Morphologic studies of lymphocyte nuclei in follicular and diffuse mixed small- and large-cell (lymphocytic-histiocytic) lymphoma

Original Contributions Morphologic Studies of Lymphocyte Nuclei in Follicular and Diffuse Mixed Small- and Large-Cell [Lymphocytic-Histiocytic] Lymphoma IRVING DARDICK, MD, FRCPC, DOUGLAS R, CALDWELL, BS, DAVID MOHER, MS, AND MAHA JABI,MD, FRCPC Twelve examples of mixed small- and large-cell lymphoma (eight follicular, one follicular and diffuse, and three diffuse) were investigated morphometrically using plastic-embedded tissue in order to study nuclear characteristics of lymphocyte populations in this form of non-Hodgkin's lymphoma (NHL) and to test morphologic bases for current NHL classification systems. This s t u d y illustrates that there are many inaccuracies, illusions, and misconceptions in the morphologic criteria currently used to classify mixed small- and large-cell lymphoma. A principal finding was that lymphocyte nuclear profiles in mixed-cell lymphomas tend to be smaller in size (P < .005) and more irregular in shape (P = .0001 ) than the morphologically similar counterparts in germinal centers of lymph nodes with reactive hyperplasia. Intercase comparison of mixed small- and large-cell lymphomas revealed a considerable range of mean nuclear area values, some of which were within the size range of normal, small lymphocytes. At the magnifications used for morphometric assessment, a high proportion of lymphocyte nuclear profiles had shallow invaginations, but only a limited number of profiles (4% to 14%) had deep (cleaved) indentations. Contrary to current definitions for this subtype of NHL, lymphocytes with "small" nuclei had the same proportion of the nuclear diameter occupied by nuclear invaginations as lymphocytes with "large" nuclei and, in fact, mean nuclear invagination depth was shallower in "small" nuclei than in "large" nuclei. Furthermore, regardless of whether it is nuclear area or shape that is evaluated, lymphocytes in mixedcell lymphoma d o n o t separate i n t o two populations of smallcleaved and large noncleaved cells. Morphometry reveals that only four of the 12 examples of mixed small- and large.cell lymphoma had a proportion of the lymphocytes in the size range of fully transformed germinal center lymphocytes that exceeded 25%, and none of the cases approached 50% even though the population of lymphocyte nuclei appearing "transformed," and therefore "large," ranged from 28% to 57%. Such results indicate that the large, noncleaved and cleaved component, as seen in histologic sections of mixed small- and large-cell lymphoma, d o n o t have nuclei of uniform size and many, in fact, are not actually large. The morphometric findings indicate reasons for the p o o r

From the Departments of Laboratory Medicine and Research. Ottawa Civic Hospital. the Canadian Reference Centre for Cancer Pathology, and the Department of Pathology, University of Ottawa. Revision accepted for publication October 13. 1987. Key words: lymphoma, lymphocyte, morphometry, nucleus. Address correspondence and reprint requests to Irving Dardick, MD. Head. Section of Electron Microscopy, Department of Laboratory Medicine. 1053 Carling Ave. Ottawa. Ontario K1Y 4E9. 9 1988 by W.B. Saunders Company. 0046-8177/88/1908-000355.00/0

observer reproducibility in classifying this subtype of NHL. HUM PATHOL 19:889--901. 9 1988 by W.B. Saunders Company.

In reassessing standards for segregating malign a n t l y m p h o m a o f m i x e d small- and large-cell (mixed-cell) type from other non-Hodgkin's lymphomas (NHLs), it must be realized that this category of NHL has the poorest observer reproducibility.l-3 Perhaps this is due to the general nature of the principal histologic criterion, ie, judging the number of large lymphoid cells4 and the lack of a standard morphologic definition 3,5,6 for this subtype. Judging the proportion of large cells in the mixed-cell subtype of NHL has proven difficult, 3 largely due to the inability of the human eye to assess small, but significant, differences in size and configuration of lymphocyti~ nuclei.3.7 U n f o r t u n a t e l y , this situation is compounded by the limited number of clinicopathologiC studies and the lack of detailed morphologic studies in the mixed form of NHL. Furthermore, as will be. come evident in the results of the current study, there are inaccuracies in our present Cytologic criteria and illusions for the estimation of nuclear size and shape t h a t perhaps are responsible for the difficulties in classifying mixed-cell lymphoma. I n addition to the mixture of cell types in the classification category of mixed small- and large-cell lymphoma, other factors also create heterogeneity in this form of NHL. Based on immunologic characteristics, there are at least two types of diffuse mixed lymphoma, ie, B and T cell, 5,s and even nodular mixed lymphoma may have two basic B cell forms. 6 On morphologic grounds, it is possible to define four categories: a mixed type with small-cleaved cells and large cells with cleaved and/or noncleaved nuclei; a mixed form with small lymphocytes and large cells with "transformed" prolymphocytic and blast-like appearance; a mixed, small and large "immunoblastic" cell type; and a mixed unclassified type. 5 In designating NHLs as mixed cell type, the proportion of small cells to large cells is the principal criterion; however, this facet has also not been standardized. In its simplest form, a mixture of small and l a r g e lymphoid cells in varying numbers, but without a preponderance of either type of cell, c o n s t i t u t e s a mixed lymphoma. 4,s Other reports have set stricter limits on 889

HUMAN PATHOLOGY

Volume 19, No. 8 [August 1988]

of each case were photographed on 35-mm film using an oil immersion (• 100) objective. 22,26 Segments of a stage micrometer (10-1xm intervals) were also photographed on each film to determine the final magnification of the measured image. The resulting negative images were projected onto a digitizing pad (Summagraphics, Fairfield, CT) at a final magnification of • 2,500. Using a cross-hair cursor, the outlines of all complete lymphoid nuclear profiles (with the exception of glancing sections) were traced, and certain other parameters measured and recorded. These included nuclear diameter, area, perimeter, shape factor, and the depth of profile indentations. Nuclear profile shape was measured as the c o n t o u r index, 27 a size-independent calctilation relating perimeter to area in which a circle has a value of 3.54. Increasing nuclear profile irregularity results in higher values. As detailed previously, 22,26,28 all nuclear profiles with indentations deeper than 0.4 p~m were designated as "invaginated," while those with deeper indentations (~>one third of the nuclear diameter) were segregated as "cleaved." This allowed calculation of the percentage of profiles with these features. Sample sizes of nuclear profiles measured for individual cases varied from 735 to 2,544, with a mean (-+SD) of 1,080 -+ 469. In order to appreciate the size and form of nuclear profiles in the various lymphoid cell subpopulations in cases of malignant lymphoma, mixed small-and large-cell type, each profile was segregated by m o r p h o t y p e 22,26,29-3z at the time it was measured morphometrically. This was done by a subjective evaluation of the interplay of the size of the nuclear profile, the relative prominence of the nucleolus, and the distribution of the condensed chromatin. Nuclear profiles in mixed-cell lymphoma corresponding in appearance to the small (resting or unstimulated) lymphocytes of peripheral blood or the mantle zone of follicular centers in reactive hyperplasia (large chromatin clumps and inconspicuous nucleolus) were labeled as morphotype 1. Slightly larger nuclei, likely including most of the small-cleaved cells, with partially disaggregated chromatin and a visible nucleolus (so that they resemble the partially transformed lymphocytes within follicular centers) were designated as morphotype 2. Larger-appearing nuclei with the disaggregated chromatin pattern and enlarged nucleolus characteristic of the fully transformed lymphocyte of follicular centers were labeled as morphotype 3 (for examples, see fig. 5). The mean parameters for lymphocyte nuclei with the above characteristics in the mantle zones and follicular centers from six normal human lymph node biopsies, which had been measured previously, 28 were used for comparison with the 12 examples of mixed-cell lymphoma. To obtain certain general information about the nuclear parameters of the lymphoid populations in mixed small- and large-cell lymphoma, whether based on nuclear size or morphotype, pooled samples were used. For this purpose, a random set of approximately 150 lymphocyte nuclear profiles from each of the 12 cases of mixed-cell lymphoma were selected from the total population measured previously for each case and combined for statistical analysis. The pooled sample consisted of 1,961 nuclear profiles. As well, intercase comparison of the proportion of variously sized lymphoid nuclei in mixed-cell lymphoma was facilitated by establishing arbitrary nuclear area classes (NACs). These are equivalent to nuclear volume classes used previously for the analysis of germinal center populations in reactive hyperplasia. 22 In the current study, measured nuclear profiles in mixed-cell lymphomas were distributed into four NACs, based on the size distribution of lymphocytes in follicular centers. NAC I represented the

the c o n t e n t o f large cells at either 30% to 70%, 5 25%to 50%, 3 o r 30% to 50% 6 o f the l y m p h o i d p o p ulation in categorizing a m i x e d l y m p h o m a , w h e t h e r diffuse or follicular. Assessing this type o f criterion still r e m a i n s subjective, b u t even the m o r e objective a p p e a r i n g a p p r o a c h o f B e r a r d a n d colleagues 9,1~ has not p r o v e n r e p r o d u c i b l e in selectively categorizing those follicular center cell l y m p h o m a s with c o m p o nents o f small a n d large l y m p h o c y t e s ? M o r p h o l o g i c studies o f N H L that include at least s o m e e x a m p l e s o f m i x e d l y m p h o m a , a n d in which various aspects o f l y m p h o c y t e nuclei are characterized, are limited.11-22 A l t h o u g h s o m e o f these studies involve m e a s u r e m e n t o f n u c l e a r p a r a m e t e r s 14,22 or u l t r a s t r u c t u r a l features o f mixed-cell lymphoma,11-13 only a few studies a t t e m p t to define s o m e aspects o f the structural f e a t u r e s o f the a b n o r m a l f o r m o f neoplastic lymphocytes comprising mixed-cell l y m p h o m a s . 14,22 T h i s m o r p h o m e t r i c investigation o f 12 e x a m p l e s o f follicular (nine) a n d diffuse (three) m i x e d smalla n d large-cell l y m p h o m a studies l y m p h o i d p o p u l a tions in this f o r m o f N H L in c o m p a r i s o n with lymphocytes o f g e r m i n a l centers a n d m a n t l e zones o f reactive hyperplasia. Such studies are not only relevant to the relationship o f certain f o r m s o f N H L to follicular c e n t e r l y m p h o c y t e s , b u t are also r e q u i r e d to elucidate reasons f o r the p o o r o b s e r v e r reproducibility in m i x e d l y m p h o m a s . 1-~

MATERIALSAND METHODS Cases for morphometric assessment in this series were chosen from those in which there was a consensus diagnosis (at least two thirds of the independent observers) for malignant lymphoma, mixed small-and large-cell, follicular or diffuse 4 (nodular or diffuse mixed lymphocytic-histiocytic, 23,24 centrocytic-centroblastic 25) by the Lymphoma Panel of the Canadian Reference Centre for Cancer Pathology or members of the Canadian Reference Centre for Cancer Pathology staff, and in which previously fixed wet tissue was available. All such cases were reassessed and only those cases were accepted in which the population of small and large lymphoid cells were judged to be approximately equal, the principal criterion used in current classifications for mixed small-and large-cell lymphoma. 4,23,24 Although this is a subjective approach to quantitating proportions of lymphoid cells, it is the usual one used by surgical pathologists and must be used to assess the validity of microscopical observations relative to the morphometric assessments. In this series, 11 of the 12 cases involved a lymph node at various sites and one case presented as a retroperitoneal mass. Eight cases were follicular (follicular areas measured in seven cases and interfollicular zone measured in one case), one was follicular and diffuse (latter zone measured), and three were diffuse mixed-cell lymphomas. Tissues available for embedding were either fixed in buffered formalin (seven cases) or B5 (five cases). Portions of tissue were dehydrated in graded alcohols, embedded in glycol methacrylate resin (Sorval embedding medium; Dupont, Wilmington, DE), sectioned at 1-p,m thickness on glass knives, and stained with hematoxylin-eosin. For morphometry, random areas within follicular or diffuse regions

890

MIXED-CELLLYMPHOMA [Dardick et al]

size range for the majority of small (unstimulated) normal lymphocytes (0 to 18.6 I,l,m2); NACs II (18.6 to 29.4 ~m 2) and III (29.4 to 38.5 b~m2) represented the size range of most partially transformed normal lymphocytes; and NAC IV (>38.5 ~m 2) represented the size of nuclear profiles for the majority of fully transformed normal lymphocytes. The data analysis was carried out using the statistical analysis system (SAS). s3 Comparisons between the lymphoma and normal data were achieved using independent t tests. The data was also assessed for normality using the Kolomogorov D statistic and the appropriate transformations were made. Data transformation is recommended for several reasons, such as for achieving normality because many statistical tests used for analyzing data are based on a normal distribution, s4,~5

70-

60-

50-

e9

40E

RESULTS ~0"

Nuclear Parameters

9 :

8

8~ | 0

Figures 1 a n d 2 provide comparative data for nuclear parameters between mantle zone and follicular center lymphocytes in lymph nodes with reactive hyperplasia and lymphocytes in the 12 cases of follicular and diffuse mixed-cell lymphoma. In the case of m e a n nuclear area (Fig 1), the similar range of m e a n values for the small ("mature" or unstimulated) lymphocyte in mantle zones and follicular centers (morphotype 1), and the gradual enlargement of the partially t r a n s f o r m e d (morphotype 2) and fully transf o r m e d ( m o r p h o t y p e 3) lymphocytes o f follicular centers was evident. In the case of mixed-cell lymphomas, the m e a n nuclear area of m o r p h o t y p e 1 lymphocytes in all of the examples was well below that for the normal c o u n t e r p a r t (Fig 1). Similarly, the majority o f m e a n area values for lymphocytes in mixed-cell l y m p h o m a with m o r p h o t y p e 2 nuclei, and all of the values for m o r p h o t y p e 3 nuclei, is smaller than for their morphologic c o u n t e r p a r t in follicular centers (Fig 1). F u r t h e r m o r e , the range of m e a n nuclear area for m o r p h o t y p e 2 and 3 lymphocytes in mixed-cell l y m p h o m a was similar. T h e spread of m e a n values for nuclear area did not differ regardless of whether l y m p h o m a s were fixed in formalin or B5 (Fig 1). Differences between the m e a n nuclear areas of germinal center a n d l y m p h o m a lymphocytes for all three morphotypes are highly significant (P < .005). As reflected by the m e a n nuclear contour index, the shape of lymphocyte nuclei, whether in mantle zones or in germinal centers, was not too variable and all results fell below a c o n t o u r index of 4.0 (Fig 2). In contrast, in mixed-cell lymphomas, all the mean contour index values were >4.2. It was evident also that the span o f m e a n c o n t o u r index values was similar for lymphocyte nuclei in mixed-cell lymphomas regardless o f the m o r p h o t y p e or the type of fixative used (Fig 2). Differences between the m e a n nuclear cont o u r indices of germinal center and l y m p h o m a lym.phocytes for all three morphotypes are highly significant (P = .0001). Table 1 provides nuclear parameter data derived by pooling all of the lymphocytes measured in each example of mixed-cell lymphoma, regardless of the

z z

9

8

$|

20"

9

:o 9

10'

0

I~1

GC1

Ml~l

GC2

MI~2

GC3

Ml'3

Scatter diagram of mean nuclear profile area of lymphocytes in mantle zones [MI] and germinal centers [GCl to GC3] of lymph nodes with reactive hyperplasia and in mixed lymphocytic-histiocytic lymphoma [MLI to ML3]. The numerals refer to the cytologic appearance of nuclei in the subsets of lymphocytes observed in reactive hyperplasia and lymphomas, ie, morphotypes 1, 2, and 3 [see Materials and Methods]. Germinal center lymphocyte nuclei exhibit a gradual increase in the range of size from morphoIype I to 3, a feature that is not observed in lymphocyte nuclei in mixed-cell lymphoma. Morphotype I lymphocyte nuclei in mixedcell lymphomas are abnormally small compared with their cytologic counterpart in mantle zones and germinal centers. Open circles represent formalin-fixed lymphomas; circles with a central dot are B5-fixed cases. FIGURE 1.

nuclear m o r p h o t y p e . In the case o f m e a n nuclear profile area, there is a considerable variation with a range between 15.6 + 7.2 (case no. 3) a n d 28.7 -+ 18.5 I,L m 2 (case no. 9), a difference of 84%. However, it is important to note that in terms o f m e a n nuclear diameter, the d i f f e r e n c e b e t w e e n the two extremes amounts to 1.6 b~m. Many o f the values for the pooled m e a n nuclear area in Table 1 are within the range of m e a n n u c l e a r areas f o u n d for small lymphocytes (morphotype 1) in l y m p h n o d e biopsies with reactive hyperplasia (Fig 1). In mixed-cell l y m p h o m a s (Table 1), there was a fairly consistent degree of nuclear irregularity, with m e a n c o n t o u r indices r a n g i n g from 4.34 -+ 0.57 (case no. 3) to 4.77 + 0.6 (case no. 9). T h e nuclear parameters with the least variation were the m e a n d e p t h o f invagination, as m e a s u r e d f r o m sectional profiles, and the percentage of such profiles 891

HUMAN PATHOLOGY

Volume 19, No, 8 [August 1988]

5,0-

0

4.8-

~

0

0 0 0

|

4.6-

4.4-

|

|

|

cc

~

| c~

o

8

ec

0 o

a

3 ~ z

4.2-

0 (J z <

4,0.

3,8'

9

|..

$

"

I. I"

3,6-

3,4

M1 GC1 I~L1 G(~2 I~L2 GC3 M'L3 FIGURE 2. Scatter diagram of mean nuclear contour index of normal [M and GC] and neoplastic [ML] lymphocytes [abbreviations and symbols are the same as in fig. 1). Nuclear contour indices in normal [MI, GC1 to GC3] lymphocytes are fairly regular in shape, as reflected by a value <4.0 in all cases. In mixed lymphocytichistiocytic lymphoma [ML1 to ML3], the majority of cases have nuclear profiles that are considerably more irregular than normal lymphocytes [mean values >4.2]. Note also the similar range of values for all three of the nuclear morphotypes in mixed-cell lymphomas,

exhibiting an indentation >/0.4 ixm (Table 1). Even though there was more than a threefold difference in the proportion of lymphocyte nuclear profiles designated as cleaved (indents /> 1/3 of the nuclear diameter), this nuclear feature only ranged from a low of 4% (case no. 7) to a high of 14% (case no. 12) (Table 1). Four of the mixed small- and large-cell lymphomas (cases no. 1, 2, 9, and 12) were selected for more detailed analysis of nuclear characteristics (Table 2; Figs 3 and 4) and for correlation of the nuclear parameters with the morphologic features in histologic sections (Fig 5). In lymphomatous tissues embedded in plastic and sectioned thinly, it is easier to appreciate both the variation in size and shape and the basic a p p e a r a n c e (morphotypes) of lymphocyte nuclei within and between the cases of mixed small- and 892

large-cell lymphomas (Fig 5). Many nuclear profiles are irregular in shape and have minor indentations of the nuclear membrane (easier to appreciate at the x 2500 magnification used for morphometric measurements), but only a few nuclear profiles bear indentations deep enough to class as "cleaved" (Fig 5). Although many nuclear profiles of morphotype 3 lymphocytes appear to be r o u n d e r and more regular than morphotype 1 and two profiles in the four cases of mixed-cell lymphoma (Fig 5), mean nuclear contour indices for the three morphotypes are quite similar (Table 2). Variations in nuclear profile contour and size between the examples of mixed-cell lymphoma are readily appreciated when the data in Table 2 and the illustrations in Fig 5 are compared, although such differences are obviously subtle at the magnifications usually used to screen histologic sections. The four cases had results for nuclear profile area and contour index (Table 2) at the extremes of the ranges for these parameters displayed in Figs 1 and 2. As well, due to the smaller portions of lymphomatous tissue embedded in glycol methacrylate, the region available for morphometry varied from case to case. Table 2 indicates that three (cases no. 1, 9, and 12) of the four lymphomas were follicular in paraffin-embedded tissues. Within the tissue portion in plastic sections, cases no. 1 and 9 had follicular areas available for morphometric analysis. In the follicular and diffuse example (case no. 2), only diffuse regions were evident in plastic sections. Although case no. 12 only had interfollicular regions available, note that all nuclear parameters for morphotype 1 to 3 lymphocytes in case no. 12 paralleled those for the lymphocytes in the follicular and diffuse regions of cases no. 1, 2, and 9 (Table 2). This suggests that a considerable part of the lymphocytes in the interfollicular regions of this case may be neoplastic. Table 2 also includes data on the mean depth of nuclear profile invaginations and the ratio of this parameter to the mean nuclear diameter (data not shown but reflected in the mean nuclear area) in order to assess morphologic aspects of cleaved nuclei in mixed-cell lymphomas. Lymphocytes with morphotype 1 nuclei, and mean areas ranging from 11.1 -+ 3.5 /.tm2 to 12.9 -+ 5.7 txm 2, had significant but shallower mean invagination depths than the lymphocyte populations with m o r p h o t y p e 2 and 3 nuclei and mean areas ranging from 18.5 -+ 8.1 /xm 2 to 31.3 -+ 18.1 txm2 (Table 2). The similarity in the ratio of invagination depth to nuclear diameter in morphotype 1 nuclei from case to case is noteworthy (Table 2). Nuclear indentations in such nuclei occupied a greater proportion of the nuclear diameter (23% to 24%) than morphotype 2 and 3 nuclei (18% to 22%). In these four examples of mixed-cell lymphomas, morphotype 2 and 3 nuclei had a very similar range of mean invagination depth and the proportion of the nuclear profile occupied by an indentation (Table 2). Such results indicate that the morphotype 2 lymphocyte nuclei (corresponding to most of the small-

MIXED-CELLLYMPHOMA(Dardick et al) TABLE 1. Comparative Values for Nuclear Parameters of Lymphocytes in Mixed Small- and Large-Cell Lymphoma

Case No. 1 2 3 4 5 6 7 8 9 10 11 12

Mean A r e a (txm2) * 27.6 17.9 15.6 19.1 22.8 20.3 19.0 20.6 28.7 18.3 19.9 28.5

-+ • • -+ • • • • • • • •

17.0 8.7 7.2 11.2 13.2 12.2 11.9 11.9 18.5 10.7 13.4 18.0

Mean P e r i m e t e r (b~m) 22.6 18.1 16.9 19.5 20.9 19.0 19.1 19.7 24.5 19.2 19.8 24.3

• 7.2 • 5.1 • 4.5 • 6.5 • 6.8 • 5.9 • 5.2 • 6.0 -+ 8.8 • 5.7 • 7.2 • 9.1

• SD. :~ Only n u c l e a r profiles with invaginations > 0.4

Mean Invagination D e p t h (b~m)$

Mean Contour Index 4.49 4.40 4.34 4.64 4.52 4.39 4.48 4.47 4.77 4.63 4.60 4.72

• 0.63 • 0.59 • 0.57 • 0.66 • 0.60 • 0.57 • 0.53 • 0.60 • 0.68 • 0.61 • 0.61 -+ 0.73

1.1 1.1 1.0 1.0 1.2 1.2 0.9 1.1 1.3 1.1 1.1 1.3

• 0.5 • 0.5 • 0.4 • 0.4 • 0.7 • 0.5 -+ 0.5 • 0.5 • 0.7 • 0.5 -+ 0.5 • 0.7

Percentage of lnvagination Profiles

Percentage of Cleaved Profiles

N

73 70 64 71 71 69 76 74 80 79 74 83

8 7 7 5 9 7 4 10 12 9 11 14

1,000 1,000 1,000 833 866 1,004 2,544 1,000 981 1,000 1,000 735

* Mean

I~m.

cleaved cells) do not differ significantly from the morphotype 3 lymphocyte nuclei (the counterpart of large noncleaved and cleaved cells) in mixed-cell lymphoma. Frequency distributions of individual nuclear profile area and contour index (Figs 3 and 4) in the four cases of mixed-cell lymphoma tabulated in Table 2 and illustrated in Fig 5 aid in investigating the morphology of the small- and large-cell populations in this type of lymphoma. The concept of a dual population of small-cleaved cells and large, predominantly noncleaved cells should produce a distinct bimodal population when nuclear area and shape are measured morphometrically. Neither in the four cases illustrated in Fig 5 nor in any of the other examples of mixed-cell lymphoma has it been possible to demonstrate a bimodal population either based on the nuclear area of sectional profiles (Fig 3) or their contour index (Fig 4). The frequency histograms (Figs 3A and 4A) indicate that both nuclear profile area and contour index have unimodal distributions with positive

skewness. Since skewness may indicate variability in scores and not necessarily the existence of a secondary population, the data was log (common logarithm) transformed. As evident in Figs 3B and 4B, the frequency distributions approach that of a normal unimodal distribution, both for nuclear area and contour index for all four lymphoma cases. Lymphocyte Population Data

Distribution of the individual nuclear profile measurements into the four predetermined NACs allowed assessment of the proportion of large cells in mixed-cell lymphoma (Table 3). If NACs I and II represent the smaller cell compartment (0 to 29.4 I,zm 2 and NACs III and IV represent the larger lymphocyte component (>30 ixmZ), then only four examples (cases no. 1, 5, 9, and 12) had a population of large nucleated lymphocytes >25% and none of the cases approached 50% of the total lymphocyte population (Table 3).

TABLE 2. Mean Nuclear Parameters of Lymphocytes in Various Architectural Areas of Follicular Mixed Small- and Large-Cell Lymphoma

Nuclear M o r p h o t y p e (%)

Mean Area (+ SO) (txm 2)

Mean Invagination D e p t h (_+ SD) (Ixm)

Ratio M e a n Invagination D e p t h to Mean Diameter

4.52 + 0.58 4.50 • 0.62 4.48 -+ 0.65

0.98 • 0.32 1.12 + 0.51 1.16 • 0.56

0.23 0.19 0.18

4.61 • 0.60 4.78 +- 0.68 4.80 • 0.70

1.01 + 0.53 1.35 -+ 0.75 1.35 • 0.71

0.23 0.20 0.21

4.31 • 0.55 4.41 • 0.55 4.44 + 0.66

0.98 • 0.35 1.05 + 0.44 1.19 + 0.57

0.24 0.20 0.22

4.76 • 0.70 4.78 + 0.70 4.68 • 0.59

0.97 + 0.37 1.23 • 0.51 1.33 +- 0.74

0.23 0.19 0.20

Mean Contour I n d e x (• SD)

Case no. 1--Follicular (follicular area m e a s u r e d ) 1 (6.2) 11.8 • 3.6 2 (41.1) 25.2 --- 13.7 3 (52.7) 31.3 -- 18.1 Case no. 9---Follicular (follicular area m e a s u r e d ) 1 (11.5) 12.9 --- 5.7 2 (49.8) 30.9 • 18.4 3 (38.5) 30.6 • 18.9 Case no. 2 - - F o l l i c u l a r a n d diffuse (diffuse a r e a m e a s u r e d ) 1 (19.1) 11.5 m 4.0 2 (49.5) 18.5 • 8.1 3 (35.0) 20.5 • 9.7 Case no. 12--Follicular (interfollicular area m e a s u r e d ) 1 (7.6) 11.1 • 3.5 2 (35.4) 29.6 --- 15.3 3 (56.1) 30.1 • 19.5

893

HUMAN PATHOLOGY

Volume 19, No. 8 [August 1988]

CASE 1

CASE

20[

200

A

2

20c

200

B

B

A

160

~

120

140

~" 80

80

4O

40

i

,

20 . o ' ~

, J

AREA

i

120

\

f

80 ~00 ~2o 140 9

05

(#m 2)

CASE

10 15 LOG AREA (~rrl 2)

i 2O

I

20

I

I

[

60

80

05

I

40 AREA ( F m 2 )

9

CASE 200

8C

40

i I t 60 80 100 AREA (~-m 2)

I 120

r 140

12

~2s

80

~ 40

I 20

16C

120

J 20

I 15

20C

160

80

I 10

LOG AREA ( ~ r n 2)

4C

0 O5

10 15 LOG AREA (,o.m 2 ]

I 20

2O

I 40

I 60

/ 80

I 100

AREA ( F m 2)

I 120

I 140

0

0'5

,,o LOG AREA

1'5

2'0

( / t i n 2)

FIGURE 3. Frequency distribution of lymphocyte nuclear profile areas in the four examples of mixed small- a n d large-cell l y m p h o m a [cases no. fl [n = fl,000], 9 [n = 9811, 2 [n = fl,0001, a n d 12 [n = 7381} that are illustrated in fig. 5 and have morphometric data tabulated in Table 2. Note the unimodal distribution of lymphocyte nuclear profile areas in both the nontransformed [A] a n d log-transformed [B] data in all four cases. Values for the three nuclear morphotypes were p o o l e d in each case for these histograms.

By analyzing the morphometric data obtained for the morphotype 3 nuclei in mixed-cell lymphoma, ie, those with the appearance of transformed lymphocytes, it was possible to obtain further information relating to the size characteristics of the so-called "large-cell" lymphoid population (Table 4). The proportion of morphotype 3 nuclei in the 12 examples varied from 28.3% to 57%, with mean nuclear areas ranging from 18.5 + 9.1 p , m 2 t o 31.3 +-- 18.3 ~ m 2 (Table 4). When the proportion of morphotype 3 nuclei that had a nuclear profile area >40 Ixm2 (mean values for histiocyte nuclei are also in this range and follicular center morphotype 3 nuclei have mean areas greater than this value 18,28) was calculated, none of the cases exceeded 50% in this category and one case only had 11% of the morphotype 3 nuclei in this size range (Table 4). It is important to note that the percentage of nuclear profiles with invaginations and clefts in morphotype 3 nuclei (Table 4) is very similar to the percentage of these nuclear features, on a case by case basis, for the entire population of lymphoid cells measured (Table 1). The information in Tables 3 and 4 can be used to gain further insight into the relationship of the ap-

pearance of the "large-cell" component, ie, lymphocytes with morphotype 3 nuclei, to the size o f the nuclei in such lymphocytes. This was accomplished by linear regression analysis and calculation of regression coefficients for mean nuclear area values against the percentage of NAC III and IV lymphocytes (Table 3; Fig 6A) in one instance and against the percentage of lymphocytes with morphotype 3 nuclei (Table 4; Fig 6B) in the other. Matching a pair of values that were both based on nuclear profile area, ie, mean nuclear area v percent NAC III and IV (Fig 6A), resulted in a regression equation with a high correlation coefficient (r = .960). Such results indicate a hi.ghly significant relationship (P < .001). However, pairing mean nuclear area with the percentage of morphotype 3 nuclei (Fig 6B) resulted in a different distribution of the matched pairs compared with Fig 6A and also a considerably lower correlation coefficient (r = .508). This data indicates that there is not a significant relationship (P > .05) between nuclear area and morphotype, that lymphocytes with morphotype 3 nuclei are not uniform in size, and that all such nuclei are not in fact "large." By definition, 4,23-25 mixed small- and large-cell 894

MIXED-CELL LYMPHOMA [Dardick et al) CASE

200

200

200

A

B

A ~Z

CASE 2

1

200

B

160

160

120

2O

20

120

80

~0i

8s

8

40

\

, c~ ~ (5 o (5 LOG CONTOUR INDEX

CONTOUR INDEX

O O (5 (5 LOG CONTOUR INDEX

o CONTOUR INDEX

CASE

CASE 9

12

20C

20r

200

160

16(

16C

f60

12[

12C

12C

120

8s

8c

2OO

~J

I 40

10

40

,i CONTOUR INDEX

o

i o o O (5 o (5 LOG CONTOUR INDEX

i

i

i

o,

.~

o CONTOUR INDEX

LOG CONTOUR INDEX

FIGURE 4, Frequency distribution of lymphocyte nuclear profile contour index [shape factor] for the same four examples of mixed smaD and large-cell lymphoma. Again, there is no evidence of a bimodal population in either the nontransformed [A] or the log-transformed [B] data, but a continuum from relatively regular profiles [low contour index value] to those with irregular shapes [high contour index value]. The distribution of measured nuclear profiles reflects the mean contour indices for the three individual morphotypes for each case [Table 2), but the values for all lymphocytes in each case were pooled to obtain these histograms.

that were virtually identical (Fig 7), and therefore no dual populations of large and small lymphoid cells with markedly differing nuclear profile shape were identifiable.

lymphoma should be composed of a small-cell population with irregularly shaped nuclei and, therefore, a high nuclear contour index. In contrast, the large-cell component primarily consists of a population of lymphocytes with noncleaved nuclei, and thus, a relatively low nuclear contour index. The data in Table 5, derived from pooled samples of the 12 cases of mixed-cell lymphoma, indicated that the mean contour index for the lymphocyte nuclei <20 Ixmz (representing the small-cell population) was similar to that for nuclei >20 Ixm~ (representing the large-cell population). Despite differences in mean invagination depth between the two populations, the direct relationship of nuclear size to invagination depth was revealed by the similar ratios of invagination depth to nuclear diameter in the small and large subpopulations (Table 5). There was also little difference between the percentage of invaginated and cleaved nuclear profiles (Table 5). The frequency distribution of nuclear contour indices for the two subsets of lymphocyte nuclear profiles produced unimodal curves

DISCUSSION

The precepts for segregating mixed small- and large-cell lymphoma in any of the widely used classifications of NHL have not been substantiated by experimental investigation. Attempts by Metter et aP to quantitate the number and type of large lymphoid cells as a guide to diagnosing subtypes of NHL, and to judge nuclear size, illustrated the variability and subjectivity of the criteria used by pathologists in deciding which lymphocytes are considered large and cleaved or noncleaved. Perhaps it is such problems that underlie the poor diagnostic reproducibility in NHL. ~-s In view of the focus of this report, it is important to realize that of the principal subtypes of 89,5

HUMAN PATHOLOGY

Volume 19, No, 8 [August 1988]

FIGURE 5. Representative fields of glycol methacrylate sections from [A] a follicular area of case no. I [nodular lymphoma), [B] a follicular area of case no. 9 [nodular lymphoma], [C] a diffuse region from case no. 2 [nodular and diffuse lymphoma]; and [D] an interfollicular region from case no. 12 [nodular lymphoma]. The numerals adjacent to lymphocyte nuclei on micrographs [A] and [B] indicate examples of the cytologic details of morphoiypes 1,2, and 3 [arrows indicate examples of "cleaved" profiles]. By comparing the nuclear characteristics in the micrographs with the morphometric values in Table 2, it becomes evident that it is difficult to accurately judge similarities or differences in nuclear size and shape in the heterogeneous populations in mixed-cell lymphoma. At this magnification, it is possible to detect the generally smaller size and more regular shape of lymphocyte nuclei in case no. 2 [C] compared with the other three examples [A, B, and D}. Note the degree of abnormaliiy of nuclear profiles in the lymphocytes in the interfollicular areas of the nodular lymphoma of case no. 12 both in the micrograph [D] and data in Table 2. [Hematoxylin-eosin stain', approximate magnification •

896

MIXED-CELL LYMPHOMA [Dardick et al)

FIGURE 5. [continuea]

cell lymphomas. 26,32 In all three of these subtypes, lymphocytes exhibit mean nuclear areas significantly smaller than those of normal lymphocytes, a feature likely resulting from an abnormal, and possibly deficient, expression of the structural proteins present in the nucleus; such proteins appear to be a principal factor responsible for nuclear size. 22,31 However, despite the "large-cell" appearance of lymphocytes with nuclei resembling those of the fully transformed lymphocytes (morphotype 3) in follicular centers in reac-

NHL, the mixed small- and large-cell type usually has the worst degree of interobserver reproducibility; various studies report a range of 26% to 61%.l-3 It is necessary to judge the results of the present study against the diagnostic difficulties inherent in current NHL classifications and the limited information underlying the concepts for these classifications. Lymphocyte nuclear profiles in mixed small- and large-cell lymphoma have a feature in common with lymphoid populations in small-cleaved cell and large897

HUMAN PATHOLOGY

Volume 19, No. 8 (August 1988)

TABLE 3.

Distribution (%] of Lymphocyte Nuclear Profile Size in Follicular and Diffuse Mixed Small- and Large-Cell Lymphoma

Case No.

NAC I

NAC II

NAC III

NAC IV

NAC I n and IV

1 2 3 4 5 6 7 8 9 10 11 12

34.2 57.9 76.5 56.4 47.1 59.0 57.1 51.4 33.2 64.0 58.9 34.7

30.3 32.3 18.1 26.7 27.8 21.9 32.6 32.9 29.2 25.7 24.6 26.7

17.0 6.9 3.6 10.3 13.1 10.1 6.6 9.3 14.5 5.4 8.8 16.7

19.4 2.9 1.8 6.6 12.0 9.1 3.8 6.4 23.1 4.9 7.7 21.8

36.4 9.8 5.4 16.9 25.1 19.2 10.4 15.7 37.6 10.3 16.5 38.5

NOTE. NAC refers to nuclear area class, the equivalent of a previously used set of nuclear volume classes. 22'32 NAC I is a size range of nuclear profile area from 0 I~m~ to 18.6 Fl,m2; NAC II ranges from 18.6 Ixm~ to 29.4 p,m2; NAC III ranges from 29.4 txm to 38.5 I~m2; NAC IV is > 38.5 Izm2.

tive lymphoid tissue, many examples of mixed-cell lymphoma have this type of lymphocyte nucleus falling within the size range of the small (unstimulated or morphotype 1) lymphocytes of mantle zones and follicular centers (Fig 1). The illusion of large size in lymphocyte nuclei described as the "large-cell" component in NHLs is produced by a combination of the chromatin distribution in such nuclei (markedly disaggregated) and the smaller than usual nuclear profiles of many of the small (morphotype 1) lymphocytes (Fig 1). Even the lymphocyte nuclei in mixedcell lymphomas classed as morphotype 2, which likely include most of the "small-cleaved" population as usually defined, fall within the size range of the morphotype 3 nuclei, and are also larger than the morphotype 1 nuclei in this N H L (Fig 1). These features have an important bearing on the percentage of nuclear profiles in NHLs that are actually within the size range of the nucleus of transformed lymphocytes in germinal centers. If the cytologic criterion of a large-cell population in the range of 25% to 50% is necessary to qualify a lesion as mixed small- and large-cell lymphoma, 3 then it is the abnormality of nuclear size that also accounts for only four of the examples in this series TABLE 4.

falling within the appropriate range for such a diagnosis, and the fact that none of the cases approach 50% of the lymphoid cells with a nuclear profile area >30 la,m 2 (Table 3). Such results emphasize that the usual coupling of increasing nuclear size and condensed chromatin redistribution that occurs during normal lymphocyte transformation 22,31,36 does not necessarily occur in malignant lymphomas. This aspect was also reflected by the linear regression analysis of mean nuclear profile area matched with the percentage of morphotype 3 nuclei per case, a relationship that is not a direct one (Fig 6). Furthermore, the percentage of the morphotype 3 nuclear profiles >40 I,l,m 2 (normal morphotype 3 lymphocytes in follicular centers have mean nuclear areas ranging from 45 la,m 2 t o 65 Ij,m 2 ) 28 did not exceed 50% in any of the 12 cases morphometrically analyzed, and was less than 25% in one half of the cases (Table 4). Such results do not necessarily imply misclassification of this group of mixed small- and large-cell lymphomas, for another morphometric study of NHLs has also reported the percentage of large cells in eight diffuse cases of this subtype to be 21.6 + 6.6. 37 One current classification system for N H L 38,39 suggests that neoplastic lymphocytes fall into four main categories based on nuclear size and shape, ie, small- and large-cleaved, and small and large noncleaved cells. Such morphologic criteria, especially the small-cleaved and large noncleaved types, are particularly appropriate to mixed small- and largecell lymphoma. T h e basis for such a hypothesis is readily testable, since measurements of lymphocyte nuclear profile size and shape should both reveal a bimodal population in this subtype of NHL. However, as shown in this report, a variety of analyses of the morphometric data (including log transformation to ensure the data has a normal distribution) always indicate a unimodal population. This does not mean that a certain percentage of nuclei in mixed-cell lymphoma is not distinctly irregular in contour or deeply grooved ("cleaved"), but simply that there is a continu u m of nuclear profiles from round and regular (low contour index) to complexly shaped (high contour index). Similarly, lymphocyte nuclear profile area, when measured from histologic sections, does not

Nuclear Profile Data for Morpho~pe 3 Lymphocytes in Mixed Small- and Large-Cell Lymphoma

Case No.

N

T y p e 3 (%)

1 2 3 4 5 6 7 8 9 10 11 12

527 350 384 357 397 284 925 388 378 409 459 419

52.7 35.0 38.4 42.9 45.8 28.3 36.4 38.8 38.5 40.9 45.9 57.0

Mean Area (Ixm~ -+ SD) 31.3 20.5 18.5 22.6 27.9 26.1 22.9 24.1 30.1 22.9 22.8 30.1

-+ + -+ -+ -+ -+ -+ -+ -+ -+ + +

18.8 9.7 9.1 10.6 14.4 14.6 10.9 14.0 18.9 13.4 23.9 19.5

898

Profiles > 40 p,m 2 (%)

I n v a g i n a t i o n s (%)

C l e a v e d (%)

46.1 13.7 11.2 20.7 39.3 33.7 20.5 25.5 41.5 19.1 23.9 39.9

72.3 69.7 71.6 72.2 73.0 70.4 74.8 72.9 84.4 78.9 69.3 82.8

4.7 6.9 7.6 2.2 10.6 6.3 4.6 5.2 13.0 9.8 9.6 12.6

MIXED-CELL LYMPHOMA [Dardick et al]

.s A 1./ A

E ::L

30

| /"

<[ <[ n,.

./

25 7 |

i,i

/

/ |

y-18.13 + 0.34x r= 0 . 9 6 0 p< 0.001

f

30 m

1

Z Z

J

J

C

./

t'

25

/

|

m >

| m

J

|

|

|

y - 1 3 . 7 9 + 0.27X r=0.508

p>0.05

I

I

i

I

10

20

30

40

~o N A C m

@!2 j /

/

i I| 7

3 //2| .//|

1@ |

4|

:E 15 0

/

9•

/

|

t"

20

,I

./

./

/

/|

..~ Z Z

35

FB

20

+ 'Ira'

i

l

l

I

30

40

50

60

15

~oTYPE 3 NUCLEI

FIGURE 6. Scatter diagrams plotting lymphocyte mean nuclear a r e a [l~m 2] in mixed lymphocytic-histiocytic lymphoma against [A] the percentage of lymphocyte nuclei in NAC III a n d IV a n d [B] the percentage of lymphocytes with morpholype 3 nuclei in this disease. The regression line and correlation coefficient in A indicate that in each case of mixed-cell lymphoma there is a high correlation between lymphocyte mean nuclear a r e a a n d the proportion of lymphocytes with nuclei in classes III a n d IV, In B, the regression line and correlation coefficient indicate a poor relationship between lymphocyte nuclear a r e a a n d the physical appearance of the nuclei, ie, not all lymphocyte nuclei in mixed-ceil lymphomas with the appearance of transformed lymphocytes are actually large. The numbers on both scattergrams represent each of the 12 cases in Tables I through 5,

The complexity of lymphocyte populations, in terms of nuclear size, form, and chromatin organization, in mixed-cell lymphoma, as well as the variations from case to case, are likely causes for the diagnostic difficulties in this subtype of NHL. As indicated by the detailed analysis of the selected four cases in this report, subtleties in lymphocyte nuclear form are difficult to appreciate in routine histologic sections. Furthermore, pathologists vary in their ability to recognize such differences. 3 Inaccurate judgement of nuclear characteristics in histologic sections, ~,7 and the considerable overlap of lymphocyte nuclear parameters between the main subtypes of follicular center cell lymphoma, 4~ are therefore two major factors underlying the problem o f diagnostic consistency. Unfortunately, classification of NHL will continue to be plagued with poor observer reproducibility until the inaccuracies, illusions, and misconceptions in the criteria used for the classifying mixed small- and largecell lymphoma are appreciated. Cleary and Sklar 42 have reviewed the impact of sound experimental approaches in molecular biology leading to major advances in the diagnosis of NHL. However, these investigators note the continuing reliance on the review of histologic material in order to subtype NHL for therapeutic and prognostic purposes. This should

simply reveal discrete samples of "small" and "large" lymphoid cells. Direct measurements of nuclear indentations also support the view that mechanisms responsible for nuclear irregularity apply equally to all neoplastic lymphocytes in mixed-cell lymphoma, as well as in other follicular center cell lymphomas. 22,26,4~ The concept of a separate "cleaved" and "noncleaved" phase of lymphocyte transformation in lymph node germinal centers ~s,39 is not supported by experimental evidence, 29,3~ and therefore is not necessarily reflected in such tumors as mixed-cell lymphoma. When populations of lymphocytes are separated into two subclasses on the basis of nuclear profile size in mixed-cell lymphoma, the small subset has a shallower mean invagination depth than the larger sized population (a reversal of what has been hypothesized), while the proportion of nuclear profiles bearing significant indentations is similar in these two subsets (Table 5). The similarity of the ratio of invagination d e p t h to mean nuclear diameter in the segregated populations in mixed-cell lymphoma (Table 5) indicates that as nuclear size increases, invaginations gradually deepen proportionally. Normal lymphocytes also reflect this structural alteration in nuclear form. 3~

TABLE 5. Lymphocyte Nuclear Parameters From Twelve Pooled Cases of Mixed Small- and Lar~le-Cell Lymphoma in Which Lymphocytes are Separated Into Small [< 20 i~m2] and Large [> 20 i~m~] Sets

Data Set Small Large

Mean Area

Mean Contour

Mean Invagination D e p t h (-+ SD)

Ratio Invagination Depth of Nuclear Diameter

No.

(Ixm2 + SD)

I n d e x (-+ SD)

Invaginated Profiles (%)

Cleaved Profiles (%)

1,117 844

13.1 - 4.0 32.4 + 13.0

4.56 -+ 0.61 4.63 -+ 0.66

0.73 -+ 0.57 1.06 -+ 0.76

0.17 0.16

72.2 83.4

14.6 10.4

899

HUMAN PATHOLOGY

Volume 19, No. 8 [August 1988]

30-

20o O.

10-

0-

3.6

4.0

4.4 4.8 5.2 Nuclear Contour Index

5.6

6.0

Histogram distribution of contour indices of small [<20 i~m2] and large [>20 t.~m2] nuclear profiles of lymphocytes in mixed lymphocytic-histiocytic lymphomas. Random samples [approximately 150 from each of the 12 cases] of morphometrically measured nuclear profiles were pooled [total sample = 1,961], and subsequently separated into the two subpopulations [small and large] before sorting individual nuclear profiles on the basis of their contour index. Note that the small [open bars] and large [solid bars] nuclei have identical distributions. This indicates that small lymphocytes do not have more irregularly shaped or indented [cleaved] nuclei than large lymphocytes. FIGURE 7.

c h a l l e n g e p a t h o l o g i s t s to a p p l y r i g o r o u s s c i e n t i f i c p r i n c i p l e s to u n d e r s t a n d i n g t h e s t r u c t u r a l aspects o f neoplastic lymphocytes in NHL.

Acknowledgment. Supported by a grant from the National Cancer Institute of Canada and the Moe Levin Fam~ ily Foundation. T h e authors appreciate the technical assistance of Diane Jeans and the manuscript preparation by Marilyn McDougall.

REFERENCES

1. Woodruff RD: Reviewing histologic diagnosis of lymphoma: Comparison of original and review diagnoses in 269 cases. Arch Pathol Lab Med 105:573, 1981 2. Velez-Garcia E, Durant J, Gams R, et ah Results of a uniform histopathologic review system of lymphoma cases. A ten-year study from the Southeastern Cancer Study Group. Cancer 52:675, 1983 3. Metter GE, Nathwani BN, Burke JS, et al: Morphological subclassification of follicular lymphoma: Variability of diagnoses among hematopathologists, a collaborative study between the repository center and pathology panel for lymphoma clinical studies. J Clin Oncol 3:25, 1985 4. Non-Hodgkin's lymphoma. Pathologic Classification Project. National Cancer Institute sponsored study of classification of non-Hodgkin's lymphomas. Summary and description of a working formulation for clinical use. Cancer 49:2112, 1982 5. Foucar K, Armitage JO, Dick FR: Malignant lymphoma, diffuse mixed small and large cell: A clinicopathologic study of 47 cases l Cancer 51:2090, 1983 6. Grogan TM, Hicks MJ, Jolley cs, et al: Identification of two major B cell forms of nodular mixed lymphoma. Lab Invest .51:504, 1984 7. Hall TL, Fu YS: Applications of quantitative microscopy in tumor pathology. Lab Invest 53:5, 1985 900

8. Nathwani BN, Metter GE, Gams RA, et ah Malignant lymphoma, mixed cell type, diffuse. Blood 62:200, 1983 9. Glick JH, Barnes JM, Ezdinli EZ, et ah Nodular mixed lymphoma: Results of a randomized trial failing to confirm prolonged disease-free survival with COPP chemotherapy. Blood 58:920, 1981 10. Mann RB, Berard CW: Criteria for the cytologic subclassification of follicular lymphomas: A proposed alternative method. Hematol Oncol 1:187, 1983 11. Levine GD, Dorfman RF: Nodular lymphoma: An ultrastructural study of its relationship to germinal centers and a correlation of light and electron microscopic findings. Cancer 35:148, 1975 12. Rilke F, Pilotti S, Carbone A, et al: Morphology of lympathic cells and their derived tumours. J Clin Pathol 31:1009, 1978 13. SaidJW, Hargreaves HK, Pinkus GS: Non-Hodgkin's lymphomas: An uhrastructural study correlating morphology with immunologic cell type. Cancer 44:504, 1979 14. Eiomoto T, Mitsui T, Kikuchi M: Uhrastructure of adult T-cell leukemia/lymphoma. Virchows Arch (Cell Pathol) 38:189, 1981 15. Abbott CR, Blewitt RW, Bird CC: Quantitative analysis of non-Hodgkin's lymphoma. J Clin Pathol 35:135, 1982 16. Crocker J, Jones EL, Curran RC: Study of nuclear diameters in non-Hodgkin's lymphomas. J Clin Pathol 35:954, 1982 17. Crocker J, Jones EL, Curran RC: Study of nuclear sizes in the centers of malignant and benign lymphoid follicles. J Clin Pathol 36:1332, 1983 18. van der Valk P. Mosch A. Kurver PJ, et ah Morphometric characterization of 52 B cell non-Hodgkin's lymphomas. J Clin Pathol 36:289. 1983 19. Tosi P. Cintorino M. Luzi P. et al: A morphometric semiautomated method for analyzing cell nuclei in lymph node sections from non-Hodgkin's lymphoma. Significance of data. Exp Path 24:237. 1983 20. Tosi P. Leoncini L. Spina D. et al: Morphometric nuclear analysis of lymphoid cells in center cell lymphomas and in reactive germinal centers. Am J Pathol 117:12. 1984 21. Zardawi IM. Firman DW, Sims TA: Stereological analysis of non-Hodgkin's lymphomas. Pathology 16:387. 1984 22 Dardick I. Dardick AM. Caldwell DR et al: NonHodgkin's lymphoma classification: Uhrastructural morphometric studies for the quantification of nuclear compartments in situ. HUM PATHOL 16:1047, 1985 23. Rappaport H. Winter WJ, Hicks EB: Follicular lymphomas: A re-evaluation of its position in the scheme of malignant lymphoma, based on a survey of 253 cases. Cancer 9:792, 1956 24. Rappaport H: Nomenclature and classification. Tumors of the hematopoletic system. In Atlas of Tumor Pathology, section III. fascicle 8. Washington DC. Armed Forces Institute of Pathology, 1966, p 10 25. Lennert K: Classification of non-Hodgkin's lymphomas. Malignant lymphomas other than Hodgkin's disease. Berlin, Springer-Verlag, 1978, p 83 26. Dardick I. Caldwell DR, McCaughey WTE. et al: Nuclear morphologic and morphometric analyses of large noncleaved cell and immunoblastic non-Hodgkin's lymphomas. HUM PATHOL 15:965, 1984 27. Schrek R: Ultrastructure of blood lymphocytes from chronic lymphocytic and lymphosarcoma cell leukemia. J Natl Cancer Inst 48:51, 1972 28. Dardick I, Dardick AM: Morphonietry of normal human lymphoid tissues: Nuclear parameters for comparative studies of lymphoma. Arch Pathol Lab Med 108:190, 1984 29. Dardick I, Setterfield G, Hall R, et al: Nuclear alterations during lymphocyte transformation. Relationship to the heterogeneous morphologic presentations of non-Hodgkin's lymphomas. AmJ Pathol 103:10, 1981 30. Dardick I, Sinnott NM, Hall R, et al: Nuclear morphology and morphometry of B lymphocyte transformation. Implications for follicular center cell lymphomas. Am J Pathol 111:35, 1983

MIXED-CELL LYMPHOMA (Dardick et al) 31. Setterfield G, Hall R, Bladon T, et al: Changes in structure and composition of lymphocyte nuclei during mitogenic stimulation. J Ultrastruct Res 82:264, 1983 32. Dardick I, CaldweU DR, Bailey DB, et ah Nuclear morphologic and morphometric analyses of nodular poorly differentiated lymphocytic lymphoma: Assessment of small cleaved nuclei. HUM PATHOL 16:1187, 1985 33. SAS Institute, Inc: SAS User's Guide: Statistics (ed 5). Cary, NC, SAS Institute, 1985 34. Gore SM: Assessing methods--Transforming the data. Br MedJ 283:548, 1981 35. Armitage P: Statistical Methods in Medical Research. Oxford, Blackwell, 1971 36. A1-Hamdani MM, Atkinson ME, Mayhew TM: Ultrastructural morphometry of blastogenesis. I: Transformation of small

lymphocytes stimulated in vivo with dinitrochlorobenzene. Cell Tissue Res 200:495, 1979 37. Marchevsky AM, Klapper E, GilJ: Computerized classification of nuclear profiles in non-Hodgkin's lymphomas. Am J Clin Pathol 87:561, 1987 38. Lukes RJ, Collins RD: Immunologic characterization of human malignant lymphomas. Cancer 34:1488, 1974 39. Lukes RJ, Collins RD: New approaches to the classification of the lymphomata. BrJ Cancer 31:1, 1975 (suppl II) 40. Dardick I, Caldwell DR: Follicular centre cell lymphoma: Morphological data relating to observer reproducibility. Cancer 58:2477, 1986 41. Payne CM, Hicks MJ, Kim A; Ultrastructural morphomettic analysis of normal human lymphocytes stimulated in vitro with mitogens and antigens. Am J Pathol 120:263, 1985 42. Cleary ML, Sklar J: DNA rearrangements in nonHodgkin's lymphoma. Cancer Surveys 4:331, 1985

901