2401

Differential Values of Ki-67 Index

and Mitotic Index of Proliferating
Cell Population
An Assessment of Cell Cycle and Prognosis in Radiation
Therapy for Cervical Cancer
Takashi Nakano, M.D., and Kuniyuki Oka, M.D.

Background. Little is known about correlations between the growth fraction determined immunohistochemically with Ki-67 antibody and radiation response
or prognosis after radiation therapy.
Methods. The prognostic value of the growth fraction determined by Ki-67 index and the mitotic index of
proliferating cell population (pMI) were assessed in 45
cervical cancers treated with radiation therapy. The specimens from the cervix before radiation therapy were immunohistochemically stained with anti-Ki-67 antibody.
Results. The mean Ki-67 index and pMI for all patients were 36.0% and 2.74%, respectively. The patients
with a Ki-67 index of 33% or greater showed significantly
better histologic response to radiation at 30 Gy than those
with less than 33%. The mean Ki-67 index for patients
with good prognosis was significantly higher than for patients with tumor recurrence or metastasis later. Further, the mean values of pMI for patients with good prognosis were significantly lower than for patients with recurrence or metastasis. The 3-year survival rate for
higher Ki-67 index (233%) was significantly better than
lower Ki-67 index (less than 33%) (90.9% versus 34.8%; P
< 0.001). However, the 3-year survival rate for higher
pM1 (23.5%) was significantly poorer than lower pMI
(less than 3.5%) (8.3% versus 81.8%; P < 0.001).
Conclusions. These results suggested that tumors
with a high growth fraction showed a good prognosis
with radiation therapy. In addition, the inverse prognostic correlation between the Ki-67 index and pMI suggested that both indices have independent values on radiFrom the Hospital Division, National Institute of Radiological
Sciences, Chiba, Japan.
The authors thank Dr. K.R. Trott, Radiation Biology, London
University, for manuscript preparation and suggestions and K. Ando,
Ph.D., for comments.
Address for reprints: Takashi Nakano, M.D., Hospital Division,
National Institute of Radiological Sciences, 4-9-1 Anagawa, Inageku, Chiba-shi 263, Japan.
Accepted for publication May 17, 1993.

ation response and prognosis after radiation therapy.

Cancer 1993; 72: 2401-8.
Key words: Ki-67 index, growth fraction, mitotic index,
prognosis, cervical cancer, radiation therapy.

Extensive efforts have been made to determine the

prognosis of patients with various cancers after radiation therapy in terms of tumor proliferative paramet e r ~ . ' - Recently,
~
the growth fraction has been determined immunohistochemically with the use of the Ki67
which recognizes a nuclear antigen
expressed in all phases of the cell cycle except GO. Ki-67
index, which is a percentage of Ki-67-positive tumor
cells, indicates a growth fraction and indicates a different nature of cell proliferation from the mitotic or labeling indices. We have already reported the difference
between the growth fraction and mitotic index during
radiation therapy for cervical cancer.' The percentage
of Ki-67 index was highly variable and inversely correlated with prognosis in some turn or^.^^'^-'^ This was attributed to the association of the Ki-67 index with
various characteristics of malignancy including nodal
status, tumor grading, and labeling index. However,
little is known about correlations between the growth
fraction and radiation response or prognosis after radiation therapy. No significant correlation between the Ki67 index and prognosis has been reported in cervical
cancer15treated with preoperative radiation therapy or
in prostatic cancer treated with hormonal therapy.16
Moreover, in radiation therapy, cancers with higher proliferative activity tend to possess higher radiation sensitivity.'' In contrast, tumor cells in the GO phase were
reported to show radiation resistance." Hence, contrary
to previous observations, the possibility remains that
the tumor with highly proliferative cells indicated by a

2402

CANCER October 15, 1993, Volume 72, No. 8

Table 1. Ki-67 Index, Mitotic Index, and Mitotic Index of Proliferating Cell Population by Stage and Histologic Subtype

Stage
1
2
3
4
Histologic subtype
Kera tinizing
Large cell nonkeratinizing
Small cell
Total

PMI

Ki-67 index

Mitotic index

patients

Percent SD

Percent SD

Percent SD

2
8
32
3

69.6 f 22.9
30.0 k 5.9
36.4 f 15.9
2 5 . 8 f 2.7

1.37 f 0.72
0.67 f 0.33
0 85 f 0.70
0.73 f 0.32

1.9 & 0.41

2.34 +. 1.30
2.88 t 2.89
2.87 +. 1.30

7
31
7
45

27.8
38.3
33.9
36.0

1.54 +_ 0.63
0.67 f 0.54
0.86 & 0.61
0.83 f 0.63

No.of

f 16.5
f 14.2
& 23.2
f 16.1

6.59 +. 2.69
1.90 2 1.76It
2.59 t 1.60
2.74 +- 2.51

,t

SD: standard deviation; pMI: mitotic index of proliferating cell population.

* P < 0.001.
t P i0.005.
$ P i0.01

high Ki-67 growth fraction may be sensitive to radiation

and may show good prognosis.
The mitotic index of tumor cells in vitro usually is
proportional to the cell proliferating speed, but the index of tumor cells in vivo is biased by the presence of
quiescent cells whose population is larger than the cycling cell population in many cancer^.^,*',^^ Hence, it is
important to assess proliferation speed and growth
fraction of tumors separately. The mitotic index specifically of the proliferating cell population can express the
relative cell cycle speed and can be estimated by the
counting mitotic index and Ki-67 index as described
below.
The purpose of the current study, therefore, was to
determine the correlation between the Ki-67 or mitotic
index and the histologic response to radiation therapy
or prognosis, including failure patterns in radiation therapy for cervical cancer. In addition, the mitotic index of
the proliferating cell population (pMI) is introduced
and its value assessed for predicting the radiation response and determining prognosis.
Materials and Methods
Forty-five patients with invasive squamous cell carcinoma of the uterine cervix who received radiation therapy alone at National Institute of Radiological Sciences
Hospital (Chiba, Japan) from 1988-1989 participated in
this study. All patients were followed for more than 3
years. Clinical stages and histologic subtypes are summarized in Table 1. The clinical staging and histologic
classification were based on the criteria of the International Federation of Gynecology and obstetric^'^ and
World Health Organization classifications.

Radiation Therapy Protocol

Patients were treated with a combination of external
and high-dose rate intracavitary irradiation. Details of
the protocol were cited elsewhere.l External whole
pelvis irradiation was performed with anterio-posterior
and posterio-anterior parallel opposed ports, with a
dose of 1.8 Gy per fraction, five times per week, to a
total dose of 30.6 Gy. This was followed by a central
shielding pelvis field, with a dose of 2 Gy per fraction,
five times per week, to a total dose of 20 Gy. Along with
the central shielding irradiation, these patients also received intracavitary irradiation by remote afterloading
system (RALS) using 6oCosources. They received four
insertions (one per week) with fraction dose of 5.5-6.0
Gy at point A, with the total doses were ranging from
22-24 Gy at point A.

Histopathologic Study
All specimens were excised from cervical tumors before
and at 30 Gy during radiation therapy. All the fresh
biopsy specimens were divided into two: one was fixed
with 10% formaldehyde solution for conventional hematoxylin and eosin staining and the other specimen
was quickly frozen for Ki-67 immunostaining. The specimens were cut with a cryostat in 6 pm thickness, airdried, and fixed with cold 4% paraformaldehyde solution for 30 minutes. Then the sections were reacted
with anti-Ki-67 monoclonal a n t i b ~ d y (DAKO-PC;
~,~
Dako, Copenhagen, Denmark) for 1 hour at room temperature. The sections were followed by reaction with
biotinylated anti-mouse immunoglobulin G for 30 minutes and avidin-biotin complex2*(Vector Laboratories,
Burlingame, CA) for 30 minutes. The sections were

Ki-67 Index and Modified Mitotic IndexlNakano and Oka

reacted with 3,3'-diaminobenzidine tetrahydrochloride

(DAB, Dojin Chemicals, Tokyo) solution with 0.01%
(w/v) hydrogen peroxide23 for 2-5 minutes at room
temperature and counterstained with hematoxylin.
Control staining were done by incubating with phosphate-buffered saline instead of anti-Ki-67 antiserum.

2403

Table 2. Histologic Grades for Radiation Effect

Grade

I1

Assessment of Ki-67 Index and Mitotic Index

A

More than 1000 tumor cells per specimen were counted

on three X200 color photographs for calculation of the
Ki-67 index. The Ki-67 index was estimated by the percentage of Ki-67-positive cancer cells among all the
counted tumor cells.
Counting mitotic figures is the traditional and most
easily performed method for assessing cell proliferation. However, this method suffers from lack of standardization and repr~ducibility.~~
To overcome these
drawbacks in the current study, the mitotic index was
assessed by counting 2000 cancer cells for each specimen from randomly selected fields, using three or more
X200 color photographs of the hematoxylin and eosin
staining.

The mitotic index of tumors in vivo inevitably suffers

from bias of the growth fraction of tumors. Hence, the
index does not directly represent the proliferating activity of the cell population. The pMI can be calculated by
the mitotic index divided by the growth fraction as the
Ki-67 index as shown in the equation below:

111

IV

Characteristics are noted in tumor cells, but tumor

structures are not destroyed. There is no defect
in tumor nests as a result of the disappearance of
tumor cells.
In addition to characteristic cellular changes, tumor
structures have been destroyed as a result of the
disappearance of tumor cells. However, a variable
number of viable cells remains.
Destruction of tumor structures is mild: viable
tumor cells are observed frequently.
Destruction of tumor structures is severe: few viable
tumor cells are seen.
Significantly altered, presumably nonviable tumor
cells are present singly or in small clusters, and
few viable cells are seen.
No tumor cells remain in any section (local cure).

change of tumor cells, were classified as poor response.

Grade IIB and greater were defined as the finding of
marked degeneration, but still viable-looking cells in
one-third or less of the tumor nests and were classified
as good response.

Statistical Analysis

Calculation of pMI

Histoloeic findines

M/(P
P/(P

+ Q) - Mitotic index
+ Q) Ki-67 index
-

where M, P, and Q indicate the mitotic cell population,

the proliferating cell population, and the quiescent cell
population, respectively.

Assessment of Histologic Response to Radiation

Therapy
Early radiation response was assessed with punch
biopsy specimens at 30 Gy. Histologic evaluation of the
radiation effect on cancer cells was specified in the tumor nests and was determined using the histologic
grading system proposed by Shimosato et al. (Table
2).25In the current study, radiation responses of Grade I
and IIA, which represent no or a little morphologic

All results of the various indices were statistically analyzed with either the chi-square test or F-test. Correlations among the Ki-67 index, pMI, and mitotic index
were analyzed with the F-test. The cumulative diseasefree survival rates were statistically analyzed with Peto
log-rank test.26
Results

Ki-67-positive staining substance was located only in

the nuclei of cancer cells, especially in chromatins and
nucleoli, which showed diffuse nuclear staining and
dot staining (Fig. 1). Ki-67-positive cancer cells were
observed specifically in the peripheral or middle areas
of most of the keratinizing type (Fig. 1, top) and dispersedly in cancer nests of most of the nonkeratinizing cell
types (Fig. 1, bottom).
Forty-five patients who were diagnosed as squamom cell carcinomas participated in the current study.
The Ki-67 index, mitotic index, and pMI are shown by
stages and histologic subtypes in Table 1. The mean
Ki-67 index for all patients was 36.0% (range, 12.285.8%). The mean mitotic index was 0.83% (range,
0.01-2.45%). The mean pMI was 2.74% (range, 0.0310.16%). There was no significant correlation between

2404

CANCER October 25, 2993, Volume 72, No. 8

stages and the above three parameters. However, the

keratinizing type showed significantly higher mitotic
index and pMI than the nonkeratinizing cell types.
Histologic radiation responses at 30 Gy of radiation
therapy are shown in Table 3. Twenty-four patients
were examined for their histologic radiation response.
The patients with a Ki-67 index of 33% or greater
showed significantly better histologic response than
those with less than 33%. However, the mitotic index
and pMI showed no correlation to the response and, in
fact, demonstrated an inverse tendency compared with
the Ki-67 index.
Correlations of the Ki-67 index, mitotic index, or
pMI and patterns of failure are shown in Figure 2, left,
center, and right, respectively. The mean Ki-67 index of
the patients with good prognosis was significantly

Figure 1. (Top) Ki-67 immunostaining for keratinizing type of

squarnous cell carcinoma of the cervix (original magnification XZOO).
The Ki-67-positive cells are observed in relatively outer parts of the
cancer nests. (Bottom) Ki-67 immunostaining for large cell
nonkeratinizing type of squamous cell carcinoma of the cervix
(original magnification X200). The Ki-67-positive cells are observed
diffusely throughout the cancer nest.

Table 3. Histologic Radiation Response by Ki-67 Index,

Mitotic Index, and Mitotic Index of Proliferating Cell
Population
Histologic
response

Good

Total

Pvalue

7
6

7
4

14
10

NS

9
4

18
6

NS

Poor
Ki-67 index (Yo)
< 33
t 33
Mitotic index (YO)
<1
t l
pMI (%)
< 3.5
2 3.5

pMI: mitotic index of proliferatinn cell population; NS: not significant

higher than those with recurrence or metastasis (P <

0.01). However, the mitotic index and pMI were inversely associated with local control and prognosis. A
similar correlation was also observed in 32 patients
with Stage I11 disease (data not shown).
The cumulative 3-year disease-free survivals for all
patients (including patients with Stage I11 disease) by
Ki-67 index, mitotic index, and pMI are shown in Figure
3. The cut-off values of the patients compared were
33% for the Ki-67 index, 1% for mitotic index, and
3.5% for pMI. The 3-year survivals for the Ki-67 index
of 33% or greater were significantly better than those
below 33% (90.9% versus 34.8% for all, 88.2% versus
26.7% for Stage 111; P < 0.001) (Fig. 3, top left). The
3-year survival rates for mitotic index of 1%or greater
were somewhat lower than those below 1% (50.0%
versus 74.1% for all, 42.9% versus 72.2% for Stage 111;
P < 0.06) (Fig. 3, right). The 3-year survival rates for
patients with pMI of 3.5% or greater were significantly
lower than those below 3.5% (8.3% versus 81.8% for
all, 10.0% versus 81.8% for Stage 111; P < 0.001) (Fig. 3,
bottom left).
Three-year cumulative local control rates were calculated using Ki-67 index, mitotic index, and pMI. The
local control rate for the Ki-67 index of 33% or greater
was significantly higher than that of less than 33%
(95.5% versus 60.1%; P < 0.01). The local control rate
for mitotic index of 1% or greater was significantly
lower than that of less than 1%(57.7% versus 92.4%; P
< 0.01)). The local control rate for pMI of 3.5% or
greater was significantly lower than that of less than
3.5% (21.191~
93.8%; p < 0.001).
Prognosis by the Ki-67 index and mitotic index or
PMI is shown in Figure 4. No correlation between the
Ki-67 and mitotic indices was noted, whereas a slight

Ki-67 Index and Modified Mitotic Index/Nakano and Oka

%
90

.
.
...
.
..

80

70

60

Q
'CI

-e m
?40

t P<.O1 1 P<.05 1
:

42.3

0::

30

2405

*:
%

20
10

.
.
..

&26

.
..
.

.O

..
t

Rec

Meta

n=27

n=9

n=9

Tumor Status

NED
n=27

5.74

1.07

NED

.. .. ...

i
I

.
.

.
.

1.36

Rec
n=9

Meta
n= 9

Tumor Status

..

.
.
...
.:.
....
NED

n=27

..

4.24

.
L

Rec
n= 9

Meta
n=9

Tumor Status

Figure 2. (Left) Correlation of Ki-67 index and patterns of failure after radiation therapy for cervical cancer. The mean Ki-67 index was
significantly higher in patients with good prognosis than in those with recurrence ( P < 0.01) or metastasis ( P < 0.05). (Center) Correlation of
mitotic index and patterns of failure after radiation therapy for cervical cancer. The mean mitotic index was significantly lower in patients
with good prognosis than in those with recurrence ( P < 0.01) or metastasis (P < 0.05). (Right) Correlation of pMI and patterns of failure after
radiation therapy for cervical cancer. The mean pMI was significantly lower in patients with good prognosis than in those with recurrence
( P < 0.001) or metastasis ( P < 0.001).

trend of a pMI decrease with an increase in the Ki-67

index was observed (r = 0.4). Most of the patients with
recurrence or metastasis had tumors with a low Ki-67
index and a high mitotic index or a high pMI. The discrimination of prognosis was more significant by functions of the pMI and Ki-67 index than by functions of
the mitotic and Ki-67 indices.
Discussion
The current study demonstrated that the Ki-67 index
was positively correlated with prognosis and early radiation response in radiation therapy for cancer of the
cervix. This contradicts the findings in lymphoma,'
breast cancer,12,13and soft tissue sarcoma,I4where the
Ki-67 index was inversely correlated with prognosis.
The results of other authors were explained by the hypothesis that the higher proliferative activity of cancer
cells shows a more malignant nature of the disease and
results in a high frequency of recurrence or metastasis.
However, no significant correlation has been reported
in the outcome of cervical cancer,I5 where the results

were not conclusive because of the inconsistent background of the patients and the small number of patients
studied. Prostatic cancer after hormonal therapy also
showed no correlation.16
However, a more complex prognostic value of the
Ki-67 index was observed in non-Hodgkin disease,"
where a higher Ki-67 index was associated with good
prognosis for high-grade malignancy but with poor
prognosis for low-grade malignancy. The current study
may suggest that the proliferating cells measured by
Ki-67 immunohistochemistry possessed high radiation
sensitivity and, consequently, the tumors were easier to
eradicate. Many investigations26-28
have suggested that
GO cells, which are the major component of tumors
with a low growth fraction, were resistant to radiation.
One of the reasons is thought to be the higher repair
activity of GO cells from potentially lethal damage
caused by irradiation. Another possibility is that tumors
with a low growth fraction are associated with hypoxic
tumor cells, and the radioresistant hypoxic cell compartment results in low local control characteristics.
The proliferative activity determined by the mitotic
index showed a negative correlation with prognosis in

CANCER October 25, 1993, Volume 72, No. 8

2406

100 7

............

t'l

- 1-..........
-__
-. --

KI 2 3 3 %
K I < 33 %

.... - - -

al

Stage III

.w

.2

50-

,
,

;--;lLx

rl

I
,

I
I

-...........
_ _,______
_ ,...

11

l
I

rll.

___

"% ]All

MI<l%
I
%
....... M I < l %

' ]

Stage

I
W.-J.

3
0

:.,
0

0-

1
2
3
4Year
After Completion Treatment

M
100 n:

nJ

Figure 3. (Top left) The cumulative disease-free survival rates by

Ki-67 index for all and Stage 111 patients. Patients with a Ki-67 index
of 33% or greater showed significantly better survival than those
with less than 33% in both all and Stage 111 patients (P < 0.01 for
both groups). (Right) The cumulative disease-free survival rates by
mitotic index for all and Stage 111 patients. The patients with a
mitotic index of 1%or greater showed significantly poorer survival
than those with less than 1% in all patients ( P < 0.05). A similar
trend was observed in Stage 111 patients (P < 0.06). (Bottom left) The
cumulative disease-free survival rates by pMI for all and Stage 111
patients. The patients with pMI of 3.5% or greater showed
significantly poorer survival than those with less than 33% in both
all and stage 3 disease ( P < 0.001 for both groups).

1
2
3
4 Year
After Completion Treatment

the current study. A similar correlation had been reported in other cancers."-'4 However, this appears to
contradict the positive prognostic value of the Ki-67
index. In addition, no correlation between the Ki-67
and mitotic indices was noted in the current study,

whereas a positive correlation between the Ki-67 index

and the mitotic or labeling indices had been observed in
other cancer^.^,^^-^^ Moreover, our previous study demonstrated different changes that the Ki-67 and mitotic
indices during radiation therapy for cervical cancer

31
0
.+

%
0 Cont

Rec
W Meta

..

.H*

0
+
f 1

lo

0 Cont

Rec
Meta

0
0

o
0 0 0
.............. ..........
...........
............--o---

................ c"'o-@o

.,0

n
-

n w ,#!$
10 20 30 40 50 60 70 80

Ki-67 Index

Q""

90 100

10

20

30

....0
.....

40 50 60 70 80 90 100
Ki-67 Index

Figure 4. (Left) Correlation among prognosis, Ki-67 index, and mitotic index. There was no significant correlation between Ki-67 index and
mitotic index (r = 0.1; Y = 0.69 0.0046X). (Right) Correlation among prognosis, Ki-67 index, and mitotic index in proliferating cell
population. There was a weak correlation between Ki-67 index and pMI (r = 0.4; Y = 5.14 - 0.065X).

Ki-67 Index and Modified Mitotic Index/Nakano and Oka

were often disparate. These results suggest that the

Ki-67 and mitotic indices are not always interdependent.
Growth fraction and cell cycle time are independent major tumor proliferation characteristics. Generally, the mitotic index, labeling index after thymidine
labeling or bromodeoxyuridine incorporation, and the
potential doubling time are used for accumulating information on cell kinetics. These proliferation indicators assess gross proliferation activity mixed with the
growth fraction and cell cycle speed of the tumors, but
they can not derive information from these two proliferative activities separately because these indices do not
differentiate the involvement of the GO cell fraction.
The current study introduced the pMI to solve these
problems. pMI excludes the quiescent cell population
by use of Ki-67 immunostaining. High pMI suggested a
shorter cell cycle time of the proliferative cell population, as the mitotic phase duration tends to be rather
constant compared with other cell cycle phases.29In the
current study, high pMI indicated a significantly poor
prognosis and was of greater prognostic reliability than
the mitotic index. Hence, the faster cell production rate
of the critical tumor cell population suggested by the
high pMI may be one of the major causes of recurrence
or metastasis. With regard to local recurrence, this
might be due to fast repopulation of surviving cells. As
for metastasis, however, it might be related to a higher
risk of tumor cell invasion into the lymphatics by the
more aggressively proliferating tumor cells.30
The current study indicated that tumors of the keratinizing cell type possess a rather high pMI and an
average or somewhat lower growth fraction. Similarly,
in head and neck cancer,31well-diff erentiated squamous cell carcinoma (such as verrucous carcinoma)
showed a high labeling index, short potential doubling
time, and high cell loss factor. Therefore, contrary to the
common perception, differentiated tumor may have a
higher cellular proliferative activity than some undifferentiated tumors.
Sasaki et aL3 and Kame1 et al.33reported that the
labeling index in the proliferating cell population could
be obtained by dividing it by the Ki-67 index for various
cancers. They observed a wide variety in the labeling
index of the proliferating cell population. These methods, however, require the administration of a toxic
agent in patients or culture of excised tumor cells, and
the labeling index would be affected by the perfusion
activity under either condition. Additionally, no assessment has been made in respect to the clinical value of
these indices. Both the Ki-67 index and pMI can be obtained directly from specimens from patients without
carrying out any other procedures.

2407

Recent studies on tumor proliferation during radiation therapy suggested that r e p ~ p u l a t i o nor
~~
acceler,~~
ated p r ~ l i f e r a t i o nof
~ ~surviving
,~~
tumor cells reduced
the chances of local control by radiation therapy. It is,
however, not easy to measure these phenomena under
routine conditions. In cancer clinics, then, the Ki-67 index and pMI may serve as effective predictive parameters for potential proliferative activity of tumors and are
expected to contribute to the individualization of radiation therapy.
References
1. Sasaki T, Sakka M. Implications of thymidine labeling index in
the growth kinetics of human solid tumors. ]up ] Cancer Res
1981; 721181-8,
2. Begg AC, Hofland I, Moonen L, Bartelink H, Schraub S, Bontemps P, et al. The predictive value of cell kinetic measurements
in a European trial of accelerated fractionation in advanced head
and neck tumors: an interim report. Int ] Radiat Oncol Biol Phys
1990; 19:1449-53.
3. McNally NJ. Can cell kinetic parameters predict the response of
tumours to radiotherapy? Int ] Radiaf Oncol B i d Phys 1989;
56:777-86.
4. Dyson JED, Joslin CAF, Rothwell RI, Quirke P, Khoury GG, Bird
CC. Flow cytofluorometric evidence for the differential radioresponsiveness of aneuploid and diploid cervix tumours. Radiother Oncol 1987; 8:263-72.
5. Rutgers DH, van der Linden PM, van Peperzeel HA. DNA-flow
cytometry of squamous cell carcinomas from the human uterine
cervix: the identification of prognostically different subgroups.
Radiother Oncol 1986; 7:249-58.
6. Gerdes J, Schwab U, Lemke H, Stein H. Production of a mouse
monoclonal antibody reactive with a human nuclear antigen
associated with cell proliferation. Znt ] Cancer 1983; 31:13-20.
7. Gerdes J, Lemke H, Baisch H, Wacker H, Schwab U, Stein H.
Cell cycle analysis of a cell proliferation-associated human nuclear antigen defined by the monoclonal antibody Ki-67. ] Imrnunol 1984; 133:1710-5.
8. Grogan TM, Lippman SM, Spier CM. Independent prognostic
significance of a nuclear proliferation antigen in diffuse large
cell lymphomas as determined by the monoclonal antibody Ki67. Blood 1988; 71:1157-60.
9. Nakano T, Oka K. Transition of Ki-67 index of uterine cervical
tumors during radiation therapy. Immunohistochemical study.
Cancer 1991; 68:517-23.
10. Hall PA, Richards MA, Gregory WM, dArdenne AJ, Lister TA,
Standsfeld AG. The prognostic value of Ki-67 immunostaining
in non-Hodgkins lymphoma. ] Pathol 1988; 154:223-36.
11. Yamada Y, Murata K, Kamihira S, Atogami S, Tsukasaki K,
Sohda H. Prognostic significance of the proportion of Ki-67 positive cells in adult T-cell leukemia. Cancer 1990; 67:2605-9.
12. Wintzer H-0, Zipfel I, Schulte-Monting J, Hellerich U, von
Kleist S. Ki-67 immunostaining in human breast tumors and its
relationship to prognosis. Cancer 1991; 67:421-8.
13. Stefan0 DD, Mingazzini PL, Scucchi L, Donnetti M, Marinozzi
V. A comparative study of histopathology, hormone receptors,
peanut lectin binding, Ki-67 immunostaining, and nucleolar organizer region-associated proteins in human breast cancer.
Cancer 1991; 67:463-71.
14. Ueda T, Aozasa K, Tsujimoto M. Prognostic significance of Ki67 reactivity in soft tissue sarcomas. Cancer 1989; 63:1607-ll.

2408

CANCER October 25, 2993, Volume 72, No. 8

15. Cole DJ, Brown CD, Crossley E, Alcock CJ, Gatter KC. Carcinoma of the cervix uteri: an assessment of the relationship of
tumour proliferation to prognosis. Br ] Cancer 1992; 65:783-5.
16. Raymond WA, Leong AS-Y, Bolt JW, Milios J, Jose JS. Growth
fractions in human prostatic carcinoma determined by Ki-67 immunostaining. ] Pafhol 1988; 156:161-7.
17. Frindel E, Tubiana M. Radiobiology and cell cycle. In: Baserga R,
editor. The cell cycle and cancer. New York: Dekker, 1971:389447.
18. Wallen AC, Ridinger DN, Dethlefsen LA. Heterogeneity of Xray cytotoxicity in proliferating and quiescent murine mammary
carcinoma cells. Cancer Res 1985; 45:3064-8.
19. International Federation of Gynecology and Obstetrics. Annual
report on the results of treatment in carcinoma of the uterus,
vagina and ovary. vol. 16. Sweden: Radiumhemmet, 1979.
20. Poulsen HE, Taylor CW, Sobin LH. Histological typing of female genital tract tumors. no. 13. Geneva: World Health Organization, 1975.
21. Arai T, Nakano T, Morita S, Sakashita S, Nakamura YK, Fukuhisa K. High dose rate remote afterloading intracavitary radiation therapy for cancer of the uterine cervix: a 20-year experience. Cancer 1992; 69:175-80.
22. Hsu SM, Raine L, Fanger H. Use of avidin-biotin-peroxidase
complex (ABC) in immunoperoxidase techniques: a comparison
between ABC and unlabeled antibody (PAP) procedures. ] Histochem Cytochem 1981; 29:577-80.
23. Graham RC, Karnovsky MJ. The early stages of absorption of
injected horseradish peroxidase in the proximal tubules of
mouse kidney. Ultrastructural cytochemistry by a new technique. ] Histochem Cytochem 1966; 14:291-303.
24. Clinical application of morphologic and immunocytochemical
assessments of cell proliferation [editorial]. A m ] Clin Pathol
1992; 97:4-13.
25. Shimosato Y, Oboshi S, Baba K. Histological evaluation of effects of radiotherapy and chemotherapy for carcinomas. ]up ]
Clin Oncol 1971; 1:19-35.

26. Pet0 R, Pike MC, Armitage P. Design and analysis of randomized clinical trials requiring prolonged observation of each patient. B r ] Cancer 1977; 5:l-39.
27. Rodriguez A, Alpen EL, Mendonca M, Deguzman RJ. Recovery
from potentially lethal damage and recruitment time of noncycling clonogenic cells in 9L confluent monolayers and spheroids.
Radiat Res 1988; 114:5151-7.
28. Mendonca MS, Rodriguez A, Alpen EL. Quiescence in 9L cells
and correlation with radiosensitivity and PLD repair. Radiat Res
1989; 117:433-47.
29. Pardee AB. A restriction point for control of normal animal cell
proliferation. Proc Natl Acad Sci U S A 1974; 71:1286-90.
30. Tubiana M, Courdi A. Cell proliferation kinetics in human solid
tumors: relation to probability of metastatic dissemination and
long-term survival. Radiother Oncol 1988; 15:l-18.
31. Dische S, Saunders MI, Bennett MH, Wilson GD, McNally NJ.
Cell proliferation and differentiation in squamous cancer. Radiother Oncol 1989; 15:19-23.
32. Sasaki K, Matsumura K, Tsuji T, Shinozaki F, Takahashi M. Relationship between IabeIing indices of Ki-67 and BrdUrd in human malignant tumors. Cancer 1988; 62:989-93.
33. Kame1 OW, Franklin WA, Ringus JC, Meyer JS. Thymidine labeling index and Ki-67 growth fraction in lesions of the breast. A m ]
Pathol 1989; 134~107-13.
34. Tubiana MLH. Gray medal lecture: cell kinetics and radiation
oncology. Int ] Radiat Oncol Biol Phys 1982; 8:1471-89.
35. Maciejewski B, P-Bayer G, Trott KR. The influence of the number of fractions and of overall treatment time on local control
and late complication rate in squamous cell carcinoma of the
larynx. Znt ] Radiat Oncol Biol Phys 1983; 9:321-8.
36. Withers HR, Taylor JMG, Maciwjewski B. The hazard of accelerated tumor clonogen repopulation during radiotherapy. Act Oncologica 1988; 27:131-46.
37. Trott KR, Kummermehr J. Accelerated repopulation in tumours
and normal tissues. Radiother Oncol 1991; 22:159-60.