Malignant Lymphoma

1. The Histology and Staging of 473 Patients at the National Cancer Institute

7OM ANDERSON, MD,’ BRUCE A. CHABNER, MD,t ROBERT C. YOUNG, MD.t COSTAN W. BERARD, MD,* A. J. GARVIN, MD,§ RICHARD M. SIMON, PHD,t AND VINCENT T. DEVITA JR, MDt

A retrospective review was performed of 473 consecutive patients with malignant lymphoma referred to the National Cancer Institute. All patients had their clinical and pathologic material reviewed and where necessary reclassified. Using a modification of the Rappaport system, 180 (38.1%) patients had a nodular lymphoma, 293 (61.9%) had a diffuse lymphoma. Nodular lymphoma patients usually pre- sented with lymphadenopathy; diffuse lymphoma patients often presented with extranodal disease, particularly those patients with poorly differentiated lymphocytic, “histiocytic”, and Burkitt’s lym- phoma. Median age for all patients was 46 years, but was lower for diffuse poorly differentiated lymphocytic, diffuse undifferentiated, and Burkitt’s lymphoma patients. Women were more likely to have a nodular rather than diffuse lymphoma, P < 0.05. Analyzed by clinical staging, most nodular lymphoma patients had CS 111 (66.1%) disease, while most diffuse lymphoma patients had either C S 111 (31.1%) or CS IV (38.9%) disease. Analyzed by pathologic staging, most nodular lymphoma patients had PS 111 (34.8%) or PS IV (49.4%) disease, most diffuse lymphoma patients had PS 1V (56.3%) disease. Only 7% of all lymphoma patients had PS I disease, only 14.9% were PS 11. Systemic symptoms were more likely to be found in diffuse lymphoma and/or advanced stage patients.

Cancer 50:2699-2707, 1982.

HE DEVELOPMENT of more effective diagnostic, stag- T ing and therapeutic tools has produced profound changes in the management of patients with malignant lymphoma. With the introduction of the Rappaport histologic32 classification system, prognostic implica- tions of the various subgroups of disease began to be appreciated. The application of staging proce- d u r e ~ ~ , ~ ~ . ~ ~ - ~ ~ * ~ ~ developed in the study of Hodgkin’s dis- ease,34 helped to define different patterns of disease spread and to verify the impact of the Rappaport system. Finally, improvements in therapy established the fact that certain subgroups of patients who previously had a very poor prognosis could be effectively treated for potential 1.26.34 or at the very least disease-free reniissions lasting years could be achieved in some pa- tients. Other subgroups could be treated with a variety oft herapeutic regimens and experience long-term, albeit not disease-free, sur~ival; l .~*~-’ indeed, in selected pa- tients, it may be possible to manage certain patients --

Medical College of Wisconsin, Milwaukee, Wisconsin. f National Cancer Institute, Bethesda, Maryland. $. St. Jude’s Children’s Research Hospital, Memphis Tennessee. 5, Medical University of South Carolina, Charleston, South Carolina. Address for reprints: Bruce A. Chabner, MD, Medicine Branch,

NC‘I, Building 10, Room 12N226 9000 Rockville Pike, Bethesda, MD 20205.

Accepted for publication October 14. 198 1.

initially without cytotoxic therapy,31 in an approach now under active investigation.

The introduction of newer histologic classification ~ y ~ t e m ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ’ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ has extended the observations of R a p p a p ~ r t , ~ ~ Jones et Chabner et ~ f . , ~ . ” Schein et ~ z f . , ~ ~ and others2-’1.26 by demonstrating that the ap- plication of cell-surface markers and knowledge gained by a variety of in vitro techniques can identify further subgroups of patients based upon the developmental origin of the malignant clones of cells, the biology of the disease processes, and the response to various therapeu- tic regimen^.^.^.^^ However, the introduction of newer and more complicated technologies and terminologies has contributed at least temporarily to a feeling of con- fusion regarding the validity of previously reported data.4*29

In order to assess the impact of staging and thera- peutic advances, as well as to uniformly establish the histologic classification of a large population of patients with malignant lymphomas studied at a single institu- tion, a retrospective review was undertaken of all pa- tients with malignant lymphoma referred to the Na- tional Cancer Institute between July 1953 and May 1975. As part of this review, a reclassification of all avail- able biopsy material was performed using the Rappaport classification system,32 as well as the systems proposed

0008-543)</82/12 1512699 $ I .25 0 American Cancer Society

2699

27 00 CANCER December 15 1982 Vol. 50

TABLE I . Histologic Classification: Rappaport System

Proportion of Proportion of Histologic Admission prior Admission after all nodular all diffuse Proportion of all

type to 1968* 1968 patients patients patients

NPDL 28 (16.6%) 60 (19.7%) 88 (48.9%) - 18.6% NML 19(11.2%) 57 (18.8%) 76 (42.2%) - 16.1% NHL 2 (1.2%) 14 (4.6%) 16 (8.9%) - 3.490

100%

DWDL 15 (8.9%) 24 (7.9%) - 39 ( I 3.3%) 8.2% DPDL 29 (17.2%) 35 ( I 1.5%) - 64 (21.8%) 13.5% DML I 1 (6.5%) 22 (7.2%) - 33 ( I 1.3%) 7.0% DHL 34 (20. I % ) 57 ( I 8.8%) - 91 (31.170) 19.2% DUL 12 (7.1%) 15 (4.9%) - 27 (9.2%) 5.7% Burkitt’s 19 ( I 1.2%) 20 (6.6%) - 39 ( I 3.370) 8.2%

100%

All nodular 49 (29%) 131 (43%) - - 180 (38.1%) All diffuse 120 (71%) 173 (57%) - - 293 (6 1.9%)

10&% 100% 100%

* Introduction of staging procedures, combination chemotherapy and aggressive radiotherapy in 1968 provides a convenient point of reference.

NPDL: nodular poorly differentiated lymphoma; NML: nodular

by Dorfman,13 Bennett, et Lennert, et a/.,25 and Lukes and Collins.27 This article summarizes the rela- tionship between the histologic subtypes of the Rappa- port classification and the clinical characteristics of dis- ease in this large retrospective series. Other reports have addressed the value of the newer pathologic classifica- tion systems,I6 the prognostic significance of nodular histiocytic lymphoma,” of divergent histologies at di- agnosis,I5 or of histologic “progression” from “favor- able” to “unfavorable” forms of disease.22

Materials and Methods

Between July 1, 1953 and May 15, 1975, 521 con- secutive patients with malignant lymphoma were eval- uated and treated at the National Cancer Institute. The clinical records and histologic sections of all patients were reviewed and each patient’s initial diagnostic bi- opsy and subsequent pretreatment staging material was classified according to a modified Rappaport system. Thirty-seven patients were excluded from analysis be- cause referral to the National Cancer Institute had been delayed until greater than 12 months from the time of diagnosis. An additional 1 1 patients were excluded on histologic grounds due to unclassifiable or inadequate tissue, leaving 473 patients with adequate histologic and clinical data for analysis.

Pretreatment clinical evaluation of patients with ma- lignant lymphomas underwent considerable evolution during the period of time covered by this study. Early patients had only careful clinical evaluations, routine laboratory studies, e.g., liver function tests, chest x-rays,

mixed lymphoma; N H L nodular histiocytic lymphoma; DWDL: dif- fuse well differentiated lymphoma; DPDL: diffuse poorly differentiated lymphoma; DML: diffuse mixed lymphoma; DHL: diffuse histiocytic lymphoma; DUL: diffuse undifferentiated lymphoma. . 0 .

intravenous pyelograms, etc., and selective pathologic sampling techniques, e.g., bone marrow aspirates and percutaneous liver biopsy; by the late 1960s, coincident with the introduction of aggressive systemic combina- tion che‘motherapy and/or extended-field radiotherapy, patients were subjected to radionuclide studies, e.g., liver-spleen, bone, and gallium scans, lymphangiogra- phy and systematic pathologic staging techniques, in- cluding bilateral iliac bone marrow aspirates and biop- sies, peritoneoscopy with multiple directed liver biop- sies, and laparotomy in selected patients. The impact of such a systematic staging approach has been presented by Chabner, et

The Wilcoxon rank-sum test was used to test whether the presence of B-symptoms was associated with stage.24 All other significance levels are derived from contin- gency chi-square tests. All significance levels are of the two-sided type.

Results

The male/female ratio was 1.6. Median age was 46. Four-hundred-thirty-eight patients (92.8%) were white and 30 (6.4%) were black. The remaining 53 had other racial backgrounds (0.8%). One hundred seventy-nine patients (37.8%) had documented B-symptoms, i.e., fe- ver, night sweats, and/or 210% loss of body weight.

Table 1 summarizes the breakdown of the Rappaport histologic subgroups. Two hundred-ninety-three pa- tients (6 ! .9%) were classified as having diffuse lympho- mas. The largest subgroup was composed of 9 1 patients with diffuse histiocytic lymphomas, representing 3 1 .1 %

No. 12 MALIGNANT LYMPHOMA I * Anderson et al. 270 1

TABLE 2. Site of Diagnostic Biopsy

Histologic type -

Site NPDL NML NHL DWDL DPDL DML DHL DUL Burkitt’s Total

Lymph node 80 64 16 31 36 22 41 15 1 1 316 Skin, SC tissue 2 3 0 1 6 2 7 I 1 23 Small bowel.

appendix 1 0 0 1 1 0 8 1 6 18 GI tract-Other 0 0 0 0 I I 1 2 2 7 Spleen 0 0 0 1 0 0 3 0 0 4 Liver 0 I 0 I 0 I 0 0 0 3 Gonad 0 0 0 0 0 0 4 0 4 8 GU Tract 0 0 0 0 0 0 2 0 0 2 Lung. pleura 0 0 0 0 0 I 2 0 0 3 Bone 0 0 0 0 5 1 3 1 1 I I Bone marrow I 0 0 0 3 0 1 0 0 5 Tonsil 0 0 0 0 0 I 3 3 0 7 Salivary gland 1 0 0 0 0 0 0 0 0 1

other 0 0 0 0 0 0 2 0 1 3 Eye 0 I 0 0 0 0 0 0 0 1 Breast 0 I 0 0 I 0 0 I 0 3 Unknown 3 6 0 4 1 1 4 14 3 13 58

Total 88 76 16 39 64 33 91 27 39 473

-

Oronasopharynx-

of all patients with diffuse histologic types and 19.2% of all cases in the entire study. The next largest groups included nodular poorly differentiated lymphocytic lym- phoma (88/180 patients, 48.9% of the nodular lympho- mas, and 18.6% of the total), nodular mixed lymphoma (76 patients, 42.2% of nodular lymphomas, and 16.1% of all patients), and diffuse poorly differentiated lym- phocytic lymphoma (64 patients, 2 1.8% of diffuse lym- phomas, and 13.5% of the total).

As part of the review of the biopsy material, patients with nodular lymphomas had their biopsy specimens characterized in terms of the prominence of the nodu- larity, and the proportion of the tissue biopsy which displayed a nodular pattern of growth. One-plus prom- inence of nodularity was present in 83 cases, 2+ in 77 cases, 3+ in 9 cases, 4+ in 5 cases. In six cases the prominence could not be characterized. Ninety-one of the 170 cases had 2+ to 4+ nodularity, these being the cases that are most easily recognized as being nodular lymphomas. When independently assessed by the pro- portion of the node that had a nodular pattern, 12 cases had 25% of their node involved with a nodular growth pattern, 41% of the cases had 50% involvement, 18 of the cases had 75% involvement, and 103 of the cases had the entire lymph node involved with a nodular growth pattern. Again, six cases could not be well char- acterized. The prognosis of these lymphomas with vari- able degrees and/or proportions of nodularity is dis- cussed in Part I1 of these manuscripts.

The sites of the diagnostic biopsies are shown in Table 2. The vast majority of patients (316, 66.8%) had their diagnosis made based upon results of a lymph node bi-

opsy. Twenty-three patients had diagnostic biopsies of skin or subcutaneous tissue, 18 had small intestinal and/ or appendiceal biopsies, and 1 1 had biopsies of bone. Age distribution by histologic subgroups is shown in Table 3. Overall median age was 46 years (range; 1-8 1). Median age for patients with Burkitt’s lymphoma was eight years. Patients with diffuse undifferentiated lym- phoma and diffuse poorly differentiated lymphocytic lymphoma had a median age of 34 years, and the fre- quency within age groups suggests a bimodal distribu- tion. Patients with other histologic types had a peak in- cidence in the 40-70-years age range. Also summarized in Table 3 are the sex ratios of patients in the various subgroups. While the overall male/female ratio was 1.6, the ratio was lower in the nodular forms of lymphoma (1.2), and was higher in the diffuse forms of disease (1.84) (P < 0.05).

Table 4 summarizes therapy received by patients prior to referral to the National Cancer Institute. Reflecting referral policies and patterns, 388 patients (82.1%) had received no therapy other than a diagnostic biopsy prior to referral; 22 (4.7%) had undergone surgical resection; 16 (3.4%) had received radiotherapy; nine (1.9%) had been on corticosteroid therapy; 12 (2.5%) had been given chemotherapy; and 19 (4.0%) had received some com- bination of therapies.

Tables 5 and 6 summarize the clinical and pathologic stages, respectively, of the 473 evaluable patients by his- tologic subtypes. The clinical stages were based upon the initial diagnostic biopsy plus the extensive but nonin- vasive evaluation techniques utilized during each era as outlined in the previous section. On the basis of these

2702 CANCER December 15 1982 VOl. 50

TABLE 3. Age and Sex Distribution by Histologic Subgroups

Age

Histologic type <I0 10-19 20-29 30-39 40-49 50-59 60-69 >70 Median Sex ratio (M/F)

Nodular NPDL 0 0 8 9 38 16 16 1 47 45/43 (1.05) NML 0 0 3 15 25 23 7 3 47 45/31 (1.45) NHL 0 0 I I 6 5 3 0 50 9/7 (1.29)

99/81 (1.22)

DWDL 0 0 0 4 9 11 13 2 52 3019 (3.33) DPDL 12 9 8 7 2 12 9 0 34 39/25 (1.56)

13 7 I 52 15/18 (0.83) DML 1 2 1 1 7 DHL 1 5 12 9 20 19 19 6 49 60/31 (1.94) DUL 3 5 4 3 3 2 4 3 34 18/9 (2.0)

1 0 0 0 0 8 28/11 (2.54) Burkitt’s 19 14 5

190/103 (1.84)

Diffuse

data, there were 46 patients (9.7%) with Clinical Stage (CS) I disease, 78 (16.5%) with CS I1 disease, 210 (44.4%) with CS 111 disease, and 139 (29.4%) with CS IV disease. A large proportion of patients with diffuse forms of malignant lymphomas presented with symptomatic and/or clinically identifiable extranodal disease (Table 2); 38.9% of such patients were eventually shown to have Stage IV disease. What often appeared to be a localized extranodal primary was usually found to be a manifes- tation of widespread disease. Such extranodal lesions are a much less common presentation in patients with the nodular lymphomas; only 25 patients ( I 3.9%) had clin- ically identifiable Stage IV disease prior to staging pro- cedures. Each of the nodular forms of lymphoma differs in its clinical presentation from its diffuse counterpart in the Rappaport classification, as the nodular subgroups tend to be predominantly CS 111 whereas the diffuse lymphomas tend to be more evenly distributed through- out the various stages, NPDL versus DPDL, P < 0.005, NML versus DML, P < 0.05, NHL versus DHL, P < 0.025. The most common clinical stage of diffuse lym- phoma patients was usually CS IV, although patients with DWDL resemble nodular lymphoma patients in

that they are usually found to have CS 111 disease (23/ 39, 59.0%); the tendency of DWDL to present as CS 111 disease is significantly different from other diffuse forms of lymphoma (P < 0.025).

Table 6 summarizes the pathologic stage (PS) of the patients after completion of all staging procedures. Only 33 patients (7.0%) were found to have PS I disease. None of the patients with DWDL and only three patients (3.4%) with NPDL were PS I; other subgroups were somewhat more likely to have PS I disease, e.g., DML 12.1%, DUL 11.5%, DPDL 9.4%, and NML 8.1%. In contrast to the frequent localized presentation of Burk- itt’s lymphoma in Africa, Stage I disease was identified in only two American patients (5.3%) with Burkitt’slym- phoma.

Seventy patients (14.9%) were classified as PS I1 dis- ease. As in PS I patients, the diffuse forms of lymphoma were more likely lo be limited to regional lymph nodes than the nodular forms of lymphoma, with the notable exception of DWDL (two patients, 5.1 %). Only five pa- tients (5.7%) with NPDL were PS 11; other cell types with a low incidence of PS I1 disease included NHL, one patient (6.3%), NML, 1 1 patients (14.9%), and DPDL,

TABLE 4. Summary of Therapy Prior to Referral to the National Cancer Institute

Therapy Admission prior Admission after

to 1968t 1968 Total (%)

None* Surgical resection Radiotherapy Steroids Single agent chemotherapy Multiple agent chemotherapy Combination of above Unknown

129 (76.8%) 8 (4.8%)

4 (2.4%) 1(4.2%) I (0.6%)

10 (6.0%) 1 (0.6%)

9 (5.4%)

259 (86.9%) 14 (4.7%) 7 (2.3%) 5 (1.7%) 3 ( I .O%) I (0.3%) 9 (3.0%) 6 (2.0%)

388 (82.1%) 22 (4.7%) 16 (3.4%) 9 (1.9%)

10 (2.1%) 2 (0.4%)

19 (4.0%) 7 (1.5%)

-~ ~~

* Includes patients with excisional biopsies. t Introduction of staging procedures, combination chemotherapy,

and aggressive radiotherapy in 1968 provides a convenient point of reference.

No. 12 MALIGNANT LYMPHOMA I * Anderson et al. 2703

TABLE 5. Relationship Between Histologic Subtype and Clinical Stage

Clinical Stage

Histologic type I 11 Ill IV

Nodular NPDL 4 (4.5%) 8 (9.1%) 61 (69.3%) 15 (17.0%) NML 9 (11.8%) 1 I (14.5%) 48 (63.2%) 8 (10.5%) NHL 2 (12.5%) 2 (12.5%) 10 (62.5%) 2 (12.5%)

All nodular patients 15 (8.3%) 21 ( 1 1.7%) I 19 (66. I % ) 25 (13.9%)

0 2 (5.1%) 23 (59.0%) 14 (35.9%) DWDL DPDL 6 (9.4%) 10 (15.6%) 22 (34.4%) 26 (40. I % ) DML 6 ( 1 8.2%) 8 (24.2%) 1 1 (33.3%) 8 (24.2%) DHL 10 (10.9%) 18 (19.8%) 23 (25.3%) 40 (44.0%) DUL 5 (18.5%) 8 (29.6%) 6 (22.2%) 8 (29.6%) Burkitt’s 4 (10.3%) 11 (28.2%) 6 (15.4%) 18 (46.2%)

All diffuse patients 31 (10.6%) 57 (19.5%) 91 (31.1%) 114 (38.9%)

Total 46 (9.7%) 78 (16.5%) 2 10 (44.4%) 139 (29.4%)

Diffuse

* Percentages indicate proportion of patients within a specific Rappaport histologic subtype.

eight patients, (12.5%). In contrast, other subgroups of diffuse lymphoma were not infrequently PS 11, e.g., DML 7 patients (21.1%), DHL 20 patients (22.0%), DUL eight patients (30.8%), and Burkitt’s lymphoma, eight patients (21.1%). Overall only 9.6% of all patients with nodular lymphomas had PS I1 disease compared to 18.2% of the patients with diffuse lymphomas.

A total of 114 patients (24.3%) were found to have PS 111 disease. In contrast to the high percentage of dif- fuse lymphoma patients presenting with localized PS I- I1 disease, the nodular lymphoma patients often pre- sented as widespread disease albeit limited to lymphatic sites. The most common histologic subgroups with PS 111 were NML (39.2%) and NPDL (33.0%); somewhat less frequent were patients with NHL (25.0%), DWDL (23.1%), DUL (23.1%) and DML (2 1.2%). Burkitt’s lym-

phoma was uncommonly found to be Stage 111 (three patients, 7.9%). Among adults the most common forms of diffuse lymphoma were the subtypes least likely to present with Stage 111 disease-DPDL, 11/64 patients (17.2%) and DHL, 16/91 patients (17.6%). These his- tologic subtypes tended to have PS IV disease. Overall, 34.8% of all nodular lymphomas were PS 111, compared to only 17.9% of diffuse lymphomas.

Careful staging procedures demonstrated the frequent presence of extralymphatic involvement (PS IV) in all patients with malignant lymphomas, regardless of his- tologic subgroup. The prevalence of PS IV disease ranged from 34.6% in DUL and 37.8% in NML to 60.9% in DPDL and 7 1.8% in DWDL. In DHL, the most com- mon subgroup of diffuse lymphoma, 52.7% were PS IV, as were 58.0% of NPDL, the most common form of

TABLE 6. Relationship Between Histologic Subtype and Pathologic Stage

Pathologic Stage

Histologic type I I1 Ill IV

Nodular 5 1 (58.0%) NPDL 3 (3.4%) 5 (5.7%) 29 (33.0%)

NML 6 (8. I % ) 1 1 (14.9%) 29 (39.2%) 28 (37.8%) NHL 2 (12.5%) 1 (6.3%) 4 (25.0%) 9 (56.3%)

All nodular patients I 1 (6.2%) 17 (9.6%) 62 (34.8%) 88 (49.4%)

Diffuse DWDL 0 2 (5.1%) 9 (23.1%) 28 (71.8%) DPDL 6 (9.4%) 8 (12.5%) 11 (17.2%) 39 (60.9%) DML 4 (12.1%) 7 (21.2%) 7 (21.2%) 15 (45.5%) DHL 7 (7.7%) 20 (22.0%) 16 (1 7.6%) 48 (52.7%) DUL 3 (11.5%) 8 (30.8%) 6 (23.1%) 9 (34.6%) Burkitt’s 2 (5.3%) 8 (21.1%) 3 (7.9%) 25 (65.8%)

All diffuse patients 22 (7.6%) 53 (18.2%) 52 ( 1 7.9%) 164 (56.3%)

Total 33 (7.0%) 70 (14.9%) 114 (24.3%) 252 (53.7%)

* Percentages indicate proportion of patients within a specific Rappaport histologic subtype.

2704 CANCER December 15 1982 Vol. 50

TABLE 7. Comparison of Pathologic Stage Demonstrating Impact of Modern Staging Procedures

< I968 >I967

Nodular PS I 3 (6.4%) 8 (6.1%)

I11 22 (46.8%) 40 (30.5%) 11 7 (14.9%) 10 (7.6%)

IV 15 (3 1.9%) 73 (55.7%)

Total 47 131

Diffuse PS I 9 (7.6%) 13 (7.5%) I1 26 (22.0%) 27 (15.6%) 111 27 (22.9%) 25 (14.5%) IV 56 (47.5%) 108 (62.4%)

Total 1 I8 173

nodular lymphoma. Again in contrast to the African experience, 65.8% of patients with Burkitt's lymphoma were PS IV. Of the entire 473 patients studied, only in the NML subgroup was another stage more common than PS IV; 29 patients (39.2%) were PS I11 and 28 patients (37.8%) were PS IV. In only three subgroups was PS IV disease less than 50%; NML (37.8%), DUL (34.6%), and DML (45.5%); 53.7% of all patients were PS IV (nodular lymphoma patients 49.4%, diffuse lym- phoma patients 56.3%). As expected, categorization by clinical or pathologic staging demonstrated differences. Forty-six patients had CS I disease but only 33 had PS I disease; 78 patients were classified as CS 11, but only 70 had PS I1 disease. The most dramatic differences were noted in Stages I11 and IV, where discrepancies are most likely to have a major impact upon therapeutic deci- sions. Two hundred-ten patients (44.4%) were classified as having CS 111 disease, the most common clinical stage. However, only 114 patients (24.3%) were PS 111 after completion of all appropriate biopsies. In contrast, the prevalence of PS Stage IV disease was found to be 53.7%, not the 29.1% indicated by clinical evaluation alone. The largest changes in Stage IV disease were in the nod- ular lymphomas, where only 13.9% were CS IV com- pared to 49.4% PS IV. While the frequency of PS IV disease was even higher in the diffuse lymphomas (56.3%), 114 patients (38.9%) had been identified as Stage IV during clinical evaluations. As previously re- ported, osteolytic bone lesions, CNS involvement, and gastrointestinal tumors are more common in the diffise than in the nodular l y m p h o m a ~ ; ~ , ' ~ , ' ~ * ~ ~ . ~ ~ these sites of extralymphatic involvement are also usually easily rec- ognizable clinically.

Table 7 summarizes the consequences of modern staging techniques, including bipedal lymphangiogra- phy, bone marrow aspirates and biopsies, percutaneous liver biopsies and multiple liver biopsies under direct

vision at peritoneoscopy. Between 197 1 and 1974 PS I- 111 patients also underwent a staging laparotomy; the yield from this procedure after careful evaluation by the techniques noted above has been reviewed by Chabner et al." Prior to 1968, only 26% of patients with malig- nant lymphomas had bipedal lymphangiograms. Sub- sequent to 1968, 66% of all patients had this staging procedure; 34% did not undergo pretreatment lym- phangiography because of contraindications such as chronic obstructive pulmonary disease, massive retro- peritoneal involvement, poor overall medical status, etc. or documented Stage IV disease. Between 1968 and May 1975, 133 patients (66.5%) had a positive lymphangio- gram.

It is of interest that the introduction of modem staging techniques did not change the proportion of patients in PS I disease. However, substantial numbers of patients with CS I1 or CS 111 disease were upstaged by these pro- cedures to PS 111 or PS IV disease. These results are similar to those reported in other s e r i e ~ . ' ~ . ~ ~

Table 8 summarizes the relationships between his- tologic subgroups, pathologic stages of disease, and the presence or absence of B symptoms. A total of 180 pa- tients (38.9%) had documented B symptoms. B symp- toms were more commonly present in patients with dif- fuse malignant lymphomas. Only 43 (24.6%) of 175 patients with nodular lymphomas had B symptoms, compared to 137 (47.4%) of 288 patients with diffuse lymphomas (P < 0.001). The presence or absence of B symptoms appeared to correlate with pathological stage for diffuse lymphoma patients in that only 22.2% of Stage I patients had B symptoms, compared to 45.3% in Stage 11, 45.1% in Stage 111, and 52.5% in Stage IV (P < 0.05).

Discussion

This retrospective analysis serves as a large-scale, long- term data base which confirms many of the previous observations made on smaller groups of patients. While newer histologic classifications have been intro- duced,3,4, 13,16.18.25,27.28.35 the Rappaport scheme32 can be profitably applied to patients with malignant lympho- mas. It can discriminate between important subgroups of patients, and as we will demonstrate in the following article, it has prognostic significance. Of the 473 eva- luable patients, 293 (6 1.9%) had diffuse and 180 (38.1 %) had nodular lymphomas. During the period covered by this analysis, the nodu1ar:diffuse ratio of patients referred to this institution has shifted. Prior to 1968 only 29% of patients had nodular lymphomas; subsequent to 1968, nodular lymphomas were present in 43% of the patients.

The nodular lymphomas have been predominantly poorly differentiated lymphocytic (88 patients, 48.9%)

No. 12 MALIGNANT LYMPHOMA I - Anderson el al. 2705

TABLE 8. B-Symptoms, Histologic Type, Pathologic Stage*

Stage ~~

I I1 111 IV Total

Histologic type NPDL NML N H L DWDL DPDL DML DHL DUL Burkitt’s All nodular All diffuse

Total

I 13 016 112 010 016

217 (29%) 013 11.2 2/11 (18.2%) 5/22 (22.2%)

7/33 (2 I .2%)

214

015 2/10 111 112 2.8 317 (43%) 9/20 (45%) 418 (50%) 518 3/16 (18.8%)

24/53 (45.3%)

27/69 (39.1%)

4/28 (14%) 9/29 (31%) 013 119 (11%) 5/10 (50%) 217 (29%)

12/16 (75%) 316 (50%)

13/60 (2 1.7%) 2315 I (45. I%)

013

3611 1 I (32.4%)

16/51 (31%) 7/28 (25%) 219 (22%)

12/28 (43%) 18/39 (46%) 7/15 (47%)

30147 (64%) 619 (67%)

12/24 (50%) 25/88 (28.4%) 851162 (52.5%)

109/249 (43.8%)

* Results expressed as number patients with B-symptoms/number of patients whose charts could be retrospectively analyzed for the pres-

and mixed lymphocytic-histiocytic (76 patients 42.4%) subtypes. Only 16 patients (8.9%) had nodular histio- cytic lymphoma. These patients had a poorer prognosis compared to those with other nodular lymphomas. Among the diffuse lymphomas, the histiocytic (91 pa- tients, 3 l . 1%) and poorly differentiated lymphocytic subgroups (64 patients, 21.8%) were most common. DWDL patients had a relatively favorable prognosis but made up only 13.3% of diffuse lymphomas and only 8.2% of the total population.

As shown in Table 2, 316 patients (66.8%) had their diagnosis established by lymph node biopsy. However, a large variety of other sites occasionally provided orig- inal pathologic specimens. With rare exceptions the ex- tranodal presentations occurred in patients with diffuse lymphomas, although nodular forms of lymphomas were found in 5/23 patients (2 1.7%) with positive biop- sies of skin or subcutaneous tissue. The frequency of diagnostic extranodal presentations of disease in the diffuse lymphomas is one explanation for their high per- centage of CS IV and PS IV disease (Table 5 ) after care- ful staging evaluation.

The median age of the population was 46, markedly different from the median age of 32 observed in our experience with Hodgkin’s disease.I2 An older popula- tion may be one of the most important factors in the increased incidence of significant morbidity observed in reports of staging laparotomies in such patient popula- tion. Were it not for an arbitrary age limit imposed for study purposes (originally 65, more recently 70), this median age might well have been even higher. The lower ages of the patients with Burkitt’s lymphoma is not sur- prising, although the median age is not much different from that observed in African series, 6/39 patients (1 5.4%) were 220 years old.* Notably, the median ages

21/87 (24.1%) 18/73 (24.6%) 4/15 (26.7%)

14/39 (35.9%) 25/63 (39.7%) 14/33 (42.4%) 53/90 (58.9%) 13/26 (50%) 18/37 (48.6%) 431 I75 (24.6%)

1371288 (47.6%)

1801463 (38.9%)

ence or absence of B-symptoms. In ten patients the records failed to mention the presence or absence of B-symptoms.

of 34 for patients with diffuse poorly differentiated lym- phocytic and undifferentiated lymphomas were signifi- cantly less than those of the other subgroups.

Because of specific admission requirements to the National Cancer Institute, most patients (82.1%) had not been treated prior to amval. This fact is of potential importance since intervening therapy could possibly al- ter the “natural history” of disease and influence sub- sequent evaluation. Most other series include a prepon- derance of previously treated patients. As reported by Jones el dZ2 intervening therapy is associated with sig- nificant changes in histologic subtypes when biopsy tis- sue is only avaiiable at the time of relapse.23 Tables 5 and 6 demonstrate that, after careful assessment, most patients with nodular lymphomas have widespread dis- ease, although apparently confined clinically to the lym- phatic system (66.1% CS 111). In contrast the most com- mon stage of diffuse lymphomas, even without system- atic sampling procedures, i.e., before 1968, was Stage IV (38.9%). Modem systematic staging evaluations have demonstrated that 53.7% of all lymphoma patients can be documented to have PS IV disease; in every subgroup but nodular mixed lymphoma, PS IV was the most com- mon single stage of disease. Next most common was PS 111 (24.4%); only 22.0% ofpatients were ultimately found to have Stage I or I1 disease. Interestingly, the patients in “unfavorable” subgroups, i. e., diffuse histologies, made up the majority of these patients, an observation consistent with other report^.'^,^^ Many of the patients with nodular lymphomas had widespread disease limited to lymphatic tissues (34.8% PS 111) in contrast to diffuse lymphomas (17.9% PS 111).

Because of the implications of accurate clinical and pathologic staging, it is important to emphasize the stag- ing procedures utilized. Systematic combination che-

2706 CANCER December 15 1982 Vol. 50

motherapy and aggressive radiotherapy were introduced at the NCI in 1968, at approximately the same time that ly mphangiography and staging biopsies became routine; this date has therefore been utilized in assessing the im- pact of staging and treatment procedures. Lymphangi- ography and staging biopsies were uncommon prior to 1968, although bone marrow aspirations were more fre- quently performed. Seventy-four percent of the patients had at least one pretreatment bone marrow aspirate prior to 1968; subsequent to the recognition that in- volvement of the bone marrow was common in certain subgroups of malignant lymphoma, this procedure was even more widely utilized. Since 1968, only nine pa- tients (3%) did not have a pretreatment bone marrow aspirate. Positive aspirates were obtained in 28.8% of patients undergoing bone marrow aspirations over the entire period covered by this review. The recognition that lymphomatous involvement of the bone marrow was often focal led to the performance of bone marrow biopsies as well as aspirates. Only 29% of patients had marrow biopsy prior to 1968. After that time 92% of patients had at least one pretreatment biopsy, and more recently bilateral iliac crest biopsies have been per- formed. Together with the aspiration technique, bone marrow biopsy has been found to be the most frequently positive tissue-sampling procedure in the entire staging evaluation. Since 1968, 100 patients (36.1%) were found to have at least one positive bone marrow biopsy.

Recognition that hepatic involvement is not uncom- mon in asymptomatic previously untreated patients, even with normal liver function tests, also stimulated more active sampling of the liver. The utilization of percutaneous liver biopsy, and peritoneoscopy with multiple liver biopsies allowed reasonably accurate as- sessment of hepatic involvement. With rare exception, neither procedure had been utilized in the patients ad- mitted prior to 1968 (Table 6). In the patients admitted after 1968, percutaneous liver biopsies were positive in 25.3% of patients undergoing biopsy. Peritoneoscopy was utilized in patients who had had a negative or un- successful percutaneous liver biopsy, or who had a con- traindication to percutaneous liver biopsy. Positive bi- opsy specimens utilizing this technique were obtained in 25.8% of patients undergoing peritoneoscopy. The percentage yield of all these sampling techniques has been previously analyzed by histologic subgroups; re- sults of this analysis including the utilization of staging laparotomies have been reported by Chabner et a1.9*10

These procedures have made important contributions to the current understanding of the biology of malignant lymphomas. They can also be readily repeated to eval- uate the efficacy of therapy and determine the need for additional therapy.

Finally, Table 7 details the relationships between his- tologic subgroups, pathologic stages of disease, and the presence or absence of B symptoms. In our series of patients, “B’ symptomatology was more commonly seen in the diffuse lymphomas (P < 0.001), which are generally considered to have an unfavorable progn~sis.’~ It is notable that women had lymphomas of the more favorable nodular subgroups more frequently than men ( P < 0.05) and therefore would be expected to have bet- ter survival characteristics. This is of interest because of the report by Cabanillas et al.,’ which indicated the sex of the patient was one of the most important prognostic determinants identified and was believed to be even more important than histologic subgroup or pathologic stage.

REFERENCES

1. Anderson T, Bender RA, Fisher RI et al. Combination che- motherapy in non-Hodgkin’s lymphoma: Results of long-term follow- up. Cancer Treat Rep 1977; 61:1057-1066.

2. Bagley CM, DeVita VT, Berard CW, Canellos GP. Advanced lymphosarcoma: Intensive cyclical combination chemotherapy with cyclophosphamide, vincristine, and prednisone. Ann Intern Med 1972;

3. Bennett MH, Farrer-Brown G. Henry K, Jellife AM. Classifi- cation of non-Hodgkin’s lymphomas. Lancet 1974; 2:405-406.

4. Berard CW. Reticuloendothelial system: An overview of neo- plasia. In: Rebreck JW, Berard CW, Abell MR, eds. International Academy of Pathology Monograph No. 16: The Reticuloendothelial System. Baltimore: Williams and Wilkins, 1975; 301-3 17.

5. Bloomfield CD, Goldman A, Dick F, Brunning RD, Kennedy BJ. Multivariate analysis of prognostic factors in the non-Hodgkin’s malignant lymphomas. Cancer 1974; 33:870-879.

6. Bloomfield CD, Kersey JH, Brunning RD, Gajl-Peczalska KJ. Prognostic significance of lymphocytic surface markers and histology in adult non-Hodgkin’s lymphoma. Cancer Treat Rep 1977; 61:963- 970.

7. Cabanillas F, Burke JS, Smith TL, Moon TE, Butler JJ, Rod- riguez V. Factors predicting for response and survival in adults with advanced non-Hodgkin’s lymphoma. Arch Intern Med 1978; 138:4 13- 418.

8. Carbone PP, Berard CW, Bennett JM, Ziegler JL, Cohen MH, Gerber P. Burkitt’s tumor. Ann Intern Med 1969; 70:8 17-832.

9. Chabner BA, Johnson RE, DeVita VT et a/. Sequential staging in non-Hodgkin’s lymphomas. Cancer Treat Rep 1977; 6 1:993-998.

10. Chabner BA, Johnson RE, Young RC et al. Sequential non- surgical and surgical staging of non-Hodgkin’s lymphoma. Ann Intern Med 1976; 85:149-154.

11. Coltman CA, Luce JK, McKelvey EM, Jones SE, Moon TE. Chemotherapy of non-Hodgkin’s lymphoma: Ten years’ experience in the Southwest Oncology Group. Cancer Treat Rep 1977; 61: 1067- 1078.

12. DeVita VT, Simon RM, Hubbard SM ef a/. Curability of Hodg- kin’s disease with chemotherapy. Ann Intern Med 1980; 92:587-595.

13. Dorfman, RF. Classification of non-Hodglun’s lymphomas. Lancet 1974; 1:1295-1296.

14. Fisher RI, DeVita VT, Johnson BL, Simon R, Young RC. Prog- nostic factors for advanced diffuse histiocytic lymphoma following treatment with combination chemotherapy. Am J Med 1977; 63: 177- 182.

15. Fisher RI, Jones RB, DeVita VT e? al. Natural history of ma- lignant lymphomas with divergent histologies of staging evaluation. Cuncer 1981; 47:2022-2025.

16. Gamin AJ, Simon RM. Young RC, DeVita VT, Berard CW. The Rappaport classification of non-Hodgkin’s lymphomas: A closer

76:227-234.

No. 12 MALIGNANT LYMPHOMA I - Anderson el al. 2707

look using other proposed classifications. Semin Oncol 1980; 7:234- 243.

17. Gehan EA. A generalized Wilcoxon test for comparing arbi- trarily singly-censored samples. Biometrika 1965; 52:203-233.

18. Gerard-Marchant R, Hamlin I, Lennart K, Rilke F, Stansfeld AG, VanUnnik JAM. Classification of non-Hodglun’s lymphomas. Lancet 1974; 2:406-408.

19. Goffinet DR, Castellino RA, Kim H et al. Staging laparotomies in unselected previously untreated patients with non-Hodgkin’s lym- phomas. Cancer 1973; 32:672-68 1.

20. Goffinet DR, Warnke R, Dunnick NR et al. Clinical and sur- gical (laparotomy) evaluation of patients with non-Hodgkm’s lym- phomas. Cancer Treat Rep 1977; 61:981-992.

2 I . Jaffe ES, Braylan RC, Nanba K, Frank MM, Berard CW. Func- tional markers: A new perspective on malignant lymphomas. Cancer Treaf Rep 1977; 61:953-962.

22. Jones RB, Hubbard S , Osborne C et a/ . Histologic conversion in non-Hodghn’s lymphoma: Evolution of nodular lymphomas to diffuse lymphomas. J Clin Inwst 1978; 6 I :437A.

23. Jones SE, Fuks Z , Bull M et al. Non-Hodgkin’s lymphomas: IV. Clinicopathologic correlations in 405 cases. Cancer 1973; 3 1:806- 823.

24. Lehman EL. Nonparametrics. San Francisco: Holden-Day Inc, 1975.

25. Lennert K, Mohri N, Stein H, Kaiserling E. The histopathology of malignant lymphoma. Br J Haematol(Supp1ement) 1975; 3 I: 193- 203.

26. Lister TA, Cullen MH, Beard MEJ ef al. Comparison of com- bined and single-agent chemotherapy in non-Hodhn’s lymphoma of favorable histologic type. Br Med J 1978; 1533-537.

27. Lukes RJ, Collins RD. Immunologic characterization of human malignant lymphomas. Cancer 1974; 34: 1488-1 503.

28. Math6 G, Rappaport H, OConor GT, Torloni H. Histological and cytological typing of neoplastic diseases of hematopoietic and lym-

phoid tissues. In: WHO International Histological Classification of Tumors. No. 14. Geneva: World Health Organization, 1976.

29. Nathwani BN. A critical analysis of the classification of non- Hodgkin’s lymphomas. Cancer 1979; 44:347-384.

30. Osborne CK. Norton L. Young RC et a/ Nodular histiocytic lymphoma: An aggressive nodular lymphoma with potential for pro- longed disease-free survival. Blood 1980; 56:98-103.

3 I . Portlock CS, Rosenberg SA. No initial therapy for Stage 111 and IV non-Hodgkin’s lymphomas of favorable histologic types. Ann Intern Med 1979; 9O:lO-13.

32. Rappaport H. Tumors of the hematopoietic system. Atlas of Tumor Pathology, Section 3, Fascicle 8. Washington, D.C., Armed Forces Institute of Pathology, 1966.

33. Reimer RR, Chabner BA, Young RC, Reddick R, and Johnson RE. Lymphoma presenting in bone. Results of histology, staging, and therapy. Ann Intern Med 1977: 8750-55.

34. Rosenberg SA. Validity of the Ann Arbor staging classification for the non-Hodgkin’s lymphomas. Cancer Treat Rep 1977; 61: 1023- 1028.

35. Schein PS, Chabner BA, Canellos GP, Young RC, Berard CW, DeVita VT. Potential for prolonged disease-free survival following combination chemotherapy of non-Hodgkin’s lymphoma. Blood 1974;

36. Strauchen JA, Young RC, DeVita VT, Anderson T, Fantone JC, Berard CW. Clinical relevance of the histopathological subclassi- fications of diffuse “histiocytic” lymphoma. N Engl J Med 1978;

37. Veronesi U, Musumeci R, Pizzetti F, Gennari L, Bonadonna G. The value of staging laparotomy in non-Hodgkm’s lymphomas. Cancer 1974; 33:446-459.

38. Young RC, Howser DM, Anderson T, Fisher RI. Jaffe E, DeVita VT. Central nervous system complications of non-Hodgkin’s lymphoma. Am J Med 1979; 66:435-443

43:181-189.

299:1382-1387.