(CANCER RESEARCH 53. 61-66. January 1. 19931

Serum Fragment of Cytokeratin Subunit 19 Measured by CYFRA 21-1

Immunoradiometric Assay as a Marker of Lung CancerJean-Louis Pujol,1, Jean Grenier, Jean-Pierre Daurès,Alain Daver, Henri Pujol, and François-Bernard Michel

Clinique líesMaladies Respiratoires. Universitétie Mtmtpellier. Hôpital Aiguelonfìue.3405V Montpellier Cedex ¡J-L P.. F-B. M.]; Centre Régionalde Lune contre le Cancer,Laboratoire de Raditi-Anal\ses, Ì4W4Montpellier Cedex ¡J.G.. H. P.¡:Départementde l'Information Médicale.Hôpital Gaston Doumergue. 30000 Nîmes¡J-P.D./: and Centre

Paul Papin. 2. Rue Moll. 4W36 Angers ¡A.D.I. France

ABSTRACT

Cytokeratin 19 is a subunit of cytokeratin intermediate filament expressed in simple epithelia and their malignant counterparts. Therefore, itis expressed by respiratory epithelium cells and has been detected in lungcancer specimens. An immunoradiometric assay was used to detect afragment of the cytokeratin 19, referred to as CYFRA 21-1, in the serumof 165 patients with histologically proved lung cancer ( 128 non-small cell

and 37 small cell lung cancers). This prospective study was conducted toevaluate the reliability of this immunoradiometric assay and to identifythe relationship between serum CYFRA 21-1 and different features of lung

cancer including prognosis. The minimal detectable concentration detected by this assay was 0.06 ng/ml. The reliability of the immunoradiometric assay was demonstrated by the linear relationship between CYFRA21-1 measurement and dilution of the serum, the reproducibility of the

dosage in intraassay and interassay, and the high sensitivity of the methodin discriminating low CYFRA 21-1 concentrations. Using a threshold of

3.6 ng/ml, sensitivity and specificity were 0.52 and 0.87, respectively. Thesensitivity of the marker was highest in squamous cell carcinoma andlowest in small cell carcinoma. In non-small cell lung cancer patients, the

marker varied significantly according to both stage of the disease(Kruskal-Wallis, 13.7; P < 0.005) and performance status (Kruskal-Wallis, 9.16; P < 0.05) inasmuch as a high serum CYFRA 21-1 level was

associated with advanced stages, mediastinal lymph nodes, and poor performance status. Consequently, the marker was significantly lower inpatients who were operated upon when compared with unresectable ones.Lung cancer patients with serum CYFRA 21-1 over 3.6 ng/ml proved to

have a significantly shorter overall survival than those with a normalserum level (log rank, P = 0.007; Wilcoxon, /' - 0.001). The negative

prognostic effect of CYFRA 21-1 was highly significant in squamous cell

carcinomas whereas it was nonsignificant for the other histologies. InCox's model analysis, performance status, stage grouping, and CYFRA

21-1 were the only significant determinants of survival. This study sup

ports the use of the serum fragment of cytokeratin subunit 19 CYFRA21-1 as an independent prognostic marker of squamous cell carcinoma of

the lung.

INTRODUCTION

Treatment of lung cancer is probably one of the great challenges ofmedical oncology owing to an increasing incidence in both men andwomen and poor prognosis (1, 2). SCLC2 have neuroendocrine bio

logical properties usually associated with a high sensitivity to chemotherapy (2-^). Surgery offers the best probability of cure for NSCLC

but it cannot be proposed for locally advanced and metastatic stages;therefore, other modality treatments such as radiotherapy-chemother

apy association have been proposed (5). Because metastatic disease isthe first cause of failure for both SCLC (6) and NSCLC (7), researchfor new treatment modalities is an important goal. Such an experimental approach requires an accurate definition of groups of patients

Received 7/2/92: accepted 10/20/92.The costs of publication of this article were defrayed in part by the payment of page

charges. This article must therefore be hereby marked advertisement in accordance with18 U.S.C. Section 1734 solely (o indicate this fact.

1To whom requests for reprints should he addressed.2 The abbreviations used are: SCLC. small cell lung carcinoma: NSCLC. non-small

cell lung carcinoma: MoAb. monoclonal antibodies: SQC. squamous cell carcinomas;MDC. minima] detectable concentration; ROC, receiver operating characteristic; St. cytokeratin standard; LDH, láclatedehydrogenase; NX. nodal stage \.

and prognosis. The three main prognostic factors able to help treatment decision are histology (mainly SCLC versus NSCLC), stage ofdisease, and performance status (8, 9). The search for new clinical orbiological variables as putative independent prognostic factors maypotentially improve the management of lung cancer patients.

Cytokeratins are intermediate filaments which are part of the cy-

toskeleton of both normal epithelia and their malignant counterparts(10). Immunohistochemical studies using broad-spectrum cytokeratin

antibodies have demonstrated the expression of cytokeratins in bothSCLC and NSCLC suggesting a common endodermal lineage(11, 12). In contrast to other intermediate filaments, cytokeratins area family of polypeptides, some of them having cell type specificity(10). Cytokeratin 19 is an acidic (type I) subunit expressed in allsimple epithelia and in carcinomas such as lung cancer which arisefrom them.

A fragment of cytokeratin subunit 19 can be measured in serum bya new immunoradiometric assay, the CYFRA 21-1 using two mouseMoAb KS 19-1 and BM 19-21. Thus, this cytokeratin 19 fragment isreferred to as CYFRA 21-1. The aims of the present prospective studyare (a) to evaluate the reliability of serum CYFRA 21-1, (b) to identify

any relationship between the marker and other clinical and biologicalparameters in both SCLC and NSCLC and (c) to analyze the prognostic significance of a high CYFRA 21-1 level at time of diagnosis.

MATERIALS AND METHODS

Patients. One hundred sixty-five consecutive lung cancer patients referred

to the Montpellier University Hospital between February 1990 and February1992 were prospectively entered in the study (Table 1). All patients had

pathologically confirmed lung cancer. Among them were 37 SCLC. 85 SQC,28 adenocarcinomas, and 15 large cell carcinomas as defined by the WHOclassification (13). Performance status was estimated according to the EasternCooperative Oncology Group and the percentage of weight loss during theprevious 4 months was recorded. Staging was carried out by exhaustive procedure according to the 4th edition of the Union Internationale Contre leCancer tumor-nodes-metastasis classification (14) and the American Thoracic

Society map of regional pulmonary nodes (15). Staging of NSCLC was established according to Mountain's stage grouping (16); for SCLC, limited

disease was defined as disease confined to one hemithorax including mediastinal lymph nodes and/or ipsilateral supraclavicular lymph nodes; extensivedisease was defined as a disease more advanced than limited SCLC. For allpatients staging procedure included clinical examination, standard chest roent-

genography. computed tomographic scan of chest and upper abdomen, fiberoptic bronchoscopy, liver sonography. and bone scanning. Brain computed tomographic scan was done systematically for SCLC and if clinically requiredfor NSCLC.

Controls. The serum CYFRA 21-1 level was measured in 104 patients with

nonmalignant pulmonary diseases (20 infectious diseases. 75 chronic obstructive pulmonary diseases, and 9 miscellaneous).

Treatment Decision. Each patient was discussed by a medical panel composed of thoracic surgeons, chest physicians, radiologists, and medical oncologists. NSCLC patients with stage I or II disease or with moderate locallyadvanced NSCLC (Ilia with resectable node metastasis) underwent surgery inan attempt to achieve complete resection. Other NSCLC patients with performance status <2 and distant metastasis (stage IV) or gross mediastinal involve

ment (stage Illb and stage Ilia with more than 2 ipsilateral mediastinal lymph

61

Research. on February 5, 2021. © 1993 American Association for Cancercancerres.aacrjournals.org Downloaded from

SERUM CYFRA 21-1 IN LUNG CANCER

Table 1 Patients' characteristics

No.M/FMean age (SD, range)

NSCLC

Performance stalus0-12-3

Slage groupingl und IIIllaIllbIV

Wt loss (%)AbsencePresence

SCLC

Disease extentLimited diseaseExtensive disease

Wt loss (%)0>5

Performance status0-12-3

165146/19

62 (10,31-88)

128

4(1

17273549

7355

37

21In

2116

289

node métastases)were eligible for chemotherapy. Best supportive care, including palliative radiation therapy when needed, was given to patients with advanced stage and poor performance status. SCLC patients underwent chemotherapy [usually an alternating regimen containing cisplatin (17)]. Treatmentwas decided upon according to routine clinical and biological findings andwithout knowledge of the serum CYFRA 21-1 level. Hence, treatment was not

considered as a prognostic variable in this study.CYFRA 21-1 Immunoradiometric Assay. A blood sample was taken from

each patient at presentation, and Ihe serum was separated and stored at -190°C

until tested.CYFRA 21-1 (Centocor Diagnostics. Malvern, PA: Cis Biointernational.

Gif/Yvette, France) is a solid-phase immunoradiometric assay based on thetwo-site sandwich method. In this method the cytokeratin 19 was recognizedby two mouse MoAb. KS 19-1 and BM 19-21. directed against two different

epitopes of a fragment of cytokeratin subunit 19. which is referred to as serumCYFRA 21-1. These MoAb were obtained by immunization against the MCF7

cell line.MoAb KS 19-1 -coated polystyrene spheres were incubated with 200 ul of

patient serum, control serum, or standard curve (composed of the followingconcentrations of cytokeratin 19: 0, 3, 8, 25, and, 50 ng/ml). for 20 h at 2-8°C.

Afterwards, the solid phase was washed with distilled water and then incubatedwith 0.85 uCi/ml of '"I-labeled BM 19-21 for 3 h at 2-8°C. Finally, the solid

phase was washed again with distilled water in order to cancel the nonfixedlabeled reagents.

Radioactivity was counted in a well-type gamma counter (Auto-Gamma:

Packard Instrument Company. Downers Grove. IL) and expressed in cpm. Thecalculated concentration of cytokeratin 19 was expressed ¡nng/ml.

Control of the Immunoradiometric Assay. The sensitivity of the methodwas assessed by determining the MDC. This method used two cytokeratin 19standards containing, respectively. 0 (cytokeratin free. StO) and 3 ng/ml (Stl):these 2 standards were measured in 6 different assays (Table 2): in each of the6 assays, the same total radioactivity was introduced for both standards. Thebound/total radioactivity was calculated for StO and Stl. The MDC in ng/ml

was determined as:

X x

where A is MDC in ng/ml: X is [standard Stl] - [standard StOJ = 3 ng/ml: Y ismean Stl (cpm) - mean StO (cpm): and y = 2 SD of StO (cpm).

The control of the reliability of the method of measurement was done usingthe following tests: (a) test of linearity of the relationship between consecutivedilutions of a serum of known high cytokeratin 19 level and results of thetitration; (b) assessment of intraassay reproducibility by measuring 7 differentsera twice during the same assay: (c) assessment of interassay reproducibilityby measuring 2 different sera in 10 different assays: (a1)analysis of the standard

curve in the low values in order to evaluate the difference of signal given bythe StO and the St l.

Total LDH assays were done following the Deusch Chemical Societyrecommendations by measuring its activity using pyruvate as a substrate(Bio-Merieux, France).

The upper limits of normal values were as follows: LDH, 330 units/liter:alkaline phosphatase. 220 units/liter; leukocytes 8000/ul. The lower limits ofnormal values were 32 g/liter for albumin and 135 mmol/liter for serumsodium.

All sera samples were assayed blind of clinical information.Statistics. ROC curves were constructed using both patient and control

subject serum CYFRA 21-1 levels ¡nan attempt to establish a sensitivity-

specificity relationship; areas under the ROC curve were calculated (18). Theserum tumor marker was not distributed normally: thus, for each patient subset,results were expressed as median, and variation was expressed as interquartilerange. Nonparametric statistical analyses were used: differences between twoindependent groups were determined by means of the Mann-Whitney U test;

differences between more than two groups were determined by means of theKruskal-Wallis one-way analysis of variance: P < 0.05 was considered as

significant. Survival was defined as the time from the date of sampling to thedate of death. Survival data were updated in February 1992 and none of thepatients were lost from sight during follow-up. Probability of survival was

estimated by the method of Kaplan and Meier (19). Single variable survivalanalyses were done by means of Wilcoxon and log rank tests and multivariateregression was done with the Cox's model (20). There were 12 variables in the

model. Serum CYFRA 21-1 was examined as normal (<3.6 ng/ml) or elevated

(>3.6 ng/ml) as were the other biological variables (LDH, serum sodium,alkaline phosphatase. albumin, and leukocytes) according to their respectiveupper normal limits. Age groups were defined as less than or equal to and over50 years old. Other variables were stage of the disease, sex, histology, weightloss, and performance status. Survival was analyzed using the SAS software

package.

RESULTS

Sensitivity and Reliability of the Immunoradiometric Assay.According to the data in Table 2 the calculated MDC was 0.06 ng/ml.The test of linearity showed a linear relationship between the titra-tions of CYFRA 21-1 and the dilutions of a serum sample with a

high cytokeratin 19 concentration (Fig. 1). Seven sera were measuredtwice during the same assay (intraassay). The results of the first andsecond measurements were, respectively (ng/ml): 12, 11.4; 27, 30; 6,7.5; 3, 2; 4.4, 4.8; 45, 41 ; 3.4, 3.2. The coefficients of variation of 2sera tested in 10 different assays (interassays) were 7 and 9.8%,respectively.

Analysis of the standard curve in the low values demonstrated thatthe difference between 0 and 3 ng/ml was 700 cpm, a value sufficientto prevent overlap of results.

Tumor Marker Distribution. The median and interquartile rangeinterquartile range, of serum CYFRA 21-1 was significantly higher in

Table 2 Results of 6 different inuminoradiotnetric assays performed in orderto determine the minimal detectable concentratitm

StO Stl

Assay123456T" (cpm)226.473219.399208.757241.441207.643198.408cpm358976718364B/T0.020.040.040.030.040.03cpm1.9472.6481.9391.7731.7761.633B/T0.991.200.930.730.850.82

" B, bound radioactivity; T, total radioactivity.

Research. on February 5, 2021. © 1993 American Association for Cancercancerres.aacrjournals.org Downloaded from

SERUM CYFRA 21-1 IN LUNG CANCER

>SO y •¿�»8.78 X - 0.17•¿�0.998; p •¿�0.002

1:16 1:32 1:64

Serum dilutions

Fig. l. Dilution test of a serum with a high level of cytokeratin I1).

level differed when the histológica! type was considered (Kruskal-Wallis test KW = 7.69; P < 0.05). Serum CYFRA 21-1 level was

significantly higher in SQC when compared with each of the otherhistologies (Mann-Whitney U test, P < 0.05).

CYFRA 21-1 and Clinical Variables at Presentation. In NSCLC,the CYFRA 21-1 level, was significantly higher in patients with ametastatic disease (7.4; interquartile range, 3.0-21.0 ng/ml) in com

parison with the limited disease patient subgroup (3.8; interquartilerange. 1.9-8.5 ng/ml; Mann-Whitney U test, P < 0.001). When the

marker was analyzed according to stage grouping, a significant increase of the serum level was observed from stage I to stage IV(Kruskal Wallis. 13.7; P < 0.005; Fig. 4) but there was no significantdifference when stage Ilia and stage Illb were compared (Mann-Whitney U test, 500; nonsignificant). CYFRA 21-1 levels were lower

in patients who were operated upon when compared with inoperablepatients (2.4; interquartile range. 1.25-5.0 versus 5.8; interquartilerange, 2.7-13.0 ng/ml; Mann-Whitney U test, P < 0.001). In addition.

1.0

0.76

0.6

g 0.28

0.26 0.6 0.76 1.0

False-Positive Rate (1 - Specificity)

Fig. 2. Receiver operating eharaeteristie eurve for serum CYFRA 2I-1. O. squanmuscell carcinoma, area under Ihe concentration curve = 0.83 ±0.03 (SE); •¿�non-small celllung cancer, area under the concentration curve = 0.80 ±0.03.

cancer patients (4.3; interquartile range, 2.3-9.5) when compared withcontrol subjects (1.05; interquartile range, 0.8-1.8; Mann-Whitney

U test, P < 0.001). Areas under ROC curves were 0.83 ±0.03, 0.80±0.03, and 0.74 ±0.05 for the SQC, NSCLC. and SCLC groups,respectively (Fig. 2). Using ROC curves, a threshold of 3.6 ng/mlwas chosen as the upper limit of normal values. The sensitivity andspecificity of this cutoff in the different patient subsets are shown inTable 3.

CYFRA 21-1 and Histology. The median (interquartile range)serum CYFRA 21-1 for SCLC, SQC. adenocarcinomas, and large cellcarcinomas were 3.4 (1.8-7.0), 5.6 (2.6-13.0), 2.9 (1.85-7.10), and3.8 (1.8-7.0) ng/ml, respectively (Fig. 3). The serum CYFRA 21-1

Table 3 Sensitivity and specificity of CYFRA 21-1 using a 3.ft-n\>/ml upper limitof normal mines

Subsets Sensitivity Specificity

Non-small cell

SQCSmall cellAll

0.560.630.460.52

0.890.910.890.87

knuka Willi! p no'.

3.6

ADE SQC LCC SCLC

Fig. 3. CYFRA 21-1 distribution according to histology. . median value: columns.interquartile range. ADE. adenocarcinoma: LCC. large cell carcinoma.

24222018161412108642n-Krutk.lW.lli. p.0.006-----nn

TTq_ji tí—

.-3.6

I & II lila 1Mb IVFig. 4. CYFRA 21-1 distribution according to stage grouping in non-small eel

cancer. , median value; columns, interquartile range.lung

Research. on February 5, 2021. © 1993 American Association for Cancercancerres.aacrjournals.org Downloaded from

SERUM CYFRA 21-1 IN LUNG CANCER

patients who presented mediastinal lymph nodes (N2 or N3) demonstrated higher serum CYFRA 21-1 levels (5.0; interquartile range,2.6-12.5 ng/ml), when compared with patients without mediastinallymph nodes (NO or Nl, 3.2; interquartile range, 1.75-8.0 ng/ml;Mann-Whitney U test, P < 0.05).

In NSCLC, the distribution of CYFRA 21-1 varied significantlyaccording to the performance status of NSCLC patients (Kruskal-Wallis 9.16; P < 0.05; Table 4). In SCLC, CYFRA 21-1 did notsignificantly differ between limited (3.0; interquartile range, 1.8-5.4ng/ml) and extensive disease (4.55; interquartile range, 2.1-11.5 ng/ml; Mann-Whitney U test, P = 0.07).

Survival. Univariate analysis showed that lung cancer patientswith serum CYFRA 21-1 over 3.6 ng/ml proved to have a significantly

shorter overall survival than those with a normal serum level (logrank, P = 0.007; Wilcoxon. P = 0.001 ; Fig. 5). The negative prognosticeffect of CYFRA 21-1 was highly significant in SQC (log rank, P <

0.001; Wilcoxon, P < 0.001) whereas it was nonsignificant for theother histologies. Separate survival analyses showed a significanteffect of stage group, weight loss, performance status, serum albumin,LDH, and CYFRA 21-1 (Table 5). With Cox's model analysis, per

formance status, stage grouping, and CYFRA 21-1 were the only

significant determinants of survival (Table 6).

DISCUSSION

New strategies of chemotherapy intensification are under evaluation in SCLC patients in an attempt to circumvent secondary chemore-

sistance (3, 4). NSCLC is a group of different histologies having alower chemosensitivity when compared with SCLC. Surgery of stageI and II NSCLC offers the best probability of cure (5, 21) but in themore advanced stages combined modality treatments are required

Table 5 Significant prognostic factors in the entire population

Median survivaFactor and level(dayslStageI

andIIIliaHlbIVPerformance

status9-12-4Wt

lossNowtloss0-5%>5%Albumin

(g/liter)>32<32Láclate

dehydrogenase(units/liter)<330>330CYFRA

21-1(ng/ml)<3.6>3.6420:?4155407148402159191367191398252402213f

1WilcoxonLogrank0.0001

0.00010.0001

0.00010.0018

0.00150.0049

0.0370.0041

0.01770.001

0.007

Table 6 Regression analvsis results: estimated relative risk for significant variablesin = 165)

VariablePerformance

indexStage of diseaseCYFRA 21-1Regression

coefficient(ß±SE)1.03

±0.220.90 ±0.290.49 ±0.21P<0.0001

0.00240.021Relative

risk2.81

2.471.64

Table 4 CYFRA 21-! distribution according to performance status (PS)in non-small cell lung cancers

PSOPS1PS2PS3Median2.7"

4.36.0

10.0Interquartile

range(ng/ml)1

.3-5.02.2-9.52.7-16.04.6-84.0n10

78319

1Kruskal-Wallis = 9.16; P < 0.05.

£ O.B

t 0.6"o

i 0.4

Z5 0.2s

Log-rink p -0.007Wilcoxon p -0.001

400days

CYFRA <•3.6 nQ/mL

600

CYFRA > 3.6 ng/mL

800

median «urvlval % c«n«ortd

<•3.6 ng/HL

> 3.6 ng/mL•¿�0

86

40«

fit

SS

34

Fig. 5. Probability of survival of all patients with normal and elevated pretreatmentserum CYFRA 21-1 level.

owing to the frequency of microscopic metastatic disease (7, 22-24).

Thus, the treatment of both SCLC and NSCLC lung cancer is still anexperimental approach. In this setting, an awareness of significantvariables able to predict metastatic spread and poor prognosis isrecommended (25).

Some important prognostic factors of lung cancer have already beenreported. For NSCLC, the two main variables which determine survival are the stage of the disease and the performance status (8). Inseveral studies, tumor markers such as tissue polypeptide antigen(26, 27), LDH (8), or carcinoembryonic antigen (26, 27) have beendemonstrated as significant prognostic factors in univariate analyses.Other biological properties of NSCLC have been investigated as putative indicators of prognosis. In different multivariate studies aneu-

ploidy (28), lack of expression of blood group A antigen (29) orexpression of the neural cell adhesion molecule (30) have been independently demonstrated as prognostic factors; however, correlationand interactions between these variables are not known. For SCLC,several studies have demonstrated the univariate influence of diseaseextent, performance status, LDH, and alkaline phosphatase on survival (9, 31-33); moreover, some authors have suggested that the

poorest prognosis is seen in SCLC patients with bone marrow métastases (34), elevated serum neuron specific enolase (35), or carcinoembryonic antigen (26).

Intermediate filaments are part of the cytoskeleton which provideimportant information on the cell origin (36). They are classified intofive tissue specific protein families. The cytokeratin family is expressed by all epithelial cells (10) including endocrine cells of thedispersed neuroendocrine system (37) and, it appears, therefore, to bea general specific and useful marker of the epithelial differentiation.Interestingly, the expression of cytokeratins is not lost by epithelialcells during malignant transformation (10, 36), a phenomenon which

64

Research. on February 5, 2021. © 1993 American Association for Cancercancerres.aacrjournals.org Downloaded from

SI-RUM CYI-RA 21-1 IN l.UNG CANCI-R

contrasts with the well-known phenotypic instability of cancer cells

(38). Antibodies raised against some common antigenic determinantsof all cytokeratins are therefore useful in typing malignant tumorswith poor histologie-aà f̄eatures of differentiation (36). Immunohis-

tochemical studies using these broad-spectrum cytokeratin antibodies

have clearly demonstrated that all four histológica! types of lungcancer expressed cytokeratins (II, 12, 39); in particular, the classicSCLC type which have been shown to coexpress both cytokeratin andneurofilament whereas the variant type only express cytokeratin (40).This supports the hypothesis of a common epithelial lineage of allhistológica! types of lung cancer. A cytokeratin is a heterotypic tet-

ratner of protofiluments composed of two polypeptidcs: one acidictype I subunit; and one basic type II subunit. Each type of epitheliaand their malignant counterpart express a specific cytokeratinpolypeptide pattern (10, 41); simple epithelia, including pseudostrat-

ified epithelia. express cytokeratins 7, 8. 18, and 19. Selective antibodies raised against simple epithelium type of cytokeratin have beenshown to react with all histologies of lung cancers (42, 43). Althoughcytokeratins are part of the cytoskeleton some fragments might bereleased in the serum owing to cell lysis or tumor necrosis, whichgives support to the present research on cytokeratin subunit 19 fragment as a tumor marker of lung cancer.

The immunoradiometric assay described in this study is able todetect the serum fragment of cytokeratin subunit 19, referred to asCYFRA 21-1, with both sensitivity and reliability. The threshold of

3.6 ng/ml was determined using the ROC curve analysis whichshowed that this threshold allowed the best sensitivity-specificity

relationship. Our study has demonstrated a correlation between serumCYFRA 21-1 and stage of the disease in NSCLC but not in SCLC.

Moreover, in NSCLC, the marker varied significantly according tonodal status. These data suggest that CYFRA 21-1 might be consid

ered as a marker of tumor mass. Thus, patients who present a highlevel of serum CYFRA 21-1 might require a careful search of métas

tases. Despite these results it is not possible to predict operability ofNSCLC using the serum level of the marker owing to the importantoverlap of the distribution in stage Ilia (usually marginally resectable)and stage Illb- (unresectable) NSCLC groups. The reason for this

result might be that the distinction between stage Ilia and stage Illbmainly depends on anatomic extension rather than tumor mass. Thelow incidence of a high serum CYFRA 21-1 level in the stage I-II

group clearly indicates that the marker is of no help in screening orearly diagnosis of the disease. Because CYFRA 21-1 is correlated

with both disease extent and performance status it is not surprisingthat univariate analysis showed that patients with a high serum levelproved to have a significantly shorter survival. The negative effect ofCYFRA 21-1 on survival was particularly significant in SQC whereas

this effect was not demonstrated in the other histologies. Interestingly.Cox's model demonstrated that the prognostic significance brought by

CYFRA 21-1 is independent and adds information to the well-knownand above-mentioned prognosis factors. This study supports the use ofthe serum fragment of cytokeratin subunit 19 CYFRA 21-1 as an

independent prognostic marker of squamous cell carcinoma of thelung.

ACKNOWLEDGMENTS

The authors wish to thank Jo Baissus for help in preparing Ihe manuscript.

REFERENCES

1. Stjemswurd. J.. and Stanley, K. Etiology, epidemiology and prevention of lung cancer.Lung Cancer, 4: 11-24. 1988.

2. Hansen. H. H. Lung cancer. In: H. M. Pinedo, D. L. Longo, and B. A. Chabner (eds.).Cancer Chemotherapy and Biological Response Modifiers, pp. 222-240. Annual 10.Amsterdam: Elsevier/North-Holland BiomédicalPress, 1988.

3. Klastersky. J. Commentary: intensive chemotherapy in small cell lung cancer. Eur. J.Cancer Clin. Oncol., 25: 933-937, 1989.

4. Hansen. H. H., and Kristjansen. P. E. G. Chemotherapy of small cell lung cancer. Eur.J. Cancer. 27: 342-349, 1991.

5. Mulshine, J. L.. Glatstein, E.. and Ruckdeschel. J. C. Treatment of non-small cell lungcancer. J. Clin. Oncol., 4: 1704-1715. 1986.

6. lannuzzi. M. C.. and Scoggin. C. H. State of the art: small cell lung cancer. Am. Rev.Rcspir. Dis.. 134: 593-608. 1986.

7. Cox. J. D.. and Yesner, R. A. Causes of treatment failure and death in carcinoma ofthe lung. Yale J. Biol. Med.. 54: 201-207. 1981.

8. O'Connel. J. P.. Kris. M. G.. Gralla. R. J.. Groshen, S., Trust, A., Fiore. J. J.. Kelsen.

D. P., Heelan, R. T., and Golhey. R. B. Frequency and prognostic importance ofpretrcatment clinical characteristics in patients with advanced non-small cell lungcancer treated with combination chemotherapy. J. Clin. Oncol.. 4: 1604-1614. 1986.

9. Osterlind. K.. [hde, D. C.. Ettinger. D. S.. Gralla. R. J.. Karrer. K.. Krauss. S., Maurer.L. H.. Rorth. M-, Sorenson. S., and Vincent. R. Staging and prognostic factors in smallcell carcinoma of the lung. Cancer Treat. Rep.. 67: 3-9, 1983.

10. Moll, R.. Franke. W. W.. Schiller. D. L.. Geiber. B., and Krepier, R. The catalog ofhuman cytokeratins: patterns of expression in normal epithelia. tumors and culturedcells. Cell, 31: 11-24, 1982.

11. Broers, J. L. V.. Rot, M. K., Oostendorp. T., Huysmans, A., Wagenaar. S. S., Wiersma-

van Tilburg. A. J. M.. Vooijs. P, and Ramaekers, F. C. S. Immunohistochemicaldetection of human lung heterogeneity using antibodies to epithelial, neuronal andneuroendocrine antigens. Cancer Res., 47: 3225-3234, 1987.

12. Pujol, J. L.. Simony, J., Laurent. J. C., Richer, G., Mary. H.. Bousquet, J.. Godard. Ph..and Michel, F. B. Phenotypic heterogeneity studied by immunohistochemistry andaneuploidy in non-small cell lung cancers. Cancer Res., 4V: 2797-2802, 1989.

13. World Health Organization. The World Health Organization Histológica! Typing ofthe Lung Tumors. Ed. 2. Am. J. Clin. Pathol., 77: 123-136. 1982.

14. Sobin. L. H.. Hermanek. P.. and Hutter. R. V. P. TNM classification of malignanttumours. Ed. 4. Geneva: International Union Against Cancer. 1987.

15. Tisi. G. M.. Friedman, P. J.. Peters. R. M.. Pearson. G.. Carr. D.. Lee, R. E.. andSelawry, O. American Thoracic Society: clinical staging of primary lung cancer. Am.Rev. Respir. Dis., 125: 659-64. 1982.

16. Mountain. C. F. A new international staging system for lung cancer. Chest, 4(s):225s-233s, 1986.

17. Pujol. J. L., Demoly, P., Gaulier. V.. Romieu. G.. Dan Aouta. M., Stenger. R.. Parrai.IL, Tadlaoui, O., Marcenac, A. M., Benahmed. S. S.. Michel, F. B.. and Godard. P.Phase II study of alternating combination chemotherapy in small cell lung cancer.Lung Cancer, 7: 285-294, 1991.

18. Beck, J. R.. and Shultz. E. K. The use of relative operating characteristic (ROC)curves in test performance evaluation. Arch. Pathol. Lab. Med.. Ill): 13-20, 1986.

19. Kaplan. E. L.. and Meier. P. Nonparametric estimation from incomplete observations.J. Am. Slat. Assoc., 53: 457^81. 1958.

20. Cox. D. R. Regression models and life tables. J. R. Statist. Soc. B.. 34: 187-202,

1972.21. Gail. M. H., Eagan, R. T, Feld, R.. Ginsberg, R.. Goodell, B., Hill, L.. Holmes, E. C,

Lukeman. J. M.. Mountain. C. F.. Oldhani, R. K., Pearson, F. G., Wright, P. W., andLake. W. H. Prognostic factors in patients with resected stage I non-small cell lungcancer. Cancer (Phila.), 54: 1802-1813. 1984.

22. Pearson, F. G.. DeLarue, N. C.. lives, R.. Todd. T. R. J., and Oxiper. J. D. Significance(if positive superior mediastinal nodes identified at mediastinoscopy in patients withresectable cancer of the lung. J. Thorac. Cardiovasc. Surg., 83: 1-11, 1982.

23. Pujol, J. L.. Rossi, J. F.. Le Chevalier, T.. Daurès.J. P.. Rouanet. P.. Douillard, J. Y.,Dubois. J. B.. Amagada. R.. Mary, H.. Godard, P., and Michel. F. B. Phase 11pilotstudy of neoadjvam ifosfamide, cisplatin. and etoposide in locally advanced non-small cell lung cancer. Eur. J. Cancer. 26: 798-801, 1990.

24. Tubiana. M. The growth and progression of human tumors: implications for management strategy. Radiothcr. Oncol.. 6: 167-184. 1986.

25. World Health Organization. WHO Handbook for Reporting the Results of CancerTreatment, WHO Offset Publication No. 48. Geneva: WHO. 1979.

26. Buccheri. G. F. Violante, B., Sartoris, A. M., Ferrigno, D., Curcio, A., and Vola. F.Clinical value of a multiple biomarker assay in patients with hronchogcnic carcinoma.Cancer (Phila.), 57: 2389-23%. 1986.

27. Salvati. F.. Cruciani, A. R., Flore, F.. De Angelis. G.. Pigorini. F.. Antilli. A.. Pau, F..Munno, R., and Cipri, A. Plasma carcinocmhryonic antigen and tissue polypeptideantigen levels in lung cancer: correlation with cell types and stage. Cancer. Detect.Prev., 8: 111-114, 1985.

28. Zimmerman, P. V. Hawson, G. A. T.. Bint. M. H.. and Parsons. P. G. Ploidy as aprognostic determinant in surgically treated lung cancer. Lancet, 2: 530-533, 1987.

29. Lee. J. S., Ro. J. Y., Sahin. A. A.. Hong. W. K.. Brown. B. W.. Mountain, C. F. andHittelman. W. N. Expression of blood-group antigen A: a favorable prognostic factorin non-small cell lung cancer. N. Engl. J. Med.. 324: 1084-1090. 1991.

30. Bcrendsen, H. H.. De Leij. L.. Poppema. S.. Postmus. P. E.. Boes. A.. Sluiter. H. J..and The. H. Clinical characterization of non-small-ccll lung cancer and tumors showing neuroendocrine differentiation features. J. Clin. Oncol.. 7: 1614-1620. 1989.

31. Rawson, N. S., and Peto. J. An overview of prognostic factors in small cell lungcancer. A report from the subcommittee for the management of lung cancer of theunited Kingdom co-ordinating committee on cancer research. Br. J. Cancer, 31:597-604, 1990.

32. Osterlind. K.. and Andersen. P. K. Prognostic factors in small cell lung cancer:multivariate rmxJel based on 778 patients treated with chemotherapy with or withoutirradiation. Cancer Res., 46: 4189-4194. 1986.

33. Albain. K. S.. Crowley, J. J., LeBlanc. M., and Livingston, R. B. Determinants ofimproved outcome in small cell lung cancer: an analysis of the 2580 patient Southwest Oncology Group data base. J. Clin. Oncol.. K: 1563-1574. 1990.

65

Research. on February 5, 2021. © 1993 American Association for Cancercancerres.aacrjournals.org Downloaded from

SERUM CYFRA 21-1 IN LUNG CANCER

34. Van Den Brande. P. Serum láclatedehydrogenase in small cell lung cancer. N. Engl. Wamaar, S. O. Cytokeralin and neurofilament in lung carcinomas. Am. J. Pathol., 116:¡.Med.. 320: 61, 1989. 363-369. 1984.

35. Jorgensen. L. G.. Osterlind. K., Hansen. H. H., and Cooper, E. H. The prognostic 40. Broers. J. L. V.. Carney. D. N.. De Ley, L., Vooijs, G. P., and Ramaekers, E C. S.influence of serum neuron specific enolase in small cell lung cancer. Br. J. Cancer. 58: Differential expression of intermediate filament proteins distinguishes classic from805-807, 1988. varianl small-cell lung cancer cell lines. Proc. Nati. Acad. Sci. USA. 82: 4409^1413.

36. Osbom. M, and Weber, K. Intermediate filaments: cell-type-specific markers in 1985.differentiation and pathology. Cell, 31: 303-306, 1982. 41. Moll, R., Krepier, R., Franke, W. W. Complex cytokeratin polypeptide patterns

37. Hoefler, H., and Denk, H. Immunohistochemical demonstration of cytokeratins in observed in certain human carcinomas. Differentiation, 23: 256-269, 1983.gastrointestinal carcinoids and their probable precursors cells. Virchows Arch. B Cell 42. Debus, E., Moll, R.. Franke, W. W., Weber, K., and Osbom, M. ImmunohistochemicalPalhol., 403: 235-240, 1984. distinction of human carcinomas by cytokeratin typing with monoclonal antibodies.

38. Nicolson. G. L. Tumor cell instability, diversification, and progression to the meta- Am. J. Pathol., 114: 121-130, 1984.static phenotype: from oncogene to oncofetal expression. Cancer Res., 47: 1473- 43. Blobel, G. A.. Moll. R., Franke, W. W., and Vogt-Moykopf, I. Cylokeratins in normal

1487, 1987. and lung carcinomas. I. Adenocarcinomas, squamous cell carcinomas and cultured39. Van Muijen. G. N. P., Ruiter. D. J., Van Leeuwen, C. Prins, F A.. Reitsema, K., and cell lines. Virchows Arch. B Cell Pathol., 45: 407-^»29,1984.

66

Research. on February 5, 2021. © 1993 American Association for Cancercancerres.aacrjournals.org Downloaded from

1993;53:61-66. Cancer Res   Jean-Louis Pujol, Jean Grenier, Jean-Pierre Daurès, et al.   21-1 Immunoradiometric Assay as a Marker of Lung CancerSerum Fragment of Cytokeratin Subunit 19 Measured by CYFRA

  Updated version

  http://cancerres.aacrjournals.org/content/53/1/61

Access the most recent version of this article at:

   

   

   

  E-mail alerts related to this article or journal.Sign up to receive free email-alerts

  Subscriptions

Reprints and

  .pubs@aacr.orgDepartment at

To order reprints of this article or to subscribe to the journal, contact the AACR Publications

  Permissions

  Rightslink site. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC)

.http://cancerres.aacrjournals.org/content/53/1/61To request permission to re-use all or part of this article, use this link

Research. on February 5, 2021. © 1993 American Association for Cancercancerres.aacrjournals.org Downloaded from