\

Endometrial CarcinomaSteroid Hormones and Receptors in Relation

to Proliferation

J.

Karin Boman

University of Umeå 1993

UMEÅ UNIVERSITY MEDICAL DISSERTATIONSNew Series No. 378 ISSN 0346-6612 ISBN 91-7174-794-0

Endometrial CarcinomaSteroid Hormones and Receptors in Relation

to Proliferation

AKADEMISK AVHANDLINGsom med vederbörligt tillstånd av Rektorsämbetet vid Umeå Universitet för avläggande av doktorsexamen i

medicinsk vetenskap vid Umeå Universitet kommer att offentligen försvaras i Onkologiska Klinikens föreläsningssal, JK-byggnaden, 2 tr, Norrlands

Universitetssjukhus, den 11 juni kl 09.00av

Karin Boman, leg. läk.

Fakultetsopponent: Docent Bo Frankendal

From the Departments of Oncology and Obstetrics and Gynaecology, University of Umeå, Sweden Editor: the Dean of the Faculty of Medicine

AbstractThe significance of the hormonal milieu for endometrial changes is as well-known as its link with endometrial carcinoma. Unopposed oestradiol treatment is shown to increase the incidence for this cancer. Obesity leads to elevated levels of oestrogens and is a risk factor for endometrial carcinoma. An association between high tumour proliferation and prognosis is a general feature of human cancer. Tumour growth can be expressed as proliferation rate and flow cytometry (FCM) is a sensitive and reproducable method to estimate S-phase fraction (SPF) and ploidy level. Both parameters have been shown to correlate with prognosis. Sex steroid hormone levels were analysed together with clinical parameters, SPF, and receptors in established endometrial carcinoma.

The study consisted of postmenopausal women with endometrial adeno­carcinoma. H ormones were analysed in 127 patients, 99 were analysed for FCM and 60 for oestrogen and progesterone receptors. RIA technique was used for hormone assay of oestrone, oestradiol, progesterone, androstenedione and testosterone plasma levels. The receptors were analysed with an immunohistochemical method, and SPF and ploidy level by flow cytometry.

A wide range of oestrogen concentrations was found. Some patients had levels comparable to fertile women. Strong correlations were found between body mass index, weight and depth of uterine cavity. No relations were found between receptors and SPF, apart from oestrogen- receptor positive tumours having a lower SPF when compared with receptor negative tumours. The influence of oestradiol on tumour proliferation expressed as SPF was ambiguous. SPF was increased with higher oestradiol levels in the group of peri-diploid, well-differentiated tumours, while a negative correlation was found for the peri-diploid, moderately differentiated tumours. The aneuploid and poorly differenti­ated tumours had a high SPF regardless of oestradiol concentration. The association between progesterone concentration and SPF was of a more general nature. Progesterone above a threshold level was related to a lower SPF in well-differentiated and moderately differentiated tumours. Thus endogenous progesterone seems to play a role in controlling the tumour’s proliferation activity, in contrast to oestradiol, that had a role which did not appear to relate to proliferation activity in any specific direction. The only stimulative association was seen in well-differentiated tumours, but SPF was still below the mean value for all diploid tumours.

UMEÅ UNIVERSITY MEDICAL DISSERTATIONSNew Series No. 378 ISSN 0346-6612 ISBN 91-7174-794-0

Endometrial CarcinomaSteroid Hormones and Receptors in Relation

to Proliferation

•Z 3 ^

Karin Roman

From the Departments of Oncology and Obstetrics and Gynaecology University of Umeå, Sweden

Editor: the Dean of the Faculty of Medicine

© 1993 Karin Boman ISBN 91-7174-794-0 Printed in Sweden

by Solfjädern Offset AB Umeå 1993

To

with best regards from the author

AbstractThe significance of the hormonal milieu for endometrial changes is as well-known as its link with endometrial carcinoma. Unopposed oestradiol treatment is shown to increase the incidence for this cancer. Obesity leads to elevated levels of oestrogens and is a risk factor for endometrial carcinoma. An association between high tumour proliferation and prognosis is a general feature of human cancer. Tumour growth can be expressed as proliferation rate and flow cytometry (FCM) is a sensitive and reproducable method to estimate S-phase fraction (SPF) and ploidy level. Both parameters have been shown to correlate with prognosis. Sex steroid hormone levels were analysed together with clinical parameters, SPF, and receptors in established endometrial carcinoma.

The study consisted of postmenopausal women with endometrial adenocarci­noma. Hormones were analysed in 127 patients, 99 were analysed for FCM and 60 for oestrogen and progesterone receptors. RIA technique was used for hormone assay of oestrone, oestradiol, progesterone, androstenedione and testosterone plasm a levels. The receptors w ere analysed w ith an immunohistochemical method, and SPF and ploidy level by flow cytometry.

A wide range of oestrogen concentrations was found. Some patients had levels comparable to fertile women. Strong correlations were found between body mass index, weight and depth of uterine cavity. No relations were found between receptors and SPF, apart from oestrogen-receptor positive tumours having a lower SPF when compared with receptor negative tumours. The influence of oestradiol on tumour proliferation expressed as SPF was ambiguous. SPF was increased with higher oestradiol levels in the group of peri-diploid, well-differentiated tumours, while a negative correlation was found for the peri-diploid, moderately differen­tiated tumours. The aneuploid and poorly differentiated tumours had a high SPF regardless of oestradiol concentration. The association between progesterone concentration and SPF was of a more general nature. Progesterone above a threshold level was related to a lower SPF in well-differentiated and moderately differentiated tumours. Thus endogenous progesterone seems to play a role in controlling the tumour’s proliferation activity, in contrast to oestradiol, that had a roll which did not appear to relate to proliferation activity in any specific direction. The only stimulative association was seen in well-differentiated tumours, but SPF was still below the mean value for all diploid tumours.

KEY WORDS: Endometrial carcinoma, oestrogens and progesterone hormones and receptors, S-phase fraction, ploidy level.

PapersThis thesis is based on the following papers which will be referred to by theirRoman numerals:

I. Boman K., Bäckström T., Gerdes U. and Stendahl U.: Oestrogens and clinical characteristics in endometrial carcinoma. Anticancer Research 10: 247-252 (1990)

II. Boman K., Strang P., Bäckström T., Gerdes U. and Stendahl U.: Influence of oestrogen on growth of endometrial carcinoma. Anticancer Research 10:173- 176 (1990)

III. Boman K., Bäckström T., Gerdes U. and Stendahl U.: Progesterone androstenedione testosterone and clinical characteristics in endometrial carcinoma. Anticancer Research 11: 2163-2166 (1991)

IV. Boman K., Strang P., Bäckström T. and Stendahl U.: The influence of progesterone and androgens on the growth of endometrial carcinoma. Cancer 1993; 71: (In press)

V. Boman K., Mäentausta O., Bäckström T., Strang P. and Stendahl U.: Sex steroid hormones and receptors in relation to S-phase fraction and ploidy level in endometrial carcinoma. (Manuscript)

Reprints were made with the permission of the publishers.

ContentsAbbreviations....................................................................................................... 8

In tro d u c tio n ............................................................................................................9THE NORMAL ENDOMETRIUM........................................................................ 9Postmenopausal Endocrinology.................................................... 9The fertile endometrium...................................................................................10The postmenopausal endometrium................................................................11ENDOMETRIAL CARCINOMA......................................................................... 11Epidemiology..................................................................................................... 11Endocrinology................................................................................................... 12Prognostic factors.............................................................................................. 13Ploidy and proliferation: background.............................................................13Chromosomes and ploidy level....................................................................... 15Flow cytometry and ploidy.............................................................................. 15Flow cytometry and S-Phase fraction (SPF)....................................................15Other Methods................................................................................................... 16Receptors............................................................................................................17

A im s o f th e s tu d y ...............................................................................................18M aterial a n d m e th o d s ...................................................................................... 19

Patients................................................................................................................19Clinical parameters............................................................................................19Hormone assays................................................................................................20Flow cytometry..................................................................................................21Steroid receptors...............................................................................................22Data analysis...................................................................................................... 22

R esu lts .....................................................................................................................23Hormone concentrations and clinical parameters(Papers I and in)...............................................................................................23Hormone concentrations and flow cytometric analysis (FCM)(Papers n and IV)..............................................................................................26Hormone concentrations, FCM and sex steroid receptors(Paper V ) ........................................................................................................... 29

G e n e ra l D isc u ss io n ...........................................................................................31C o n c lu s io n s ..........................................................................................................35A c k n o w le d g e m e n ts .......................................................................................... 36R e fe re n c e s .............................................................................................................37P a p e rs I-V ............................................................................................................. 47

AbbreviationsBMI body mass indexFCM flow cytometrySPF S-phase fractionFSH follicle stimulating hormoneLH luteineizing hormoneU labelling indexMI mitotic indexEl oestroneE2 oestradiolT testosteroneA androstenedioneP progesteroneER oestrogen receptorER+ oestrogen receptor positivePR progesterone receptorPR+ progesterone receptor positiveM mitosis

8

Introduction

THE NORMAL ENDOMETRIUM Postmenopausal EndocrinologyThe mean age for menopause varies in different parts of the world. The current estimate in Sweden is 51 years.1 With a life expectancy of 80 years,2 women can expect to live a substantial part of their life as postmenopausal.

The phases and the function of the normal fertile endometrium are directly connected with hormone production of the ovaries that takes place in the theca and granulosa cells of the maturing follicle and in corpus luteum. These, in turn, are regulated by two pituitarian hormones, follicle stimulating hormone (FSH) and luteineizing hormone (LH). Oestradiol dominates during the proliferative phase, and progesterone during the secretory phase.

Deficit of ovarian hormones, secondary to the nearly complete depletion of ovarian follicles, leads to menopause and associated systemic effects. Simply stated, the endocrine profile of the menopause is that of hypoestrogenism with relative androgen excess.

In postmenopausal women there is a major decrease in both oestradiol and oestrone concentrations. During the reproductive years more than 90 per cent of these hormones is produced by the ovaries,5 but after menopause the ability to synthesise oestradiol from pregnenolone in the ovaries is decreased as a result of the aromatase system disappearing with the granulosa cells. Oestradiol drops to less than 10 per cent of the premenopausal level and the concentration of circulating oestradiol is less than 50 pmol/l.4-6 The major source of the remaining oestradiol is peripheral conversion of oestrone to oestradiol.7

The decrease of oestrone to approximately thirty per cent of premenopausal levels is significantly smaller than that for oestradiol. Oestrone thus becomes the predominant oestrogen, with a mean concentration of approximately 130 pmol/ 14-6, 8 oestrone is primarily derived from peripheral conversion of androstenedione,9 ,10 and the mean conversion rate is more than doubled after menopause, from 1.2 to 2.8 per cent.n »12 The rate is also correlated to body mass index (BMI), as the most efficient conversion of androstenedione to oestrone takes place in adipose tissue.15 In postmenopausal obese women this rate rises to 3.9 per cent.10 In some reports chronological age has been considered to be an additional factor for an increased conversion rate,10,14 but this is contradicted in others.15

Androstenedione is reduced to about 50 per cent of premenopausal levels, and is mainly produced by the adrenal glands, with an ovarian contribution of approximately 20-30 per cent.12’ ^

With the postmenopausal decrease of oestradiol level, the negative feedback on the gonadotrophins disappears. High levels of LH stimulate the remaining ovarian stroma to produce androgens, but in contrast to androstenedione, testosterone is only marginally lowered after menopause. The ovarian secretion

9

of testosterone increases and contributes with approximately 50 per cent, while peripheral contribution decreases because androstenedione, as a source of peripheral conversion, is reduced.12, 16

Progesterone synthesis diminishes during the perimenopausal period, mani­fested by corpus luteum insufficiency. Postmenopausal concentrations are equal to levels seen in the follicular phase (0.6- 3.5 nmol/1).16,17 Progesterone is now mainly secreted from the adrenals, but there is still some production from the ovaries.18

The fertile endometriumStudies of the fertile endometrium have demonstrated that the hormone profile of the follicular phase is characterised by an average oestradiol level between 150 and 370 pmol/1, and with a pre-ovulatory peak value higher than 690 pmol/1. Progesterone values during the follicular phase are 1.0 - 4.4 nmol/1, with a minor pre-ovulatory rise. The LH surge induces the luteal phase with the next rise of progesterone. The concentration reaches 15-42 nmol/1 during this phase.19-21

Proliferation rate can partly be estimated by using morphometrical or biochemical methods to estimate, e.g., mitosis activity, glandular size, or amount of DNA and RNA.22 During the secretory phase a decrease in proliferation can be seen from the number of mitosis and from the amount of DNA + RNA. This starts already two days before the LH-peak, thus simultaneously with the pre-ovulatory progesterone rise, and before the substantial rise in the progesterone concentra­tion that occurs two days after the LH-peak.20,22 Ferenczy has reported a similar development; between days 8-10 and 11-14 in the menstrual cycle there is a decrease in DNA synthesis.25 The proliferative activity starts early in the menstrual cycle, and is shown to be at a maximum level before oestradiol reaches its preovulatory peak.24 The onset of proliferation differs in different cells and layers, and is first seen in the glandular epithelium in the functional layer.25

In the normal menstrual cycle, the concentrations of the oestrogen and progesterone receptors undergo characteristic variations. The expression is regulated by oestradiol and progesterone.25-27 Oestradiol stimulates the synthesis of both receptors, whereas progesterone depletes them.28

The highest content of endometrial oestrogen receptors occurs in the epithelial cells during the proliferative phase, and persists until the mid to late secretory phase, when it decreases.25, 27, 29 Progesterone receptor concentration remains low in the early proliferation phase and reaches a maximum level in the late proliferation phase,26 and then gradually disappears during the late secretory phase.27 Flow cytometric studies on the normal fertile endometrium have shown that the mean S-phase fraction (SPF) for proliferative endometrium is significantly higher than for secretory endometrium. In hyperplastic endometrium, the SPF mean value falls in between those values, while carcinomatous endometrium has the highest SPF value.50, 51

10

The postmenopausal endometriumAfter menopause the endometrium is senile atrophic due to the lack of oestrogen, but might also be weakly proliferative. It is very rarely secretory. Cystic, adenomatous or atypical hyperplasia and cancer in situ may develop, partly depending upon the endocrine milieu, and will be discussed.

ENDOMETRIAL CARCINOMA

EpidemiologyEndometrial cancer rates are highest in Northern Europe, North America; intermediate in Southern Europe and Latin America; and low in Africa and Asia.32

In Sweden, approximately 950 new cases are diagnosed every year, and endometrial cancer is second to ovarian cancer in frequency among gynaecologi­cal cancers. Eighty-five percent are diagnosed after 55 years of age when 98 per cent of the women are postmenopausal. The median age of incidence is 66 years.2

There are some well-known endogenous risk factors identified with endome­trial adenocarcinoma: Obesity, nulliparity, early menarche, late menopause are examples of variation of normal physiology,53"37 and polycystic ovaries and granulosa- theca cell tumours are examples of diseases.38"40

Both diabetes mellitus and hypertension are associated with endometrial cancer, but it is unclear whether the association is independent of the effects of obesity and age.33, 35, 36>41 However, remaining correlations have been shown when corrected for these two factors.35

Among exogenous risk factors, unopposed oestrogen has received most attention. In epidemiological studies, Persson has shown that oestrogen treatment involves an increased risk for endometrial carcinoma unless opposed by progesterone.42’ 43 Numerous retrospective studies have correlated the use of exogenous oestrogen to a 2- to 12-fold increased risk of developing endometrial carcinoma.41’44-47 Some of these studies have recived methodological criticism, but the consensus is that there is a true causal relationship between oestrogen treatment and an increased risk of endometrial carcinoma. The risk is dependent on duration, dose and habitus. Non obese, non diabetic and normotensive women are at higher risk, suggesting that the additional oestrogen effect is less crucial to obese, diabetic and hypertensive women.44 Whether this depends on a threshold effect or a different metabolism has not yet been answered.

Exogenous hormones administered as combination oral contraceptives are associated with a 50 per cent reduction in incidence compared with non-users.47- 49 The protective effect seems to be higher with increasing doses of progesterone and decreases with years of discontinuation.4̂ The protective effect was also higher among nulliparous.49

First degree relatives more often have endometrial cancer. There is also an over representation of breast cancer among those women. Thus specific inherited factors, including adipositas, might exist but this can also be due to coincidences associated with environmental risk factors clustered in families.50

11

EndocrinologyUnopposed oestrogen causes endometrial proliferation and hyperplasia, includ­ing atypical hyperplasia.51 This condition is a premalignant stage for progression to endometrial adenocarcinoma.52*54

Conditions with unopposed oestrogen are anovulatory cycles in the polycystic ovarian syndrome, during climacterium when the cyclical progesterone fails, and oestrogen substitution without sufficient addition of progesterone.

Oestrogen can be considered as a cocarcinogen, probably necessary but not by itself enough to cause development of cancer.55 Yet, endometrial adenocar­cinoma occasionally occurs in ovariectomized women,56 exemplifying that cancer can develop under hypoestrogenic conditions.

Several studies have focused upon the increased risk for endometrial carcinoma after unopposed oestrogen therapy,45’ 46’ 57 and also upon whether patients with cancerous endometrium have an inadequate endogenous hormone balance that might predispose for malignant changes. The results are contradic­tory. Some early studies found elevated levels of oestrogens.58-60 When patients and controls were matched for weight and age, two conditions that affect the conversion rate of androstenedione to oestrone,12 the difference disappeared.10’ 61-65 However this result has been contradicted by others.64 Enhanced aromatase activity in vitro in adipose tissue from endometrial cancer patients has been demonstrated.15’ 65’ 66

The relationship between androgens and endometrial adenocarcinoma has also been difficult to evaluate since C-19 steroids can be converted either to oestrogens or steroids with an androgenic profile through different steroid pathways. A study of androgen concentrations where endometrial cancer was compared with a control group did not show any significant difference.67

Obesity is associated with decreased levels of sex hormone binding globulin (SHBG),68’ 69 resulting in more unbound oestrogen, mainly oestradiol, which is the most potent oestrogen. Simultaneously, there is, per se, an increased oestrogen pressure due to the increased conversion rate.

No difference in progesterone concentration has been shown between endometrial cancer and control groups and low concentration has not been associated with an increased risk of cancer.60’70-72 After menopause, progester­one concentrations are comparable with the low levels of the follicle phase in fertile women and elevated oestrogen levels causes a constant ”anovulatoric state”. The importance of having a progesterone concentration in balance with oestrogen has been shown when studying hormone substitution therapy. In studies with cyclic high-dose therapy, 8 of 22 developed hyperplasia, and 1 of 24 during sequential low-dose therapy.75

The influence of endogenous hormones on the growth of endometrial carcinoma in vivo is one of the topics of this thesis.

12

Prognostic factorsThe prognosis for a patient with endometrial adenocarcinoma is dependent on the malignancy of the tumour and duration of disease and can be expressed in the classical prognostic parameters: clinical stage, histopathological grade and myométrial invasion. These factors have been the base for choice of therapy. Clinical stage (used in these papers) is a strong and independent factor. However, as 90 % of the patients are in stages I and II the majority of deaths are in this stage. Myométrial invasion has thereby become very important for choice of therapy in early stages. In many studies histopathological grade has been correlated to survival,129’ 159 but the estimation is subjective.

There is a strong need for objective reproducible methods of compiling a prognosis. As a result, different methods for measurements have been developed to express proliferation rate. During the last decade, interest has been focused upon flow cytometric measurements, since the method is sensitive and reproduc­ible. Ploidy level.74-77 and S-phase fraction77-79 have been shown in several studies to be of prognostic value and in multivariate analysis even to be superior to traditional parameters, except stage77 or grade.79 The background and its relation to other methods are discussed below.

Sex steroid receptors also contain prognostic information and have been of clinical interest for hormone treatment with gestagen. This is discussed later.

Ploidy and proliferation: backgroundAn association between a high tumour proliferation rate and prognosis is a general feature of human cancer.77’ ^ 98 In order to estimate proliferation, several morphometric and biochemical methods have been developed.

The first histological base for development of the different measuring methods was laid when Miescher isolated ”nuclein” from the cell nuclei.80 This later led to the discovery of DNA as the main constituent of the chromosomes and as carrier of the genetic code information. In 1936, it was found that the amount of DNA doubled during the cell cycle.81

The cell cycle is the name given to the series of events that occur in a proliferating cell (Fig. 1):1. G 1-phase represents the time between the end of mitosis and the beginning of the S-phase. Most of the specialised functions of the cell are carried out during Gl (and GO).2. S-phase: The time during which the DNA is duplicated.3. G2-phase: The phase while the cell prepares itself for mitosis, with the number of chromosomes doubled.4. Mitosis (M) or cell division.

13

s

G2

G2 M

Figure 1. The cell cycle with its different phases and DNA content.

This occurs provided that the mother cell has doubled its DNA contents. In a normal cell the doubling is from 2c to 4c (lc = 1 set of chromosomes, n=23). Immediately before division the chromatin condenses into chromosomes, each of which divides in two for distribution to each daughter cell. This phase is usually easy to identify in a light microscope. After the mitosis, the cell continues the next cycle or enters a resting phase, GO.

Individual cells of human solid tumours have a mean duration in the cell cycle of 2 days, but there is a wide variety.82 The cell can chose to re-enter the cycle or rest in GO.83 A recent theory is that the cell may also rest in the S-phase, thus enter a SO phase.84’85 The explanation would be nutrient deprivation and/or cell- crowding phenomena.

The growth rate of a tumour depends upon three main factors: proportion of cells in the active cell cycle, the duration of the cell cycle and the amount of cell loss, i.e. necrosis and apoptose.8^

14

Chromosomes and Ploidy levelThe chromosomal changes can be characterised as numerical and structural. Numerical changes indicate the gain or loss of chromosomes. Structural changes with rearrangements of the chromatin may be of several varieties, e.g. translocation and deletion. Chromosome studies on hyperplastic and atypical endometrium seldom show aneuploidy,87 and studies on tumourous endometrium show that a fairly large part, 42%, retains a diploid chromosome set.88 The pattern is related to histopathologic differentiation, and deviations are more common in the undifferentiated tumours.89

Ploidy measurements, regardless of method (chromosomal, single cell or FCM) show that the majority of endometrial tumours are near diploid, and their generally favourable prognosis may be associated with this characteristic.74,77 ’90' 92 Ploidy also relates, to some extent, to the degree of differentiation.74’ 95> 94

Flow cytometry and ploidyIn the FCM measurement, it is the deviation of DNA content that can be

estimated. When the cell has the DNA content of a normal set of chromosomes, 2c, it is diploid, and divergence is regarded as aneuploidy. Lindahl et al. evaluated 166 endometrial cancers, stages I and II, and showed that DNA ploidy had a much more marked prognostic value than the degree of differentiation, myométrial invasion and receptor content concentration.75 Britton et al. evaluated 256 endometrial cancers. The recurrence rate was 10% in the diploid tumours compared to 34% in the aneuploid, and in their multivariate analysis only histological subtypes and DNA ploidy remained significant.76

Flow cytometry and S-Phase fraction (SPF)Cytometric investigations are of importance both for prognostic and descriptive purposes to increase the understanding of basic tumour biology. FCM is one of the methods for studying cell proliferation and today probably the one used most. The method reveals the ploidy level and the proportions of cells in the different phases, Gl, S and G2 + M. It is a static measurement of the number of cells in S- phase and described in a histogram. FCM has the advantage of being an objective and rapid technique. FCM also has good reproducibility, given that the conditions are optimal.95

After staining the cell suspension with a DNA binding fluorescence the DNA content in each cell is measurable. The SPF is calculated from the number of cells between the G l and G2 peaks. This calculation is easily done in the diploid case, but might be more difficult in certain aneuploid tumours as there may be overlapping aneuploid peaks. SPF is known to have prognostic value in a variety of tumours: Urinary bladder,96 breast cancer,97 cervical,98 ovarian99 and endome­trial carcinoma.77’ 100

The impact of SPF as a prognostic independent factor has been shown. Stendahl et al. found SPF to be, next to clinical stages III and IV, the strongest

15

predictor of outcome.77 Rosenberg et al. found that only SPF/ploidy and FIGO grade (considering the histological subgroup uterine papillary serous carcinoma) were independent prognostic variables.79 Geisinger et al. also found SPF to carry prognostic information.78

Other MethodsOther methods for evaluation of proliferative properties are described below.

Thymidine labelling index (LI).The description of the thymidine labelling index in the 1950s by Taylor et al. provided a new approach to the study of growth kinetics. The added tritiated thymidin (3H-TdR) is incorporated in the newly synthesised DNA and detected either by autoradiography or by scintigraphy. The method is still in use and measures both the duration of all the phases in the cell cycle and the proportion of cells in S-phase and mitosis. Disadvantages are the multistep technique and the time required. Good correlation has been reported between LI and FCM,101 LI and MI.102 Several studies have shown the method’s value in describing the biological course of events.105»104

Mitotic index (MI).The condensation of the chromatin directly prior to mitosis enables us to visualize this part of the cell cycle. The method depends on the investigator’s capability to recognise the mitosis. Contradictory results on reproducibility, both intra- and interindividually, have been reported. The mitotic index relies on the fixation of the cells to make the mitosis last and prevent the fission from taking place.105 Measurements of MI are facilitated by adding a dye, which better visualises mitosis.10̂ Several studies have reported MI to be an important predictor for survival in different malignancies.107' 109

Bromodeoxyuridine labelling indexBromodeoxyuridine is a pyrimidine analogue of thymidine which is incorporated into DNA by competitive antagonism with thymidine. It can be detected by monoclonal antibodies, and used both in vivo and in vitro. The method is able to identify different cell-cycle phases, cell kinetic parameters such as S-phase duration time, and the potential tumour doubling time (Tp0t) can be calculated. It is a faster technique than LI, and does not need radioactive labelling, thus being more attractive for in vivo use. A good correspondence between this technique and other FCM methods has been described.110

Proliferation-associated antigensProliferating cell nuclear antigen (PCNA) is a protein identical to cyclin which plays a critical roll in the initiation of cell proliferation as an auxiliary protein for DNA polymerase S.

Ki-67 is a nuclear antigen which is expressed in cycling cells but not in quiescent cells. Correlation between Ki-67 and SPF has been demonstrated.111

16

ReceptorsAn abundance of receptors could be another reason for increased hormonal stress. Compared to the normal endometrium, the hyperplastic one has a denser representation of progesterone receptors.112 Others found levels comparable to those of the normal proliferative endometrium.25»115 Usually there is a decreased receptor representation in the endometrial adenocarcinoma tissue,25»112»ll4»115 and also a tendency for a lower receptor content with decreased differentia­tion.116» 117

The receptor might also be low or non-functioning even when present, leading to inadequate response to present hormone levels. Thus tumours may be receptor positive but non-responding.118 Primary endometrial resistance to progesterone has been described, but is probably rare.119

Receptors as a prognostic factor are not as frequently used in endometrial cancer as in breast or prostate cancer, but receptor presence seems to be favourable for prognosis.78»120-125 Some studies indicate that the progesterone receptor has a higher degree of prognostic value than oestradiol receptors, or is perhaps the only receptor with a sufficient prognostic value.117» 120» 126-128 However, the results are not unambiguous.

17

Aims of the study

To describe the relation between clinical parameters, associated with endome­trial adenocarcinoma, and plasma levels of the sex steroid hormones: oestrone, oestradiol, progesterone, androstenedione and testosterone.To study the distribution and interrelationship between the plasma concentra­tion of oestrone, oestradiol, progesterone, androstenedione and testosterone in endometrial adenocarcinoma patients.To investigate the endogenous concentration and balance of oestradiol and progesterone and their relation to flow cytometric SPF and ploidy in endome­trial adenocarcinoma.To investigate the SPF, ploidy level and receptor content in endometrial carcinoma tissue in relation to oestradiol and progesterone plasma concentra­tion.

18

Material and methods

PatientsThe patient material in Papers I and III consisted of 128 consecutive postmenopausal women with adenocarcinoma of the endometrium, admitted to the Department of Gynaecologic Oncology, University Hospital, Umeå between July 1984 and July 1986. The median age was 67 years (range 50-94). To be defined as postmenopausal, two years of amenorrhea before clinical symptoms of malignancies was required.

All patients were examined under anaesthesia and fractional uterine curretage was performed. According to the clinical FIGO system, the clinical staging of the patients was as follows: 111 (87%) of the patients were in stage 1,10 (8%) patients in stage II, 3 (2%) patients in stage m and 4 (3%) in stage IV. According to the WHO classification,129 42 (33%) tumours were well differentiated, 67 (52%) moderately differentiated and 19 (15%) poorly differentiated . All specimens were reviewed by the same pathologist, who confirmed the diagnosis of endometrial adenocar­cinoma.

The patient material in Papers II and IV consisted of the same 128 patients as discussed in Papers I and DI. Ninety-nine tumour samples were available for FCM analysis. The difference was due either to necrotic or insufficient samples or to prior surgery.

The patient material in Paper V, where also receptor analyses were performed, consisted of 60 patients taken from the later part of the bigger consecutive material from Papers I and ID.

The patient material analysed for FCM (n=99) and for receptors (n=60) was compared with the consecutive group (n=128) as regards hormone concentra­tions, grade, age, BMI, SPF and ploidy level. No significant differences were found between the groups (Table 1).

Clinical parametersThe clinical parameters investigated were: Clinical stage of disease, histopathology, body mass index (BMI) weight, depth of uterine cavity, age, pregnancy, parity, years of menstruation, menopausal age and years, hypertension and diabetes mellitus.

19

Table 1. Mean horm one levels and clinical parameters in the patient groups from paper I+m, n+IV and V. No significant differences betw een m aterials w ere found.

Analysis____________________________Hormone FCM_________ Receptor

Number 127 99 60Oestrone (pmol/1) 247 263 284Oestradiol (pmol/1) 85+ 87 98Progesterone (nmol/1) 0.51 0.51 0.55Androstenedione (nmol/1) 2.9 2.9 2.8Testosterone (nmol/1) 1.2 1.2 1.2Peridiploid (%) - 80 77Aneuploid (%) - 20 23S-phase fraction (%) - 11.7* 12.1Depth of uterine cavity (cm) 7.7 7.7 8.0Age (years, median) 67 68 68BMI (kg/m2) 29 29 29Weight (kg) 73 74 73Stage I (%) 87 87 85Grade 1 (%) 33 27 32Grade 2 (%) 52 56 47Grade 3 (%) 15 17 22

+ n=126 patients. ̂ n^TÖ patients

Hormone assaysPretreatment blood samples for oestrone, oestradiol, progesterone, androstenedione and testosterone assays were collected between 10 a.m. and 1 p.m. After centrifugation, the plasma was frozen directly in small aliquots and stored at -70 °C until analysed. Plasma levels of oestrone, oestradiol, testosterone, androstenedione and progesterone were measured in duplicate samples by radioimmunoassay (RIA) after celite chromatography for each patient. The RIA method used has been described elsewhere. *3° All samples from the same patient were assayed in duplicate in the same run. The accuracy of each method is expressed as the coefficient of the intra-assay variation, which has been calculated as the variation between duplicates of the same sample within the same assay. If the mean value differed more than 30% between the duplicates, the sample was reassayed before

20

it was accepted. In order to express the accuracy of the method in terms of coefficient of inter-assay variation, the same sample was analysed in triplicate for each assay run. The middle sample in the assay had only 50% of the volume compared with the other two, thereby enabling parallelism to be controlled.

Antisera for oestrone used were obtained from Bio Mérieux (Charbonnierers, France) and for oestradiol from Miles-Yeda (Rehvat, Israel). Interassay coefficients of variation were 12.7% and 13% and intraassay 10.5 and 10.0% for oestrone and oestradiol, respectively. Antisera used for progesterone and androstenedione were obtained from Endocrine Sciences, Tarzana, California and for testosterone from Chemical Laboratory, The Faculty of Veterinary Medicine, University of agricultural Sciences, Uppsala, Sweden). Interassay coefficients of variation were 7.5%, 9.9% and 10.4% and intraassay 5.5%, 8.0% and 8.4% for progesterone, androstenedione and testosterone, respectively.

Hormone levels from postmenopausal healthy women were used as reference to define dividing points between ”low” and ”high” plasma hormone concentra­tion.131 A concentration below the mean of the reference value (60 pmol/1) was considered as low and concentration equal and above the mean +2 SD was considered as high (120 pmol/1).

Flow cytometrySamples for flow cytometry were taken at the second curretage. The specimens were preserved in a 12% DMSO (dimethylsulphoxide) solution and instantly frozen to -70°C. Before preparation the biopsies were thawed at room tempera­ture for 60-90 minutes.

A suspension of cell nuclei was obtained using a combined mechanical and enzymatic technique, described by Tribukait et al.132 Cells were fixed in absolute ethanol and digested with RNAse and pepsin. The suspended nuclei were washed and stained with ethidium bromide. The DNA content of a mean number of 30 000 cell nuclei was analysed from each tumour biopsy. For the analysis an ICP11 flow cytometer, with a flow rate of up to 1000 cells/sec was used. The excitation and emission wavelengths were 455-490 nm and 590-630 nm, respectively.

The tumour DNA values were expressed as relative values, where the DNA content of normal diploid G0/G1 cells was given the value 2.0c. Human lymphocytes with a diploid DNA content (2.0c) and with a coefficient of variation of 2-3%, were used as an external standard. The modal DNA value of each sample was defined by the most prominent G0/G1 peak. Ploidy levels of 1.8-2.2c were regarded as near diploid (peridiploid) tumours, with other values considered to be aneuploid.

Diploid and peridiploid tumours in this study were considered as one group, since these groups together have a better prognosis than tumours outside this range.77

When possible, SPF was calculated as the proportion of cells between the Gl and G2 peaks, corrected for background fluorescence, according to the simplified method described by Baisch et al.133

21

Steroid receptorsAn immunohistochemical localisation of oestrogen and progesterone receptors was performed using commercial reagent sets (Abbott Laboratories, Diagnostic Division, North Chicago, IL) according to the instructions of the manufacturer. The results were evaluated as specific staining: not present (-), weak (+) , moderate (++) and strong (+++). All three grades of staining have been interpreted as receptor-positive.

Data analysisThe Mann-Whitney U-test or the Kruskal-Wallis test was used when non- continuous parameters, such as histopathology, clinical stage, ploidy level and receptor content were analysed for differences in hormone concentration and other parameters. Continuous clinical parameters, such as hormone concentra­tions, SPF and age, were studied by using Spearman’s rank correlation or normal correlation coefficient analysis. Chi-square and Fisher’s exact tests were used for comparing frequencies.

Hormone concentrations of all patients were studied and described. The distribution patterns were described in histograms, in skewness to reflea the degree of symmetry in the distribution, and kurtosis to measure the amount of data in the tail.

Mean, ±SE, was used if not otherwise stated.

Table 2. Statistical m easurem ents o f horm one concentrations.

Skew- Kur-Mean ±SE Median Min Max ness tosis

El Cpmol/1) 247 17 191 18 1196 2.1 5.5E2 (pmol/1) 85 8.3 59.5 5 628 3.5 15T (nmol/1) 1.2 0.046 1.04 0.17 3.3 1.0 1.5A (nmol/1) 2.9 0.13 2.52 0.9 9.8 1.8 4.6P (nmol/1) 0.51 0.024 0.50 0.03 1.8 0.83 2.4

El: Oestrone. E2: Oestradiol. T: Testosterone. A: Androstenedione. P: Progesterone.

22

Results

Hormone concentrations and clinical parameters (Papers I and III)From the 128 patients, 126 oestradiol assays were obtained together with 127 assays from each of oestrone, progesterone, androstenedione and testosterone. Values for mean, ±SE, median, range, skewness and kurtosis are given in Table2. The distributions of the hormones are shown in Figs. 2-6. Oestradiol and oestrone had a skewed distribution, both with a tail towards high concentrations. The same tendency was noted for androstenedione.

Awide range of values was obtained for the oestrogens: oestrone 118-1196 and oestradiol 5-628 pmol/1. Eighteen patients (14%) had an oestradiol concentration above 120 pmol/1, a value corresponding to levels seen in proliferation phase in fertile women. This was arbitrarily regarded as a high value in our analysis. Progesterone had a range between 0.03-1.75 nmol/1. The concentrations (>1 nmol/1) are comparable to values seen during the proliferative phase.

The correlations between and p-values for different hormone concentrations are shown in Table 3 and Fig. 7. Oestrone correlated with oestradiol, testosterone, androstenedione and progesterone. Oestradiol correlated with oestrone, testosterone and androstenedione. Progesterone correlated with oestrone. Androstenedione correlated with oestrone, oestradiol and testosterone. Testosterone correlated with androstenedione, oestrone and oestradiol.

Table 3. Interrelation betw een the plasm a horm one concentrations.

El E2 T A PEl - p=0.0001 p=0.002 p-0.03 p^O.OlE2 rsÄ0.63 - p=0.003 p=0.02 nsT rs=0.28 rs=0.27 - p*0.02 nsA rs=0.19 rs=0.20 rs=0.20 - nsP rs=0.22 rs=0.04 rs=0.06 rs=-.08 -

rs = Spearman rank correlation coefficient, ns =not significant.

The hormone concentrations were analysed against the clinical parameters mentioned in the section on Material and Methods. The results are summed up in Table 4. BMI, weight and depth of uterine cavity were the features which most consistently correlated with oestrone and oestradiol. A negative correlation between androstenedione and age was found. There were significant relations between oestrogens and diabetes mellitus and oestradiol and hypertension, respectively.

23

sc<L>

OJxSz

50

40

30

20

10

0 ■̂ ^ . p . f p rrT^rn200 400 6000

SaCÖa

JJ

30

20

10

0400 800 12000

Oestradiol Oestrone

§CÖa

<DXSDz

20

10

00 ,8 1,2 1,6

Progesterone

- I T T -

0 1 2 3 4 5 6 7 8 9 10Androstenedione

Sccôa

20

10 -

£BPz

\Ål i t

0 ,5 1 1,5Testosterone

2 2,5 3 3,5

Figures 2-6. Frequency distribution of sex steroid hormone plasma concentra­tions. For values see Table 2.

24

Acetate CH,V I

Cholesterol C = O

Pregnenolone

Progesterone

OH

TestosteroneAndrostenedione

OH

HOHO Oestrone Oestradiol

Figure 7. A schematic figure of the steroid synthesis, with arrows showing the pathways, and dotted lines showing the significances found between the hormone assays of P, A, T, El and E2. Significance values are shown in Table 3.

25

Table 4. Correlations and significances betw een investigated clin ical param eters and horm one concentrations.

El_______E2_______ P________ A________TGrade 1-3 ns ns ns ns nsStage I-IV ns ns ns ns nsDepth of uterine cavity 0.43 0.31 ns ns nsResidual tumour ns ns ns ns nsAge 0.18 ns ns 0.22 nsBMI 0.43 0.52 ns ns nsWeight 0.36 0.44 ns ns nsNo. of pregnancies p<0.05 p<0.05 ns ns nsNo. of parities ns p<0.05 ns ns nsAge of menarche ns ns ns ns nsAge of menopause ns ns ns ns nsYears of menstruation ns ns ns ns -0.20Menopausal years ns ns ns -0.25 0.20Hypertension ns p<0.05 ns ns nsDiabetes mellitus* pO.OOl p<0.01 ns ns ns

El:Oestrone. E2: Oestradiol. T:Testosterone. A: Androstenedione. PProgesterone. P<0.05: rs=0.18-0.20. p<0.01: rs=0.22-0.25. pO.OOl: rs=0.31-0.52.+The diabetes mellitus group had a significantly higher BMI, p<0.05.

Hormone concentrations and flow cytometric analysis (FCM) (Papers II and IV)Ninety-nine tumours were available for flow cytometric analysis. Seventy-nine tumours (80 %) were peridiploid and 20 (20 %) were grossly aneuploid. DNA values between 1.7 c and 4.0 c were found. S-phase fraction (SPF) was evaluable in 78 (79 %) of the samples. The mean value was 11.7%; range 4.0% -34.9%. Aneuploid tumours had a higher mean SPF than peridiploid tumours, 16.4% and 10.5%, respectively (p< 0.001).

There was no difference between peridiploid and aneuploid tumours in mean hormone levels of oestrone, oestradiol, progesterone, androstenedione and testosterone, and no correlation was found between hormone concentrations and SPF when the whole group was studied.

The material was further studied in subgroups stratified for tumour grading and ploidy level.

26

Table 5. D efin ition o f cutoff levels that produced significant differences betw eensubgroups,1* defined according to ploidy level and tum our grading.

Diploid & Aneuploid Diploidnmol/1 0.6 0.7 0.8 0.9 0.6 0.7 0.8 0.9

W+M+P ns ns ns p<0.04 ns ns p<0.05 p<0.01W+M ns ns p<0.01 *o A O Ö ns ns p<0.01 p<0.01W ns ns p<0.03 - ns ns p<0.05 -M ns ns ns - ns ns - -P ns ns ns - ns ns - -

W: well differentiated; M: moderately differentiated; P: poorly differentiated. +The subgroups W, M and P were subjected to stepwise analysis at the cutoff levels of progesterone shown in the table. Where few observations were made, no statistical value is given.

There was no linear correlation, but when progesterone concentration exceeded 0.8 nmol/1 a lower S-phase fraction was demonstrated in the subgroups of: (1) well (2) well and moderately (3) diploid: well, (4) diploid: well and moderately and (5) diploid: well, moderately and poorly differentiated tumours (Table 5 and Figs. 8-9). When poorly differentiated and/or aneuploid tumours were excluded, the correlation between higher progesterone concentration and lower SPF became more obvious.

The following four subgroups were analysed for oestradiol: (1) aneuploid: well, moderately and poorly differentiated, n=l6 (20%), (2) diploid: well- differentiated, n=20 (26%), (3) diploid: moderately differentiated, n=34 (44%) and (4) diploid: poorly differentiated tumours, n=8 (10%). The aneuploid group showed a high SPF which did not relate to oestradiol concentration, and neither did the group diploid: poorly differentiated tumours. The diploid: well differen­tiated tumours analysed as a whole group showed no significant linear correla­tion, but the subgroup with low oestradiol concentration, <60 pmol/1, had lower SPF than the group with oestradiol concentration above 60 pmol/1, p< 0.01. Still the SPF was below the mean of the peridiploid group (9.6% and 10.5% respectively). Diploid: moderately differentiated tumours showed a negative correlation between oestradiol concentration and SPF, p< 0.05, rs= 0.40 (Fig. 10).

The SPF, on one hand, was similar in the subgroup consisting of patients with high oestradiol (>120 pmol/1) and low progesterone (<0.5 nmol/1,6 of 78 patients) compared with the complementary subgroup. The subgroup with low oestradiol (<60 pmol/1) and high progesterone (>0.8 nmol/1, 6 of 78 patients) showed, on the other hand, a significantly lower S-phase fraction, 7.3% and 12.1%, p<0.005).

27

12—1

COC

<uC/3S3X!Ph

11 -

10 -

9 -

8 -

7 -

75 66 • 51 • 58 39 26

34 16

“1------ 1 1 1 1 1 1 1----10.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9

Progesterone, nmol/1

S+

Qsco4—>O

r2boC/2

-C

12 - !

11 -

10 -

9 -

8 -

7 -

52 46 4034

2623 17

10

i0.1

i I I I I I I I0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9

Progesterone, nmol/1Figures 8-9. Mean S-phase fraction values at a progesterone concentration exceeding the presented level. Each observation includes the cumulative number of patients, shown inside the figure. The X axis shows the lowest progesterone concentration included. The Y axis shows the mean S-phase fraction. Fig. 8 includes all tumors. Fig. 9 shows the values when only well-differentiated and moderately differentiated peridiploid tumors are included.

28

A D:W D:M D:P

Figure 10. SPF (±SE) in the groups defined according to ploidy level, tumor grading and oestradiol level. The oestradiol concentration in the left bar is <60 pmol/1 and in the right bar >60 pmol/1. A: aneuploid; D:W diploid: well differentiated, D:M diploid: moderately differentiated and D:P diploid: poorly differentiated. 0i0=p<0.05 (^"^)=p< 0.01.

Oestrone, androstenedione and testosterone were analysed in relation to SPF in the same manner, without finding any correlations or differences between the groups.

Hormone concentrations, FCM and sex steroid receptors (Paper V)Sixty tumours were analysed for oestrogen and progesterone receptors. Fifteen out of the 60 tumour samples were receptor positive, and 45 were receptor negative. Nine were positive for oestrogen receptors (ER+) and 12 for progester­one receptors (PR+). Six tumours were stained for both ER+ and PR+. Six biopsies showed only PR+ and thus lacked ER+ and 3 were only ER+ (Fig. 11).

The following parameters were analysed in the different receptor positive subgroups against the receptor negative one: Ploidy level, SPF and plasma concentrations of oestrone, oestradiol, androstenedione, progesterone and testosterone. The results of ploidy level, SPF, and levels of oestradiol and progesterone are shown in Table 6.

29

ER+ Both PR+

Figure 11. The distribution of oestro­gen and progesterone receptor positivity in the 15 tumours stained for receptor presence.

Table 6. SPF, ploidy level, estradiol and progesterone plasm a concentra­tion In the different receptor groups. The receptor positive groups w ere com pared w ith the receptor negative group.

S-Phase FractionSPF

No. %

Ploidy levelAneu

No. ploidPeri­diploid

HormonesE2 P pmol/1 nmol/1

Ree neg+ 37 13 42 11 31 ( 74%) 104 0.55ER+ or PR+ 13 9.3* 14 2 12 ( 86%) ns nsER+and PR+ 4 9.1 5 0 5 (100%) ns nsER+ 7 8.1* 8 0 8 (100%) ns nsPR+ 10 10.1 11 2 9( 82%) ns nsER+ PR- 3 6.7 * 3 0 3 (ioo%) ns nsPR+ ER- 6 10.7 6 2 4( 67%) 124* ns

+The receptor negative group was used as a control group.SPF was evaluable in 50 tumours and ploidy level in 56 tumours. E2: oestradiol; P: progesterone.(*)=p<0.05.

Fifty-six tumours were evaluated for ploidy level and 50 for SPF. A low SPF was seen in the groups with oestrogen receptors, p<0.05 (Table 6). No other differences in SPF were noted. There was no significant difference in ploidy level between the different receptor subgroups. Several groups were small. It is noteworthy that all ER+ tumours were peridiploid, while in the group (PR+ER-), 2 of 6 were aneuploid.

Among the tumours positive for progesterone receptors, only (PR+ER-) showed a significantly higher concentration of oestradiol than the receptor- negative group (p<0.05). As to the oestrogens, this was the only difference found between the groups. There were no significant differences seen for progesterone (Table 6).

30

General Discussion

In Papers I and in the clinical parameters, analysed against the hormone concentrations, demonstrated that BMI and depth of the uterine cavity were the parameters that most consistently correlated to both oestrone and oestradiol. BMI and its correlation to the oestrogens is reported elsewhere,15, 65,154 and further discussed below.

Oestrone levels correlated with age, otherwise there was no association between oestrogens and age. This result is consistent with some previous findings,15, 6l* 154 but contradicts others.155 The negative correlation between androstenedione and age is in agreement with what has earlier been shown.154 The reason why the concentrations of oestrogens do not decrease, while that of androstenedione does, could be related to the increase in aromatase activity during ageing.14,136 There was no significant change of BMI with age.

Parity is of interest in endometrial carcinoma and it has been found that nulliparous and low parity women have a higher frequency of endometrial carcinoma.54,157 In this study, both oestrone and oestradiol showed significant positive correlation with the total number of pregnancies but the number of children born showed only a significant positive correlation with oestradiol. No obvious explanation for this result was found.

Endometrial carcinoma is known to be more common in nulliparous and low parity women.54,157

There were significantly higher values of both oestrone and oestradiol in the diabetic patients, but BMI was also significantly higher in this group. Similar correlations have been shown earlier.65 The hypertensive patients demonstrated significantly higher oestradiol levels than normotensive ones. In this case, BMI was comparable between groups. There was no evident explanation of these relations and it would appear that they require further study.

Previous investigations have reported higher levels of oestrogens in well differentiated tumours.158 The present investigation was unable to support this. The difference might depend upon the materials used. The present study includes only postmenopausal women and the oestradiol concentration was less than half of that in the other study.

Progesterone concentration was investigated in relation to the same param­eters and no correlations were found. In the etiology of endometrial cancer, progesterone has not been connected with any of the risk factors mentioned. Theoretically a deficit could be as crucial as an abundance of oestrogen, but no differences concerning progesterone concentrations between patients and con­trols have been shown.60, 70-72

Oestrogen has long been considered as an important risk factor for developing endometrial carcinoma.42 Both oestrone and oestradiol values were distributed over a wide range, with a considerable proportion in the range of premenopausal ovulating women. There were also patients with low concentrations, which supports the opinion that hormone concentrations cannot be the only factor for development of endometrial carcinoma.56 The highest progesterone concentra­

31

tions in these patients were well below the levels seen during the secretory phase of fertile women, where progesterone exerts its main function.

Non-ovarian sources assume major roles in oestrogen production in meno­pause. Adipose tissue has been identified and described as an important source of oestrone.61 The presence of aromatase enzymes in adipose tissue has previously been shown in several publications,^’ 156 and thus aromatase enzymatic activity is dependent on the amount of adipose tissue.10 This indicates the presence of a rate-limiting step. The finding of a strong positive correlation between BMI and oestrone concentration gives further credence to this statement.

In Paper ID, the analysis of different hormone concentrations showed several interrelations. The correlation between androstenedione and oestrone (p<0.05) is in agreement with earlier observations,154 but there are also reports where no correlation was found.6 The correlation between progesterone and oestrone appears to have attracted less interest.

There was a highly significant correlation between oestrone and BMI but none between BMI and androstenedione or progesterone. That oestrone depends on BMI is understandable, as the synthetic enzymes are situated in the adipose tissue, especially if the aromatase enzyme is a rate-limiting step. If so, a relation between the precursors and the product will become weaker and might not even be seen in plasma. This is in contrast to a situation where aromatase has unlimited capacity, when a strong relation between precursors and products would be expected. However, if there is a correlation between a precursor and the product (progesterone and androstenedione versus oestrone), as was the case in this study, it is likely that this correlation indicates the importance of the precursor. The correlation between progesterone and oestrone was even stronger than between androstenedione and oestrone, which might indicate that the peripheral synthesis of oestrone also has progesterone as a precursor. That would further stress the importance of an increased BMI.

In Paper II, the material was analysed on the basis of endogenous hormones and their known stimulative and antiproliferative effect on the normal en­dometrium. The proliferative activity was evaluated by FCM, a technique that measures the number of cells in the synthesis phase, i.e. the S-phase fraction (SPF). Several studies have shown the prognostic value of SPF for endometrial adenocarcinoma.77’ 79’ 100

The stimulative effect of oestrogen was only seen in the diploid well- differentiated tumour group and with a threshold effect. The response is similar to that of the fertile endometrium to oestradiol, in which a certain threshold gives the maximum effect on proliferative parameters such as the number of glandular mitosis and the amount of DNA + RNA.22-24 Diploid and well-differentiated tumours with an oestradiol concentration >60 pmol/1, still had an SPF-value below the mean value of the whole diploid group. SPF in the poorly differentiated and aneuploid tumours was high, regardless of the oestradiol concentration.

Uterine depth and size probably have minor or neglectable prognostic value. The uterine depth of cavity correlated with oestradiol concentration which, by itself, influences uterine growth. There was no correlation between SPF and depth of the uterine cavity in this material. In previous studies, multivariate analysis did not demonstrate any correlation between depth and survival, while SPF was

32

highly correlated to survival.77, 140 Depth of uterine cavity still influences the subdivision of clinical stage, but recent information indicates that uterine size is not an independent risk factor, but rather relates to cell type, grade and myométrial invasion.141 Delayed diagnosis and a tumour location preventing drainage of the cavity are other reasons for uterine enlargement.

In the association between progesterone concentration and SPF a more general influence could be seen in the different subgroups (Paper IV). SPF decreased significantly when progesterone reached a concentration of 0.8 nmol/ 1 if the aneuploid or poorly differentiated tumours were excluded. Progesterone thus related to SPF even in groups where the tumours have more malignant characteristics, while oestradiol only related to the diploid, well-differentiated group. The diploid, moderately differentiated tumours, showed a negative correlation with oestradiol and SPF. The receptor content did not differ between groups and does not explain the negative correlation. The mean SPF was equal to the mean of all tumours, 11.7%.

The association between progesterone concentration and SPF was also seen in the endogenous hormone balance. The group with low oestradiol and high progesterone had a lower SPF than the others, while there was no difference between high oestradiol and low progesterone and the complementary group. Thus also in the hormone balance, the antiproliferative effect of the progesterone can be seen through lower SPF. It is known from previous studies that the SPF, as measured by FCM, is correlated to survival.77

In Paper V, it is shown that SPF was significantly lower in the oestrogen receptor positive groups (ER+), (ER+ or PR+) and (ER+PR-), than in the receptor negative group. These groups display a combination of favourable factors, as they are also receptor-positive and mainly peridiploid; all in the groups of (ER+) and (ER+PR). But no significant difference concerning ploidy and receptor content was shown. The groups were throughout small.

Of the receptor positive specimens, 6 were positive only for progesterone receptors. This (PR+ER-) representation has also been found by others. 120> 142 This pattern deviates from the normal endometrium, containing both oestrogen and progesterone receptors, but with a later decrease for PR during the secretory phase.27, 112, 145 The explanation could be that PR is less sensitive to tumour transformation and better resists histological dedifferentiation than ER. It has been shown that poorly differentiated tumours were more likely to be (PR+ER-), than well-differentiated ones.144

Results on the prognostic value of receptors in the endometrial carcinoma have been contradictory, but some reports show that PR per se has better prediction value.117,120’12̂ 128 it seems reasonable to conclude that progesterone receptor positivity by itself indicates a favourable prognosis. The receptor may persist even in the dedifferentiated tumours and consequently the tumour is still adjustable to endogenous antiproliferative monitoring through progesterone. In contrast to oestradiol, the progesterone concentration showed a more consistent effect on the growth rate of endometrial adenocarcinoma. Even in (ER+) tumours, SPF was significantly lower than in the receptor negative tumours.

An immunohistochemical technique was used for receptor analysis. The use of this technique results in a lower number of endometrial cancer specimens being

33

classified as receptor-positive.145»146 This might be due to different sensitivities of the biochemical and immunohistochemical assays. Alternatively, tissue homogenates used for biochemical assays may contain a substantial amount of benign tissue, which may elevate the apparent receptor concentration. The different results concerning the prognostic value of receptors could depend on these differences.

Oestrogens can be regarded as cocarcinogenous because of their effect on developing hyperplasia which, in turn, may be a premalignant stage to endome­trial carcinoma. In this investigation on invasive carcinoma, oestradiol did not influence the proliferative rate (Paper II). It might be critical only during tumour induction or in a fairly intact endocrine target, as discussed above.

A hormone-poor state might possibly be more precarious. Progesterone above a threshold level was associated with decreased SPF (Paper IV). Experiments on nude mice have shown that growth under oestrogen-poor conditions leads to development of less differentiated tumours and a tendency for enhanced tumour growth.147 Oestradiol, within physiological levels or opposed, might no longer be a main risk or driving factor in established cancer, and low hormone levels, including both low oestrogen and progesterone, might be more critical for continued tumour development. If so, a restrictive attitude towards combined substitution treatment is unnecessary, and to some extent it has already changed. Patients treated for stage I endometrial adenocarcinoma were given substitution therapy because of severe postmenopausal symptoms. The recurrence rate was lower in the substituted group, and the conclusion was that postoperative oestrogen replacement therapy was not contraindicated in selected low-risk patients.148“150

Two important questions remain: Should there be a substitution-free period after treatment and should progestagen be added? There are no reports in favour of an interval between treatment and replacement therapy. Progesterone relates to decreased growth rate and a supplement in physiological doses ought to be beneficial in the event of occult metastasis.

34

Conclusions

Oestradiol and oestrone concentrations are strongly correlated with body mass index, weight and depth of uterine cavity. Increased oestradiol concen­tration is correlated to both diabetes mellitus and hypertension, both diseases being known to covariate with endometrial adenocarcinoma. The hormone concentrations were not correlated to histopathological grade or clinical stage of disease, two of the most established prognostic factors.A wide range of oestrone and oestradiol plasma hormone concentrations was seen in the patients. A considerable number had oestradiol concentrations in the same range as seen in the proliferative phase of the fertile endometrium. Interrelations were found between the two oestrogens, between oestrogens and androstenedione, and between oestrone and progesterone. The latter correlation could indicate a wider peripheral conversion, giving further importance to the amount of adipose tissue.The influence of endogenous oestradiol on proliferation was ambiguous. Increased SPF was seen only in well-differentiated, peridiploid tumours, while moderately differentiated, peridiploid tumours had a negative correlation.The association between progesterone concentration and SPF was of a more general nature. Endogenous progesterone above a certain level related to lower SPF. The combination of high progesterone and low oestradiol was related to a decreased SPF, while high oestradiol and low progesterone did not affect SPF.Presence of oestrogen receptors did not increase proliferation, but SPF was significantly lower in oestrogen receptor positive than in the receptor negative tumours. SPF did not differ between progesterone receptor positive tumours and receptor negative tumours.

35

Acknowledgements

I wish to express my sincere gratitude to:Professor Torbjörn Bäckström for his inspiring attitude and constructive tutorship in leading me through the sex steroid jungle. His creative and enthusiastic approach to research makes me proud to have had him as supervisor.Professor Ulf Stendahl, for his strong support and absolute conviction that this was something I could manage. Without his interest and encouragement this thesis would never have been completed.Peter Strang for his generous way of introducing me to, and sharing his deep knowledge of, flow cytometric measurements — and for being a good friend! Professor Olle Kjellgren for his curt but interested way of commenting the manuscripts — and for generously sharing his vast knowledge.The Department of Physiology and Center for Reproductive Biology are acknowledged for providing laboratory facilities for hormone assays, and especially Mrs. Agneta Andersson for her skilful work.My co-authors, Olli Mäentausta and Ulf Gerdes, for stimulating and positive collaboration.The staff of the Department of Gynaecological Oncology for continuous support and for always being ready to help. And my colleagues: Helena Persson, my tutor in clinical work, who is always willing to share her experi­ence; Mona Ridderheim for her stimulating and interesting cooperation; and Göran Edbom for his friendship and always being ready to start an intense discussion.Monica Forsgren and Karin Gladh, for everyday help with things big and small.Khama Rogo for valuable comments and for revising the English text in my first manuscripts.Bengt Hållberg for his statistical advice and stimulating teaching.Nigel Rollison for skilful help in revising the English text.Rune Hagberg for his art that always gives me pleasure and which adorns my papers in this thesis.Claes for all the time he saved me through his Mac assistance. His patience with my work and with me was even more impressive than his computer skills.Christofer and Karolina for their understanding and all the times they took care of themselves and my part of the housework. Sometimes I am almost astonished by the fact that I am the mother of these three wonderful people.

This work has been supported financially by Lions Research Foundation, University of Umeå and the Swedish Cancer Research Foundation. I am most grateful for this support.

36

References

1. Sjöstedt S. Sex och preventivmedel efter 40 (in Swedish). Journalen 1981; 1:14.

2. The Cancer Registry. Cancer incidence in Sweden. National Board of Health and Welfare (Socialstyrelsen), Stockholm; 1989.

3. Judd HL. Hormonal dynamics associated with the menopause. Clin. Obstet. Gynecol. 1976; 19:775.

4. Judd HL, Judd GE, Lucas WE, Yen, SS. Endocrine function of postmenopausal ovary: Concentration of androgens and estrogens in ovarian and peripheral vein blood. J. Clin. Endocrinol. Metab. 1974; 39:1020.

5. Greenblatt RB, Colle ML, Mahesh VB. Ovarian and adrenal steroid production in the postmenopausal woman. Obstet Gynecol 1976; 47:383.

6. Vermeulen A, Verdonck L. Sex hormone concentrations in post-menopausal women. Clinical Endocrinology, 1978; 9:59.

7. Judd HL, Shamonki IM, Frumar AM, Lagasse LD. Origin of serum estradiol in postmenopausal women. Obstetrics & Gynaecology 1982; 6:680.

8. Rader MD, Flickinger GO, De Villa Jr GG, Mikuta JJ, Mikhail G. Plasma estrogens in postmenopausal women. Am. J. Obstet. Gynecol. 1973; 116:1069.

9. Grodin JM, Siiteli PK, Macdonald PC. Source of estrogen production in postmenopausal women. JCE &M 1973; 36:207.

10. MacDonald PC, Edman CD, Hemsell DL, Porter JC, Siiteri PK. Effect of obesity on conversion of plasma androstenedione to oestrone in postmenopausal women with and without endometrial cancer. Am J Obstet Gynecol 1978; 130:448.

11. Siiteri PK, MacDonald PC. Role of extraglandular estrogen in human endocrinology. In: Handbook of Physiology (vol.II), 615. Baltimorel973.

12. Chang RJ, Judd HL. The ovary after menopause. Clin Obstet and Gynecol. 1981; 24:181.

13. Forney JP, Milewich L, Chen GT, et al. Aromatization of androstenedione to estrone by human adipose tissue in vitro. Correlation with adipose tissue mass, age, and endometrial neoplasia. J. Clin. Endocrinol. Metab. 1981; 53:192.

14. Hemsell DL, Grodin JM, Brenner PF, Siiteri PK, Macdonald PC. Plasma precursors of estrogen. Correlation of the extent of conversion of plasma androstendione to estrone with age. J. Clin. Endocrinol. Metab. 1974; 38:476.

15. Meldrum DR, Davidson BJ, Tataryn IV, Judd HL. Changes in circulating steroids with aging in postmenopausal women. Obstet Gynecol 1981; 57:624.

16. Vermeulen A. The hormonal activity of the postmenopausal ovary. J Clin Endocrinol Metab 1976; 42:247.

17. Donnez J, Langerock S, Thomas K. Peritoneal fluid volume and 17 beta- estradiol and progesterone concentrations in ovulatory, anovulatory and postmenopausal women. Obstet Gynecol 1982; 59:687.

18. Bäckström T, Mählck CG, Kjellgren O. Progesterone as a possible marker for “nonendocrine” ovarian malignant tumors. Gynecol Oncol 1983; 16:129.

37

19. Abraham GE, Maroulis GB, Marshall JR. Evaluation of ovulation and corpus luteum function using measurements of plasma progesterone. Obstet. Gynecol. 1974; 44:522.

20. Guerrero R, Aso T, Brenner PF, Zekan C, Landgren B-M, Hagenfeldt K, Diczfalusy E. Studies on the pattern of circulating steroids in the normal menstrual cycle. Acta Endocrinol (Copenh). 1976; 81:133.

21. Landgren B-M, Undén A-L, Diczfalusy E. Hormone profile of the cycle in 68 normally menstruating women. Acta Endocrinol (Copenh) 1980; 94:89.

22. Johannisson E, Landgren B-M, Rohr HP, Diczfalusy E. Endometrial morphol­ogy and peripheral hormone levels in women with regular menstrual cycles. Fertility and Sterility 1987; 48:401.

23. Ferenczy A, Betrand G, Gelfand M. Proliferation kinetics of human en­dometrium during the normal menstrual cycle. Am J Obstet Gynecol 1979; 133:859.

24. Key TJA, Pike MC. The dose-effect relationship between ‘unopposed’ oestrogens and endometrial mitotic rate: Its central role in explaining and predicting endometrial cancer risk. Br J Cancer 1988; 57:205.

25. Bergeron C, Ferenczy A, Shyamala G. Distribution of estrogen receptors in various cell types of normal, hyperplastic, and neoplastic human endometrial tissue. Lab Invest 1988; 58:338.

26. Bergeron C, Ferenczy A, Toft DO, Schneider W, Shyamala G. Immunocytochemical study of progesterone receptors in human endometrium during the menstrual cycle. Lab Invest 1988; 59:862.

27. Garcia E, Bouchard P, de Brux J, Berdah J, Frydman R, Schaison G, et al. Use of immunocytochemistry of progesterone and estrogen receptors for en­dometrial eating. J Clin Endocrinol Metab 1988; 67:80.

28. Jänne O, Kontula K, Vihko R, Feil P, Bardin CW. Progesterone receptor and regulation of progestin action in mammalian tissues. Med Biol 1978; 56:225.

29. Tseng L, Gurpide E. Estradiol and 20alfa- dihydroprogesterone dehydrogenase activities in human endometrium during menstrual cycle. Endocrinology 1974; 94:419.

30. Lindahl B, Alm P, Killander D, Långström E, Tropé C. Flow cytometric DNA analysis of normal and cancerous human endometrium and cytological- histopathological correlations. Anticancer Research 1987; 7:781.

31. Thornton JG, Quirke P, Wells M. Flow cytometry of normal, hyperplastic and malignant human endometrium. Am J Obstet Gynecol 1989; 161:487.

32. Muir C, Waterhouse J, Mack T, et al. (eds.) Cancer incidence in five continents, vol V. Lyon, France, IARC Scientific Publication 88,1987.

33. Wynder EL, Escher GC, Mantel N. An epidemiological investigation of cancer of the endometrium. Cancer 1966; 19:489.

34. MacMahon B. Risk factors for endometrial cancer. Gynecol Oncol 1974; 2 :122 .

35. Elwood JM, Cole P, Rothman KJ, Kaplan SD. Epidemiology of endometrial cancer. J Natl Cancer Inst 1977; 59:1055.

36. Pettersson B. Risk Factors for endometrial carcinoma. Thesis, Acta Universitatis Upsaliensis, 6. Uppsala, 1985.

38

37. Parazzini F, Vecchia C, Bocciolone L, Franceschi S. The epidemiology of endometrial cancer. Gynecol Oncol 1991;4l:l.

38. Lucas WE. Causal relationships between endocrine-metabolic variables in patients with endometrial carcinoma. Obstet and Gynecolog Survey 1974; 29:507.

39. Fechner RE, Kaufman RH. Endometrial adenocarcinoma in Stein-Leventhal syndrome. Cancer 1974; 34:444.

40. Gusberg S, Kordon P. Proliferative endometrial response to theca-granulosa cell tumors. Am J Obstet Gynecol 1971; 111:633.

41. Hulka BS, Fowler WC, Kaufman DG, et al. Estrogen and endometrial cancer: Cases and two control groups from North Carolina. Am J Obstet Gynecol 1980; 137:92.

42. Persson I. The risk of endometrial and breast cancer after estrogen treatment. A review of epidemiological studies. Acta Obstet Gynecol Scand; Suppl. 1985; 130:59.

43. Persson I. Hormone replacement therapy and the risk of endometrial cancer. Acta Oncol 1989; 28:583.

44. Smith DC, al. e. Estrogens and endometrial cancer in a retirement community. N Engl J Med 1975; 293:1164.

45. Ziel HK, Finkle WD. Increased risk of endometrial carcinoma among users of conjugated estrogens. N Eng J Med 1975; 293:1167.

46. Mack T, Pike MC, Henderson BE, Pfeiffer RI, Gerkins VR, Arthur M, et al. Estrogens and endometrial cancer in a retirement community. N Engl J Med 1976; 294:1262.

47. Kelsey JL, LiVolsi VA, Holford TR, et al. A case-control study of cancer of the endometrium. Am J Epidemiol 1982; 116:333.

48. Weiss NS, Sayvetz TA. Incidence of endometrial cancer in relation to the use of oral contraceptives. N Engl J Med 1980; 302:551.

49. Henderson BE, Casagrande JT, Pike MC. The epidemiology of endometrial cancer in young women. Br J Cancer 1983; 47:749.

50. Lynch HT, Krush AJ, Larsen AL, Magnuson CW. Endometrial carcinoma: Multiple primary malignancies, constitutional factors and heredity. Am J Med Sci 1966; 252:381.

51. Gusberg SB. Hormone dependence of endometrial carcinoma. Obstet. Gynecol. 1967; 30:287.

52. Sommers SC, Meissner WA. Endocrine abnormalities accompanying human endometrial cancer. Cancer 1957; 10:516.

53. Cambell PE, Barter RA. The significance of atypical hyperplasia. J. Obstet Gynaecol Br Commonw 1961; 68:668.

54. Gusberg SB, Milano C. Detection of endometrial cancer and its precursors. Cancer 1981; 47:1173-

55. King RJB, Whitehead MI. Estrogen and progestin effects on epithelium and stroma from pre- to postmenopausal endometria: application to clinical studies of the climacteric syndrome. In: Steroids and endometrial cancer. Ed. Jasonni VM, et al. New York: Raven Press, 1983: 105.

56. Moghissi KS. Oral contraceptives and endometrial and cervical cancer. J Tox and Environ Health 1977; 3:243.

39

57. McDonald TW, Annegers JF, OTallon WM, Dockerty MB, Malkasion Jr GD, Kurland LT. Exogenous estrogen and endometrial carcinoma: Case-control and incidence study. Am J Obstet Gynecol 1977; 127:572.

58. Hausknect R, Gusberg S. Estrogen metabolism in patients at high risk for endometrial carcinoma. Am J Obstet Gynecol 1973; 116:981.

59- Benjamin F, Deutsch S. Plasma levels of fractionated estrogens and pituitary hormones in endometrial carcinoma. Am J Obstet Gynecol 1976; 126:638.

60. Oettinger M, Samberg I, Levitan Z, Eibschitz I, Sharf M. Hormonal profile of endometrial cancer. Gynecol obstet Invest 1984; 17:225.

61. Judd H, Lucas W, Yen S. Serum 17ß-estradiol and estrone levels in postmenopausal women with and without endometrial cancer. J Clin Endocrinol and Metab 1976; 43::272.

62. Reed MJ, Hutton JD, Beard RW, Jacobs HS, James VHT. The conversion of androstenedione to oestrone in women with endometrial cancer. J. Steroid. Biochem. 1979; 11:905.

63. Judd H, Davidson B, Frumar A, Shamonki I, Lagasse L, Ballon S. Serum androgens and estrogens in postmenopausal women with and without endometrial cancer. Am J Obstet Gynecol. 1980; 136(7):859.

64. Petterson B, Bergström R, Johansson EDB. Serum oestrogens and androgens in women with endometrial carcinoma. Acta Universitatis Upsaliensis 6: 1- 17, 1985.

65. Schindler AE, Ebert A, Friedrich E. Conversion of androstenedione to oestrone by human fat tissue. J. Clin. Endocrinol. Metab. 1972; 35:627.

66. Yamaki J, Yamamoto T, Okada H. Aromatization of androstenedione by normal and neoplastic endometrium of the uterus. J Steroid Biochem 1985; 22(1):63.

67. Poortman J, Andriesse R, Donker GH, Thijssen JHH. Adrenal androgens and endometrial cancer. In: Ed. Jasonni VM, et al. Steroids and Endometrial Cancer. Raven Press, New York, 1983:137.

68. Davidson BJ, Gambone JC, Lagasse LD, Castaldo GL, Hammond PK, Siiteri . PK, et al. Free estradiol in postmenopausal women with and without

endometrial cancer. J Clin Endocrin and Metab 1981; 52(3):404.69. Anderson DC. Sex-hormone-binding globulin. Clin. Endocrinol. 1974; 3:69.70. Brody S, Carlström K, von Uexküll A-K, Lagrelius A, Lunell N-O, Rosenborg

L. Peripheral hormone levels and the endometrial condition in postmenopausal women. Acta Obstet Gynecol Scand 1983; 62:525.

71. Dell’Acqua S, Lucisano A, Acampora MG, Parlati E, Cinque B, Maniccia E, et al. Ovarian and peripheral levels of C21, C19 and C18 steroids in postmenopausal women with and without endometrial cancer. In: Ed. Jasonni VM, et al. Steroids and Endometrial Cancer. Raven Press, New York, 1983: 227.

72. Bond A, Diver M, Davis C. Plasma progesterone concentration in women with and without adenocarcinoma of the endometrium. AmJ Obstet Gynecol 1987; 156(2):437.

73- Whitehead MI. The effects of estrogens and progestogens on the postmenopausal endometrium. Maturitas 1978; 1:87.

40

74. Iversen OE. Flow cytometric deoxyribonucleic acid index: A prognostic factor in endometrial carcinoma. Am J Obstet Gynecol 1986; 155:770.

75. Lindahl B, Alm P, Femö M, Killander D, Långström E, Norgren A, et al. Prognostic value of flow cytometrical DNA measurements in stage H I endometrial carcinoma: Correlations with steriod receptor concentration, tumor myométrial invasion, and degree of differentiation. Anticancer Res 1987; 7:791.

76. Britton LC, Wilson TO, Gaffey TA, et al. DNA ploidy in endometrial carcinoma: major objective prognostic factor. Mayo Clin Proc 1990; 65:643.

77. Stendahl U, Strang P, Wagenius G, Bergström R, Tribukait B. Prognostic significance of proliferation in endometrial adenocarcinoma: A multivariate analysis of clinical and flow cytometric variables. Int J Gynecol Pathol 1991; 10:271.

78. Geisinger KR, Homesley HD, Morgan TM, Kute TE, Marshall RB. Endometrial adenocarcinoma: A multiparameter clinicopathologic analysis including the DNA profile and the sex steroid hormone receptors. Cancer 1986; 58:1518.

79. Rosenberg P, Wingren S, Simonsen E, Stål O, Risberg B, Nordensköld B. Flow cytometric measurements of DNA index and S-Phase on paraffin-embedded early stage endometrial cancer: An important prognostic indicator. Gynec Oncol 1989; 35:50.

80. Miescher F. Z Physiol Chem 1871; 4:441.81. Caspersson T. Über den chemischen Aufbau des Strukturen des Zellkernes.

Skand Arch Physiol 1936; suppl 8: 73.82. Tubiana M, Malaise EP. Comparison of cell proliferation kinetics in human

and experimental tumors: respons to irridiation. Cancer treat. Rep. 1976; 60:1887.

83. Lajtha LG. On the concept of the cell cycle. J Cell Comp Physiol 1963;Suppl 1:143.

84. Drewinko B, Li Ying Yang, Barlogie B, Trujillo JM Cultured human tumour cells may be arrested in all stages of the cycle during stationary phase: demonstration of quiscent cells in Gl, S and G2 phase. Cell Tissue Kinet. 198417:453.

85. Allison DC, Ridolpho PF, Anderson S, Bose K. Variations in the 3H thymidine labelling of S-phase cells in solid mouse tumors. Cancer Res 1985; 45:6010.

86. Steel GG. Cell loss as a factor in the growth rate of human tumours. Eur. J. Cancer 1967; 3:381.

87. Bonilla-Musoles F. Chromosomen bei adenomatüer hyperplasie und adenocarcinom des endometrium. Z Geburtshilfe Gynecol 1971; 175:218.

88. Cervenka J, Konlischer L. Chromosomes in human cancer. Springfield, ed. Thomas CCT, 1973: vol III.

89. Slot E. Cytogenetics of carcinoma of the endometrium. Neoplasma 1980; 27:483.

90. Baker MC. A chromosome study of seven near-diploid carcinomas of the corpus uteri. Br J Cancer 1968; 22:683.

91. Atkin NB. Prognostic value of cytogenetic studies of tumors of the female genital tract. In: Advances in Clinical Cytology; eds Koss LG, Coleman DV. Masson Publishing USA Inc., New York 1984: 103. (voi 2).

41

92. Moberger B, Auer G, Forsslund G, Moberger G. The prognostic significance of DNA measurements in endometrial carcinoma. Cytometry 1984; 5:430.

93. Atkin NB. Prognostic significance of ploidy level in human tumors. I. Carcinoma of the uterus. Journal of the National Cancer Institute 1976; 56(5):909.

94. Wagenius G, Strang P, Boman K, Tribukait B, Stendahl U. Residual myométrial invasion after intracavitary irradiation of endometrial adenocar­cinoma stages I and II. In Vivo 1988; 2:243.

95. Heiden T, Strang P, Stendahl U, Tribukait B. The reproducibility of flow cytometric analyses in human tumors. Methodological aspects. Anticancer Research 1990; 10:49.

96. Tribukait B. Diagnostic and prognostic significance of modal DNA value and proportion of S-phase cells in human carcinoma of the bladder. In: Quantative Image Analysis in Cancer Cytology and Histology. Eds. Mary JY, Rigaut JP. Amstedam: Elsevier, 1986: 315.

97. Fisher B, Gunduz N, Constantino J, Fisher ER, Redmond C, Mamounas EP, et al. DNA flow cytometric analysis of primary operable breast cancer. Relation to ploidy and S-phase fraction to outcome of patients in NSABP B-04. Cancer 1991; 68:1465.

98. Strang P, Stendahl U, Bergström R, Frankendahl B, Tribukait B. Prognostic flow cytometric information in cervical squamos cell carcinoma*. A multivariate analysis of 307 patients. Gynecol Oncol. 1991; 43:3.

99. Kallioniemi O-P, Punnonen R, Mattila J, Lehtinen M, Koivula T. Prognostic significance of DNA index, multiploidy and S-phase fraction in ovarian cancer. Cancer 1988; 61:334.

100. Geisinger KR, Kute TE, Marshall RB, Homesley HD, Morgan TM. Analysis of the relationships of the ploidy and cell cycle kinetics to differentiation and the female sex steroid hormone receptors in adenocarcinoma of the endometrium. Am J Clin Pathol 1986; 85:536.

101. Silvestrini R, Costa A, Veneroni S, DelBino G, Persici P. Comparative analysis of different approaches to investigate cell kinetics. Cell Tissue Kinet. 1988; 21:123.

102. Åkerman M, Brandt I, Johnsson A, Olsson H. Mitotic activity in non- Hodgkin's lymphomas. Relation to the Kiel classification and to prognosis. Br J Cancer 1987; 55:219.

103. Meyer JS, Prey MU, Babcoock DS, McDewitt RW. Breast carcinoma cell kinetics, morphology, stage and host characteristics. A thymidine labelling study. Lab Invest 1986; 54:41.

104. Nordqvist S. The synthesis of DNA and RNA in human carcinomatous endometrium in short-term incubation in vitro and its response to oestradiol and progesterone. J Endocr 1970; 48:29.

105. Graem N, Helweg-Larsen K. Mitotic activity and delay in fixation of tumour tissue. Acta Path Microbiol Scand 1979; 87:375.

106. Vasko J, Malmström P-U, Wester K, Busch C, Taube A. Prognostic value of standardized determination of mitotic frequency in bladder cancer using image analysis. J Urogen Pathol 1991; 1:53.

42

107. Baak JPA, van Dop H, Kurver PHJ, Hermans JO. The value of morphometry to classic prognosticators in breast cancer. Cancer 1985; 56:374.

108. Ronan S, Han M, Das Gupta T. Histologic prognostic indicators in cutaneous malignant melanoma. Semin. Oncol. 1988; 15:558.

109. Tsujimoto M, Aozada K, Ueda T, Morimura Y, Komatsubara Y, Doi T. Multivariate analysis for histologic prognostic factors in soft tissue carcino­mas. Cancer 1988; 62:994.

110. Sahani K, Tribukait B, Einhorn N. Comparative study of proportion of S-phase cells in ascites and pleural effusions in ovarian carcinoma using antibromodeoxyuridine monoclonal antibody and DNA flowcytometry. Acta Oncolog 1989; 28:705.

111. Schwartz B, Pinkus G, Bacus S, Toder M, Weinberg D. Cell proliferation in non-Hodgkin's lymphomas: Digital image analysis of Ki-67 antibody stain­ing. Am J Pathol 1989; 134:327.

112.Jänne O, Kauppila A, Kontula K, Syrjälä P, Vihko R. Female sex steroid receptors in normal, hyperplastic and carcinomatous endometrium. The relationship to serum steroid hormones and gonadotropins and changes during medroxyprogesterone acetate administration. Int. J. Cancer 1979; 24:545.

113. Elliot PM, Krozowski Z, Jansen RPS. Cytosol oestradiol and progesterone receptors in endometrial hyperplasia and adenocarcinoma. Aust and New Zeeland J. of Obstet and Gynecol 1980; 20:199.

114. Young PC, Ehrlich CE, Cleary RE. Progesterone binding in human endome­trial carcinomas. Am. J. Obstet. Gynecol. 1976; 125:353.

115. Neumannova M, Kauppila A, Vihko R. Cytosol and nuclear estrogen and progestin receptors and 17 beta-hydroxysteroid dehydrogenas activity in normal and carcinomatous endometrium. Obstetrics and Gynecology 1983; 6l(2):181.

116. Kauppila A, Isotalo HE, Kivinen ST, Vihko R. Prediction of clinical outcome with estrogen and progesterone receptor concentration and their relation­ships to clinical and histopathological variables in endometrial cancer. Cancer Research 1986; 46:5380.

117. Utaaker E, Iversen OE, Skaarland E. The distribution and prognostic implications of steroid receptors in endometrial carcinomas. Gynecologic Oncology 1987; 28:89.

118. Brunner N, Osborne CK, Spang-Thomsen M. Endocrine therapy of human breast cancer grown in nude mice. Breast Cancer Res and Treat 1987; 10:229.

119. Keller DW, Wiest WG, Askin FB, Johnson LW. Pseudocorpus luteum insufficiency: a local defect of progesterone action on endometrial stroma. J Clin. Endocrin. and Metab. 1979; 48:127.

120. Creasman WT, Soper JT, McCarty Jr KS, McCarty Sr KS, Hinshaw W, Clarke- Pearson DL. Influence of cytoplasmic steroid receptor content on prognosis of early stage endometrial carcinoma. Am J Obstet Gynecol 1985; 151:922.

121. Creasman WT, McCarty KS, Barton TK, McCarty Jr KS. Clinical correlates of estrogen- and progesterone-binding Pproteins in human endometrial adeno­carcinoma. Obstet and Gynecol 1980; 55(3>363.

43

122. Martin JD, Hähnel R, McCartney AJ, Woodings TL. The effect of estrogen receptor status on survival in patients with endometrial cancer. Am. J. Obstet. Gynecol 1983; 147:322.

123. Liao BS, Twiggs LB, Leung BS, Yu WCY, Potish RA, Prem KA. Cytoplasmic estrogen and progesterone receptors as prognostic parameters in primary endometrial carcinoma. Obstet Gynecol 1986; 67(4):463.

124. Vihko R, Alanko A, Isomaa V, Kauppila A. The predictive value of steroid hormone receptor analysis in breast, endometrial and ovarian cancer. Med. Oncol and Tumor Pharmacother. 1986; 3:197.

125. Palmer DC, Muir IM, Alexander AI, Cauchi M, Bennett RC, Quinn MA. The prognostic importance of steroid receptors in endometrial carcinoma. Obstet and Gynecol. 1988; 72:388.

126. Lindahl B, Alm P, Fernö M, Grundsell H, Norgren A, Tropé C. Relapse of endometrial carcinoma related tro steroid receptor concentration, staging, histologic grading and myométrial invasion. Anticancer Research 1986; 6:1317.

127. van der Putten HWHM, BaakJPA, KoendersTJM, Kurver PHJ, StolkHG, Stolte LAM. Prognostic value of quantitative pathologic features and DNA content in individual patients with stage I endometrial adenocarcinoma. Cancer 1989; 63:1378.

128. Kleine W, Maier T, Geyer H, Pleiderer A. Estrogen and progesterone receptors in endometrial cancer and their prognostic relevance. Gynecologic Oncology 1990; 38:59.

129. Annual report on results of treatment of gynecological cancer, FIGO. Statement of results obtained in 1979-1981, inclusive. Eds. Pettersson F, Kolstad P, Ludvig H, Ulfelder H. Radiumhemmet, Stockholm, 1988.

130. Bäckström T, Carlström K, von Schoultz B, Toivonen J. Effect of progester­one, administered via intravaginal rings, on serum concentrations of oestradiol, FSH, LH and prolactin in women. J Reprod Fert. 1982; 64:53.

131. Mählck CG. Plasma estradiol in women with epithelial ovarian carcinoma. Acta Obstet Gynecol Scand; Suppl 137, 1986.

132. Tribukait B. Flow cytometry in surgical pathology and cytology of tumours of the genito-urinary tract. In: Advances in Clinical Cytology, vol. 2, p 163. Eds. Koss LG, Coleman D. Masson, New York,1984.

133. Baisch H, Göhde W, Linden W. Analysis of PCP-data to determine the fraction of cells in various phases of cell cycle. Radiat Environ Biophys 1975; 12:31.

134. Jensen J, Riis BJ, Hummer L, Christiansen C. The facts of age and body composition on circulating serum oestrogens and androstenedione after the menopause. Br J Obstet Gynaecol 1985; 92(3):260.

135. Baddawy SZ, Elliot LJ, Elbadawi A, Marshall LD. Plasma levels of oestrone and oestradiol - 17ß in postmenopausal women. Br J Obstet Gynaecol 1979; 86(1):56.

136. Simpson ER, Merrill JC, Hollub AJ, Graham-Lorence S, Mendelson CR. Regulation of estrogen biosynthesis by human adipose cells. Endocrine Reviews 1989; 10(2):136.

137. Kjellgren O. Epidemiology and pathophysiology of corpus carcinoma. Acta Obstet Gynecol Scand; Suppl 65, 1977; 77.

44

138. Lindahl B, Alm P, Femö M, Norgren A, Tropé C. Plasma steroid hormones, cytosol receptors, and thymidine incorporation rate in endometrial carci­noma. Am J Obstet Gynecol 1984; 149:607.

139. Tiitinen A, Forss M, Aho I, Vesterinen E, Nieminen U. Endometrial adenocar­cinoma: Clinical outcome in 881 patients and analysis of 146 patients whose death was due to endometrial cancer. Gyn Oncol 1986; 25:11.

140. Wagenius G, Bergström R, Strang P, Gerdes U, Rogo K, Tribukait B, et al. Prognostic significance of flow cytometric and clinical variables in endome­trial adenocarcinoma stages I and II. Anticancer Research 1992; 12:725.

141. Creasman WT, Morrow CP, Bundy BN, et al. Surgical pathologic spread patterns of endometrial cancer (a GOG study). Cancer 1987; 60:2035.

142. Geisinger KR, Marshall RB, Kute TE, Homesley HD. Correlation of female sex steroid hormone recptors with histologic and ultrastructural differentiation in adenocarcinoma of the endometrium. Cancer 1986; 58:1506.

143. Levy C, Robel P, Gautray JP, de Brux J, Verma U, Descomps B, et al. Estradiol and progesterone receptors in human endometrium normal and abnormal menstrual cycles and early pregnancy. Am J Obstet Gynecol 1980; 136:646.

144. Quinn MA, Pearce P, Fortune DW, Koh SH, Hsieh C, Cauchi M. Correlation between cytoplasmic steroid receptors and tumour differentiation and invasion in endometrial carcinoma. B J Obstet and Gynecol 1985; 92:399.

145. Mutch DG, Soper JT, Budwit-Novotny DA, Cox EB, Creasman WT, McCarty Sr KS, et al. Endometrial adenocarcinoma estrogen receptor content: Association of clinicopathologic features with immunohistochemical analy­sis compared with standard biochemical methods. Am J Obstet Gynecol 1987; 157:924.

146. Segreti EM, Novotny DB, Soper JT, Mutch DG, Creasman WT, McCarty KS. Endometrial Cancer: Histologic correlates of immunohistochemical localiza­tion of progesterone receptor and estrogen receptor. Obstet Gynecol 1989; 73:780.

147. Horvath G, Fernö M, Baldetorp B, Cameron R, Ranstam J. Estradiol induced changes in tumor growth and steroid receptor content in a heterotransplanted human endometrial adenocarcinoma. In Vivo 1991; 5:401.

148. Lee RB, Burke TW, Park RC. Estrogen replacement therapy following treatment for stage I endometrial carcinoma. Gynecol Oncol 1990; 36:189.

149. Creasman WT. Estrogen replacement therapy: is previously treated cancer a contraindication? Obstet gynecol 1991; 77(2):308.

150. Baker D. Estrogen replacement therapy in patients with previous endome­trial carcinoma. Comp Ther 1990; 16:28.

45