma. Cancer, 39, 17151729. Loehr, WJ, Mujahed, Z & Zahn, FD (1969). Primary lymphoma of the gastrointestinal tract; a review of 100 cases. Annals" of Surgery, 170, 232-238. Maor, MH, Maddux, B, Osborne, BM, Fuller, LM, Sullivan, JA, Nelson, RS et al. (1984). Stages IE and IIE non-Hodgkin's lymphomas of the stomach. Cancer, 54, 2330-2337, Naqvi, MS, Burrows, L & Kark, AE (1969). Lymphoma of the gastrointestinal tract: Prognostic guides based on 162 cases. Annals" of Surgery, 170, 221-231. National Cancer Institute (NCI) (1982). Sponsored Study of Classifications of Non-Hodgkin's Lymphomas. Summary and description of the working formulation for clinical usage. Cancer, 49, 2112-2135. Paulson, S, Sheehan, RG and Stone, MJ (1983). Large cell lymphomas of the stomach: improved prognosis with the resection of all intrinsic gastrointestinal disease. Journal of Clinical Oncology, 1,263269. lymphomas. Recent Advances in Haematology, 2, 263-267. Peters, MW, Hasselback, R & Brown, TC (1968). The natural history of lymphomas related to clinical classification. In Proceedings of the International Confererence on Leukaemia-Lymphoma, ed. Zarafonetis, C, pp. 357-371. Lea & Febinger, Philadelphia. Rao, AR, Kagan, AR, Potyk, D, Nussbaum, H, Chart, P, Hintz, BLet al. (1984). Management of gastrointestinal lymphoma. American Journal of Clinical Oncology, 7, 213-219. Rosenberg, SA, Diamond, HD, Jaslowitz, B & Craver, LF (1961). Lymphosarcoma: a review of 1209 cases. Medicine, 40, 31-84. Rosenfelt, F & Rosenberg, SA (1980). Diffuse histiocytic lymphoma presenting with gastrointestinal tract lesions. The Stanford Experience. Cancer, 45, 2188-2193. Saraga, P, Hurliman, J & Ozzello, L (1981). Lymphomas and pseudolymphomas of the alimentary tract. Human Pathology, 12, 713723. Sheridan, WP, Medley, G & Brodie, GN (1985). Non-Hodgkin's lymphoma of the stomach: A prospective pilot study of surgery plus chemotherapy in early and advanced disease. Journal of Clinical Oncology, 3, 495-500. Shimm, DS, Dosoretz, DE, Anderson T, Linggood, RM, Harris, NL & Wang, CC (1983). Management of gastric lymphoma. Cancer, 52, 2044-2048. Skudder, PA & Schwartz, SI (1985). Primary lymphoma of the gastrointestinal tract. Surgery, Gynecology and Obstetrics, 160, 5-8. Weaver, DK & Batsakis, JG (1964). Primary lymphomas of the small intestine. American Journal of Gastroenterology, 42, 620-625. Weingard, DN, Decosse, JJ, Sherlock, P, Straus, D, Lieberman, PH & Filippa, DA (1982). Primary gastrointestinal lymphoma, A 30 year review. Cancer, 49, 1258-1265.