prediction of occult neck metastases in laryngeal ... · pendent predictors of occult nodal disease...

Vol. 2, 1801-1808, October 1996 Clinical Cancer Research 1801

Prediction of Occult Neck Metastases in Laryngeal Carcinoma: Role

of Proliferating Cell Nuclear Antigen, MIB-1, and E-Cadherin

Immunohistochemical Determination

Alessandro Franchi, Oreste Gallo, Vieri Boddi,

and Marco Santuccit

Institute of Anatomic Pathology [A. F., M. S.]. Otolaryngology Clinic[0. 0.], and Institute of General Pathology [V. B.], University of

Florence, Viale 0. B. Morgagni 85, 50134 Florence, Italy

ABSTRACT

The aim of this study is to investigate the predictive

value of proliferative activity assessment and E-cadherin

expression by means of immunohistochemistry in identify-

ing patients with laryngeal squamous cell carcinoma at a

high risk for occult node metastasis. Thirty consecutive

patients treated for laryngeal carcinoma with false clinically

negative nodes (occult metastases, pN+) between the years

1980 and 1990 were selected for this study. A group of 30

cases with negative cervical lymph nodes (pN-) having a

similar anatomic site and tumor size distribution was used

as control. In each case, several histological parameters,

including grade, pattern of invasion, number of mitosis (X 10

high-power field), tumor inflammatory infiltrate, and tumor

sclerosis, were assessed. Proliferative activity was deter-

mined using immunohistochemical staining for proliferating

cell nuclear antigen (PCNA) and MIB-1. Other putative

prognostic factors investigated at the immunohistochemical

level were the cell adhesion molecule E-cadherin and two

oncoproteins, p53 and c-erbB-2. In pN+ cases, the expres-

sion of PCNA and MIB-l was significantly higher than in

the pN- group. Moreover, a significant loss of E-cadherin

expression was observed in carcinomas with occult metas-

tases. No differences in p53 and c-erbB-2 oncoproteins were

found between pN+ and pN- cases. Among the other path-

ological parameters examined, only histological grade was

significantly associated with the presence of occult metasta-

ses, but on mubtivariate analysis, this relationship was lost.

We conclude that PCNA, MIB-1, and E-cadherin are inde-

pendent predictors of occult nodal disease in laryngeal

squamous cell carcinoma, and their immunohistochemical

determination could be useful in identifying patients with

clinically negative lymph nodes who are at considerable risk

for occult metastases and who may benefit from elective

neck dissection.

Received 3/26/96: revised 7/3/96; accepted 7/3/96.I To whom requests for reprints should be addressed. Phone: 39 55416951; Fax: 39 55 4379868.

INTRODUCTION

Preoperative identification of occult lymph node metastasis

represents a crucial point in the clinical management of patients

with laryngeal cancer for its prognostic and therapeutic impli-

cations (I). Data from literature indicate that 4-40% of patients

with laryngeal carcinoma and clinically negative neck lymph

nodes have indeed occult metastases on subsequent histological

examination (2-6). As a consequence, treatment of N0 patients

is a matter of great controversy between those who are in favor

of a “wait and see” attitude and those who prefer elective neck

dissection. Indeed, prophylactic neck dissection eliminates the

risk of late metastases in N0 cases, but exposes the patient to a

higher morbidity and higher hospital costs, which are in most

cases unnecessary ( I , 3). Another possible disadvantage of

bymphadenectomy may be the local reduction of the immuno-

logical host defense, which may facilitate the spread of residual

tumor cells. Therefore, the identification of biological tumor

factors indicative of high or low risk of occult lymph node

metastases would be extremely helpful in deciding on elective

or delayed therapeutic strategy in the treatment of the neck in N0

laryngeal cancer.

Previous clinicopathobogical studies have shown that the

presence of locoregional metastases in patients with laryngeal

carcinoma is significantly associated with some histological

features of the primary tumor, such as the degree of differenti-

ation, tumor growth pattern, presence of peritumoral desmopla-

sia, and inflammation (7-9). Additional studies confirmed that

histological grade and host immune response correlate signifi-

cantly with occult lymph node metastasis (1 , 10). However,

because of the absence of established criteria, the evaluation of

this group of histological parameters is very subjective, and

reproducibility is poor.

In recent years, the efforts of several research groups have

bed to a dramatic improvement in the understanding of the

neoplastic growth process and have furnished novel molecular

tools that may help in identifying tumors more likely to metas-

tasize, being therefore potentially useful for management and

prognostic purposes. In particular, several lines of evidence

suggest that the evaluation of tumor cell proliferative activity,

cell adhesion markers, and oncoprotein expression could allow

an estimate of the biological behavior of head and neck squa-

mous cell carcinoma. When proliferative activity was evaluated

by tumor labeling index in a series of head and neck squamous

cell carcinomas, the results indicated that lesions with a high

proliferative activity have a poorer survival (1 1). Recently,

Benazzo et a!. observed a significant correlation between pro-

liferative indices measured after in vivo bromodeoxyuridine

incorporation and tumor dimension, histological differentation,

and lymph node involvement in a group of patients with carci-

nomas of the head and neck ( 1 2). Similar results were obtained

Research. on June 12, 2020. © 1996 American Association for Cancerclincancerres.aacrjournals.org Downloaded from

http://clincancerres.aacrjournals.org/

1802 Predictors of Occult Metastases in Laryngeal Carcinoma

2 The abbreviations used are: PCNA, proliferating cell nuclear antigen;

OR, odds ratio; Cl, confidence interval.

Table 1 C linicab features of 60 pa tients with 1arynge al squam ous cell carcinoma, the obj ect of this study

Site T” Treatment

Partial Total Age (Mean) Sex M:FSupraglottic Glottic 1 2 3 4

pN+ 20 10 7 8 10 5 13 17 58.5 29:1

pN- 20 10 8 9 8 5 16 14 59.3 29:1

a TNM staging.

when proliferation was measured by flow cytometric analysis of

DNA content (13).

Loss of intercellular adhesion might be an important event

in the early phases of the metastatic process. The potential

relationship between the expression of the adhesion molecule

E-cadherin, a member of a family of calcium-dependent cell-cell

adhesion molecules also including P- and N-cadherin, and thepresence of lymph node metastasis in laryngeal carcinoma has

been the object of studies by different groups. Schipper ci’ a!.

observed that E-cadherin expression decreased with loss of

differentintion in primary cnrcinomns, and thnt lymph node

metastases expressed a lower bevel of the protein. suggesting an

important role of cadhenn loss in the metastatie process ( 14). In

a subsequent study, Mattijssen e� a!, were unable to find anysignificant relationship between the level of li-eadherin expres2sion and the presence of lymph node metastasis in a group of 50

head and neck squamous cell carcinomas ( I 5) However. these

studies only examined patients with clinical evidence of mela-

static disease. and no information is available in the literature

about the relationship between E-cadherin expression and the

presence of occult lymph node metastases,Indeed, most of the studies concerning predictors of the

behavior of carcinomas of the head and neck have consideredheterogeneous groups of lesions concerning both the site of

primarylesion and the clinical node status, The present study is

focused on a homogeneous population of patients with laryngealsquamous cell carcinoma and clinically negative neck lymph

nodes with the aim of defining prognostic parameters that per�mit the identification of patients at risk for occult metastases,We evaluated several histopathological features of primary le-

sions, and determined immunohistochemically tumor cell pro-

liferative activity and E-cadherin expressionS Moreover, westudied the expression of e-erbB-2 and p53 oneoproteins. the

products of two genes known to be involved in the development

of head and neck squamous cell carcinomas (16-18),

PATIENTS AND METHODS

Cases Studied. Thirty consecutive patients treated at the

Institute of Otorhinolaryngology of the University of Florence

between 1980 and 1990 for laryngeal squamous cell carcinomawith occult baterocervical metastases were selected for thisstudy, All patients had clinically negative laterocervical lymphnodes and underwent elective neck dissection, with subsequenthistological recognition of one or more occult metastases

(pN + )‘ A group of 30 patients treated in the same period for

laryngeal squamous cell carcinoma with clinically and histobog-

ically negative laterocervical lymph nodes (pN-) was used as

control, Clinical data of the two groups of patients are summa-ri�ed in Table 1.

Formalin-fixed, paraffin-embedded tissue samples from re-

section specimens were retrieved in each case, and serial see-

tions were stained with H&E or used for the immunohistochem-

ical studies. In each case, the following histological parameters

were assessed: histological degree of differentiation (well, mod-

erately, or poorly differentiated), pattern of invasion (pushing or

infiltrating), number of mitoses (X 10 high-power field), peritu-

moral inflammatory infiltrate (present or absent), and peritu-

moral sclerosis (present or absent).

Immunohlstochemlstry. lmmunohistochemical studies

were performed by the uvidin-biotin complex technique (Duko-

putts. High Wycombe. UK)� Proliferative activity was investi-

gated using monocbonul antibodies against Ki-67 (MIB-l. lm-

munotech, Marseille, France) and PCNA2 (PC-lO, DAKO

Corp,, Carpinteria, CA). For E-cudherin detection, the mono-

clonal antibody HECD- 1 (Nuclear Laser Medicine, Milan, Italy)was employed, For p53 and c-erhB-2 detection, the monoclonal

antibodies D07 (DAKO Corp.) and mAhI (Triton Diagnostics.

Alameda, CA) were used, MIB-1, E-cadherin, and p53 immu-

nostuining required microwave pretreatment of dewuxed and

rehydrated sections, Negative controls were performed by sub-stitution of the primary antibody with nonimmune rabbit ormouse serum,

Scoring of the immunostaining was performed by oneobserver (A, F.). who had no knowledge of patients’ clinical

status, using a standard light microscope equipped with an

eyepiece grid of 10 X 10 squares and a X40 objective, In each

case, at beast 1000 tumor cells were counted from areas withhigh and low expression of the antigen. All stained cells were

considered positive regardless of the intensity of the staining.

and the results were expressed in percentage of’ positive cells.

An arbitrary cutoff value of 20% positive cells was chosen to

separate cases with high and low expression of the antigens.

Statistical Analysis. Statistical tests were performed us-ing EGRET (Statistics and Epidemiology Research Corporation.Seattle, WA) and Stata (Stata Corporation. College Station,

TX) The relationship between clinical and pathological van-

ables was assayed by Fisher’s exact test, The correlation be-

tween the different immunohistochemical indices was evaluated

using Spearman’s correlation coefficient. The role of each pos-sible prognostic factor (univaniate analysis) and the joint effect

of all these factors (multivaniate analysis) was explored using

the multivaniate logistic regression analysis, The final results of

these analysis are the ORs and their 95% CIs, The likelihood



Clinical Cancer Research 1803

The distribution of percentages of positively stained cells

Table 2 Evaluation of prognostic parameters in 30 pN+ and 30

pN- squamous cell carcinomas of the larynx

Prognostic factors Score pN+ pN-

PCNA >20% 24 6MIB-l >20% 17 5E-cadherin �20% 15 4

C-erbB-2 Negative

�50%

>50%

21

3

6

226

2

p53 Negative

1-10%

11-60%60-100%

11

3

88

10

4

II5

Histological grade I

IIIII

S

178

1 1

181

Pattern of invasion

Number of mitoses

PushingInfiltrating

� 10

14I 619

191 118

Peritumoral inflammation AbsentPresent

1911

1812

Peritumoral sclerosis Absent

Present

18

12

12

18

ratio statistic was used to test the following hypothesis: OR =

1.0. P values <0.05 were considered significant.

RESULTS

Table 2 synoptically summarizes the results of the immu-

nohistochemical and histopathological analysis of the pN+ and

pN- groups.

All 60 tumors stained positively for PCNA, which local-ized in the nucleus, with different degrees of intensity (Fig. 1).

Some of the mitotic figures were associated with weak anddiffuse cytoplasmic staining. The PCNA score for the whole

series ranged between 6.4 and 7 1 .7% (25.6 ± 15.9, mean ±

SD). Twenty-four patients in the pN+ group had a tumor with

PCNA score greater than 20%, whereas only six patients of thepN- group had carcinomas with more than 20% of positive

cells (P < 0.001; Fisher’s exact test).

Positivity for MIB-l was observed in all cases. Only nu-

clear staining having a finely granular pattern was observed

(Fig. 2). Almost all mitotic figures were stained. The percentage

of positive cells ranged between 2 and 49. 1% (mean, I 8.9 ±

10.3). In the pN+ group, 18 cases showed more than 20% of

positive cells, whereas in the pN- group, only 5 cases had a

MIB-l score >20% (P < 0.001 ; Fisher’s exact test).

Considering the whole series of 60 cases, the percentage of

labeled cells tended to be higher using the PCNA antibody.

However, a significant correlation between PCNA and MIB-limmunoreactivity in each case was found (r�� 0.61 ; P <

0.001). Both PCNA and MIB-l scores increased as tumorsbecame less differentiated (P = 0.03 and 0.2, respectively;

Fisher’s exact test). No relationship was found between PCNA

and MIB-l indices and mitosis count.E-cadherin expression varied greatly from case to case, and

in most tumor cells, immunoreactivity was associated with the

cell membrane (Fig. 3). Cytoplasmic staining was observed in

seven cases, and four of these were poorly differentiated card-

nomas. The percentage of E-cadherin-positive cells decreased

‘ ‘ �-.I �

� � *�a’�$ � , b�’ � ) -

. #{149}*s��r#{149}O?�1r�, ‘0#{149}.

� :� � ‘� ‘-“

�. � �� -4 &� � ‘ �

� �: � � . �

� ‘ �-. a:

#{149}‘#{248}’U#{216}? � “ �

‘. ,4H,#{149}!* %� �

� 4� � ‘a #{149}�*�

t� � � , � -

. � � !�_-�..�,‘s’ � ‘-

# V A � � ‘- at#{149}%4#{149} S : ,� � � � -

-� . . #{163}. -‘-

4’ . . a � S � ..�

-.- #{149}�-‘,d��’.. 1�b. S*� Q!.p� *�

, �. .� . � J

� .4, � . -

Fig. I Diffuse nuclear immunostaining for PCNA in a poorly differ-

entiated squamous cell carcinoma.

significantly in the less-differentiated tumors (P = 0.00 1 ; Fish-

er’s exact test). Moreover, a significant association was found

between bow levels of cadherin expression (�20% of positive

cells) and the presence of occult lymph node metastasis (P =

0.006; Fisher’s exact test). When the expression of proliferation

markers (PCNA and MIB-b) was compared with E-cadherin

expression in each single case of the whole series, no significant

relationship was found. However, when groups with high and

low expression of the antigens (>20% and �20%) were com-

pared, it became evident that cases with high proliferative ac-

tivity tended to have a low expression of E-cadherin. In partic-

ular, a significant relationship was found between high

expression of PCNA and low expression of E-cadherin (P =

0.004; Fisher’s exact test).

Regarding p53 expression, 39 cases (65%) showed positive

nuclear staining for the oncoprotein (Fig. 4). No significant

differences were found in p53 levels between the pN+ and pN-

group using either a four-point scale scoring system ( - , nega-

tive; +, 1-10%; ++, 11-50%; and +++, >50%: Ref. 19) or

a single cutoff point (�20% or >20% positive cells). No rela-

tionship was evidenced between p53 expression and prolifera-

tion indices. c-erbB-2 oncoprotein was detected in 17 cases

(28.3%). In 14 instances, the immunostaining localized at the

cytoplasmic membrane bevel (Fig. 5), and in 3 there was diffusecytoplasmic staining. Six cases from the pN+ group and two

cases from the pN- group showed a number of positive cells

greater than 20%, but this difference was not significant. Sim-

ilarly, no significant differences were found using a three-point

scale scoring system (- , negative; +, �50%; and + +, >50%:

Ref. 20).




p �, , . , I :‘..�

.t 1�1�� #{149} � .

., ‘ v#,0, . # .

,.I � - . ‘ ‘.

,�,1 .�‘

t�’.t’�, ,: . ‘, _, � a. ,‘;‘3t�,�ms. .,. -� ‘ ‘.‘!

... ‘ I � .�-, , � ., ..f,-#{149}:’� ‘ ‘ “ .,� p . ,

�:/. #{149} � ;dI..!.� .;�..‘� .c-. #{149}

.. #{182}.t�, .# -,

)4. � ‘ 4 ‘‘“S ‘ � ,,�-. -. . . ,.-,-.

.:i’� �#{149}s�’. � . � ;‘�‘.

‘,.. ,. I ‘%��‘�

2 a-,. , #{149}�,, . , : #{149} ‘�‘ �“--:- � #{149}‘�, ,, ‘I

. : _� . . . � I. ,,� , #{149}, #{149}�#{149}�, . ‘. . .,, ‘.� #{149}

‘a ;�‘r’�i:’ � . � � .% . I ‘ #{149}_4._ ‘.. ‘: #{149}.. . , �. .1g.

..-.�-I’ � � , � “- -. . ‘ � .

� /0 “ ‘“s � .

-.st. � ‘ � .“,. �_‘.; 4

pI.,,,* #{149}.::!r, � #{149}

. ‘ . #{149}�-‘�.;‘ -j.: �

- �

� ‘ , M� ‘ :� - -.�. � -. . , � �- .

. .‘ ‘Si,,’ . � ‘..‘

. # ‘�. . ‘. .

Fig. 2 MIB-l nuclear immunoreactivity in a well-differentiated squa-

mous cell carcinoma. The staining is distributed predominantly at theperiphery of tumor nests.

using the antibodies against PCNA, MIB-l, E-cadherin, p53,

and c-erbB-2 in each individual tumor of pN- and pN+ groups

is reported in Fig. 6.

Among the several histopathological features analyzed,

only the histological grade had a significant relationship with

the presence of occult metastases (P 0.02; Fisher’s exact test),

as patients with poorly differentiated carcinomas had a higher

risk of occult metastatic disease.

Univariate logistic regression analysis of the different pa-

rameters showed that high levels of PCNA and MIB-l, low

levels of E-cadherin, and poor histological differentiation were

significantly associated with the presence of occult lymph node

metastases. On multivariate analysis, only PCNA, MIB-l, and

E-cadhenn expression maintained an independent status for the

prediction of occult lymph node metastases (Table 3).

Comparison of the immunohistochemical findings in

T3-T4 tumors versus T1-T2 tumors showed that the advanced

lesions had a higher proliferative activity as assessed with both

PCNA and MIB-l stainings (P = 0.009 and P = 0.07, respec-

tively; Fisher’ s exact test) and decreased E-cadherin expression

(P = 0.07; Fisher’s exact test). Levels of p53 and c-erbB-2

expression were comparable in the two groups. No significant

difference for any of the markers tested was evidenced between

gbottic and supragbottic carcinomas.

DISCUSSIONThe metastatic process is the result of the progressive

acquisition of an aggressive tumor cell behavior, characterized

by increased cell proliferation rates, modification of cell adhe-

sive properties, and genetic alterations. Accordingly, in the

present study, we observed that in laryngeal squamous cell

carcinoma, a high proliferative activity assessed by MIB-l and

PCNA immunostaining, and low expression of the adhesion

molecule E-cadherin, significantly correlated with the presence

of occult nodal metastases.

Until recently, a major limitation in evaluating the prog-

nostic significance of these markers by means of immunohisto-

chemistry, which allows the in situ identification of the antigen,

was the necessity of employing fresh tissue, in which morphol-

ogy preservation is often suboptimal, thus precluding the use of

archive material. This problem has been overcome partially by

the introduction of methods of antigen retrieval, such as micro-

wave oven pretreatment of paraffin sections, which have greatly

facilitated the study of putative prognostic markers in surgical

pathology of tumors (2 1 ). The reliability of this technique in

studying cell proliferation and cell adhesion processes has been

demonstrated clearly by several groups (21-23). A double as-

sessment of proliferative activity using both PCNA and MIB-l

immunostainings was performed in this study, because it has

been pointed out that PCNA immunostaining may suffer from

some limitations, the major one being a reduction of positivity

when fixation time exceeds 48 h (24). On the other hand, MIB-l

determination seems not to be affected by fixation time (25), but

experimental data suggest that reduction of blood supply, which

is a common event in central areas of a tumor, may determine a

modification of the expression of the antigen (26). In the present

series, PCNA index tended to be higher than MIB-l index.

Nevertheless, a significant relationship was found between the

expression of the two antigens as already observed in other

malignancies (27-29). On the other hand, no relationship was

found between the expression of PCNA and MIB-l indices and

mitosis count, possibly because these parameters evaluate dif-

ferent phases of the proliferating cell cycle.

The results of the present study indicate that tumor cell

proliferation indices assessed by immunohistochemistry corre-

late strongly with the presence of occult metastases in laryngeal

squamous cell carcinoma. A similar relationship has been found

previously in carcinomas from different sites (28, 30). Several

studies have now confirmed the adverse prognosis of laryngeal

carcinomas with high proliferative activity as well as the asso-

ciation of other prognostic parameters, such as the degree of

differentiation and pboidy status, with enhanced proliferation

(1 1, 31). Therefore, enhanced proliferative activity may not only

be important in the early phases of malignant transformation of

the squamous epithelium of the larynx (32, 33) during the

transition from dysplasia to invasive carcinoma but could also

be a feature of lesions with a higher tendency to spread and

metastasize. Our findings are discordant with those of a recent

study, in which no association was found between PCNA and

Ki-67 indices and the presence of lymph node metastasis in a

series of laryngeal carcinomas (34). A comparison of the results

is difficult, because the distribution of PCNA and Ki-67 indices

in the node-negative and node-positive groups is not indicated in

this article. In addition, the carcinomas examined belonged to

different histotypes (squamous cell, spindle cell, large cell un-

differentiated, and neuroendocrine), and patients had different

clinical N status, whereas we considered only squamous cell

carcinomas in patients with clinically negative neck lymph

nodes.



siveness (35, 36). Our results confirm that there is a significant

loss of E-cadherin expression in poorly differentiated squamous

cell carcinomas of the larynx ( 14, 15). Moreover, we observed

that loss of tumor differentiation was also accompanied by

Fig. 3 E-cadherin expression in

a well-differentiated squamous

cell carcinoma. The immuno-staining is localized at the cell

membrane of most tumor cells(A). In a moderately differenti-ated carcinoma, only a few tu-

mor cells express the antigen (B).


‘. � � � ‘#{149},�.J � , . �

. . t,,#{149},., . ‘�;“% #{149}... ,

‘ - . \�, . � . ,‘ � �

‘ ‘ S � �

�. : ‘� � ‘ � i’, . ..

‘ . , . ‘ , d� � �

� ‘‘..,: � --,“ #{149}� . �1#{149}� :‘�.

. ‘,�%#{234}� #e ‘

.‘ ..... . ‘#{149}.,� . �l . �

‘ . #{149}4’ “‘ �w . � ‘

? ‘,‘ . � #� ..# . . ‘.,bv..� :: t�1;

. ...�?‘ � )#{149}��

‘$. . � . �‘..... �‘;.. .. . .-. ,,.:‘�,, . ‘ .� � .

. ,.,.-,,‘. ..,.....- ‘#{149}#{149}__., �:. . ,. .#{149} ‘ . . . �‘�:�‘ � :�‘. �

. . (� � ..�.‘- �: ‘� �

#{149}‘�‘�;, ‘ \r,.�‘ �. I � ,. . � � ‘ �‘‘. ‘ ,�,..

,� � ‘a � , � ‘;� ‘

. . ‘‘::: � � ‘ .� � ‘ .. � . .‘�;-�.‘_ ‘.

. . �c”�” �‘�‘,: ‘�‘ � “ � . . .. ::‘ � ,..‘� ,:�, a. ,“‘.‘� .. ;,. . :‘. �. - . � ‘ .� ‘ ,d S,., ,

.. . � .,.,t � �. , ., ‘:� #{149}q�’ (. ‘ J%’’-.’ % � ‘ . ..‘ .,:‘� � t$.�%,:.??�:I, � � ., ‘. .‘

. . #{149}‘�b � g,’ � �‘. a: I ‘: S... ‘� � �#{149}:.‘‘ �: � : � � :� -

Fig. 4 pS3 oncoprotein expression in a moderately differentiated squa-

mous cell carcinoma.

Several recent studies have shown that E-cadherin expres-

sion is reduced in epithebial tumors from different sites when

compared with the epithelia of origin; these modifications have

been correlated with tumor degree of differentiation and aggres-

, -. ..z�_d#{248}, ,,

� �-... . . ‘, --...--.#{149}�-..- ..‘. t ‘

_\ . I:’ : .‘ � 4i1 #{149} .‘ �l.’ ‘� �. . �1’��” � ‘ �“ ‘, t#{149}

“V’p.�.1#f”‘ ._I ! ‘� � , � :? � ,�

. ‘ ‘�Jl”w “fY!f’ji’�#{149}�� a

,. I- � ,# ‘ ,.jt ,#{149}� a” F

� :

,.. a; � � � ,#{149},� � .�: � ;�1:. !7.Z.I: ‘

..� .� -�.-“ �� ,.* ., � �

� a

� � � . I

. . . w.� �. , � � � . � I

.. � ‘;.r.��__�i:Ic:7.! .:‘�

‘ , I-. #{149}‘‘“.. y‘. .. � . a ,�

;P�#{149}, ‘� , ., . :� ,TT �

Fig. 5 c-erbB-2 immunostaining in a well-differentiated squamous cell

carcinoma.



0

0

‘CzUQ. a-a

x

100

Vm

w

0

0

0

0

100

mIn

pN- pN+

S 100

0

B0

50C.w

0

0

0

0

0

0

a0pN- pN+

0

60

006

pN- pN+


§0

0

Fig. 6 Distribution of percent-age of cells positive for PCNA,MIB-l, E-cadherin, p53, and c-erbB-2 in a group of 30 pN-and 30 pN+ squamous cell car-

cinomas of the larynx with cut-

off line at 20%.

Table 3 Univariate and multivariate logistic regression analysis of 60 laryngeal cane er patients according to presence (pN + ) or absence (pN -)

of occult neck metastases

Prognostic parameter Categories

Univariate analysis Multivariate stepwise analysis

OR (95% CI) LRS (dfl” P value OR!� (95% CI) LRS (df)� P value

PCNA �20%C>20%

1.0016.00 (4.52-56.70) 23.13 (1) <0.001

1.006.16 (1.61-23.58) 1 1.56 (1) <0.001

MIB-1 �20%C

>20%1.00

8.64 (2.57-29.01) 14.30 (1) <0.0011.00

4.76 (1.01-22.58) 3.89 (1) 0.048E-cadherin

c-erbB-2

s20%’�>20%

Negativec�50%>50%

1.00

0.15 (0.04-0.55) 9.77 (1)

1.00 3.14(3)

0.52 (0.12-2.37)

3.14 (0.57-17.35)

0.0080.371

1.00

0.13 (0.04-0.40) 15.61 (1) <0.001

p53 Negativec1-10%

1 1-50%<50%

1.00 1.36 (4)0.68 (0.12-3.82)0.66 (0.19-2.31)1.45 (0.36-5.94)

0.850

Histological grade ICII

ifi

1.00 8.53 (3)

2.08 (0.60-7.24)17.60 (1.71-181.3)

0.036

Pattern of invasion PushingcInfiltrating

1.001.97 (0.70-5.54) 1.69 (1) 0.429

Number of mitoses �l0’

>101.00

0.58 (0.21-1.62) 1.07 (1) 0.584

Peritumoral inflammation AbsentcPresent

1.00

0.87 (0.31-2.46) 0.07 (1) 0.965Pentumoral sclerosis Absentc

Present1.000.44 (0.16-1.25) 2.42 (1) 0.299

a Computed by likelihood ratio statistic; df, degrees of freedom.b Adjusted OR of metastasis.C Reference category.




cytoplasmic localization of the antigen in a number of cases,

thus confirming the findings obtained in other carcinomas and

adenocarcinomas (37-39). These data taken together suggest

that deregulated E-cadherin expression is related to the acqui-

sition of a dedifferentiated and more aggressive phenotype by

the neoplastic cells, resulting in a higher tendency to spread to

locoregional lymph nodes (36). In contrast to our data, Mattijs-

sen et a!. found no relationship between the levels of E-cadherin

expression and the presence of lymph node metastases in a

group of 50 head and neck squamous cell carcinomas (15).

However, these authors included in their series squamous cell

carcinomas from different sites, thus limiting a direct compari-

son of the results.

In vitro studies have shown that cell-cell adhesion is related

strictly to proliferation and differentiation processes. Decreased

proliferative activity has been noted when stratification is in-

duced in cultures of normal human urothelial cells, a process

that is associated with desmosome formation and increased

expression of E-cadherin (40). Similarly, cyclic AMP and re-

baxin induce differentiation, growth inhibition, and increased

expression of E-cadherin in cultured human breast cancer cells

(41, 42). We observed that loss of cadherin expression was

accompanied by an increased PCNA and MIB-1 expression in

our series of laryngeal squamous cell carcinomas. It is therefore

possible that decreased cell-cell adhesion activity in tumor tis-

sue may be rebated not only to the invasive and metastatic

properties, but also to the growth rate of tumor cells.

Our study confirms the high frequency of p53 oncopro-

tein expression in laryngeal carcinomas, indicating that p53

gene alterations may be a common genetic event in the

development ofthese neoplasms (17, 18, 43, 44). Differently

from Field et a!. (16), we found a limited expression of

c-erbB-2 in our series. Moreover, we prevalently obtained

membrane immunostaining, whereas the above-mentioned

authors observed only cytoplasmic staining. As observed

previously, the study of the expression of p53 and c-erbB-2

oncoproteins failed to reveal any significant relationship with

the presence of lymph node metastases in laryngeal squa-

mous cell carcinoma (17, 34, 45).

The evaluation of several traditional histopathological pa-

rameters showed that only the degree of differentiation was

correlated significantly with the presence of occult lymph node

metastases in our series. However, in multivariate analysis, this

parameter failed to add independent information to identify

patients at risk for occult disease. The absence of prognostic

significance of parameters such as tumor growth pattern, peri-

tumoral sclerosis, and inflammatory reaction, which has been

correlated previously with lymph node status in laryngeal car-

cinoma by other authors (1, 7, 10), points out the difficulty in

obtaining an objective and reproducible evaluation of these

histopathobogical features.

In conclusion, our study indicates that the use of an immu-

nohistochemical determination of tumor cell proliferation and

E-cadherin expression in specimens of laryngeal squamous cell

carcinoma may help in the identification of those patients with

clinically negative lymph nodes who are at considerable risk for

occult metastases. Although the sensitivity and specificity of

these markers may still be insufficient to allow a decision of

elective neck dissection, they may supplement the results ob-

tamed by sophisticated imaging techniques (46) and may be

incorporated in a multidisciplinary approach to the definition of

the appropriate therapeutic strategy in the treatment of the neck

in these patients.

REFERENCES

I. Bocca, E., Calearo, C., De Vincentiis, I., Marullo. T., Mona, 0., andOttaviani, A. Occult metastases in cancer of the larynx and their rela-

tionship to clinical and histological aspects of the primary tumor: a

four-year multicentric research. Laryngoscope, 94: 1086-1090, 1984.

2. Van den Brekel, M. W. M, Casteljins, J. A., Croll, 0. A., Stel, H. V.,Valk, J., Van der Waal, I., Golding, R. P., Meyer, C. J. L. M., and Snow,0. B. Magnetic resonance imaging vs palpation of cervical lymph node

metastases. Arch. Otolaryngol. Head & Neck Surg., 117: 666-673,1991.

3. Gallo, 0., Boddi, V., Bottai, 0. V., Parrella, F., and Fini Storchi, 0.

The treatment of the clinically negative neck in laryngeal cancer pa-tients. Head Neck, in press, 1996.

4. Shah, J. T., Medina, J. E., Shaha, A. R., Schantz, S. T., and Marti,

J. R. Cervical lymph node metastasis. Curr. Probl. Surg., 30: 273-344,1993.

5. Friedman, M., Mafee, M. F., Pacella, B. L. Strorigl, T. L., Dew, L. L.,

and Toriumi, D. M. Rationale for elective neck dissection in 1990.

Laryngoscope, 100: 54-59, 1990.

6. Au, S., Tiwan, R. M., and Snow, 0. B. False positive and falsenegative neck nodes. Head Neck Surg., 8: 78-82, 1985.

7. McGavran, M. H., Bauer, W. C., and Ogura, J. K. The incidence ofcervical lymph node metastases from epidermoid carcinoma of the

larynx and their relationship to certain characteristics of the primarytumor. Cancer (Phiba.), 14: 55-66, 1960.

8. Geelen, 0. P., Hordijck, 0. J., Ravasz, L. A., and Terhaard, C. H.Degree of tumor differentiation as a prognostic factor in advancedlaryngeal cancer. Acta Otorhinolaryngol. BeIg., 49: 1-4, 1995.

9. Kowalski, L. P., Franco, E. L., and De Andrade Sobrinho, J. Factorsinfluencing regional lymph node metastasis from laryngeal carcinoma.

Ann. Otol. Rhinol. & Laryngol., 104: 442-447, 1995.

10. Ghouri, A., Zamora, R. L., Session, D. 0., Spitznagel, E. L., and

Harvey, J. E. Prediction of occult neck disease in laryngeal cancer by

means of a logistic regression statistical model. Laryngoscope, 104.

1280-1284, 1994.

1 1 . Chauvel, P., Courdi, A., Gioanni, J., Vallicioni, J., Santini, J., andDemard, F. The labelling index: a prognostic factor in head and neckcarcinoma. Radiother. Oncol., 14: 231-237, 1989.

12. Benazzo, M., Mevio, E., Occhini, A., Franchini, 0., and Danova,

M. Proliferative characteristics of head and neck tumors: in t’ivo eval-uation by bromodeoxyuridine incorporation and flow cytometry. ORL J.

Otorhinolaryngol. Relat. Spec., 57: 39-43, 1995.

13. Wolf, 0. T., Fisher, S. 0., Truelson, J. M., and Beals, T. F. DNA

content and regional metastases in patients with advanced laryngeal

squamous cell carcinoma. Laryngoscope, 104: 479-483, 1994.

14. Schipper, J. H., Frixen, U. H., Behrens, J., Unger, A., Jahnke, K.,

and Birchmeier, W. E-cadherin expression in squamous cell carcinomas

of head and neck: inverse correlation with tumor differentiation and

lymph node metastasis. Cancer Res., 51: 6328-6337, 1991.

15. Mattijssen. V., Peters, H. M., Schalkwijk, L., Manni, J. J.,

Van’t Hof-Grootenboer, B., De Mulder P. H. M., and Ruiter, D. J.E-cadherin expression in head and neck squamous-cell carcinoma is

associated with clinical outcome. Int. J. Cancer, 55: 580-585, 1993.

16. Field, J. K., Spandidos, D. A., Yiagnisis, M., Gosney, J. R.,

Papadimitriou, K., and Stell, P. M. C-erbB-2 expression in squamous

cell carcinoma of the head and neck. Anticancer Res., 12: 613-620,

1992.

17. Field, J. K., Pavelic, Z. P., Spandidos, D. A.. Stambrook, P. J.,Jones, A. S., and Gluckman, J. L. The role of the p53 tumor suppressor

gene in squamous cell carcinoma of the head and neck. Arch. Otolar-

yngol. Head Neck Surg., 119: 1118-1122, 1993.




18. Maestro, R., Dolcetti, D., Gasparotto, D., Doglioni, C., Pelucchi, S.,

Barzan, L., Grandi, E., and Boiocchi, M. High frequency of p53 genealterations associated with protein over-expression in human squamouscell carcinoma of the larynx. Oncogene, 7: 1 159-1 166, 1992.

19. Ebina, M., Steinberg, S. M., Muishine, J. L., and Linnoila, R. I.Relationship of p53 overexpression and up-regulation of proliferatingcell nuclear antigen with the clinical course of non-small cell lungcancer. Cancer Res., 54: 2496-2503, 1994.

20. Kay, E. W., Barry Walsh, C. J., Cassidy, M., Curran, B., andLeader, M. C-erbB-2 immunostaining: problems with interpretation.

J. Clin. Pathol., 47: 816-822, 1994.

21. Cuevas, E. C., Bateman, A. C., Wilkins, B. S., Johnson, P. A.,

Williams, J. H., Lee, A. H. S., Jones, D. B., and Wright, B. H.Microwave antigen retrieval in immunocytochemistry: a study of 80antibodies. J. Clin. Pathob., 47: 448-452, 1994.

22. Moll, R., Mitze, M., Frixen, U. H., and Birchmeier, W. Differentialboss of E-cadherin expression in infiltrating ductal and lobular breastcarcinomas. Am. J. Pathol., 143: 1731-1742, 1993.

23. Pignatelli, M., Ansari, T. Q., Gunter, P., Liu, D., Hirano, S.,Takeichi, M., Kloppel, 0., and Lemoine, N. Loss of membranousE-cadherin expression in pancreatic cancer: correlation with lymph nodemetastasis, high grade, and advanced stage. J. Pathol., 174: 243-248,

1994.

24. Golick, M. L., and Rice, M. Optimum staining of PCNA in paraffin

sections is dependent on fixation, drying, and intensification. J. Histo-

technol., 15: 39-41, 1992.

25. Rose, D. S. C., Maddox, P. H., and Brown, D. C. Which prolifer-

ation markers for routine immunohistology? A comparison of fiveantibodies. J. Clin. Pathol., 47: 1010-1014, 1994.

26. Baisch, H., and Gerdes, J. Simultaneous staining of exponentiallygrowing versus plateau phase cells with the proliferation-associated

antibody 1067 and propidium iodide: analysis by flow cytometry. Cell

Tissue Kinet., 20: 387-391, 1987.

27. Kamel, 0. W., LeBrun, D. P., Davis, R. E., Berry, 0. J., andWarnke, R. A. Growth fraction estimation of malignant bymphomas informalin-fixed paraffin-embedded tissue using anti-PCNA/cyclin l9A2.Correlation with Ki67 labelling. Am. J. Pathol., 138: 1471-1477, 1991.

28. Oya, M., Yao, 1., Nagai, E.. and Tsuneyoshi, M. Metastasizing

intramucosal gastric carcinomas. Well differentiated type and prolifer-ative activity using proliferative cell nuclear antigen and Ki-67. Cancer(Phila.), 75: 926-935, 1995.

29. Czader, M., Porwit, A., Tani, E., Ost, A., Mazur, J., and Auer, 0.DNA image cytometry and the expression of proliferative markers(proliferating cell nuclear antigen and Ki-67) in non-Hodgkin’s lym-

phomas. Mod. Pathol., 8: 5 1-58, 1995.

30. Albers, P., Miller, 0. A., Orazi, A., Ulbright, T. M., Abbers, J.,Donohue, J. P., and Foster, R. S. Immunohistochemical assessment oftumor proliferation and volume of embryonal carcinoma identify pa-tients with clinical stage A nonseminomatous testicular germ cell tumorat low risk for occult metastasis. Cancer (Phila.), 75: 844-850, 1995.

31. Tomasino, R. M., Daniele, E., Bazan, V., Morello, V., Tralongo, V.,

Nuara, R., Nagar, C., Salvato, M., Ingria, F., Restico, S., Dardanoni, 0.,Vecchione, A., and Russo, A. Prognostic significance of cell kinetics inlaryngeal squamous cell carcinoma: clinicopathological associations.

Cancer Res., 55: 6103-6108, 1995.

32. Munck-Wikland, E., Edstrdm, S., Jungmark, E., Kuylesteerna, R.,

Lindholm, J., and Auer, 0. Nuclear DNA content, proliferating-cell

nuclear antigen (PCNA), and p53 immunostaining in predicting pro-gression of laryngeal cancer in situ lesions. mt. J. Cancer, 56: 95-99,1994.

33. Shin, D. M., Voravud, N., Ro, J. Y., Lee, J. S., Hong, W. K., and

Hittelman, W. N. Sequential increases in proliferating cell nuclearantigen expression in head and neck tumorigenesis: a potential bio-marker. J. Natl. Cancer Inst., 85: 971-978, 1993.

34. Resnick, J. M.. Uhlman, D., Niehans, 0. A., Gapany. M., Adams.

0., Knapp, D., and Jaszcz, W. Cervical lymph node status and survivalin laryngeal carcinoma: prognostic factors. Ann. Otol. Rhinol. &

Laryngol., 104: 685-694, 1995.

35. Shimoyama, Y., Hirohashi, S., Hirano, S., Noguchi, M., Shimosato,

Y., Takeichi, M., and Abe, 0. Cadherin cell-adhesion molecules inhuman epithelial tissues and carcinomas. Cancer Res., 49: 2128-2133,

1989.

36. Pignatelli, M. E-cadherin: a biological marker of tumour differen-tiation. J. Pathol., 171: 81-82, 1993.

37. Pignatelli, M., Ansari, T. Q., Gunter, P., Liu. D., Shinji, H.,Takeichi, M., Kloppeb, 0., and Lemoine, N. R. Loss of membranousE-cadherin expression in pancreatic cancer: correlation with lymph node

metastasis, high grade, and advanced stage. J. Pathol., 174: 243-248,

1994.

38. Rasbridge, S. A., Gillett, C. E., Sampson, S. A., Walsh, F. S., and

Millis, R. R. Epithelial (E-) and placental (P-) cadherin cell adhesion

molecule expression in breast carcinoma. J. Pathol., 169: 245-250,

1993.

39. Gagliardi, 0., Kandemir, 0., Liu, D., Guida, M., Benvestito, S.,Ruers, T. 0. M., Benjamin, I. S., Northover, J. M. A., Stamp, 0. W. H.,Talbot, I. C., and Pignatelli, M. Changes in E-cadherin immunoreactiv-ity in the adenoma-carcinoma sequence of the large bowel. Virchows

Archiv., 426: 149-154, 1995.

40. Southgate, J., Hutton, K. A. R., Thomas, D. F. M., and

Trejdosiewicz, L. K. Normal human urothelial cells in vitro: prolifera-tion and induction of stratification. Lab. Invest., 71: 583-594, 1994.

41. Revillion, F., Vandewalle, B., Hornez, L., and Lefebvre, J. Influ-

ence of cAMP on E-cadhenn expression and cell surface heparan sulfateproteoglycan synthesis in human breast cancer cells. Anticancer Res.,13: 1625-1630, 1993.

42. Sacchi, T., Bani, D., Brandi, M. L., Falchetti, A., and Bigazzi, M.

Relaxin influences growth, differentiation, and cell-cell adhesion ofhuman breast-cancer cells in culture. mt.j. Cancer, 57: 129-134, 1994.

43. Dolcetti, R., Doglioni, C., Maestro, R., Gasparotto, D., Barzan, L.,

Pastore, A., Romanelli, M., and Boiocchi, M. p53 over-expression is anearly event in the development of human squamous-cell carcinoma ofthe larynx: genetic and prognostic implications. mt. J. Cancer, 52:

178-182, 1992.

44. Anwar, K., Nakakuki, K., Imai, H., Naiki, H., and Inuzuka, M.

Overexpression of p53 protein in human laryngeal carcinoma. Int. J.

Cancer, 53: 952-956, 1993.

45. Nadal, A., Campo, E., Pinto, J., Mallofr#{232}, C., PalacIn, A., Arias, C.,

Traserra, J., and Cardesa, A. p53 expression in normal, dysplastic, and

neoplastic laryngeal epithelium: absence of a correlation with prognostic

factors. J. Pathol., 175: 181-188, 1995.

46. Baatenburg de Jong, R. J., Knegt, P., and Verwoerd, C. D. A.

Reduction of the number of neck treatments in patients with head and

neck cancer. Cancer (Phila.), 71: 2312-2318, 1993.



1996;2:1801-1808. Clin Cancer Res A Franchi, O Gallo, V Boddi, et al. immunohistochemical determination.role of proliferating cell nuclear antigen, MIB-1, and E-cadherin Prediction of occult neck metastases in laryngeal carcinoma:

Updated version

http://clincancerres.aacrjournals.org/content/2/10/1801

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

[email protected] 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://clincancerres.aacrjournals.org/content/2/10/1801To request permission to re-use all or part of this article, use this link



http://clincancerres.aacrjournals.org/cgi/alerts

mailto:[email protected]



prediction of occult neck metastases in laryngeal ... · pendent predictors of occult nodal disease...

Documents