Ultrastructural heterogeneity of lung carcinomas

Ultrastructural Heterogeneity of Lung Carcinomas: Representativity of Samples for Electron Microscopy in Tumor Classification WOLTER J. MOOI, MD, KOERTP. DINGEMANS, PHD, SJOERD Sc. WAGENAAR, MD, A. A. M. HART,PHD,AND C. A. WAGENVOORT, MD Histologie heterogeneity of tumors is a well-known phenomenon, which has been repeatedly studied at the light microscopic level. Electron microscopy has been advocated as an adjunct in classification of tumors which pose difficulties on light microscopic classification. However, in view of tumor heterogeneity, it might be anticipated that the problem of sample error could detract from the usefulness of electron microscopy in tumor typing. UItrastructural heterogeneity of tumors has thus far not been systematically investigated. We performed an ultrastructural randomized and blinded study of superficial and deep samples of 44 reseeted lung carcinomas, tumors which are notorious for their histologic heterogeneity. Neuroendocrine and squamous differentiation, as well as adenodifferentiation, were assessed separately and semiquantitatively in each sample. Twenty-six tumors showed more than one type of differentiation in at least one sample. However, in only two cases did the main type of differentiation differ between the two samples. A further nine cases showed one predominant differentiation type in both samples, but a similarly pronounced second differentiation type in one of the samples. Thus, in terms of ultrastructural diagnosis, the two samples showed a major discrepancy in two of 44 cases, and a minor discrepancy in nine of 44 cases. We conclude that ultrastructural heterogeneity o f lung tumors is a common occurrence, but that it only rarely leads to tota-lIy dlf,ferent ultrastruetural diagnoses. HUM PATHOL 2P. f227--1234. 9 1990 by W.B. Saunders Company.

Electron microscopy has gained increasing recognition as a valuable tool in histologic typing of tumors. 1,2 Uhrastructural detection of differentiation is often possible in tumors that appear "undifferentiated" light microscopically, thus allowing a more precise and accurate typing of these tumors. However, since electron microscopy is necessarily performed on the basis of very small samples of tissue, the question arises as to whether the ultrastructural data are representative for the whole tumor. From the Department of Pathology, University of Amsterdam, Amsterdam, The Netherlands; the Department of Pathology, St Antonius Ziekenhuis, Nieuwegein, The Netherlands; and the Department of Pathology, The Netherlands Cancer Institute, Amsterdam, The Netherlands. Accepted for publication February 9, 1990. Supported in part by Centocor Europe BV, Holland. Key words: lung carcinoma, ultrastructure, heterogeneity, classification. Address correspondence and reprint requests to Wolter J. Mooi, MD, DePartment of Pathology, The Netherlands Cancer Institute, Plesmanlaan 121, 1066 CX Amsterdam, The Netherlands. 9 1990 by W.B. Saunders Company. 0046-8177/90/2112-000855.00/0

The phenomenon of histologic heterogeneity of tumors, ie, the occurrence of different histologic types in different areas of a tumor, poses a potentially formidable problem to the diagnostic applicability of ultrastructural typing of tumors. This problem has been repeatedly mentioned in the literature, 3-5 but to our knowledge, it has never been the subject of a systematic, randomized study of different samples taken from the same tumors. Lung carcinomas are especially notorious for their histologic heterogeneity. This has been systematically investigated at the light microscopic level. 6-8 Varying with the extent of sampling and the definitions of heterogeneity used, the heterogeneous tumors constituted between 13% 7 and 66% s of the total. Heterogeneity of expression of various differentiation-related antigens in lung tumors was a prominent finding in an immunohistochemical study, 9 and it has even been suggested on the basis of immunohistochemical and ultrastructural data that all lung tumors are probably heterogeneous. I~ The superiority of electron microscopy over conventional light microscopy in the detection of differentiation in lung carcinomas has been well established. Ix-x4 A large proportion of lung carcinomas show more than one differentiation pattern ultrastructurally, even when this cannot be discerned at the light microscopic level. 1~ To investigate ultrastructural tumor heterogeneity systematically, we have collected superficial and deep samples of widely distributed tumor areas of 44 resected lung carcinomas. These 88 samples were randomized, and in each the presence and prominence of neuroendocrine and squamous differentiation, as well as adenodifferentiation, were assessed semiquantitatively on electron microscopy. On the basis of comparison of the findings in each pair of sampies, an estimate of the ultrastructural heterogeneity of these tumors was obtained.

MATERIALSAND METHODS Forty-four resection specimens of primary lung carcinomas were investigated. The tumors were classified light microscopically using World Health Orgainzation criteria ~6 as squamous carcinomas (17 cases), adenocarcinomas (14

1227

HUMAN PATHOLOGY

Volume 21, No. 12 [December 1990)

cases), large cell carcinomas (four cases), and small cell carcinomas (nine cases). The series of 44 tumors was not a consecutive series, but availability of well-preserved material for electron microscopy, rather than histologic considerations, formed the criterion for inclusion in this study. From each tumor, a sample from the outermost rim and a second sample located at a distance from the first sample and deeper within the tumor (usually 5 to 10 mm from the tumor-lung interface) were fixed in Karnovsky's fixative and processed for electron microscopy. The total of 88 grids with ultrathin sections were randomized and examined by two of the authors (W.J.M. and K.P.D.) who were not aware of the light microscopic diagnoses, which two samples formed a pair taken from the same tumor, or which samples were "superficial" or "deep" in location. The degree of neuroendocrine differentiation (N), squamous differentiation (S), and adenodifferentiation (A) of the tumor cells was assessed in each specimen. The ultrastructural differentiation criteria were largely based on the work of McDowell and associates. ]] For each of the differentiations, a score ranging from 0 to 3 was given: 0 indicated absence of the differentiation, 1 indicated small traces, 2 indicated moderate development, and 3 indicated marked development of the differentiation ultrastructurally. In detail, criteria for each score were as follows: Neuroendocrine differentiation: 0, no dense core granules found; 1, occasional dense core granules present in rare cells; 2, dense core granules regularly found, but in a minority of cells; and 3, dense core granules frequent in most cells. Squamous differentiation: 0, no tonofibrils, desmosomes small; 1, occasional small tonofibrils, often attached to desmosomes which are small or moderately sized; 2, large tonofibrils or small but frequent tonofibrils, prominent desmosomes, alternatively, distinct intercellular bridging (not necessarily associated with prominent desmosomes); and 3, large and frequent tonofibrils, prominent desmosomes and/or keratinization, intercellular bridging may be absent. Adenodifferentiation: 0, no intercellular or intracellular lumina, inconspicuous endoplasmic reticulum, Golgi complex, and microvilli, no granules identifiable with certainty as exocrine granules; 1, rare lumen formation, or more organelle-rich cytoplasm (especially rough endoplasmic reticulum and Golgi complex), rare secretory granules, microvilli regularly seen; 2, intercellular or intracellular lumina lined by microvilli regularly seen, organelle-rich cytoplasm, small or moderate numbers of secretory granules; and 3, extensive lumen formation or papillary structures, abundant microvilli, alternatively, organelle-rich cytoplasm with abundant exocrine granules. All 88 samples were scored, yielding for each sample a concise code, eg, N0-S3-A0. The ultrastructural diagnosis was based on the highest score of a sample or on the two highest scores when these were equal (eg, N0-S3-A2 = squamous carcinoma; N0-S2-A2 = adenosquamous carcinoma). Thus, some of the information on the score was lost in the ultrastructural diagnosis. For each of the differentiations, a coefficient of consistency between the two samples' K was assessed. 17 This coefficient equals 1 when identical scores are found in each pair of samples, and it equals 0 when the scores in each pair are independent from each other. Whether there was a systematic difference in scores between superficial and deep samples was investigated using the Sign test (which assesses the frequency with which a superficial sample is assigned a higher or lower score than

the deep sample) and the Wilcoxon matched pairs signed ranks test (which assesses whether the average score of all superficial samples differs significantly from the average score of all deep samples18). Both tests yield a P value which indicates statistical significance when P < 0.05. Discrepancies between light and electron microscopic diagnoses were also investigated. A discrepancy was considered to exist when the predominant differentiation in at least one of the two samples differed from the light microscpic diagnosis of the tumor. Clinical follow-up data were collected when there was absence of neuroendocrine granules in tumors considered small cell carcinomas light microscopically or when neuroendocrine granules were present in tumors considered to be non-small cell carcinomas.

RESULTS T h e results o f the u l t r a s t r u c t u r a l scoring o f all 88 samples are given in T a b l e 1. T h e scoring on the basis o f t h e c r i t e r i a g i v e n p r o v e d to be s u r p r i s i n g l y straightforward, a n d only rarely did the two investigators not i m m e d i a t e l y r e a c h a g r e e m e n t . E x a m p l e s o f various samples and their scores are given in Figs 1 t h r o u g h 5. T h e only serious p r o b l e m e n c o u n t e r e d was the evaluation o f small e l e c t r o n - d e n s e granules e n c o u n t e r e d in six samples; in s o m e o f these the granules strongly r e s e m b l e d n e u r o e n d o c r i n e granules, but they w e r e located in the apical r a t h e r t h a n basal cytoplasm o f t u m o r cells with obvious adenodifferentiation, suggesting that they were exocrine granules (Fig 6). I n o t h e r samples, they were less regular in size a n d s h a p e t h a n those normally seen in small cell l u n g carcinomas. I n six samples, a score o f "N = ?" was given. In two o f the f o u r t u m o r s in which this p r o b l e m arose the o t h e r s a m p l e gave unequivocal results, but in the o t h e r two t u m o r s the same p r o b l e m was e n c o u n t e r e d again. T h e s e f o u r t u m o r s could not be included as p a r t o f the statistical analysis o f the results, a n d will t h e r e f o r e be discussed separately below. In only o n e t u m o r (case no. 17) were unmistakable n e u r o e n d o c r i n e granules p r e s e n t in one sample, but absent in the other. I n 26 o f 40 tumors, m o r e t h a n o n e differentiation was seen in at least o n e s a m p l e (eg, N0-S2-A1). T h e r e was no systematic d i f f e r e n c e in p r o m i n e n c e o f the three differentiations b e t w e e n the superficial a n d the d e e p s a m p l e (Sign test: P N = 1.00, P S = 0.42, P A = 0.61; Wilcoxon m a t c h e d pairs signed r a n k s test: P N = 0.50, P S = 0.41, P A = 0.55). C o m p a r i s o n o f the scores o f superficial a n d d e e p samples is p r o v i d e d in T a b l e 2. T h e same score in superficial a n d d e e p samples was p r e s e n t in the following n u m b e r s o f tumors: n e u r o e n d o c r i n e differentiation: 35 o f 40 cases (88%, K = 0.62); s q u a m o u s differentiation: 30 o f 44 cases (68%, K = 0.56); a n d a d e n o d i f f e r e n t i a t i o n : 29 o f 44 cases (66%, K = 0.52). T h e p r e d o m i n a n t d i f f e r e n t i a t i o n was the s a m e in b o t h samples in 29 of 40 t u m o r s (73%, K = 0.62; see also T a b l e 3). O f the r e m a i n i n g cases, nine showed o n e p r e d o m i n a n t differentiation in b o t h samples, but a second, equally p r e d o m i n a n t differentiation in o n e

1228

HETEROGENEITYOF LUNG CARCINOMAS [Mooi et al]

TABLE 1. Ultrastructural Assessment of Differentiation of

Superficial and Deep Samples of 44 Lung Carcinomas Electron Microscopy Case No.

Superficial Light Microscopy

Deep

N

S

A

N

S

A

0 0 0 0 0 2 0 0 0 0 0 0 0 0 0 0 2

3 2 3 2 2 1 1 1 3 2 3 0 2 1 3 2 0

0 0 0 0 1 0 1 1 0 0 0 1 0 2 0 1 2

0 0 0 0 0 3 0 0 0 0 0 0 0 0 0 0 0

2 2 3 1 3 1 2 1 3 2 3 0 2 2 2 2 1

0 0 0 0 0 0 2 1 0 0 0 1 0 2 2 2 1

0 0 ? 0 0 0 0 0 0 0 0 0 0 0

0 2 0 0 0 1 0 1 2 2 1 0 2 1

3 1 2 1 3 3 3 3 3 1 3 3 1 3

0 0 ? 0 0 0 0 0 0 0 0 0 0 0

1 2 0 0 0 1 0 1 1 2 1 0 0 2

3 1 2 1 3 3 3 3 3 0 3 3 2 1

0 0 0 ?

0 1 1 1

1 1 2 2

0 0 0 0

0 2 1 3

1 0 1 1

0 3 2 3 ? 2 ? 2 1

2 0 1 0 1 0 1 0 1

1 0 0 0 0 0 1 0 0

0 3 2 3 ? 1 1 1 2

2 0 1 0 1 0 1 1 1

0 0 1 0 0 1 0 0 0

Squamous carcinoma 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 Adenocarcinoma 18 19 20 21 22 23 24 25 26 27 28 29 30 31 Large cell c a r c i n o m a 32 33 34 35

In eight cases, a discrepancy was f o u n d between the light microscopic and ultrastructural diagnosis of at least one sample (cases no. 6, 14, 17, 19, 27, 30, 31, and 36; see Table 1). It is of interest that four of these tumors showed the same discrepancy with the light microscopic diagnosis in b o t h samples. However, these discrepancies were usually not absolute, since the differentiation type corresponding to the light microscopic diagnosis was present in all but one of these cases (albeit less p r o n o u n c e d than the predominant ultrastructural differentiation type). In two non-small cell carcinomas (cases no. 6 and 17), n e u r o e n d o c r i n e granules were detected. Case no. 6 was diagnosed light microscopically as a poorly differentiated squamous carcinoma. Ultrastructurally, squamous differentiation was found, but neuroendocrine granules were also present in both samples. T h e patient, a 64-year-old male, remains well and without clinical evidence of disease 21/2 years after lobectomy for the lung tumor, which was staged as T1-N1-M0. Case no. 17 was diagnosed light microscopically as a moderately differentiated squamous carcinoma. Ultrastructurally, all three differentiations were found, including n e u r o e n d o c r i n e differentiation in one of the two samples. T h e patient, a 67-year-old male, u n d e r w e n t a lobectomy for a T2N l-M0 tumor; this patient has been lost to follow-up. Case no. 36 was diagnosed light microscopically as a small cell carcinoma, but in both untrastructural samples n e u r o e n d o c r i n e granules were absent. However, in b o t h samples, s q u a m o u s d i f f e r e n t i a t i o n was present, notably intercellular bridging, which was subtle but unmistakable ultrastructurally. T h e patient, a 69-year-old man, is well and without clinical evidence of disease 21/2 years after p n e u m o n e c t o m y for the lung tumor, which was staged as T1-N2-M0.

Small cell c a r c i n o m a 36 37 38 39 40 41 42 43 44

sample (eg, case no. 14:N0-S1-A2 versus N0-S2-A2). For the purpose of this study, such a discrepancy will be t e r m e d " m i n o r ultrastructural heterogeneity". T h e two remaining cases showed different predomin a n t differentiations in the two samples (cases no. 30 and 31, see Table 1). This will be termed "major ultrastructural heterogeneity". Interestingly, all cases which showed ultrastructural heterogeneity belonged to the 26 tumors which had more than one differentiation in at least one sample; that is to say, no cases were encountered with, for example, pure squamous differentiation in one sample and pure adenodifferentiation in the other.

DISCUSSION

To o u r knowledge, this is the first systematic semiquantitative study o f t u m o r heterogeneity at the ultrastructural level. We chose lung carcinomas for our investigation since histologic heterogeneity of these tumors is c o m m o n and has been well documented at the light microscopic level. 4-6 O u r objective was to assess to what extent such heterogeneity interferes with ultrastructural lung t u m o r classification, which is normally p e r f o r m e d on the basis of one small sample. At this point, it should be recognized that the term "heterogeneity', has been used in different ways: some have used it, like us, to indicate differences between different areas of the same tumor, ~ whereas others have used the term to indicate the presence of more t h a n one type of differentiation occurring in the same t u m o r cells, s Some have even used the term "heterogeneity" to indicate differences between different tumors of the same (World Health Organization-defined) subclass of t u m o r s ? 7 T h e problem o f

1229

HUMAN PATHOLOGY

Volume 21, No. 12 [December 1990)

FIGURE 1. Neuroendocrine differentiation. [Left] Small, electron-dense neuroendocrine granules are seen in practically all tumor cells. Therefore, the n e u r o e n d o crine differentiation of this sample was scored as N3. [Electton micrograph, magnification x 5,500.] [Right] Small cytoplasmic extension in which dozens of densecore granules are present. Some of these are slightly irregular in shape, but their regular size, a n d the fact that they are located in a cytoplasmic extension, argue strongly in favor of a n e u r o e n d o c r i n e nature of these granules. Neuroendocrine granules were common in this sample. Score: N3. [Electron m i c r o g r a p h , magnification x 19,500.]

sample error in surgical pathology is the consequence of the first of these three phenomena. From a diagnostic point of view, the results given above can be rearranged as follows: 41 samples showed only one type of differentiation; 36 of these were from "homogeneous" tumors. Thus, when only one differentiation was seen, the chance that a second sample would show the same predominant differentiation was 36 in 41, ie, about 90%. The other 39 samples showed more than one type of differentia.tion; only 22 of these were from "homogeneous" tu-

mors. Consequently, when more than one type of differentiation was seen in one sample, the chance that a second sample showed the same predominant differentiation fell to 22 in 39, ie, roughly 55%. This indicates that, on the basis of one sample, the investigator can obtain a rough idea as to what the chances are that a second sample will yield the same diagnosis. The main conclusion from our figures is that ultrastructural lung tumor typing yields a consistent diagnosis between two widely separated samples in about three quarters of cases, and only minor discrepancies

FIGURE 2. Squamous differentiation. Prominent desmosomes are present between tumor cells, and tonofibrils are seen attached to the desmosomes and also d e e p e r within t h e cytoplasm. The desmosomes are located on small but ultrastructurally distinct intercellular bridges. Score: $2. [Electron micrograph, magnification x 16,732.]

1230

HETEROGENEIIYOF LUNG CARCINOMAS [Mooi et al]

FIGURE 3. Squamous differentiation. [Left] Large desmosomes with tonofibrils attached to them, and running in the cytoplasm of lwo adjacent tumor cells. No intercellular bridging is present, but the desmosomes are prominent enough to warrant a score of $3. [Electron micrograph, magnification x 67,300.] [Right] Cell membrane with thickening of the cytoplasmic side, indicative of keratinization.All samples in which keratinized tumor cells were seen were scored as $3. [Electron micrograph, magnification x 86,800.]

FIGURE 4. Adenodifferentiation. [Left] Tumor cells show organelle-rich cytoplasm with abundant RER and many electron-dense secretory granules, too large and unequal to represent neuroendocrine granules. Two small intercellular lumina with microvitli are visible [top and bottom]. The abundance of rough endoplasmic reticulum and secretory granules warrants a score of A3. [Electron micrograph, magnification x 7,100.] [Right] Tumor cells with abundant mucin granules which characteristically become confluent and which are extruded into an intercellular lumen lined by microvilli. Score: A3. [Electron micrograph, magnification x 7,100.]

1231

HUMAN PATHOLOGY

Volume 21, No. 12 [December 1990)

FIGURE 5. Combined squamous a n d adenodifferentiation. Two a d j a c e n t tumor cells are separated by a slit-like lumen in which numerous microvilli project. Both t u m o r cells contain much endoplasmic reticulum which is partly dilated. Numberous tonofibrils are also present in both tumor cells. Scores: $I, A2. Light microscopically, neither of the two differentiations was noted, the tumor being classified as a large cell carcinoma [case no. 34). [Electron micrograph, magnification x 15,200.]

in almost all o f the r e m a i n i n g tumors; therefore, the information obtained ultrastructurally holds true for other parts o f the tumors in the large majority of cases. O f particular interest f r o m a diagnostic point o f view is the assessment o f the presence or absence o f n e u r o e n d o c r i n e granules in lung carcinomas. ~4,1~ In the present study, heterogeneity, in this respect, was f o u n d in one case (case no. 17:N2-S0-A2 versus N0S1-A1). U n f o r t u n a t e l y , no clinical follow-up data have been retrievable in this interesting case. In f o u r tumors, difficulties were e n c o u n t e r e d in the assessm e n t o f the n a t u r e o f intracytoplasmic granules, and

in these cases we were unable to reach a verdict as to w h e t h e r n e u r o e n d o c r i n e granules were present or absent. Consequently, it should be recognized that occasionally, even in well-preserved lung t u m o r tissue, it may be impossible to d e t e r m i n e w h e t h e r the t u m o r has n e u r o e n d o c r i n e granules or not. However, in the large majority o f samples (82 out o f 88) we did not have problems in this respect. It has b e c o m e increasingly a p p a r e n t that some non-small cell lung carcinomas may show n e u r o e n d o crine granules on electron m i c r o s c o p y ? 2,2~ T h e s e tumors may show light- and electron microscopic fea-

FIGURE 6. Small intracytoplasmic granules resembling neuroendocrine granules, but l o c a t e d in the apical cytoplasm of a tumor cell. Note the close proximity of an intercellular lumen in which microvilli project. The location of these granules argues strongly against a neuroendocrine nature. In this p r o b l e m case [case no. 20), a score of "N = .~' was given, since w e felt that no conclusive assessment of the presence or absence of n e u r o e n d o c r i n e differentiation was possible. [Electron micrograph, magnification x 41,325.]

1232

HETEROGENEI[YOF LUNGCARCINOMAS[Mooi et al]

TABLE 2. Ultrastructural Scores of Superficial and Deep Samples of 44 Lung Carcinomas Deep Sample Differentiation

0

1

2

3

Total

N e u r o e n d o c r i n e differentiation Superficial sample (K = 0.62) 0

32

--

--

--

32

--

--

1

2 3 Total Squamous differentiation Superficial sample (K = 0.56) 0

1

2

1

33

2 3

10

1

2 3 Total Adenodifferentiation Superficial sample (K = 0.52) 0 1 2 3 Total

1

--

1

2

1 2 3

5 2 40

--

--

--

10

4

1

1 -11

2 -15

7 2 13

1 3 5

13 15 11 5 44

14 5 --19

2 5 3 1 11

1 3 2 -6

---8 8

17 13 5 9 44

NOTE: Scores of "N = ?" have been omitted (see text).

tures of squamous cacinoma, adenocarcinoma, or large cell carcinoma. They have been tentatively named "atypical endocrine tumors". 12 The clinical course of these tumors remains insufficiently known. Case no. 6 seems to fit this category; follow-up data in this case (21/2-year disease-free survival after resection of a T1-N1-M0 lung tumor) seem to indicate a less virulent behavior than most small cell carcinomas, but no firm conclusions can be drawn from this single case. Case no. 36 is of particular interest. This tumor was considered small cell carcinoma light microscopically, but no neuroendocrine granules were present, both samples showing predominantly squamous differentiation, with subtle but distinct intercellular bridging. The follow-up data of the patient indicate a

TABLE 3.

Electron Microscopic Diagnoses of Superficial Versus Deep Specimen Deep Sample

Superficial Sample

N

S

A

NS

NA

SA

Total

N S A NS NA SA Total

4 --1 . -5

-12 1 .

-1 11

1 --.

1 ---

-2 2

-1

1 2 7

6 15 14 1 1 3 40

.

. .

1 14

. . -12

.

. -1

NOTE: Scores o f "N = ?" have been omitted (see text). Abbreviations: N, highest score for n e u r o e n d o c r i n e differentiation; S, highest score for squamous differentiation; A, highest score for adenodifferentiation; NS, equal and highest scores for n e u r o e n d o c r i n e and squamous differentiation; NA, equal and highest scores for n e u r o e n d o c r i n e and adenodifferentiation; SA, equal and highest scores for squamous and adenodifferentiation.

less virulent behavior than a typical small cell carcinoma. The patient is free of clinical disease 21/2 years after pneumonectomy for a T1-N2-M0 tumor, which would be highly unusual for a small cell tumor of this stage. This case is similar to three cases previously reported by us.14 We would certainly caution against making a diagnosis of small cell carcinoma when the tumor definitely has no neuroendocrine granules. In our experience, lung carcinomas with the classic light microscopic histology of small cell carcinoma, and with compatible clinical course, have always contained neuroendocrine granules, provided sufficient wellpreserved material was available. I4 In conclusion, the co-occurrence of two or three differentiations is common in lung carcinomas/and shifts in prominence of these differentiations are also often seen. When they are sufficiently great, such shifts lead to histologic heterogeneity, which we found in only a minority of cases. However, it should be realized that the figures found for such heterogeneity depend very much on the criteria used and the amount of tissue sampled to assess such heterogeneity. When one ultrastructural sample shows only one type of differentiation, a second sample will probably yield the same diagnosis. Thus, used cautiously, ultrastructural assessment of differentiation may play a significant role in lung tumor typing, and fears of excessive sample error problems in ultrastructural tumor typing have not been confirmed by this first systematic semiquantitative study on the subject,

Acknowledgment. T h e e x p e r t t e c h n i c a l a s s i s t a n c e o f M.A. van den Bergh Weerman, University of Amsterdam, T h e N e t h e r l a n d s ; a n d E. V a a r k a m p , St A n t o n i u s Z i e k e n h u i s , N i e u w e g e i n , T h e N e t h e r l a n d s , is g r a t e f u l l y a c k n o w l e d g e d . REFERENCES 1. Ghadially GN: Diagnostic Electron Microscopy o f Tumours. London, England, Butterworths, 1980 2. H e n d e r s o n DW, Papadimitriou JM: Ultrastructural Appearances o f T u m o u r s . A Diagnostic Atlas (ed 2). Edinburgh, Scotland, Churchill Livingstone, 1986 3. Osborn M, Weber K: Biology o f disease. T u m o r diagnosis by intermediate filament typing: A novel tool for the surgical pathologist. Lab Invest 48:372-384, 1983 4. Erlandson RA: Diagnostic ilffmunohistochemistry o f ihum a n tumors: A n interim evaluation. A m J Surg Pathol 8:615-624, 1984 (editorial) 5. Huszar M, Suster S, Herczeg E, et ah Sclerosing hemangioma o f the lung: Immunohistochemical d e m o n s t r a t i o n o f mesenchymal origin using antibodies to tissue-specific intermediate filaments. Cancer 58:2422-2427, 1986 6. Reid JD, Carr AH: T h e validity and value o f histological and cytological classifications o f lung cancer. Cancer 14:673-677, 1961 7. Hirsch FR, Ottesen G, P o d e n p h a n t J, et al: T u m o r heterogeneity in lung cancer based on light microscopic features. A retrospective study o f a consecutive series o f 200 patients, treated surgically. Virch Arch [A] 402:147-156, 1983 8. Roggli VL, Vollme r RT, G r e e n b e r g SD, et al: Lung cancer heterogeneity: A blinded and r a n d o m i z e d study of 100 consecutive cases. HUM PATHOL 16:569-579, 1985 9. Gatter KC, Dunnill MS, Pulford KAF, et ah H u m a n lung tumours: A correlation o f antigenic profile with histological type. Histopathol 9:805-823, 1985

1233

HUMAN PATHOLOGY

Volume 21, No. 12 [December 1990]

10. Dunnill MS, Gatter KC: Ceilular heterogeneity in lung cancer. Histopathology 10:461-475, 1986 1I. McDowell EM, McLaughlin JS, Merenyl DK, et ah The respiratory epithelium. V. HistOgenesis of lung carcinomas in the human. JNCI 61:587-606, 1978 12. McDowell EM, Wilson TS, Trump BF: Atypical endocrine tumors of the lung. Arch Pathol Lab Med 105:20-28, 1981 13. Horie A, Ohta M: Ultrastructural features of large cell carcinoma of the lung with reference to the prognosis of patients. HuM PATHOL 12:423-432, 1981 14. Mooi WJ, Dingemans KP, van Zandwijk N: Prevalence of neuroendocrine granules in small cell lung carcinoma: Usefulness of electron microscopy in lung cancer classification. J Pathol 149:41-48, 1986

15. Dingemans KP, Mooi WJ: Ultrastructure of squamous cell carcinoma of the lung. Pathol Annu 19:249-273, 1984 16. World Health Organization: Histological Typing of Lung Tumours (ed 2). Geneva, Switzerland, World Health Organization, 1981 17. Cohen J: A coefficient of agreement for nominal scales. Appl Psychol Measurem 20:37-46, 1960 18. Armitage P: Statistical Methods in Medical Research. Oxford, England, Blackwell, 1971 19. Mooi WJ, van Zandwijk N, Dingemans KP, et ah The "grey area" between small cell and non-small cell lung carcinomas. Light and electron microscopy versus clinical data in 14 cases. J Pathol 149:49-54, 1986 20. Neal MH, Kosinski R, Cohen P, et ah Atypical endocrine tumors of the lung: A histologic, ultrastructural, and clinical study of 19 cases. Ht~M PATHOL17:1264-1277, 1986

1234