In most cases of small-cell lung carcinomas (SCLC) phenotypic features compatible with a neuroendo- crine differentiation status can be identified by mono-.
Clinical Characterization of Non-Small-Cell Lung Cancer Tumors Showing Neuroendocrine Differentiation Features By Hendrikus H. Berendsen, Lou de Leii, Sibrand Poppema, Pieter E. Postmus, Adriane Boes, Henk J. Sluiter, and Hauw The In most cases of small-cell lung carcinomas (SCLC) phenotypic features compatible with a neuroendocrine differentiation status can be identified by monoclonal (MOC) antibody-based immunohistological procedures. Similar features can be recognized only in a minority of non-SCLC tumors. During a period of 30 months, all diagnostic non-SCLC biopsies (141 cases) were prospectively analysed for the presence of markers indicative for neuroendocrine differentiation. In 31% of all cases, such a presence could be noticed. Neuroendocrine differentiation (50%to 100% positivestaining tumor cells) was recognized more frequently
SMALL-CELL
LUNG carcinoma (SCLC) is characterized by early and widespread metastasis. Therefore, the possibility of a curative resection is extremely rare. Most patients, however, benefit from polychemotherapeutic treatment.' Squamous-cell, adeno-, large-cell, and adenosquamous carcinomas of the lung are often referred to as non-SCLC. In this group, a resection, which may turn out to be curative, can be performed in a selected group of patients. 2 Only a subgroup of non-SCLC patients benefits from chemotherapeutic treatment. 3 It is not possible to identify this subgroup by normal histopathological procedures. The presence of several neuroendocrine differentiation-related antigens can be recognized in the majority of SCLC biopsy specimens,4 whereas most non-SCLC specimens lack such markers.5 In this study, we investigated whether the reactivity with a series of monoclonal (MOC) antibodies directed against neuroendocrine differentiation antigens could indicate a subgroup within the non-SCLC group and whether the From the Departmentsof ClinicalImmunology, Pulmonology, and Pathology, University Hospital, Groningen, The Netherlands. Submitted October 6, 1988; accepted June 9, 1989. Supported by Koningin Wilhelmina Fonds Grant GUKC 85-6. Address reprintrequests to H.H. Berendsen, MD, Department of ClinicalImmunology, University Hospital,Oostersingel 59, 9713 EZ Groningen,The Netherlands. © 1989 by American Society of ClinicalOncology. 0732-183X/89/0711-0010$3.00/0 1614
in adenocarcinoma when compared to large-cell and squamous-cell carcinoma (X1 = 9.31, 2 degrees of freedom, P < 0.01). To investigate whether the clinical behavior of these "neuroendocrine" non-SCLC cases mimics SCLC, a multivariate analysis for prognostic factors was performed. Among other prognostic factors, biopsies containing more than 50% positivestaining tumor cells with the MOC antibody-i (MOC1) were recognized as negative prognostic factors. J Clin Oncol 7:1614-1620. © 1989 by American Society of Clinical Oncology.
clinical parameters obtained from this group resembled that of SCLC. MATERIALS AND METHODS Patients Biopsy specimens were collected from 141 untreated nonSCLC patients (January 1985 to June 1987). The specimens were obtained during clinical diagnostic work-up or exploratory thoracotomy. Clinical diagnostic staging included medical history, physical examination, x-rays, and computed tomography of the chest, bronchoscopy, and mediastinoscopy, if N2 disease was suspected. Additional staging was only performed if clinical signs or results of routine laboratory investigations were suggestive for distant metastasis. If clinical diagnostic staging revealed no histological evidence of inoperability, exploratory thoracotomy was performed. According to the American Joint Committee for Cancer Staging and End Result Reporting,6 in each case clinical diagnostic staging, or in cases of exploratory thoracotomy, postsurgical pathologic staging was performed. The series included 111 primary tumor biopsies obtained on thoracotomy or bronchoscopy, and 30 extirpated regional lymph nodes and distant metastasis. Histopathological evaluation according to World Health Organization (WHO) criteria7 revealed 86 squamous-cell carcinomas, 40 adenocarcinomas, 14 large-cell carcinomas, and one adenosquamous carcinoma. Cases showing combined histologies with SCLC admixtures were not included in this study.
Immunohistochemistry One pretreatment specimen was included from every patient for immunohistochemistry. If in a certain patient both primary and metastatic tumor biopsies were available, the primary tumor biopsy was included in the study. Standard light microscopical evaluation and immunohistochemical staining were simultaneously performed on one single biopsy specimen according to a recently described
Journalof Clinical Oncology, Vol 7, No 11 (November), 1989: pp 1614-1620
RECOGNITION OF POOR-PROGNOSIS SUBGROUP procedure.8 In short, biopsies were placed in a fixative according to McLean and Nakane.' After fixation, the specimens were divided into two portions. One portion was rinsed and kept in the sodium phosphate buffer to wash out the fixative for at least 12 hours. This portion was subsequently embedded in optimal cutting temperature (OCT) compound (Miles Laboratories, Inc, Naperville, IL) and snap frozen in liquid freon and kept at -80 0C until use. Cryostat sectioning and immunostaining of the specimen were performed according to described methods.'" The other portion was further fixated in formalin and embedded in plastic" for normal histopathological work-up and light microscopical evaluation.
MOC Antibodies A series of MOC antibodies (MOC-series: MOC-I, MOC21, MOC-32, MOC-51, and MOC-52) reactive with SCLCassociated antigens have been produced and characterized.12"1 These antibodies, except for MOC-51, have been screened in both our laboratory and that of participants of the First International Workshop on Monoclonal Antibodies directed against SCLC. 4 During this workshop, MOC-1, MOC-21, MOC-32, and MOC-52 were assigned to SCLC-CD I or CDla. These clusters comprise a group of MOC antibodies ""'57reactive with normal and malignant neuroendocrine tissues including SCLC. MOC-51, which has not been clustered yet, also recognizes a neuroendocrine-related differentiation antigen." 2 The epitopes recognized by MOC-I, MOC-51, and MOC-52 are destroyed during formalin fixation and subsequently performed paraffin or plastic embedding. Although all SCLC-CD1 antigens may be closely interrelated, the antigens recognized by the different MOC antibodies are clearly different from one another (Table 1). Reactivity of the MOC antibodies on the section was scored using three categories of reactivity: no reactivity versus less than 10%, 10% to 50% and 50% to 100% positively staining tumor cells.
StatisticalAnalysis For statistical evaluation of prognostic effect of different variables on patient survival, the methods of semiparametric survival analysis were used (ie, log-rank tests and Cox model regression analysis"). The P-values below 5% were considered as significant.
RESULTS
Tumor phenotype was analyzed in 141 cases of non-SCLC by a MOC antibody-based immunoTable 1. Characteristics of Monoclonal Antibodies Used
Moab
Cluster Designation
Ig Subclass
Apparent Molecular Weight of Matching Antigen
MOC-1 MOC-21 MOC-32 MOC-52
CD I CD 1" CD 1' CD 1
IgG, IgG,o IgM IgG1
125 kDo 25 kDa 40 kDa -*
Abbreviation: Ig, immunoglobin. *Not detectable by immunoprecipitation.
1615
histochemical staining procedure. The observed staining with different MOC antibodies directed against neuroendocrine differentiation antigens in all 141 cases is summarized in Fig 1. A minority of non-SCLC tumor specimens (31%) contained tumor cells expressing one or more of the assessed neuroendocrine markers. The results of the different subgroups are shown in Fig 1. Considering any positive staining with one or more of the MOC antibodies, there was no significant difference between squamous-cell carcinoma (26.7%), adenocarcinoma (43.6%), and large-cell undifferentiated carcinoma (21.4%) (x 2 = 4.22, 2df, P < 0.2). However, if the positively staining specimens are divided into categories according to the amount of positively staining cells, significant differences can be appreciated between the various histopathological classifications, ie, if the specimens exhibiting only focal (< 10%) staining are excluded (X2 = 8.92, 2df, P < 0.02) or if specimens exhibiting less than 50% positively staining cells are excluded (X2 = 9.31, 2df, P < 0.01), then it is apparent that the high reactivity categories (10% to 100% of the cells staining positively) are present more frequently in adenocarcinoma. In an attempt to correlate clinical data with the expression of neuroendocrine markers in the biopsy specimens several clinical subclassifications have been made. First, a comparison has been made between the staining results obtained in primary tumor biopsies and those obtained in biopsies from distant metastasis and lymph nodes (Fig 1). No significant differences could be detected between these categories (one or more positively staining MOCs, X2 = 0.64, ldf, P < 0.5; < 10% staining excluded, X2 = 2.39, ldf, P < 0.9; < 50% staining excluded, x2 = 3.46, ldf, P < 0.1). Second, to assess whether the expression of neuroendocrine markers may be considered as a negative prognostic factor, a multivariate analysis was performed with regards to a series of possible prognostic factors. This series included age, stage, conventional histopathological classification, resection or not, amount of resected lung tissue (bi)lobectomie versus pneumectomie, and the immunohistochemical data (< 10%, 10% to 50%, and > 50% positive staining for the MOC1, MOC-21, MOC-32, MOC-51, and MOC-52 alone and the cumulative scores; Fig 1).
BERENDSEN ET AL
1616 -50
so
4
-0 5 S
ZO
ci a
IO
is
L r
U
AOC-I 1M-21
c0
MOC-I HOC-21 eC-32
MOC-32 MOC-51 MOC-52
All specimens (n=141)
c
so
b
SO
"
5il 40
0D0-51
OiC-52
c1
Primary tumour specimens (n=lll)
50
-
50
0
"-e
..
10
10
IOC-21 NOC-32
OC-51 IOC-52
u
h
Y II · OC-32 ·MOC-51 ·MOC-52 "
"
HElC-1
rýVA
J10
20-
MHOC-1 OC-21
0C-I
cu
Adenocarcinoma (n=40)
Squamous cell carcinoma (n=B6)
NMC-21I O-32
MOC-51] OC-52
c
Large cell undifferentiated carcinoma (n=14)
>50% Positive-staining tumor cells
[] >10% Positive-staining tumor cells 0 []Any positive staining (including
