- Email: [email protected]
Proliferating cell nuclear antigen may be superior to argyrophilic nucleolar organizer regions in predicting shortened survival of patients with non-small cell lung cancer
Surgical Oncology 1995; 4: 83-89
Proliferating cell nuclear antigen may be superior to argyrophilic nucleolar organizer regions in predicting shortened survival of patients with non-small cell lung cancer T. OYAMA, T. MITSUDOMI, T. MIZOUE,* A. OHGAMI, T. OSAKI, R. NAKANISHI AND K. YASUMOTO Departments of Surgery /I and *Clinical Epidemiology, University of Occupational and Environmental Health, School of Medicine, Kitakyushu 807, Japan
We examined proliferating cell nuclear antigen (PCNA) in 102 patients with surgically treated non-small cell lung cancer (NSCLC). PCNA labelling index (L1) tended to be higher in tumours of higher stages than those of early stages, in squamous cell carcinomas than adenocarcinomas, or in poorly differentiated adenocarcinoma than in well-differentiated. A positive correlation was observed between the PCNA LI and argyrophilic nucleolar organizer regions (Ag-NOR) count which we previously examined (r =0.31, P =0.002). In survival analysis of 79 patients who died of lung cancer, only age, stage and PCNA LI were found to be significant prognostic factors on multivariate analysis among seven potential prognostic factors including sex, age, year of operation, histological type, stage, Ag-NOR count, and PCNA L1. We conclude that PCNA may be superior to Ag-NOR in predicting shortened survival of patients with non-small cell lung cancer. PCNA staining can be performed with ease and it may be' applied in a clinical laboratory on a routine basis to help predict prognosis of NSCLC. Surgical Oncology 1995; 4: 83-89. Keywords: Ag·NOR, multivariate analysis, non-small cell lung cancer, PCNA, prognostic
factor, proliferative activity.
Prognosis and outcome in non -small cell lung cancer (NSCLC) is generally poor and mostly depends on its TNM stage [17]. However, it is also true that the biological behaviour of tumour cells varies and that it is often difficult to predict individual outcome from stage alone. Thus it is important to identify new prognostic factors which can be used to guide treatment planning. Efforts are being made to correlate var ious markers such as DNA ploidy, bromodeoxyuridine labeling and alterations in oncogenes or tumour suppresser genes with prognosis [18]. In this study, we examined 102 patients with NSCLC for PCNA staining and correlated the find ings with various clinical and pathological features. Prognostic implication of PCNA was assessed using the Cox proportional hazards modeling technique to evaluate the significance and independence of the various prognostic
INTRODUCTION Proliferating cell nuclear antigen (PCNA) was first identified using antibodies found in some patients with systemic lupus erythematosus [1, 2]. It is known that this nuclear protein is an auxiliary protein of the 36 kDa DNA polymerase-e5 [3, 4], wh ich is thought to play a significant role in maintaining the fidelity of DNA replication [5]. PCNA increases in late G1 phase and peaks in S phases of the cell cycle and is not detectable in quiescent cells [6-8]. The synthesis of PCNA correlates with the proliferating state of cells in various human tumours [9-16] . Correspondence: Tsunehiro Oyama, MD, Department of Surgery II, University of Occupatio nal and Environmental Health, School of Medicine, 1-1 Iseigaoka, Yahatanishi-ku, Kitakyushu 807, Japan.
© 1995 Blackwell Science Ltd
83
1. Oyama et al.
84
factors. We also compared the PCNA labeling index w ith the argyrophil nuclear organizer regions (AgNOR) which we examined previously (19) in terms of their correlation and prognostic value.
MATERIALS AND METHODS Surgical specimens We examined 102 of 159 consecutive Japanese patients (64.2%) with NSCLC who underwent surgical resection at the Department of Surgery II, School of Medicine, University of Occupational and Environmental Health in Kitakyushu, Japan from August 1979 to October 1987. The criteria for inclusion into the study were based solely on the availability of paraffin blocks and follow-up data. There were 71 men and 31 women ranging from 39 to 84 years (mean, 64 years) in age. Forty-one patients had stage I disease, 10 had stage II, 31 had stage ilia, 10 had stage IIlb, and 10 had stage IV disease, according to the TNM staging system (20). Histological typing of the tumours was performed according to the WHO classification [21]. There were 56 patients with adenocarcinomas, 38 ~ squamous cell carcinomas and 8 large cell carcinomas. Resected specimens were fixed in 20% formalin for 3 days and were embedded in paraffin. For "hlstoloqlcal study, sections were stained with haematoxylin and eosin.
.
PCNA staining Sections (3 JIm thick) were routinely dewaxed and immersed in methanol with 0.3% hydrogen peroxide for 30 min to block endogenous peroxidase activity. The labelled avidine-biotin technique was used for immunostaining utilizing the Labelled Strept Avidin Biotin (LSAB) kit (DAKO Japan, Tokyo). The antibody employed in this study was the mouse monoclonal IgG2a antibody against PCNA, clone PC 10 (DAKO Japan, Tokyo). A dilution of 1: 200 of the pr imary antibody w it h overnight incubation at room temperature was chosen as a condition for reaction in accordance with Hall et al. [7]. The specificity of the reaction was demonstrated in non-neoplastic lung tissue. The PCNA labelling index (L1) was determined by calculating the percentage of positive nuclei of
tumour cells among 1000 tumour cells from several high -power fields chosen at random [16]. The reproducibility of PCNA L1, which were performed by one individual (T.O.), was good with less than 5% difference among mean counts. Statistical analysis Relationship between PCNA LI and various clinical and pathological parameters was evaluated by analys is of variance (ANOVA). Survival analysis was performed for 79 patients, excluding 23 patients who died of a cause other than lung cancer. The survival rate was calculated using the Kaplan-Meier method [22]. The significance of any difference was evaluated by the log rank test [23]. The Cox proportional hazards modelling technique was used to identify independent factors that had a significant influence on overall survival [24]. Association trends were assessed by assigning scores (a, 1 or 2) to the three levels of relevant independent parameters while using indicator terms for the other independent parameters in the models. Every analysis was performed using the PROC GLM and PROC PHGLM routines of the Statistical Analysis System (SAS) program library [25,26].
RESULTS PCNA immunostaining was clearly evident as granular or homogeneous nuclear staining. PCNApositive cells were distributed in the periphery of squamous cell carcinoma tumour nests, whereas in adenocarcinoma they tended to be randomly distributed (Fig. 1). The mean PCNA ~I for all tumours was 40.7% (range, 2.6-90.2%). We analysed the various clinical or pathological factors which may affect the PCNA LI and the results are shown in Table 1. The PCNA LI of stage III or IV disease was significantly greater than that of stage II disease (P <0.05). The PCNA LI in adenocarcinoma was significantly lo wer than that in Among adenosquamous cell carcinoma. carcinomas, the PCNA LI was higher in poorly differentiated than in well-differentiated tumours. However, there was no such tendency in the case of squamous cell carcinomas. Such factors as sex, age, local invasion, nodal status and presence of ©1995 Blackwell Science Ltd, Surgical Oncology, 4: 83-89
PCNA in non-small cell lung cancer
Figure 1. peNA staining in adenocarcinoma of non-small cell lung cancer. Bar: 10 JIm.
Parameter Sex Female Male Age <60 60-70 ~70
Year of operation 1979-1983 1984-1987 T 1 2 3,4 N 0 1,2,3 M 0 1 Stage I II
ilia IIIb IV Histology Adenocarcinoma Well differentiated Moderately differentiated Poorly differentiated Squamous cell carcinoma Well differentiated Moderately differentiated Poorly differentiated Large cell carcinoma Total
n
any metastasis did not influence PCNA L1. In our previous study, Ag-NOR staining was performed in the same patients with NSCLC [19]. There was positive correlation (r = 0.31, P = 0.002) between the PCNA LI and Ag-NOR count (Fig. 2), suggesting that both tests measured similar features of tumour cells. We then examined the prognostic implication of PCNA using univariate and multivariate analysis. Seven potential prognostic parameters (sex, age, year of operation, histology, stage, Ag-NOR count, and PCNA L1) were assessed (Table 2). The cut-off values for the PCNA LI and Ag·NOR count were chosen to produce three categories with equal Table 1. Relationship between the peNA labelling index (L1) and various clinical and pathological parameters in patients with non-small cell lung cancer
Mean±SE 31 71
38.0±4.54 42.0±3.00
29 40 33 50
39.4±4.71 38.9 ±4.01 44.1 ±4.42 34 .6 ±3.48
52
46.7±3.41
34 46 22
31.0±4.2B 3B.9±3.6B 50.4±5.32
55 47
41.B±3.42 39.6±3.69
92 10
39.8±2.63 49.6±7.96
41 9 31 11 10
31.5±3.86
56 12 31 13 38 8 19 11
8 102
*P
85
237+8.23] ] ] 43.7 ±4.43 * * * 50.6±7.44 49.6±7.81 33 .4±3.21 21.1 +6.86 32.2±4.27 ] ** 47.8±6.59 51.4±3.89 52.0±7.87 54.5±5.11 45.4±6.71 41.6±8.48 40.7 ±2.50
J
**
86
T. Oyama et a1. 7 (r= 0.31, P= 0.001)
o
o
6
o o o
o
5
O+----.--r-....----r--...--r---....-""T"""-...---,
o
20
40
60
80
100
PCNAlabelling index (%)
Figure 2. Correlation between the PCNA labelling index and the Ag-NOR count (r = 0.31, P = 0.001).
numbers of patients. In univariate analysis, PCNA L1, Ag-NOR count, and stage were significantly related to shortened survival. Figure 3 depicts the survival curves of each of the three groups of patients determined, according to the cut-off points of 23.5 and 60.0. There was a statistically significant (P = 0.007) difference in the survival rate between the lower and higher group PCNA LI patients. In the multivariate analysis using Cox's proportional hazards model, age, PCNA LI and stage were independently associated with shortened survival of the patients (Table 2). The hazard ratio (HR) of death was calculated to increase to 2.47 for a PCNA LI of 23.560.0 and to 3.59 for a PCNA LI of more than 60.0, compared with those with PCNA LI of less than 23.5. There was a significant
Table 2. Univariate and multivariate analysis using a proportional hazard model of the survival of patients with nonsmall cell lung cancer Parameter
PCNA(L1,%) <23.5 23.5-60.0 ~60.0
AgNOR count <2.20 2.20-2.93
.
~2.93
Sex Female Male Age <60 60-70 ~70
Year of operation 1979-1983 1984-1987 Histology Adenocarcinoma Squamous cell Large cell Stage I II ilia and llib IV
n
Univariate analysis
Multivariate analysis
HR
95% CL
P
HR
95% CL
P
26 27 26
1.0 2.04 2.98
1.03-4.03 1.48-6.00
0.04 0.002
1.0 2.47 3.59 Trend
1.11-5.49 1.46-8.84
0.03 0.005 P= 0.006
26 27 26
1.0 2.84 3.15
1.39-5.79 1.57-6.32
0.004 0.001
1.0 2.03 1.20 Trend
0.88-4.71 0.48-3.02
0.10 0.70 P=0.93
26 53
1.0 0.88
0.50-1.53
0.64
1.0 1.20
0.61-2.34
0.60
24 33 22
1.0 1.39 1.00
0.75-2.59 0.48-2.06
0.29 0.99
1.0 2.70 1.67
1.25-5.82 0.70-3.99
0.01 0.25
37 42
0.97 1.0
0.57-1.66
0.91
0.72-2.58
0.34
47 27 5
1.0 1.09 1.46
0.62-1.92 0.57-3.76
0.76 0.43
1.0 0.67 1.24
0.33-1.35 0.42-3.70
0.26 0.70
32 6 31 10
1.0 r» 1.76 3.44 7.61
0.63-4.89 1.79-6.61 3.29-17.6
0.28 0.0002 0.0001
1.0 2.26 4.03 5.58
0.73-6.95 1.86-8.75 2.13-14.65
0.16 0.0004 0.0005
1.37 1.0
HR: hazard ratio. CL: confidence limit.
©1995 Blackwell Science Ltd, Surgical Oncology, 4: 83-89
PCNA in non-small cell/ung cancer 100
'!t
It. \
m .~
50
:>
to
~°L..
23.5> LI
t., ...\..
\·······6o:(i~·u·······_·
]
P = 0.007
23.5 - 60.0 LI
.._-;
O-+--""'--"--""T"--""'--" o
2 3 Years after operation
4
5
Figure 3. Survival curves of patients with non-small cell lung cancer according to the PCNA labelling index (PCNA L1). The difference between the lower and higher PCNA LI groups is statistically significant (P = 0.007)
(P = 0.006) trend indicating that a high PCNA LI was associated with poor survival. The Ag-NOR count, however, was not independently associated with poor prognosis of the patients in this model, although it was a significant prognostic factor in univariate analysis.
DISCUSSION
We showed that the PCNA LI tended to be higher in tumours of higher stages than in those of early stages, in squamous. cell carcinomas than in adenocarcinomas or in poorly differentiated adenocarcinoma than in well-differentiated. Tumour doubling times of squamous cell carcinoma and poorly differentiated carcinoma are usually shorter than adenocarcinoma or well-differentiated carcinoma, respectively [27]. Hence our result seems reasonable, considering that PCNA LI is associated with parameters related to proliferative activity or aggressiveness of tumour cells, such as the histological grade of differentiation [28-31], DNA ploidy [9-11], bromodeoxyuridine labelling [12], Ki67 labelling [13-15], mitotic index [16], or blood vessel invasion [11]. However, Fontanini et a/. reported that there is no relatioship between PCNA LI and tumour stage, histological type or mitotic count in 40 NSCLC [11]. The reason for disagreement is not clear but the fact that their analysis is confined to node-negative NSCLC may have some relevance. Nucleolar organizer regions (NOR) are loops of © 1995 Blackwell Science Ltd, Surgical Oncology, 4: 83-89
87
DNA that transcribe to ribosomal RNA [32] and are associated with certain proteins, including RNA polymerase I, 823, and C23. NOR are demonstrated by means of the argyrophilia of their associated proteins using the Ag-NOR technique [33-35]. The number of Ag-NORs per nucleus is also believed to correlate with cellular proliferative activity of various human malignant neoplasms. PCNA labelling is at a maximum during the S phase [6] in which there is nucleolar disaggregation with a corresponding increase in Ag-NOR [32, 36]. A positive correlation is therefore observed between PCNA labelling and the Ag-NOR count in many types of human tumours [28, 29, 37-40). Indeed, we found that it is also the case in NSCLC. We were able to show that PCNA LI was an independent parameter that was associated with increased risk of death using a Cox proportional hazards model. The model excluded the Ag-NOR count as an independent prognostic factor probably due to relatedness between PCNA LI and AgNOR. The PCNA labelling index has been reported to be associated with survival in various human tumours [29, 39]. Fontanini et a/. [11] reported that PCNA is the only significant prognostic factor among tumour stage, size, blood vessel invasion, and mitotic counts by Cox model in node-negative NSCLC and they concluded that PCNA LI is the most useful method for estimating cell proliferation. In the recent study by Ishida and his colleagues [41], PCNA LI and Ag=NOR are both independent prognostic factors in NSCLC. However, in their study, there are 142 tumours out of 211 NSCLC with PCNA LI of 5% or less. This is at odds with that in other reports including ours; mean PCNA LI is 53% and 41% in the study by Theunissen et a/. [16] and in ours, respectively, and PCNA LI ranges from 0% to 70% with a median of 15% in the study by Fontanini et a/. [11]. The reason for the discrepancy is unknown but it may be necessary to standardize the method for staining and counting of PCNA-positive cells in future.
CONCLUSIONS
A high PCNA LI is independently associated with poor prognosis in patients with NSCLC and the PCNA LI correlates positively with the Ag-NOR count in NSCLC. PCNA staining may be more
88
T. Oyama et al.
reliable in assessing aggressiveness of NSCLC than other prognostic parameters, including Ag-NDA. PCNA labelling may be used to study small specimens such as those obtained by fine-needle aspiration "[42J or by bronchoscope. We may be able to use PCNA staining even to guide treatment planning for NSCLC, because PCNA is reported to be associated with chemosensitivity in lung tumours [43J. It is now clear that human malignant tumours, including those of the lung, arise and develop through accumulation of genetic lesions of oncogenes and tumour suppressor genes. In our laboratory, efforts are now being made to correlate these genetic alterations with parameters related to proliferative activity of lung cancer cells.
9.
10.
11.
12.
ACKNOWLEDGEMENTS
13.
We are grateful to Ms M. Kiyofuji and Ms K. Nishida for their excellent technical assistance. This work was partly supported by grants from the Fukuoka Cancer Society.
14.
15.
REFERENCES 1. Tan EM. Autoantibodies to nuclear antigens (ANA): their immunobiology and med icine. Adv Immunol 1982; 33: 167-240. 2. Miyachi K, Fritzler MJ, Tan EM. Autoantibodies to a nuclear antigen in proliferating cells. J Immuno/1978; 121: 2228-34. 3. Prelich G, Tan C-K, Kostura M, et al. Functional identity of "proliferating cell nuclear antigen and a DNA polymerase-J auxiliary protein. Nature 1987; 326: 517-20. 4. Brove R, Frank R, Blundell PA, MacDonald-Bravo H. Cyclin/PCNA is the auxil iary protein of DNA polymerase-J. Nature 1987; 326: 515-17. 5. Kunkel TA, Sabatino RD, Bambara RA. Exonucleolytic proofreading by self thymus DNA polymerase-J. Proc Nat! Acad Sci USA 1987; 84: 4865-9. 6. Galand P, Degraef C. Cyclin/PCNA immunostaining as an alternative to tritiated thymidine pulse labeling" for marking S phase cells in paraffin sections from animal and human tissues . Cell Tiss Kinet 1989; 22: 383-92. 7. Hall PA, Levison DA, Woods AL, et al. Proliferating cell nuclear antigen (PCNA) irnrnunolocalization in paraffin sections: an index of cell proliferation with evidence of derequlated expression in some neoplasms. J Patho/1990; 162: 285-94. 8. Pelosi G, Zamboni G, Doglioni C, et al . Immunodetection of proliferating cell nuclear antigen
16.
17.
18.
19.
20. 21.
22.
23.
assesses the growth fraction and predicts malignancy in endocrine tumours of the pancreas . Am j Surg Patho/1992; 16: 1215-25. Garcia RI, Coltrera MD, Grown AM. Analysis proliferative grade using PCNNcyclin monoclonal antibodies in fixed, embedded tissues. Am J Pathol 1989; 134: 733-9. Carey FA, Fabbroni : G, Lamb D. Expression of proliferating cell nuclear antigen in lung cancer: a systematic study and correlation with DNA ploidy. Histopathology 1992; 20: 499-503. Fontanini G, Macchiarini P, Pepe S, et al. The expression of proliferating cell nuclear antigen in paraffin sections of peripheral, node-neqative nonsmall cell lung cancer. Cancer 1992; 70: 1520-7. VanDierendonck JH, Wysman JH, Keyzer R, et al. Cell-cycle-related staining patterns of anti-proliferating cell nuclear antigen monoclonal antibod ies. Comparison with BrdURd labeling and Ki-67 staining . Am JPathol 1991; 138: 1165-72. Tsuji T, Sasaki K, Kimura Y, et al. Measurement of proliferating cell nuclear antigen (PCNA) and its clinical application in oral cancers. Int J Oral Maxillofac Surg 1992; 21: 369-72. Quinn CM, Wright NA. The clinical assessment of proliferation and growth in human tumours: evaluation of methods and applications as prognostic variables. J Patho/1990; 160: 93-102. Fontanini G, Pingitore R, Bigini 0, et al. Growth fraction in non -small cell lung cancer estimated by proliferating cell nuclear antigen and comparison with Ki·67 labeling and DNA flow cytometry data. Am J Patho/1992; 141: 128-59. Theunissen PHMH, Leers MPG, Bollen ECM. Proliferating cell nuclear antigen (PCNA) expression in formalin-fixed tissue of non-small cell lung carcinoma. Histopathology 1992; 20: 251-5. Minna JD, Pass H, Glatstein E, Ihde DC. Cancer of the lung. In: DElVi,ta VT, Hellman S, Rosenberg SA, eds. Cancer: Principles and Practice of Oncology. Philadelphia: J. B. Lippincott, 1989,591-705. Fielding LP, Fenoglio-Preiser CM, Freedman MA. The future of prognostic factors in outcome prediction for patients with cancer. Cancer 1992; 70: 2367-77. Oyama T, Mitudomi T, Yoshida Y, et al . Nucleolar organizer region are independently associated with a shortened survival of patients with non-small cell lung cancer. Surg Oncol 1993; 2: 341-7. Mountain CF. A new international staging system for lung cancer. Chest 1986; 89 (Suppl.): 225s-233s. World Health Organization. Histological typing of lung tumours (WHO), Ed. 2. Am J Clin Pathol 1982; 77: 123-36. Kaplan E, Meier P. Non-parametric estimation from incomplete observation. J Am Stat Assoc 1958; 53: 457-81. Mantel N. Evaluation of survival data and two new rank order statistics arising in its consideration. Cancer Chemo Rep 1966; 50: 163-70.
© 1995 Blackwell Science Ltd, Surgical Oncology, 4: 83-89
PCNA in non-small cell lung cancer
24. Cox DR. Regression models and life tables. J Royal Stat Soc Series B 1972; 34: 187-220. 25. SAS Institute Inc: The PHGLM Procedure . In: SUGI Supplemental Library User's Guide: Statistics. Cary, NC: 1986: 437-66. 26. SAS Institute Inc. The GLM Procedure. In: SAS/ STAT™ User's Guide, Release 6.03 Edition. Cary, NC: 1988: 549-640. 27. Fraser RG, Pare JAP, Pare PO, et et, eds. Neoplastic Disease of the Lung. Philadelphia: Saunders, 1989. 28. Skopelitou A, Korkolopoulou P, Papanicolaou A, et al. Comparative assessment of proliferating cell nuclear antigen immunostaining in transitional cell carcinomas of the urinary bladder: correlation with other conventional prognostic pathologic parameters. Eur Uro/1992, 22: 235-40. 29. Yu CC, Fletcher CD, Newman PL, et el, A comparison of proliferating cell nuclear antigen (PCNA) immunostaining, nucleolar organizer region (AgNOR) staining, and histological grading in gastrointestinal stromal tumours. J Patho/1992; 166: 147-52. 30. Benner SE, Hong WK, Lippman SM, et al. Intermediate biomarker in upper aerod igestive tract and lung chemoprevention trials . J Cell Biochem Suppl 1992; 16G: 33-8. 31. Barbareschi M, Girlando S, Mauri FA. et al. Tumour suppresser gene products, prol iferation, and differentiation markers in lung neuroendocrine neoplasms. J Patho/1992; 166: 343-50. 32. Egan M, Crocker J. Evaluation of nucleolar organizer regions in pulmonary p~thology. Thorax 1990; 45: 225-32. 33. Goodpasture C, Bloom SE. Visualization of nucleolar organizer regions in mammalian chromosomes using silver staining. Chromosoma (Berl) 1975; 53: 37-50. 34. Howell WM, Denton TE, Diamond JR. Differential staining of the satellite regions of acrocentric chromosomes. Experientia (Basel) 1975; 31: 260-2. 35. Ploton 0, Menager M, Jeanneson P, et al.
© 1995 Blackwell Science Ltd, Surgical Oncology, 4: 83-89
89
Improvement in the stain ing and in the visualization of the argyrophilic proteins of the nucleolar organizer region at the optical level. Histochem J 1986; 18: 5-14. 36 . Crocker J, Boldy JAR, Egan MJ. How should we count AgNORs? Proposals for a standardized approach. J Patho/1989; 158: 185-8. 37. Korkolopoulou P, Patsouris E, Pangalis G, et el, A comparative assessment of proliferating cell nucleolar organizer region staining in non-Hodgkin's lymphomas: a histochemical and immunohistochemical study of 200 cases. Hum Pathol 1993; 24: 371-7. 38. Evans AT, Blessing K, Orrell JM, Grant A. Mitotic ind ices, anti·PCNA immunostaining, and AgNORs in thick cutaneous melanomas displaying paradoxical behavior. J Patho/1992; 168: 15-22. 39. Pich A. Chiusa L, Pisani P, et al. Argyrophilic nucleolar organizer region counts and proliferating cell nuclear antigen scores are two reliable indicators of survival in pharyngeal carcinoma. J Cancer Res Clin Onco/1992; 119: 106-10. 40. Miyazaki S,' Sasano H, Suzuki T, et al. Nucleolar organizer regions in human esophageal disorders: comparison with proliferating cell nuclear antigen by immunostaining. Mod Patho/1992; 5: 396-401 . 41. Ishida T, Kaneko S, Akazawa K, et al. Proliferating cell nuclear antigen expressed and argyrophlic nucleolar organizer regions as factors influencing prognosis of surgically treated lung cancer patients. Cancer Res 1993; 53: 5000 -3. 42. Wang KP, Kelly SJ, Britt JE. Percutaneous needle aspiration biopsy of chest lesions: new instrument and new technique. Chest 1988 ; 93: 993-7. 43 Ogawa J, Iwazaki M, Inoue H, et al. Immunohistochemical study of glutathione-related enzymes and proliferative antigens in lung cancer: relation to cisplatin sensitivity. Cancer 1993; 71: 2204-9.
Proliferating cell nuclear antigen may be superior to argyrophilic nucleolar organizer regions in predicting shortened survival of patients with non-small cell lung cancer T. OYAMA, T. MITSUDOMI, T. MIZOUE,* A. OHGAMI, T. OSAKI, R. NAKANISHI AND K. YASUMOTO Departments of Surgery /I and *Clinical Epidemiology, University of Occupational and Environmental Health, School of Medicine, Kitakyushu 807, Japan
We examined proliferating cell nuclear antigen (PCNA) in 102 patients with surgically treated non-small cell lung cancer (NSCLC). PCNA labelling index (L1) tended to be higher in tumours of higher stages than those of early stages, in squamous cell carcinomas than adenocarcinomas, or in poorly differentiated adenocarcinoma than in well-differentiated. A positive correlation was observed between the PCNA LI and argyrophilic nucleolar organizer regions (Ag-NOR) count which we previously examined (r =0.31, P =0.002). In survival analysis of 79 patients who died of lung cancer, only age, stage and PCNA LI were found to be significant prognostic factors on multivariate analysis among seven potential prognostic factors including sex, age, year of operation, histological type, stage, Ag-NOR count, and PCNA L1. We conclude that PCNA may be superior to Ag-NOR in predicting shortened survival of patients with non-small cell lung cancer. PCNA staining can be performed with ease and it may be' applied in a clinical laboratory on a routine basis to help predict prognosis of NSCLC. Surgical Oncology 1995; 4: 83-89. Keywords: Ag·NOR, multivariate analysis, non-small cell lung cancer, PCNA, prognostic
factor, proliferative activity.
Prognosis and outcome in non -small cell lung cancer (NSCLC) is generally poor and mostly depends on its TNM stage [17]. However, it is also true that the biological behaviour of tumour cells varies and that it is often difficult to predict individual outcome from stage alone. Thus it is important to identify new prognostic factors which can be used to guide treatment planning. Efforts are being made to correlate var ious markers such as DNA ploidy, bromodeoxyuridine labeling and alterations in oncogenes or tumour suppresser genes with prognosis [18]. In this study, we examined 102 patients with NSCLC for PCNA staining and correlated the find ings with various clinical and pathological features. Prognostic implication of PCNA was assessed using the Cox proportional hazards modeling technique to evaluate the significance and independence of the various prognostic
INTRODUCTION Proliferating cell nuclear antigen (PCNA) was first identified using antibodies found in some patients with systemic lupus erythematosus [1, 2]. It is known that this nuclear protein is an auxiliary protein of the 36 kDa DNA polymerase-e5 [3, 4], wh ich is thought to play a significant role in maintaining the fidelity of DNA replication [5]. PCNA increases in late G1 phase and peaks in S phases of the cell cycle and is not detectable in quiescent cells [6-8]. The synthesis of PCNA correlates with the proliferating state of cells in various human tumours [9-16] . Correspondence: Tsunehiro Oyama, MD, Department of Surgery II, University of Occupatio nal and Environmental Health, School of Medicine, 1-1 Iseigaoka, Yahatanishi-ku, Kitakyushu 807, Japan.
© 1995 Blackwell Science Ltd
83
1. Oyama et al.
84
factors. We also compared the PCNA labeling index w ith the argyrophil nuclear organizer regions (AgNOR) which we examined previously (19) in terms of their correlation and prognostic value.
MATERIALS AND METHODS Surgical specimens We examined 102 of 159 consecutive Japanese patients (64.2%) with NSCLC who underwent surgical resection at the Department of Surgery II, School of Medicine, University of Occupational and Environmental Health in Kitakyushu, Japan from August 1979 to October 1987. The criteria for inclusion into the study were based solely on the availability of paraffin blocks and follow-up data. There were 71 men and 31 women ranging from 39 to 84 years (mean, 64 years) in age. Forty-one patients had stage I disease, 10 had stage II, 31 had stage ilia, 10 had stage IIlb, and 10 had stage IV disease, according to the TNM staging system (20). Histological typing of the tumours was performed according to the WHO classification [21]. There were 56 patients with adenocarcinomas, 38 ~ squamous cell carcinomas and 8 large cell carcinomas. Resected specimens were fixed in 20% formalin for 3 days and were embedded in paraffin. For "hlstoloqlcal study, sections were stained with haematoxylin and eosin.
.
PCNA staining Sections (3 JIm thick) were routinely dewaxed and immersed in methanol with 0.3% hydrogen peroxide for 30 min to block endogenous peroxidase activity. The labelled avidine-biotin technique was used for immunostaining utilizing the Labelled Strept Avidin Biotin (LSAB) kit (DAKO Japan, Tokyo). The antibody employed in this study was the mouse monoclonal IgG2a antibody against PCNA, clone PC 10 (DAKO Japan, Tokyo). A dilution of 1: 200 of the pr imary antibody w it h overnight incubation at room temperature was chosen as a condition for reaction in accordance with Hall et al. [7]. The specificity of the reaction was demonstrated in non-neoplastic lung tissue. The PCNA labelling index (L1) was determined by calculating the percentage of positive nuclei of
tumour cells among 1000 tumour cells from several high -power fields chosen at random [16]. The reproducibility of PCNA L1, which were performed by one individual (T.O.), was good with less than 5% difference among mean counts. Statistical analysis Relationship between PCNA LI and various clinical and pathological parameters was evaluated by analys is of variance (ANOVA). Survival analysis was performed for 79 patients, excluding 23 patients who died of a cause other than lung cancer. The survival rate was calculated using the Kaplan-Meier method [22]. The significance of any difference was evaluated by the log rank test [23]. The Cox proportional hazards modelling technique was used to identify independent factors that had a significant influence on overall survival [24]. Association trends were assessed by assigning scores (a, 1 or 2) to the three levels of relevant independent parameters while using indicator terms for the other independent parameters in the models. Every analysis was performed using the PROC GLM and PROC PHGLM routines of the Statistical Analysis System (SAS) program library [25,26].
RESULTS PCNA immunostaining was clearly evident as granular or homogeneous nuclear staining. PCNApositive cells were distributed in the periphery of squamous cell carcinoma tumour nests, whereas in adenocarcinoma they tended to be randomly distributed (Fig. 1). The mean PCNA ~I for all tumours was 40.7% (range, 2.6-90.2%). We analysed the various clinical or pathological factors which may affect the PCNA LI and the results are shown in Table 1. The PCNA LI of stage III or IV disease was significantly greater than that of stage II disease (P <0.05). The PCNA LI in adenocarcinoma was significantly lo wer than that in Among adenosquamous cell carcinoma. carcinomas, the PCNA LI was higher in poorly differentiated than in well-differentiated tumours. However, there was no such tendency in the case of squamous cell carcinomas. Such factors as sex, age, local invasion, nodal status and presence of ©1995 Blackwell Science Ltd, Surgical Oncology, 4: 83-89
PCNA in non-small cell lung cancer
Figure 1. peNA staining in adenocarcinoma of non-small cell lung cancer. Bar: 10 JIm.
Parameter Sex Female Male Age <60 60-70 ~70
Year of operation 1979-1983 1984-1987 T 1 2 3,4 N 0 1,2,3 M 0 1 Stage I II
ilia IIIb IV Histology Adenocarcinoma Well differentiated Moderately differentiated Poorly differentiated Squamous cell carcinoma Well differentiated Moderately differentiated Poorly differentiated Large cell carcinoma Total
n
any metastasis did not influence PCNA L1. In our previous study, Ag-NOR staining was performed in the same patients with NSCLC [19]. There was positive correlation (r = 0.31, P = 0.002) between the PCNA LI and Ag-NOR count (Fig. 2), suggesting that both tests measured similar features of tumour cells. We then examined the prognostic implication of PCNA using univariate and multivariate analysis. Seven potential prognostic parameters (sex, age, year of operation, histology, stage, Ag-NOR count, and PCNA L1) were assessed (Table 2). The cut-off values for the PCNA LI and Ag·NOR count were chosen to produce three categories with equal Table 1. Relationship between the peNA labelling index (L1) and various clinical and pathological parameters in patients with non-small cell lung cancer
Mean±SE 31 71
38.0±4.54 42.0±3.00
29 40 33 50
39.4±4.71 38.9 ±4.01 44.1 ±4.42 34 .6 ±3.48
52
46.7±3.41
34 46 22
31.0±4.2B 3B.9±3.6B 50.4±5.32
55 47
41.B±3.42 39.6±3.69
92 10
39.8±2.63 49.6±7.96
41 9 31 11 10
31.5±3.86
56 12 31 13 38 8 19 11
8 102
*P
85
237+8.23] ] ] 43.7 ±4.43 * * * 50.6±7.44 49.6±7.81 33 .4±3.21 21.1 +6.86 32.2±4.27 ] ** 47.8±6.59 51.4±3.89 52.0±7.87 54.5±5.11 45.4±6.71 41.6±8.48 40.7 ±2.50
J
**
86
T. Oyama et a1. 7 (r= 0.31, P= 0.001)
o
o
6
o o o
o
5
O+----.--r-....----r--...--r---....-""T"""-...---,
o
20
40
60
80
100
PCNAlabelling index (%)
Figure 2. Correlation between the PCNA labelling index and the Ag-NOR count (r = 0.31, P = 0.001).
numbers of patients. In univariate analysis, PCNA L1, Ag-NOR count, and stage were significantly related to shortened survival. Figure 3 depicts the survival curves of each of the three groups of patients determined, according to the cut-off points of 23.5 and 60.0. There was a statistically significant (P = 0.007) difference in the survival rate between the lower and higher group PCNA LI patients. In the multivariate analysis using Cox's proportional hazards model, age, PCNA LI and stage were independently associated with shortened survival of the patients (Table 2). The hazard ratio (HR) of death was calculated to increase to 2.47 for a PCNA LI of 23.560.0 and to 3.59 for a PCNA LI of more than 60.0, compared with those with PCNA LI of less than 23.5. There was a significant
Table 2. Univariate and multivariate analysis using a proportional hazard model of the survival of patients with nonsmall cell lung cancer Parameter
PCNA(L1,%) <23.5 23.5-60.0 ~60.0
AgNOR count <2.20 2.20-2.93
.
~2.93
Sex Female Male Age <60 60-70 ~70
Year of operation 1979-1983 1984-1987 Histology Adenocarcinoma Squamous cell Large cell Stage I II ilia and llib IV
n
Univariate analysis
Multivariate analysis
HR
95% CL
P
HR
95% CL
P
26 27 26
1.0 2.04 2.98
1.03-4.03 1.48-6.00
0.04 0.002
1.0 2.47 3.59 Trend
1.11-5.49 1.46-8.84
0.03 0.005 P= 0.006
26 27 26
1.0 2.84 3.15
1.39-5.79 1.57-6.32
0.004 0.001
1.0 2.03 1.20 Trend
0.88-4.71 0.48-3.02
0.10 0.70 P=0.93
26 53
1.0 0.88
0.50-1.53
0.64
1.0 1.20
0.61-2.34
0.60
24 33 22
1.0 1.39 1.00
0.75-2.59 0.48-2.06
0.29 0.99
1.0 2.70 1.67
1.25-5.82 0.70-3.99
0.01 0.25
37 42
0.97 1.0
0.57-1.66
0.91
0.72-2.58
0.34
47 27 5
1.0 1.09 1.46
0.62-1.92 0.57-3.76
0.76 0.43
1.0 0.67 1.24
0.33-1.35 0.42-3.70
0.26 0.70
32 6 31 10
1.0 r» 1.76 3.44 7.61
0.63-4.89 1.79-6.61 3.29-17.6
0.28 0.0002 0.0001
1.0 2.26 4.03 5.58
0.73-6.95 1.86-8.75 2.13-14.65
0.16 0.0004 0.0005
1.37 1.0
HR: hazard ratio. CL: confidence limit.
©1995 Blackwell Science Ltd, Surgical Oncology, 4: 83-89
PCNA in non-small cell/ung cancer 100
'!t
It. \
m .~
50
:>
to
~°L..
23.5> LI
t., ...\..
\·······6o:(i~·u·······_·
]
P = 0.007
23.5 - 60.0 LI
.._-;
O-+--""'--"--""T"--""'--" o
2 3 Years after operation
4
5
Figure 3. Survival curves of patients with non-small cell lung cancer according to the PCNA labelling index (PCNA L1). The difference between the lower and higher PCNA LI groups is statistically significant (P = 0.007)
(P = 0.006) trend indicating that a high PCNA LI was associated with poor survival. The Ag-NOR count, however, was not independently associated with poor prognosis of the patients in this model, although it was a significant prognostic factor in univariate analysis.
DISCUSSION
We showed that the PCNA LI tended to be higher in tumours of higher stages than in those of early stages, in squamous. cell carcinomas than in adenocarcinomas or in poorly differentiated adenocarcinoma than in well-differentiated. Tumour doubling times of squamous cell carcinoma and poorly differentiated carcinoma are usually shorter than adenocarcinoma or well-differentiated carcinoma, respectively [27]. Hence our result seems reasonable, considering that PCNA LI is associated with parameters related to proliferative activity or aggressiveness of tumour cells, such as the histological grade of differentiation [28-31], DNA ploidy [9-11], bromodeoxyuridine labelling [12], Ki67 labelling [13-15], mitotic index [16], or blood vessel invasion [11]. However, Fontanini et a/. reported that there is no relatioship between PCNA LI and tumour stage, histological type or mitotic count in 40 NSCLC [11]. The reason for disagreement is not clear but the fact that their analysis is confined to node-negative NSCLC may have some relevance. Nucleolar organizer regions (NOR) are loops of © 1995 Blackwell Science Ltd, Surgical Oncology, 4: 83-89
87
DNA that transcribe to ribosomal RNA [32] and are associated with certain proteins, including RNA polymerase I, 823, and C23. NOR are demonstrated by means of the argyrophilia of their associated proteins using the Ag-NOR technique [33-35]. The number of Ag-NORs per nucleus is also believed to correlate with cellular proliferative activity of various human malignant neoplasms. PCNA labelling is at a maximum during the S phase [6] in which there is nucleolar disaggregation with a corresponding increase in Ag-NOR [32, 36]. A positive correlation is therefore observed between PCNA labelling and the Ag-NOR count in many types of human tumours [28, 29, 37-40). Indeed, we found that it is also the case in NSCLC. We were able to show that PCNA LI was an independent parameter that was associated with increased risk of death using a Cox proportional hazards model. The model excluded the Ag-NOR count as an independent prognostic factor probably due to relatedness between PCNA LI and AgNOR. The PCNA labelling index has been reported to be associated with survival in various human tumours [29, 39]. Fontanini et a/. [11] reported that PCNA is the only significant prognostic factor among tumour stage, size, blood vessel invasion, and mitotic counts by Cox model in node-negative NSCLC and they concluded that PCNA LI is the most useful method for estimating cell proliferation. In the recent study by Ishida and his colleagues [41], PCNA LI and Ag=NOR are both independent prognostic factors in NSCLC. However, in their study, there are 142 tumours out of 211 NSCLC with PCNA LI of 5% or less. This is at odds with that in other reports including ours; mean PCNA LI is 53% and 41% in the study by Theunissen et a/. [16] and in ours, respectively, and PCNA LI ranges from 0% to 70% with a median of 15% in the study by Fontanini et a/. [11]. The reason for the discrepancy is unknown but it may be necessary to standardize the method for staining and counting of PCNA-positive cells in future.
CONCLUSIONS
A high PCNA LI is independently associated with poor prognosis in patients with NSCLC and the PCNA LI correlates positively with the Ag-NOR count in NSCLC. PCNA staining may be more
88
T. Oyama et al.
reliable in assessing aggressiveness of NSCLC than other prognostic parameters, including Ag-NDA. PCNA labelling may be used to study small specimens such as those obtained by fine-needle aspiration "[42J or by bronchoscope. We may be able to use PCNA staining even to guide treatment planning for NSCLC, because PCNA is reported to be associated with chemosensitivity in lung tumours [43J. It is now clear that human malignant tumours, including those of the lung, arise and develop through accumulation of genetic lesions of oncogenes and tumour suppressor genes. In our laboratory, efforts are now being made to correlate these genetic alterations with parameters related to proliferative activity of lung cancer cells.
9.
10.
11.
12.
ACKNOWLEDGEMENTS
13.
We are grateful to Ms M. Kiyofuji and Ms K. Nishida for their excellent technical assistance. This work was partly supported by grants from the Fukuoka Cancer Society.
14.
15.
REFERENCES 1. Tan EM. Autoantibodies to nuclear antigens (ANA): their immunobiology and med icine. Adv Immunol 1982; 33: 167-240. 2. Miyachi K, Fritzler MJ, Tan EM. Autoantibodies to a nuclear antigen in proliferating cells. J Immuno/1978; 121: 2228-34. 3. Prelich G, Tan C-K, Kostura M, et al. Functional identity of "proliferating cell nuclear antigen and a DNA polymerase-J auxiliary protein. Nature 1987; 326: 517-20. 4. Brove R, Frank R, Blundell PA, MacDonald-Bravo H. Cyclin/PCNA is the auxil iary protein of DNA polymerase-J. Nature 1987; 326: 515-17. 5. Kunkel TA, Sabatino RD, Bambara RA. Exonucleolytic proofreading by self thymus DNA polymerase-J. Proc Nat! Acad Sci USA 1987; 84: 4865-9. 6. Galand P, Degraef C. Cyclin/PCNA immunostaining as an alternative to tritiated thymidine pulse labeling" for marking S phase cells in paraffin sections from animal and human tissues . Cell Tiss Kinet 1989; 22: 383-92. 7. Hall PA, Levison DA, Woods AL, et al. Proliferating cell nuclear antigen (PCNA) irnrnunolocalization in paraffin sections: an index of cell proliferation with evidence of derequlated expression in some neoplasms. J Patho/1990; 162: 285-94. 8. Pelosi G, Zamboni G, Doglioni C, et al . Immunodetection of proliferating cell nuclear antigen
16.
17.
18.
19.
20. 21.
22.
23.
assesses the growth fraction and predicts malignancy in endocrine tumours of the pancreas . Am j Surg Patho/1992; 16: 1215-25. Garcia RI, Coltrera MD, Grown AM. Analysis proliferative grade using PCNNcyclin monoclonal antibodies in fixed, embedded tissues. Am J Pathol 1989; 134: 733-9. Carey FA, Fabbroni : G, Lamb D. Expression of proliferating cell nuclear antigen in lung cancer: a systematic study and correlation with DNA ploidy. Histopathology 1992; 20: 499-503. Fontanini G, Macchiarini P, Pepe S, et al. The expression of proliferating cell nuclear antigen in paraffin sections of peripheral, node-neqative nonsmall cell lung cancer. Cancer 1992; 70: 1520-7. VanDierendonck JH, Wysman JH, Keyzer R, et al. Cell-cycle-related staining patterns of anti-proliferating cell nuclear antigen monoclonal antibod ies. Comparison with BrdURd labeling and Ki-67 staining . Am JPathol 1991; 138: 1165-72. Tsuji T, Sasaki K, Kimura Y, et al. Measurement of proliferating cell nuclear antigen (PCNA) and its clinical application in oral cancers. Int J Oral Maxillofac Surg 1992; 21: 369-72. Quinn CM, Wright NA. The clinical assessment of proliferation and growth in human tumours: evaluation of methods and applications as prognostic variables. J Patho/1990; 160: 93-102. Fontanini G, Pingitore R, Bigini 0, et al. Growth fraction in non -small cell lung cancer estimated by proliferating cell nuclear antigen and comparison with Ki·67 labeling and DNA flow cytometry data. Am J Patho/1992; 141: 128-59. Theunissen PHMH, Leers MPG, Bollen ECM. Proliferating cell nuclear antigen (PCNA) expression in formalin-fixed tissue of non-small cell lung carcinoma. Histopathology 1992; 20: 251-5. Minna JD, Pass H, Glatstein E, Ihde DC. Cancer of the lung. In: DElVi,ta VT, Hellman S, Rosenberg SA, eds. Cancer: Principles and Practice of Oncology. Philadelphia: J. B. Lippincott, 1989,591-705. Fielding LP, Fenoglio-Preiser CM, Freedman MA. The future of prognostic factors in outcome prediction for patients with cancer. Cancer 1992; 70: 2367-77. Oyama T, Mitudomi T, Yoshida Y, et al . Nucleolar organizer region are independently associated with a shortened survival of patients with non-small cell lung cancer. Surg Oncol 1993; 2: 341-7. Mountain CF. A new international staging system for lung cancer. Chest 1986; 89 (Suppl.): 225s-233s. World Health Organization. Histological typing of lung tumours (WHO), Ed. 2. Am J Clin Pathol 1982; 77: 123-36. Kaplan E, Meier P. Non-parametric estimation from incomplete observation. J Am Stat Assoc 1958; 53: 457-81. Mantel N. Evaluation of survival data and two new rank order statistics arising in its consideration. Cancer Chemo Rep 1966; 50: 163-70.
© 1995 Blackwell Science Ltd, Surgical Oncology, 4: 83-89
PCNA in non-small cell lung cancer
24. Cox DR. Regression models and life tables. J Royal Stat Soc Series B 1972; 34: 187-220. 25. SAS Institute Inc: The PHGLM Procedure . In: SUGI Supplemental Library User's Guide: Statistics. Cary, NC: 1986: 437-66. 26. SAS Institute Inc. The GLM Procedure. In: SAS/ STAT™ User's Guide, Release 6.03 Edition. Cary, NC: 1988: 549-640. 27. Fraser RG, Pare JAP, Pare PO, et et, eds. Neoplastic Disease of the Lung. Philadelphia: Saunders, 1989. 28. Skopelitou A, Korkolopoulou P, Papanicolaou A, et al. Comparative assessment of proliferating cell nuclear antigen immunostaining in transitional cell carcinomas of the urinary bladder: correlation with other conventional prognostic pathologic parameters. Eur Uro/1992, 22: 235-40. 29. Yu CC, Fletcher CD, Newman PL, et el, A comparison of proliferating cell nuclear antigen (PCNA) immunostaining, nucleolar organizer region (AgNOR) staining, and histological grading in gastrointestinal stromal tumours. J Patho/1992; 166: 147-52. 30. Benner SE, Hong WK, Lippman SM, et al. Intermediate biomarker in upper aerod igestive tract and lung chemoprevention trials . J Cell Biochem Suppl 1992; 16G: 33-8. 31. Barbareschi M, Girlando S, Mauri FA. et al. Tumour suppresser gene products, prol iferation, and differentiation markers in lung neuroendocrine neoplasms. J Patho/1992; 166: 343-50. 32. Egan M, Crocker J. Evaluation of nucleolar organizer regions in pulmonary p~thology. Thorax 1990; 45: 225-32. 33. Goodpasture C, Bloom SE. Visualization of nucleolar organizer regions in mammalian chromosomes using silver staining. Chromosoma (Berl) 1975; 53: 37-50. 34. Howell WM, Denton TE, Diamond JR. Differential staining of the satellite regions of acrocentric chromosomes. Experientia (Basel) 1975; 31: 260-2. 35. Ploton 0, Menager M, Jeanneson P, et al.
© 1995 Blackwell Science Ltd, Surgical Oncology, 4: 83-89
89
Improvement in the stain ing and in the visualization of the argyrophilic proteins of the nucleolar organizer region at the optical level. Histochem J 1986; 18: 5-14. 36 . Crocker J, Boldy JAR, Egan MJ. How should we count AgNORs? Proposals for a standardized approach. J Patho/1989; 158: 185-8. 37. Korkolopoulou P, Patsouris E, Pangalis G, et el, A comparative assessment of proliferating cell nucleolar organizer region staining in non-Hodgkin's lymphomas: a histochemical and immunohistochemical study of 200 cases. Hum Pathol 1993; 24: 371-7. 38. Evans AT, Blessing K, Orrell JM, Grant A. Mitotic ind ices, anti·PCNA immunostaining, and AgNORs in thick cutaneous melanomas displaying paradoxical behavior. J Patho/1992; 168: 15-22. 39. Pich A. Chiusa L, Pisani P, et al. Argyrophilic nucleolar organizer region counts and proliferating cell nuclear antigen scores are two reliable indicators of survival in pharyngeal carcinoma. J Cancer Res Clin Onco/1992; 119: 106-10. 40. Miyazaki S,' Sasano H, Suzuki T, et al. Nucleolar organizer regions in human esophageal disorders: comparison with proliferating cell nuclear antigen by immunostaining. Mod Patho/1992; 5: 396-401 . 41. Ishida T, Kaneko S, Akazawa K, et al. Proliferating cell nuclear antigen expressed and argyrophlic nucleolar organizer regions as factors influencing prognosis of surgically treated lung cancer patients. Cancer Res 1993; 53: 5000 -3. 42. Wang KP, Kelly SJ, Britt JE. Percutaneous needle aspiration biopsy of chest lesions: new instrument and new technique. Chest 1988 ; 93: 993-7. 43 Ogawa J, Iwazaki M, Inoue H, et al. Immunohistochemical study of glutathione-related enzymes and proliferative antigens in lung cancer: relation to cisplatin sensitivity. Cancer 1993; 71: 2204-9.