Breast Cancer Research and Treatment 16: 51-55, 1990. © 1990 Kluwer Academic Publishers. Printed in the Netherlands.

Brief communication

A breast pre-cancer test? Preliminary results based on a breast temperature rhythm abnormality during the menstrual cycle

Hugh W. Simpson

University Division of Pathology, Royal Infirmary, Glasgow, UK Key words: thermal cycle, breast pre-cancer test, menstrual cycle Summary This study documents the menstrual thermal cycle of 16 breasts considered at normal risk for breast cancer (8 women) and 15 breasts considered at high risk for breast cancer (i.e. ipsilateral N = 7 or contralateral N = 8 to a previously [locally] excised carcinoma). The surface thermometry studies were carried out for 11/2hours each evening for one menstrual cycle using a special automated instrumentation, the 'Chronobra', with the subject at home. The surface temperatures were adjusted to remove the fluctuations in arterial blood temperature during the menstrual cycle, so that they are thought to represent a breast-specific menstrual rhythm. Across subjects, the cycles were synchronised by the day of the progesterone peak obtained by radioimmunoassay of saliva collected daily. Following ovulation, the normal risk breasts exhibited a steady rise of temperature for 14 days to a well defined peak. In contrast the 'high-risk' breasts exhibited a continuous hyperthermia with smaller peaks. Linear discrimination analysis by multiple regression achieved a complete separation of the individual normal and 'high-risk' data sets. The method shows promise for the development of a non-invasive screening test for breast pre-cancer in premenopausal women.

Introduction The management of the premenopausal breast clinic patient with vague lumpiness or mastalgia is a dilemma. The question of cancer 'somewhere' is in the mind of the clinician since one-third of all breast cancers present in the premenopause [1]. A mammogram is often unhelpful since the dense fibrous breast of this age group [2] tends to obscure any neoplastic density. Where no particular lump presents, which focus of disease should be biopsled? The dilemma is made worse when there is a family history of breast cancer, since some 9% of breast cancer cases have this history [3]. The clinican under pressure from the patient to do something might consider anti-oestrogen therapy, but

ICI, the manufacturers of one of the principal agents, have specifically come out against such a use of tamoxifen because the long-term risk benefits need more evaluation [4]. Herein a method is reported which has potential for use in the clinical situation described above. The test involves the patient wearing a brassiere containing miniaturised thermometric devices in order to monitor the menstrual metabolic cycles by surface heat. The garment is worn for 11/2 hours each evening in a controlled environment in the privacy of the home. The collection of daily saliva samples for progesterone radio-immunoassay serves to synchronise cycles across individuals. As a first step the present results compare 16 normal-risk and 15 'high-risk' breasts. The latter

Address for offprints: H.W. Simpson, Department of Pathology, GlasgowRoyal Infirmary, Glasgow, G4 0SF, UK

52

HW Simpson

was defined as the breast contralateral or ipsilateral to a histologically proven cancer previously removed by a breast conservation lumpectomy. It was considered that if an objective separation of normal-risk and high-risk breasts could be made, then there might be a basis for a clinical test of 'risk'.

Subjects and methods Both the 'normal-risk' and 'high-risk' subjects were parous women aged 30-44 years. The former were mainly members of a church mothers' group. Each filled in a questionnaire requesting information on classical risk factors of breast disease in themselves. On the basis of these questionnaires, but without any clinical examination, a normal-risk group was constituted. The high-risk group were all patients who had had surgery for a carcinoma (mainly conservation lumpectomy) in this hospital (courtesy of Mr. Colin McArdle). Clinical details of these cases will be found in Table 1. Mastectomy sites, breasts treated by radiotherapy, and anovulatory cycle data (judged by lack of a progesterone peak) have been excluded.

For 11/2 hours evening breast temperature data were collected for one menstrual cycle by a special thermometric instrument custom-built for the purpose called the Chronobra. The basic fabric of this brassiere is similar to the conventional garment. However, a marsupial pouch between the cups holds a compact instrument package measuring 140 x 65 x 20 mm and weighing 176 g. In turn, this contains a hybrid chip for electronic thermometry, an oscillator to effect sampling, and a 4-K miniaturised digital semiconductor memory store. From the instrument package are subtended, between the layers of the fabric, 16 leads ending in a 'collar stud' buttonholded into the fabric. Each collar stud has fused inside it a semiconductor junction as a means of measuring temperature. The studs are disposed in the fabric to measure and store temperatures as follows: breast surface (4 x 2), shoulder (2 x 2), bra outer surface (1 x 2), and lapel (1 x 2) sensors. In the present study readings were made every 68 seconds. The conditions were standardised in terms of clothing (a zippered polo-neck ski shirt) and room temperature (20°+ 2°C). The subject was asked to sit in the same chair for one hour after supper and to follow a standard routine (e.g. reading or watching television). She also was asked to

Table 1. Details of 'high-risk' subjects

Patient (breast) number

Cancer diagnosis

1 (1, 2 )

'Adenocarcinoma'

2 (3, 4) 3 (5, 6) 4 (7) 5 (8) 6 (9, 10) 7 (11, 12) 8 (13)

'Infiltrating cancer' 'Infiltrating cancer' 'Infiltrating ductal' 'Infiltrating ductal' 'Infiltrating ductal' 'Solid cancer' R. 'Spheroidal cell carcinoma' L. 'scirrhus carcinoma' 9 (14, 15) 'Breast carcinoma'

Age at Treatment study Surgery

X-ray

Chemotherapy Other

42

Lumpectomy

No

No

42 42 42 39 29 35 37

Lumpectomy Lumpectomy Mastectomy Lumpectomy Lumpectomy Lumpectomy Bilateral Lumpectomy

No No No R. only No No R. only

40

Lumpectomy

No

xTen months after end of cytotoxic therapy. (15 breasts total) i.e. 7I + 8C.

Span (months) Ipsilateral (I) between or contralateral surgery and (C) data used study 10

I+ C

No No No Yes No No No

Tamoxifen before study No No No No No No No

4 6 4 13x 24 48 50

I+ C I+ C C C I+ C I+ C I (i.e. Left)

No

No

56

I+ C

Breast pre-cancer test

record the oral temperature using a provided clinical thermometer and to collect a 2 ml sample of saliva by dribbling into a tube. After being stored in the freezing compartment of the domestic refrigerator for the menstrual cycle, these specimens were submitted for progesterone radio-immunoassay at

Table 2. Discriminant analysis of normal and high risk breasts

Breast Known (adjusted) Predicted Original Cancer no: Risk Indicator Risk Ipsi (I) Contralateral (C) Normal Risk 1 Normal 2 Normal 3 Normal 4 Normal 5 Normal 6 Normal 7 Normal 8 Normal 9 Normal 10 Normal 11 Normal 12 Normal 13 Normal 14 Normal 15 Normal 16 Normal High Risk 1 High 2 High 3 High 4 High 5 High 6 High 7 High 8 High 9 High 10 High ll High 12 High !3 High 14 High 15 High

-

0.4340 0.4021 0.7081 0.5171 0.5127 0.4411 0.3296 0.6259 0.4650 0.6005 0.1321 0.3913 0.5434 0.4657 0.4528 0.0615

Normal Normal Normal Normal Normal Normal Normal Normal Normal Normal Normal Normal Normal Normal Normal Normal

None None None None None None None None None None None None None None None None

0.3866 0.2270 0.6210 0.5418 0.5213 0.5064 0.3802 0.2373 0.5426 0.3513 0.2696 0.1993 0.7443 0.6692 0.3795

High High High High High High High High High High High High High High High

I C I C I C C C I C I C I I C

Exclusions of high risk subjects due to mastectomy(most cases were lumpectomy) or radiotherapy to (residual) breast tissue. Based on 21 clays of (individual) oral adjusted breast temperatures straddling the salivaryprogesteronepeak. For identification purposes the high-risk cases are in the same order as Table 1.

53

the Tenovus Institute, Cardiff (courtesy Dr Attila Turkes). In the present study, the breast surface temperatures actually used were those recorded from a site just below the nipple. Previously the menstrual rhythm of breast temperature has been described and the conclusion reached that it is a measure of breast metabolism [1]. Here the method has been refined by correcting each breast temperature by the deviation for that day of the oral temperature from its average for the menstrual cycle. Specifically, if on a particular day the oral temperature was 0.25°C warmer than its average during the menstrual cycle (the normal situation in the luteal phase), then 0.25°C was subtracted from the observed breast temperature for that day. This manoeuvre is considered to compensate for the alterations of temperature of the arterial blood perfusing the breast at different phases of the cycle, thus isolating a 'breast-specific' temperature.

Results For both the normal-risk and the 'high-risk' subjects the averaged peak luteal salivary progesterone value was in excess of 400 pmol/1, representing 4-5 x the levels in the proliferative phase of the menstrual cycle. Standard errors for the control and 'high-risk' subjects were overlapping, and it was concluded that progesterone maxima were normal for both groups. The 'breast-specific' temperature cycles for all subjects were lined up on their progesterone maxima. Ovulation was considered to have occurred 8 days previously [5]. The averaged temperatures for the 16 normal-risk breasts and the 15 'high-risk' breasts are summarised in Fig. 1 with their standard errors. The 'breast-specific' temperatures of the normal-risk breasts show an almost uninterrupted rise for the 14 days after ovulation; therafter there is a relatively sharp fall. In contrast, the 'high-risk' breasts exhibit a maximum 3 days after ovulation, and 2 smaller highs 7 and 14 days later. The mean 'breast-specific' temperature of the 'high-risk' breasts is invariably warmer for the 21

H W Simpson

54 35.6.

35.4

35.2

35.0

34.8

34.6

34.4

34.2

34.0 -10 -8 -6

-4

-2

0

2

4

6

8

10

Day of menstrual cycle relative to progesterone peak

Fig. 1. Time plots of 'breast-specific' temperatures, y axis: breast-specific temperature ° C. Extra y axis at day - 8 = ovulation (see Ref 5 for choice of day - 8; progesterone peak = day 0). Means + one standard error of 'breast-specific'temperatures (see text) for normal-risk breasts (N = 16) (lower line) and 'high-risk' breasts (N = 15) (upper line). Note that the averages do not overlap; also that the normal-risk subjects (only) exhibit a steady rise of temperature for 14 days after ovulation.

days straddling the day of the progesterone maximum. The difference between the means is statistically significant P = 0.02, (overall the 16 normal breasts averaged 34.582 ° C + 0.167 (95% CI), the 9 contralatera135.168 + 0.636, and the 6 ipsilateral 34.947 + 0.612). The utility of the method to detect risk depends on whether there is sufficient in-group homogeneity for separation by, for instance, linear discrimination analysis, so that a future case of unknown

risk can be confidently allocated to one group or the other. We used the Minitab program of Pennsylvania State University for linear discrimination by multiple regression. The indicator variable Y, which is being predicted, is zero or one depending on the cancer risk state, i.e. normal or high. The independent variables X, which are the predictors, are the daily maximum breast temperatures. These cover a span from 10 days before to 10 days after the progesterone peak, i.e. there are 21 predictors. These variables are then entered into the standard discrimination procedure within the Minitab program. There is a complete discrimination of the normal-risk and high-risk groups; all the normal risk breasts exhibit positive predictors and all the 'highrisk' breasts, negative predictors. The negative predictors can be further divided into the weakly negative contralateral breasts (mean + - 0 . 3 6 5 ) and the more strongly negative ipsilateral breasts ( m e a n + - 0 . 5 4 3 ; P = 0.01 by Mann-Whitney test). The method was further tested by running the calculation with the last 3 'high-risk' cases entered as unallocated. In spite of this, all 3 were allocated to 'high-risk'.

Discussion Previously the normal menstrual cycle of breast metabolism has been reviewed; breast volume, epithelial mitoses, epithelial proliferation activity in vitro, and breast temperature all exhibit a menstrual cyclicity with a peak at about day 26 in a 28-day cycle [1]. Histopathological sample studies of 500 consecutive cancer mastectomies from this hospital have revealed that premenopausal specimens exhibit 2.6-9.5 x the amount of 'benign' focal epithelial hyperplasias as compared with age-matched autopsy controls (See Table 1 in reference 6). It is reasonable to predict that under such circumstances the growth cycles, i.e. the mammary menstrual cycle, may also be disordered. The data are consistent with this interpretation.

Breast pre-cancer test

It should be noted that the power of the discrimination analysis is due to the cumulative effect of the abnormalities along the cycle; on any one or two days the discrimination is weak. These studies require confirmation on larger numbers and an extension to women known to be in 'high-risk' states, e.g. positive family history, nulliparity, early menarche, etc., but who have not been subjected to any operative procedure. These might help to allay doubts that the discrimination obtained herein is related to operative intervention or prior chemotherapy. Such reservations are, however, difficult to sustain for the contralateral breasts, and it should be noted that only one of the patients had had chemotherapy.

Acknowledgements Dr Atilla Turkes kindly carried out the radioimmunoassay of salivary progesterone. I am grateful to Mr Colin McArdle, of the University Department of Surgery, Royal Infirmary, for allowing access to

55

his breast cancer patients. Sources of support: Tenovus Institute, Cardiff and The Robertson Trust.

References 1. Simpson HW, Griffiths K: The diagnosis of breast pre-cancer by the Chronobra. I. Background review. II. The breast pre-cancer test. Chronobiol Int 6: 355-393, 1989 2. Department of Health and Social Security: Breast cancer screening report to the Health Ministers of England, Wales and Northern Ireland. London, HMSO, Chairman: Professor Sir Patrick Forrest, 1987 3. Lynch HT, Conway T, FItzgibbon R, Schreiman J, Watson P, Marcus J, Fitzirnmons ML, Lynch JF: Age-of-onset heterogeneity in hereditary breast cancer: minimal clues for diagnosis. Breast Cancer Res Treat 12: 275-285, 1988 4. Diver JMJ, Jackson IM, Fitzgerald JD: Tamoxifen and nonmalignant indications. Lancet i: 733, 1986 5. Dyrenfurth I, Jewelewicz R, Warren M, Ferin M, Vanden Wiele RL: Temporal relationships of hormonal variables in the menstrual cycle. In: Ferin M, Halberg F, Richart RM, Vanden Wiele RL (eds) Biorhythms and Human Reproduction. John Wiley, New York, 1974, pp 171-201 6. Simpson HW, Mutch F, Halberg F, Griffiths K, Wilson D: Bimodal age-frequency distribution of epitheliosis in cancer mastectomies. Cancer 50: 2417-2422, 1982

A breast pre-cancer test? Preliminary results based on a breast temperature rhythm abnormality during the menstrual cycle.

This study documents the menstrual thermal cycle of 16 breasts considered at normal risk for breast (8 women) and 15 breasts considered at high risk f...
361KB Sizes 0 Downloads 0 Views