Inr J. Rodrnlron Oncology Bwl Phys, Vol. Printed in the US A. All rights reserved.
19, pp. 1351-1355 Copyright
0360-3016/90 $3.00 t .oO Cl 1990 Pergamon Press plc
??Original Contribution
OUTCOME ANALYSIS OF LOCALIZED GASTROINTESTINAL LYMPHOMA TREATED WITH SURGERY AND POSTOPERATIVE IRRADIATION MARY K. GOSPODAROWICZ,* SIMON B. SUTCLIFFE,* ROY M. CLARK,* ALON J. DEMBO,* BRUCE J. PATTERSON,+ PETER J. FITZPATRICK,* THERESA CHUA* AND RAYMOND S. BUSH* Princess Margaret Hospital. Toronto. Ontario, Canada One hundred thirteen patients with localized gastrointestinal lymphoma treated by surgery and postoperative irradiation between 1967 and 1985 were reviewed. At 15 years, actuarial survival of this group was 40.6%, with a cause-specific survival of 69.2% and a relapse-free rate of 64%. Two-thirds of relapses occurred at distant sites. In Stage IA and IIA patients with no residuum or with positive resection margins, (N = 90) only site of involvement and stage predicted for relapse. Age, histologic subtype group, and depth of bowel wall invasion did not affect relapse risk. In the very favorable group (Stage IA, IIA, no residuum or microscopic residuum), 8.4% of patients with stomach lymphoma relapsed compared to 25% of patients with small bowel lymphoma. The risk of early relapse was higher in those with Stage IIA small bowel lymphoma than those with Stage IA small bowel lymphoma. We continue to recommend adjuvant abdominal irradiation for patients with Stage IA, IIA completely resected stomach lymphoma and Stage IA completely resected small bowel lymphoma. We recommend combined modality therapy for patients with completely resected Stage IIA small bowel lymphoma and all other localized gastrointestinal lymphoma where visible residual disease is present. Gastrointestinal
lymphoma, Surgery, Radiation prognostic factors.
INTRODUCrION
ported in the previous study (1967-1978) was updated and the cohort of patients with Stage I and II GIL treated with surgical resection and abdominal irradiation between 1979 and 1985 was added.
The gastrointestinal tract is one of the most common sites of extranodal involvement for non-Hodgkin’s lymphoma. Primary gastrointestinal lymphoma (GIL) is a disease observed in patients presenting with gastrointestinal symptoms in whom the predominant lesion is confined to the gut with no evidence of extra-abdominal involvement. These patients are classified in the Ann Arbor system as having Stage IE and IIE lymphoma. Treatment of localized GIL remains controversial. Although some investigators recommend surgical resection followed by either radiation therapy (XRT) or systemic chemotherapy, others recommend primary chemotherapy without resection of the primary lesion (3, 7, 9, 10, 11). We have previously reported a failure analysis of 1 13 patients with localized GIL (5). A favorable group of patients for treatment with surgery followed by abdominal irradiation was identified. This group included 52 patients with Stage IA and Stage IIA with no residual or small residual (usually microscopic) bulk after surgical resection. Patients in this group had 85% failure-free rate and an 88% cause-specific survival at 10 years. To re-evaluate our earlier recommendation, the follow-up of patients re-
METHODS AND MATERIALS The recordsof all adult patientswith localizedGIL treated primarily at the Princess Margaret Hospital with surgery and postoperative irradiation between January 1, 1967 and December 3 1, 1985 were reviewed, and followup of those reviewed previously was undertaken. Staging procedures, treatment methods, and statistical methods were previously reported (5). Since 1979, patients were staged with postoperative CT scan of the abdomen in addition to the lymphogram, chest X ray, bone marrow biopsy, GI series, and laboratory tests including complete blood count, sedimentation rate (ESR), and liver function tests. In addition to the previously reported factors, we examined the prognostic value of the depth of bowel wall invasion based on the microscopic description in the pathology report. Patients were divided in three groups ac-
+ Department of Pathology. * Department of Biostatistics. Accepted for publication 2 I June 1990.
Presented at the 17th International Congress of Radiation Oncology, Paris, July. 1989. * Department of Radiation Oncology. 1351
1352
1. J. Radiation Oncology 0
Biology 0 Physics
December
1990. Volume
Table 1. Characteristics of 1I3 patients with localized GIL treated with surgical resection and postoperative XRT
19. Number
6
RESULTS of patients treated between 1967 and 113 patients with localized GIL who were treated with surgery and abdominal irradiation. Median age of this group of patients was 57 years and the age ranged from 17 to 84 years. Distribution by stage, site of presentation, and histology is shown in Table 1. Note that there were no patients with Burkitt’s lymphoma in this series. Of 70 patients with stomach lymphoma, 50 received whole abdominal irradiation to a median dose of 2500 cGy (range 1500-3000 cGy), in 20 fractions over 4 weeks, and 20 received irradiation to the upper abdomen (stomach bed with celiac and upper paraaortic nodes) to a median dose of 3500 cGy (range 2000-4000 cGy) in 20 fractions over 4 weeks. Posterior kidney shields were placed to allow a maximum dose of 2000 Gy to the kidneys. The majority (34/35) of patients with small bowel lymphoma received whole abdominal irradiation to the same dose as those with stomach lymphoma. Patients with large bowel. usually rectal, lymphoma were treated with pelvic irradiation. There were no serious late XRT complications: specifically, no patients developed perforation or malabsorption. The overall actuarial survival was 59% at 10 years and 40.6% at I5 years; cause-specific survival was 76.6% at 10 years and 69.2% at 15 years. Relapse-free rate was 64% at 10 and 15 years, with distant extra-abdominal failures representing 67.5% of all relapses (Fig. la). Since our previous analysis suggested that only asymptomatic patients with small bulk (microscopic residuum) or no residual disease after surgery should be treated with abdominal irradiation, analysis of prognostic factors and pattern of failure was limited to this group only (90 patients). Cause-specific survival for this group was 86.9%) Reviewing
No. pts.
records
1985. we identified
Stage IA IIA localized IIA extensive 19 119 Histological subtype Low grade Intermediate grade Diffuse small Follicular large Diffuse large High grade Unclassified Postoperative bulk No residuum Small Medium Large Not known Site Stomach Small bowel Large bowel
54 41 6 4 8 17
88 17 64 5 3 84 16 4 8
70 33 10
Small bulk = ~2.5 cm. residual tumor mass or no residuum; medium bulk = 2.5-5 cm. residuum: large bulk = >5 cm. residuum.
cording to the level of bowel wall invasion: those with tumor confined to the submucosa, those with extension to the muscularis propria, and those with extension to the serosa. Survival and relapse-free rates were calculated using Wilcoxon-Gehan method (4). Analysis for prognostic factors was performed using log-rank method of Peto (8).
Cause-Specific
Survival (90.13)
iii-*
6 Time (years)
9
12
15
3
6
9
12
Time (years)
Fig. 1. (a) Cause-specific survival (CSS) and relapse-free rate (RFR) for 1 13 patients treated with surgery and postoperative XRT. (b) Cause-specific survival and relapse-free rate for 90 patients with Stage IA and IIA, no residuum after surgery treated with postoperative XRT.
15
Localized gastrointestinal lymphoma 0 M. K.
1353
GOSPODAROWICZ et al.
Table 3. Prognostic factors in 90 patients with no residual disease or small residuum only after surgery (Log rank
method): prognostic factors-multivariate
analysis
Outcome variable
Age Stage Pathology Site Bulk
Death from disease p value
Relapse p value
0.17 0.05 0.14 0.02 0.52
0.5 0.03 0.3 0.009 0.06
P=O 008
i
60 -
Small Bowel (24-5)
\ \
\__k____-__-f_ 50
at 10 years and 82.3% at 15 years, and the relapse-free rate was 75.5% at 10 and 15 years (Fig. lb). The prognostic value of patient’s age, stage, histologic subtype, postoperative tumour bulk, and site of involvement were assessed in a multivariate analysis (Table 2). Note that the depth of the bowel wall invasion did not influence prognosis in patients without gross residual disease after surgical resection (Table 3). The lack of prognostic value for pathological subtype is likely due to the small number of patients presenting with low grade histology and the good overall treatment results. At I5 years the crude risk of relapse in patients with Stage IA or IIA stomach lymphoma with no residual disease or with microscopic residual disease was 8.4% (5/59 patients). The risk of relapse for Stage IA and IIA patients with small bowel lymphoma was 25% (6/24 patients). The actuarial risk of relapse by site is shown in Figure 2. Patients with Stage IIA small bowel lymphoma (4 of 14 patients relapsed) appeared to have a higher risk of early relapse (within the first 5 years post-therapy) than those with Stage IA small bowel lymphoma ( l/ 10 patients relapsed), although overall risk of relapse measured over 10 years was not different. Cause-specific survivals and relapse-free rates were excellent for patients with Stage IA and IIA stomach lymphoma as well as the small number
fi 2
0
”
4
a 6
”
8
”
10
4
5 n 12
a 14
t. 16
B ” 18 20
Time (years)
Fig. 2. Localized GIL Stage IA, HA with no residuum or small residuum post op. treated with abdominal XRT relapse-free rate by site of involvement.
of patients with Stage IA small bowel lymphoma (Fig. 3a and b). There were only six patients with completely resected rectal lymphoma treated with postoperative XRT. Only one relapsed and none died of lymphoma. The number of patients with rectal lymphoma is too small. however. to draw firm conclusions about management. In this analysis we confirmed that our previously published criteria for selection of patients with localized GIL for treatment with postoperative XRT give reproducible results in a prospectively documented and followed group of patients (Table 4). DISCUSSION Although gastrointestinal tract involvement by nonHodgkin’s lymphoma is common, favorable presentations of localized and completely resectable disease are infrequent. Failure analysis of patients with such presentations is difficult because a large number of known prognostic factors must be considered. As always, in case of relatively
Table 3. Prognostic value of depth of bowel wall invasion in patients with localized GIL with no residuum or small residuum
post surgery (Log rank method) Relapse
Depth of bowel wall invasion Submucosa Muscle Serosa
C.S.S.
No. of pts.
Univariate O/E
9 21 68
0.39 0.54 1.2
0.87 1.04 0.97
0.62 0.91 1.01
0.93 1.71 0.85
0.16 5.15
0.54 2.11
0.3 3.64
0.19 4.71
p value Overall X2 C.S.S. = Cause-specific survival. * Adjusted for influence of stage, site, and tumor O/E = observed to expected ratio for the event.
bulk.
Adjusted* O/E
Univariate O/E
Adjusted* WE
1354
1. J. Radiation Oncology 0 Biology 0 Physics
December 1990, Volume 19. Number 6
a
a
b
I I
p
I I_______-____________----IIASmall Bowel(14-3)
70
.
\
p-0016
\
L _ mm, \ \
60
t
t
0
70 - :
2
4
6
6
10
12
14
16
16
20
501. 0
’ 2
Time (years)
small numbers of patients, multivariate analysis can underestimate the importance of some factors. Almost 70% of patients with localized GIL present with large cell histological subtype. Large cell lymphomas can be cured with combination chemotherapy, and in patients presenting with generalized large cell lymphoma, approximately one-third are cured (1). Treatment of localized large cell non-Hodgkin’s lymphomas with a combination of chemotherapy and radiation gives excellent results (2). Localized GIL are increasingly treated with chemotherapy due to the belief that patients treated with radiation alone have a high relapse rate (9). The adverse prognostic value of the tumor bulk in nonHodgkin’s lymphoma is well recognized, and many patients with localized GIL present with bulky disease. One would therefore expect a high risk of occult metastatic disease associated with localized GIL. Our failure analysis, however, identifies a very low risk of distant failure for a favorable localized GIL treated with surgical resection and abdominal XRT. This suggests a different pattern of disease from nodal or other extranodal lymphomas. Indeed, in patients with bulky and unresectable GIL, it is failure to control local disease that results in poor outcome rather than distant failure. A tissue-specific homing mechanism has been implicated to explain the behavior of primary GIL (6).
Table 4. Stage I and II stomach and small bowel lymphoma treated with surgery and postoperative XRT
Favorable Unfavorable
L__________________ ’ 4
’ 6
’ 8
10
’ 12
’ 14
’
16
16
’
20
Time (years)
Fig. 3. Localized GIL Stage IA and IIA with no residuum XRT. (a) Cause-specific &.uvival; (b) Relapse-free rate.
1967-1978
IIASmall Eowel(14-3)
\
1979-1985
No. of pts.
RFR %
No. of pts.
RFR %
48 18
85.2 31.1
36 1
81.6 0
RFR = relapse-free rate at 5 years: Favorable = Stage IA and IIA, complete resection or microscopic residuum only; Unfavorable = Stage IA and IIA with gross residual disease, Stage IB, Stage IIB.
or small residuum
postoperatively
treated with abdominal
In our previous analysis we were unable to show any difference between stomach lymphoma and small bowel lymphoma or between Stage IA and Stage IIA localized presentations. With larger patient numbers and longer follow-up, we were able to show that patients with small bowel lymphoma and particularly those with Stage II small bowel lymphoma are at a higher risk of failure. Unlike patients with resectable stomach lymphoma, those with small bowel involvement are more likely to have extensive nodal involvement in the mesentery and to have more extensive disease at presentation. Accurate information regarding intra-abdominal extent of disease is not easily available with the usual staging procedures and is dependent on the description of disease at the initial laparotomy. In recognition of the higher risk of relapse for patients with small bowel lymphoma, we now recommend a combined modality approach for those patients, especially those with Stage II disease, except for those with low grade histology. The data presented above suggest that low dose 20-25 Gy postoperative XRT for patients with resected tumors provides as good results as with a higher dose therapy. There is no indication that the whole abdominal XRT gives better results than the regional XRT for stomach lymphoma: however, with the number of patients at risk, a small difference in relapse risk can be missed. Accurate information regarding completeness of surgery, extent of intraabdominal involvement, and staging is necessary prior to recommending management with postoperative abdominal XRT alone. In cases where residual disease is suspected, management with systemic chemotherapy or a combined modality approach is recommended. The definition of the subgroup of patients favorable for treatment with postoperative XRT alone was confirmed by outcome of patients treated between 1979 and 1985. With larger numbers of patients available for analysis, we were able to recognize a higher risk of relapse for patients with Stage IIA small bowel lymphoma and to exclude them from the favorable category. Although results of treatment with surgery and che-
Localized gastrointestinal lymphoma 0 M. K.
motherapy of localized GIL are excellent, we believe that the approach of surgery and low dose postoperative abdominal XRT produces excellent results and is safer, especially in older patients, and is to be preferred. The next question to be asked is whether one can identify a group of patients who are curable with surgery alone. It is likely that a large proportion of patients currently treated with postoperative XRT are cured with surgery
GOSPODAROWICZ
et al.
1355
alone. The real benefit of postoperative XRT is not known for those patients with no residual disease postoperatively. Abdominal failures are difficult to diagnose early, and there is no evidence that salvage therapy of relapsed large cell lymphoma is successful. Because of the above, and the relative lack of toxicity of low dose abdominal XRT, investigators have been reluctant to withhold postoperative therapy.
REFERENCES 1.
Coltman, C. A.; Dahlberg, S.; Jones, S. E.; Miller, T. P.: Dana, B. W.; McKelvey. E. M.; Hartstock, R. J.: Dixon,
D. 0. CHOP is curative in thirty percent of patients with large cell lymphoma: a twelve-year southwest oncology group follow-up. In: Skain, A. T., ed. Advances in chemotherapy: update on treatment for diffuse large cell lymphoma. New York: John Wiley and Sons; 1986:71-77. 2. Connors. J. M.; Klimo, P.; Fairey, R. N. Brief chemotherapy and involved field radiation therapy for limited stage histologically aggressive histology lymphoma. Ann. Intern. Med. 23:107; 1987. 3. Dragosics, B.; Bauer, P.; Radaszkiewicz, T. Primary gastrointestinal non-Hodgkin’s lymphomas. A retrospective clinicopathologic study of I50 cases. Cancer 55: 1060; 1985. 4. Gehan. E. A. A generalised Wilcoxon test for comparing arbitrarily single-censored samples. Biometrika 52:203-223: 1965. 5. Gospodarowicz, M. K.; Bush, R. S.; Brown, T. C.; Chua. T. Curability of gastrointestinal lymphoma with combined surgery and radiation. Int. J. Radiat. Oncol. Biol. Phys. 9: 3: 1983. 6. Isaacson. P.: Wright, D. Extranodal malignant lymphoma arising from mucosa-associated lymphoid tissue. Cancer 53: 2515: 1984.
7
Maor, M. H.; Maddux, B.; Osborne, B. M.; Fuller, L. M.: Sullivan, J. A.; Nelson, R. S.; Martin, R. G.; Libshitz, H. I.; Velasquez. W. S.: Bennett, R. W. Stages IE and IIE nonHodgkin’s lymphomas of the stomach. Comparison of treatment modalities. Cancer 54:2330: 1984.
8 Pete, R.; Pike, M. C.; Armitage. P.; Breslow, N. E.: Cox. D. R.; Howard, S. V.; Mantel, N.; McPherson, K.: Pete, J.; Smith. P. G. Design and analysis of randomised clinical trials requiring prolonged observation of each patient. Br. J. Cancer 35:1-39: 1977. 9. Shepherd. F. A.: Evans, W. K.; Kutas, G.: Yau, J. C.; Dang, P.; Scott, J. G.; Farquharson, H. A.: Francombe, W. H.: Bailey, D.; Baker, M. A. Chemotherapy following surgery for Stages IE and IIE non-Hodgkin’s lymphoma of the gastrointestinal tract. J. Clin. Oncol. 1988:253. 10. Shimm. D. S.; Dosoretz, D. E.; Anderson, T.; Linggood. R. M.; Harris, N. L.: Wang, C. C. Primary gastric lymphoma. An analysis with emphasis on prognostic factors and radiation therapy. Cancer 52:2044; 1983. 1 I. Shiu, M. H.: Nisce, L. Z.; Pinna. A.; Strauss, D. J.; Tome. M.; Filippa. D. A.: Lee, B. J. Recent results of multimodal therapy of gastric lymphoma. Cancer 58: 1389: 1986.
19, pp. 1351-1355 Copyright
0360-3016/90 $3.00 t .oO Cl 1990 Pergamon Press plc
??Original Contribution
OUTCOME ANALYSIS OF LOCALIZED GASTROINTESTINAL LYMPHOMA TREATED WITH SURGERY AND POSTOPERATIVE IRRADIATION MARY K. GOSPODAROWICZ,* SIMON B. SUTCLIFFE,* ROY M. CLARK,* ALON J. DEMBO,* BRUCE J. PATTERSON,+ PETER J. FITZPATRICK,* THERESA CHUA* AND RAYMOND S. BUSH* Princess Margaret Hospital. Toronto. Ontario, Canada One hundred thirteen patients with localized gastrointestinal lymphoma treated by surgery and postoperative irradiation between 1967 and 1985 were reviewed. At 15 years, actuarial survival of this group was 40.6%, with a cause-specific survival of 69.2% and a relapse-free rate of 64%. Two-thirds of relapses occurred at distant sites. In Stage IA and IIA patients with no residuum or with positive resection margins, (N = 90) only site of involvement and stage predicted for relapse. Age, histologic subtype group, and depth of bowel wall invasion did not affect relapse risk. In the very favorable group (Stage IA, IIA, no residuum or microscopic residuum), 8.4% of patients with stomach lymphoma relapsed compared to 25% of patients with small bowel lymphoma. The risk of early relapse was higher in those with Stage IIA small bowel lymphoma than those with Stage IA small bowel lymphoma. We continue to recommend adjuvant abdominal irradiation for patients with Stage IA, IIA completely resected stomach lymphoma and Stage IA completely resected small bowel lymphoma. We recommend combined modality therapy for patients with completely resected Stage IIA small bowel lymphoma and all other localized gastrointestinal lymphoma where visible residual disease is present. Gastrointestinal
lymphoma, Surgery, Radiation prognostic factors.
INTRODUCrION
ported in the previous study (1967-1978) was updated and the cohort of patients with Stage I and II GIL treated with surgical resection and abdominal irradiation between 1979 and 1985 was added.
The gastrointestinal tract is one of the most common sites of extranodal involvement for non-Hodgkin’s lymphoma. Primary gastrointestinal lymphoma (GIL) is a disease observed in patients presenting with gastrointestinal symptoms in whom the predominant lesion is confined to the gut with no evidence of extra-abdominal involvement. These patients are classified in the Ann Arbor system as having Stage IE and IIE lymphoma. Treatment of localized GIL remains controversial. Although some investigators recommend surgical resection followed by either radiation therapy (XRT) or systemic chemotherapy, others recommend primary chemotherapy without resection of the primary lesion (3, 7, 9, 10, 11). We have previously reported a failure analysis of 1 13 patients with localized GIL (5). A favorable group of patients for treatment with surgery followed by abdominal irradiation was identified. This group included 52 patients with Stage IA and Stage IIA with no residual or small residual (usually microscopic) bulk after surgical resection. Patients in this group had 85% failure-free rate and an 88% cause-specific survival at 10 years. To re-evaluate our earlier recommendation, the follow-up of patients re-
METHODS AND MATERIALS The recordsof all adult patientswith localizedGIL treated primarily at the Princess Margaret Hospital with surgery and postoperative irradiation between January 1, 1967 and December 3 1, 1985 were reviewed, and followup of those reviewed previously was undertaken. Staging procedures, treatment methods, and statistical methods were previously reported (5). Since 1979, patients were staged with postoperative CT scan of the abdomen in addition to the lymphogram, chest X ray, bone marrow biopsy, GI series, and laboratory tests including complete blood count, sedimentation rate (ESR), and liver function tests. In addition to the previously reported factors, we examined the prognostic value of the depth of bowel wall invasion based on the microscopic description in the pathology report. Patients were divided in three groups ac-
+ Department of Pathology. * Department of Biostatistics. Accepted for publication 2 I June 1990.
Presented at the 17th International Congress of Radiation Oncology, Paris, July. 1989. * Department of Radiation Oncology. 1351
1352
1. J. Radiation Oncology 0
Biology 0 Physics
December
1990. Volume
Table 1. Characteristics of 1I3 patients with localized GIL treated with surgical resection and postoperative XRT
19. Number
6
RESULTS of patients treated between 1967 and 113 patients with localized GIL who were treated with surgery and abdominal irradiation. Median age of this group of patients was 57 years and the age ranged from 17 to 84 years. Distribution by stage, site of presentation, and histology is shown in Table 1. Note that there were no patients with Burkitt’s lymphoma in this series. Of 70 patients with stomach lymphoma, 50 received whole abdominal irradiation to a median dose of 2500 cGy (range 1500-3000 cGy), in 20 fractions over 4 weeks, and 20 received irradiation to the upper abdomen (stomach bed with celiac and upper paraaortic nodes) to a median dose of 3500 cGy (range 2000-4000 cGy) in 20 fractions over 4 weeks. Posterior kidney shields were placed to allow a maximum dose of 2000 Gy to the kidneys. The majority (34/35) of patients with small bowel lymphoma received whole abdominal irradiation to the same dose as those with stomach lymphoma. Patients with large bowel. usually rectal, lymphoma were treated with pelvic irradiation. There were no serious late XRT complications: specifically, no patients developed perforation or malabsorption. The overall actuarial survival was 59% at 10 years and 40.6% at I5 years; cause-specific survival was 76.6% at 10 years and 69.2% at 15 years. Relapse-free rate was 64% at 10 and 15 years, with distant extra-abdominal failures representing 67.5% of all relapses (Fig. la). Since our previous analysis suggested that only asymptomatic patients with small bulk (microscopic residuum) or no residual disease after surgery should be treated with abdominal irradiation, analysis of prognostic factors and pattern of failure was limited to this group only (90 patients). Cause-specific survival for this group was 86.9%) Reviewing
No. pts.
records
1985. we identified
Stage IA IIA localized IIA extensive 19 119 Histological subtype Low grade Intermediate grade Diffuse small Follicular large Diffuse large High grade Unclassified Postoperative bulk No residuum Small Medium Large Not known Site Stomach Small bowel Large bowel
54 41 6 4 8 17
88 17 64 5 3 84 16 4 8
70 33 10
Small bulk = ~2.5 cm. residual tumor mass or no residuum; medium bulk = 2.5-5 cm. residuum: large bulk = >5 cm. residuum.
cording to the level of bowel wall invasion: those with tumor confined to the submucosa, those with extension to the muscularis propria, and those with extension to the serosa. Survival and relapse-free rates were calculated using Wilcoxon-Gehan method (4). Analysis for prognostic factors was performed using log-rank method of Peto (8).
Cause-Specific
Survival (90.13)
iii-*
6 Time (years)
9
12
15
3
6
9
12
Time (years)
Fig. 1. (a) Cause-specific survival (CSS) and relapse-free rate (RFR) for 1 13 patients treated with surgery and postoperative XRT. (b) Cause-specific survival and relapse-free rate for 90 patients with Stage IA and IIA, no residuum after surgery treated with postoperative XRT.
15
Localized gastrointestinal lymphoma 0 M. K.
1353
GOSPODAROWICZ et al.
Table 3. Prognostic factors in 90 patients with no residual disease or small residuum only after surgery (Log rank
method): prognostic factors-multivariate
analysis
Outcome variable
Age Stage Pathology Site Bulk
Death from disease p value
Relapse p value
0.17 0.05 0.14 0.02 0.52
0.5 0.03 0.3 0.009 0.06
P=O 008
i
60 -
Small Bowel (24-5)
\ \
\__k____-__-f_ 50
at 10 years and 82.3% at 15 years, and the relapse-free rate was 75.5% at 10 and 15 years (Fig. lb). The prognostic value of patient’s age, stage, histologic subtype, postoperative tumour bulk, and site of involvement were assessed in a multivariate analysis (Table 2). Note that the depth of the bowel wall invasion did not influence prognosis in patients without gross residual disease after surgical resection (Table 3). The lack of prognostic value for pathological subtype is likely due to the small number of patients presenting with low grade histology and the good overall treatment results. At I5 years the crude risk of relapse in patients with Stage IA or IIA stomach lymphoma with no residual disease or with microscopic residual disease was 8.4% (5/59 patients). The risk of relapse for Stage IA and IIA patients with small bowel lymphoma was 25% (6/24 patients). The actuarial risk of relapse by site is shown in Figure 2. Patients with Stage IIA small bowel lymphoma (4 of 14 patients relapsed) appeared to have a higher risk of early relapse (within the first 5 years post-therapy) than those with Stage IA small bowel lymphoma ( l/ 10 patients relapsed), although overall risk of relapse measured over 10 years was not different. Cause-specific survivals and relapse-free rates were excellent for patients with Stage IA and IIA stomach lymphoma as well as the small number
fi 2
0
”
4
a 6
”
8
”
10
4
5 n 12
a 14
t. 16
B ” 18 20
Time (years)
Fig. 2. Localized GIL Stage IA, HA with no residuum or small residuum post op. treated with abdominal XRT relapse-free rate by site of involvement.
of patients with Stage IA small bowel lymphoma (Fig. 3a and b). There were only six patients with completely resected rectal lymphoma treated with postoperative XRT. Only one relapsed and none died of lymphoma. The number of patients with rectal lymphoma is too small. however. to draw firm conclusions about management. In this analysis we confirmed that our previously published criteria for selection of patients with localized GIL for treatment with postoperative XRT give reproducible results in a prospectively documented and followed group of patients (Table 4). DISCUSSION Although gastrointestinal tract involvement by nonHodgkin’s lymphoma is common, favorable presentations of localized and completely resectable disease are infrequent. Failure analysis of patients with such presentations is difficult because a large number of known prognostic factors must be considered. As always, in case of relatively
Table 3. Prognostic value of depth of bowel wall invasion in patients with localized GIL with no residuum or small residuum
post surgery (Log rank method) Relapse
Depth of bowel wall invasion Submucosa Muscle Serosa
C.S.S.
No. of pts.
Univariate O/E
9 21 68
0.39 0.54 1.2
0.87 1.04 0.97
0.62 0.91 1.01
0.93 1.71 0.85
0.16 5.15
0.54 2.11
0.3 3.64
0.19 4.71
p value Overall X2 C.S.S. = Cause-specific survival. * Adjusted for influence of stage, site, and tumor O/E = observed to expected ratio for the event.
bulk.
Adjusted* O/E
Univariate O/E
Adjusted* WE
1354
1. J. Radiation Oncology 0 Biology 0 Physics
December 1990, Volume 19. Number 6
a
a
b
I I
p
I I_______-____________----IIASmall Bowel(14-3)
70
.
\
p-0016
\
L _ mm, \ \
60
t
t
0
70 - :
2
4
6
6
10
12
14
16
16
20
501. 0
’ 2
Time (years)
small numbers of patients, multivariate analysis can underestimate the importance of some factors. Almost 70% of patients with localized GIL present with large cell histological subtype. Large cell lymphomas can be cured with combination chemotherapy, and in patients presenting with generalized large cell lymphoma, approximately one-third are cured (1). Treatment of localized large cell non-Hodgkin’s lymphomas with a combination of chemotherapy and radiation gives excellent results (2). Localized GIL are increasingly treated with chemotherapy due to the belief that patients treated with radiation alone have a high relapse rate (9). The adverse prognostic value of the tumor bulk in nonHodgkin’s lymphoma is well recognized, and many patients with localized GIL present with bulky disease. One would therefore expect a high risk of occult metastatic disease associated with localized GIL. Our failure analysis, however, identifies a very low risk of distant failure for a favorable localized GIL treated with surgical resection and abdominal XRT. This suggests a different pattern of disease from nodal or other extranodal lymphomas. Indeed, in patients with bulky and unresectable GIL, it is failure to control local disease that results in poor outcome rather than distant failure. A tissue-specific homing mechanism has been implicated to explain the behavior of primary GIL (6).
Table 4. Stage I and II stomach and small bowel lymphoma treated with surgery and postoperative XRT
Favorable Unfavorable
L__________________ ’ 4
’ 6
’ 8
10
’ 12
’ 14
’
16
16
’
20
Time (years)
Fig. 3. Localized GIL Stage IA and IIA with no residuum XRT. (a) Cause-specific &.uvival; (b) Relapse-free rate.
1967-1978
IIASmall Eowel(14-3)
\
1979-1985
No. of pts.
RFR %
No. of pts.
RFR %
48 18
85.2 31.1
36 1
81.6 0
RFR = relapse-free rate at 5 years: Favorable = Stage IA and IIA, complete resection or microscopic residuum only; Unfavorable = Stage IA and IIA with gross residual disease, Stage IB, Stage IIB.
or small residuum
postoperatively
treated with abdominal
In our previous analysis we were unable to show any difference between stomach lymphoma and small bowel lymphoma or between Stage IA and Stage IIA localized presentations. With larger patient numbers and longer follow-up, we were able to show that patients with small bowel lymphoma and particularly those with Stage II small bowel lymphoma are at a higher risk of failure. Unlike patients with resectable stomach lymphoma, those with small bowel involvement are more likely to have extensive nodal involvement in the mesentery and to have more extensive disease at presentation. Accurate information regarding intra-abdominal extent of disease is not easily available with the usual staging procedures and is dependent on the description of disease at the initial laparotomy. In recognition of the higher risk of relapse for patients with small bowel lymphoma, we now recommend a combined modality approach for those patients, especially those with Stage II disease, except for those with low grade histology. The data presented above suggest that low dose 20-25 Gy postoperative XRT for patients with resected tumors provides as good results as with a higher dose therapy. There is no indication that the whole abdominal XRT gives better results than the regional XRT for stomach lymphoma: however, with the number of patients at risk, a small difference in relapse risk can be missed. Accurate information regarding completeness of surgery, extent of intraabdominal involvement, and staging is necessary prior to recommending management with postoperative abdominal XRT alone. In cases where residual disease is suspected, management with systemic chemotherapy or a combined modality approach is recommended. The definition of the subgroup of patients favorable for treatment with postoperative XRT alone was confirmed by outcome of patients treated between 1979 and 1985. With larger numbers of patients available for analysis, we were able to recognize a higher risk of relapse for patients with Stage IIA small bowel lymphoma and to exclude them from the favorable category. Although results of treatment with surgery and che-
Localized gastrointestinal lymphoma 0 M. K.
motherapy of localized GIL are excellent, we believe that the approach of surgery and low dose postoperative abdominal XRT produces excellent results and is safer, especially in older patients, and is to be preferred. The next question to be asked is whether one can identify a group of patients who are curable with surgery alone. It is likely that a large proportion of patients currently treated with postoperative XRT are cured with surgery
GOSPODAROWICZ
et al.
1355
alone. The real benefit of postoperative XRT is not known for those patients with no residual disease postoperatively. Abdominal failures are difficult to diagnose early, and there is no evidence that salvage therapy of relapsed large cell lymphoma is successful. Because of the above, and the relative lack of toxicity of low dose abdominal XRT, investigators have been reluctant to withhold postoperative therapy.
REFERENCES 1.
Coltman, C. A.; Dahlberg, S.; Jones, S. E.; Miller, T. P.: Dana, B. W.; McKelvey. E. M.; Hartstock, R. J.: Dixon,
D. 0. CHOP is curative in thirty percent of patients with large cell lymphoma: a twelve-year southwest oncology group follow-up. In: Skain, A. T., ed. Advances in chemotherapy: update on treatment for diffuse large cell lymphoma. New York: John Wiley and Sons; 1986:71-77. 2. Connors. J. M.; Klimo, P.; Fairey, R. N. Brief chemotherapy and involved field radiation therapy for limited stage histologically aggressive histology lymphoma. Ann. Intern. Med. 23:107; 1987. 3. Dragosics, B.; Bauer, P.; Radaszkiewicz, T. Primary gastrointestinal non-Hodgkin’s lymphomas. A retrospective clinicopathologic study of I50 cases. Cancer 55: 1060; 1985. 4. Gehan. E. A. A generalised Wilcoxon test for comparing arbitrarily single-censored samples. Biometrika 52:203-223: 1965. 5. Gospodarowicz, M. K.; Bush, R. S.; Brown, T. C.; Chua. T. Curability of gastrointestinal lymphoma with combined surgery and radiation. Int. J. Radiat. Oncol. Biol. Phys. 9: 3: 1983. 6. Isaacson. P.: Wright, D. Extranodal malignant lymphoma arising from mucosa-associated lymphoid tissue. Cancer 53: 2515: 1984.
7
Maor, M. H.; Maddux, B.; Osborne, B. M.; Fuller, L. M.: Sullivan, J. A.; Nelson, R. S.; Martin, R. G.; Libshitz, H. I.; Velasquez. W. S.: Bennett, R. W. Stages IE and IIE nonHodgkin’s lymphomas of the stomach. Comparison of treatment modalities. Cancer 54:2330: 1984.
8 Pete, R.; Pike, M. C.; Armitage. P.; Breslow, N. E.: Cox. D. R.; Howard, S. V.; Mantel, N.; McPherson, K.: Pete, J.; Smith. P. G. Design and analysis of randomised clinical trials requiring prolonged observation of each patient. Br. J. Cancer 35:1-39: 1977. 9. Shepherd. F. A.: Evans, W. K.; Kutas, G.: Yau, J. C.; Dang, P.; Scott, J. G.; Farquharson, H. A.: Francombe, W. H.: Bailey, D.; Baker, M. A. Chemotherapy following surgery for Stages IE and IIE non-Hodgkin’s lymphoma of the gastrointestinal tract. J. Clin. Oncol. 1988:253. 10. Shimm. D. S.; Dosoretz, D. E.; Anderson, T.; Linggood. R. M.; Harris, N. L.: Wang, C. C. Primary gastric lymphoma. An analysis with emphasis on prognostic factors and radiation therapy. Cancer 52:2044; 1983. 1 I. Shiu, M. H.: Nisce, L. Z.; Pinna. A.; Strauss, D. J.; Tome. M.; Filippa. D. A.: Lee, B. J. Recent results of multimodal therapy of gastric lymphoma. Cancer 58: 1389: 1986.
