ETIOLOGY, EPIDEMIOLOGY, AND PATHOLOGY OF PROSTATIC CANCER

L. M. FRANKS, MD, FCAP, FRCP PATH*

The incidence of prostatic cancer is influenced by age, race, and hormones. The effects of age and race are well established, in that prostatic cancer be- comes increasingly common with age and is more frequent in some races, e.g. Caucasians and some Negroes, and rare in the Mongoloid races, but there are wide regional variations. The reasons for this distribution are unknown but may reflect fundamental genetic changes in the cell or in the endocrine pattern, or differences in exposure to external environmental agents. Although the endocrine system influences the growth of prostatic cancer it may play no part in the primary carcinogenic process, the cause of which is unknown. The pathology of the disease is well established. Prostatic cancers begin in the outer part of the gland and involve capsule lymphatics and blood vessels at an early stage. Distant spread is very common. Prostatic cancers vary widely in their biological behavior, many having a very low growth rate (latent or inactive cancers). Factors influencing growth rate and prognosis are not well defined.

0 DISCUSS THESE THREE TOPICS I N THE TERMS T suggested by the organizers would seem to require a formidable feat of compression but my task is made easier by the fact that we know practically nothing about etiology and little about epidemiology; added to which, there has been little new knowledge about the pathology of the disease for the last 30 years. I shall therefore only indicate some outstand- ing features.

DEFINITION OF TERMS

Under the microscope and to the naked eye all cases of prostatic cancer are similar, but bi- ologically-that is in the patient-three types can be distinguished:

1. Clinical cancer: Any case in which a firm clinical diagnosis of prostatic cancer is made and confirmed by histology should be de- scribed as a clinical cancer.

2. Latent cancer: These tumors by defini- tion exist but do not become manifest, i.e. they produce no clinical evidence of disease. They are found incidentally.

3 . Occult cancer: These tumors manifest

Presented at the American Cancer Society's National Conference on Urologic Cancer, Washington, D.C., March 29-31, 1973.

Department of Cellular Pathology, Imperial Cancer Research Fund, Lincoln's Inn Fields, London WC2A SPX, England.

Received for publication June 18, 1973.

themselves by their metastases. The primary tumor remains hidden (occult).

These definitions have no direct relation- ship to size, growth rate, histologic structure, local invasion, or distant metastases. They are concerned only with the method of presenta- tion.

ETIOLOGY AND EPIDEMIOLOGY There are three etiologic factors which seem

to be closely associated with prostatic cancer: age, race and the endocrine system.

The Hormones and Prostatic Cancer Since the hormones were first shown to have

an effect on the prostate they have usually been considered to play a primary role in prostatic carcinogenesis, although the evidence for this is slight. I think that it is reasonable to assume that the main part played by the hormones is to stimulate the development and maintenance of the prostatic epithelium so that a sufficient number of cells is present in which malignant change can occur. The hor- mones may play no part in the actual process of carcinogenesis. Once a tumor has developed the neoplastic cells may remain responsive to the factors which control normal growth, provided that the cells still retain these partic- ular normal differentiated characters." It seems to be true that prostatic cancer and be- nign enlargement do not occur in prepubertal eunuchs or true eunuchoids.20 A simple expla-

1092

No. 5 PROSTATIC CANCER . Franks 1093

nation for this may be that prostatic epithe- lium does not develop to any extent in these patients.

There seems to be no direct relationship be- tween steroid hormone levels-estrogens, an- drogens, or adrenal steroids-in the blood or urine, and the development of prostatic cancer.4.18~2~ Pituitary hormones, which may possibly be involved, have not been inten- sively studied mainly because suitable meth- ods have only recently become available.23 An added drawback to endocrine studies is that most have been carried out after the disease has been diagnosed and it is probable that if there is an endocrine basis, the critical changes may have taken place many years be- fore clinical symptoms appear. All these endo- crine studies are based on the assumption that changes occur in the humoral environment, but we must also consider, the possibility that there may be a primary cellular change in re- sponsiveness to hormones. A detailed study, using modern methods (e.g. Ruokonen et a1.25), of changes in steroid metabolism by prostatic cells during aging may give useful information.

Race and Prostatic Cancer There are two other factors involved in

prostatic carcinogenesis-age and race. These have been discussed in detail.17930J3 The origi- nal papers should be consulted.

Even taking possible sources of error into account there seem to be remarkable racial and geographical differences in incidence. It has been recognized for over 40 years that the clinical disease was very rare in the yellow races.27 Mortality and morbidity data are simi- lar. The only consistent finding is the low in- cidence in Mongoloid races but there is a re- markable variation in incidence in other races. The highest rate (age adjusted morbid- ity rate) is that for Negroes in one particular area of the United States (Alameda County) of 65.3, while the Caucasians in the same area have a rate of 38.0. The regional variations are even more striking, varying from 17.1 for non-Latin Texans in 37.8 for urban inhabi- tants of Iowa. The extremes in Canada range from 17.0 in Newfoundland to 39 in Saskatch- ewan. In South and Central America, Chile has a rate of 11.3 and Colombia 23.3. Other continents show a similar variation. In Africa the rate varies from 4.4 in Uganda to 29.1 in Rhodesia. There are apparently inexplicable variations such as the rate for Bantu in Cape

Province of 19.2 as compared with 9.4 for Bantu in Johannesburg. The rate in Indians (6.5 in Bombay, 9.4 in Natal) is low, as is that in Eastern Europe (e.g. Poland, 4.6-12.8) and Israel (3.1 for non-Jews to 10.8-13.2 for Jews.) New Zealand has a consistently high rate40.00 for Europeans and 40.3 for Maoris. The Japanese have a consistently low rate (3.2-4.3). There are no comparable fig- ures for China but the rate for Chinese in Singapore is 0.9.

Significance of diferences in racial dis- tribution: It is difficult to place any reli- ance on minor differences but the incidence in Mongoloid races is so greatly and consistently lower than in other groups that it can be con- fidently accepted, even though more critical studies may introduce minor variations in de- tail. I t seems likely too that the high incidence in some groups of American Negroes and in New Zealand can probably be accepted. The significance of lesser differences, e.g. between Eastern and Western Europe, is less certain. If the differences are due to environment they should be affected by migration. Unfortu- nately, results available at present are confusing.16 Prostate cancer incidence in Japa- nese immigrants to the U.S. rose but to no- where near the incidence in Caucasians. There are no figures as yet for descendants of Japa- nese migrants. The original paper should be consulted for details and for a discussion of the influence of an “imported environment.”

Other migrants from low incidence areas, particularly from Eastern Europe, also show a rise in incidence.30 This American experience with migrants and the high incidence in two dissimilar racial groups in New Zealand sug- gests that possible environmental factors may be concerned. Yet other figures, e.g. from Hawaii, where significant environmental fac- tors appear to vary little between the races, give a widely varying incidence between the races.

The Effects of Age on Tumor Incidence The age-associated incidence of prostatic

cancer is established beyond any doubt. It is rare before the age of 50, after which time the incidence increases rapidly until the age of 80. The rate of increase then seems to slow.

Ashley3 has shown that there is a somewhat similar rate of increase for latent cancer, using 4 series of cases; 1 from Germany,14 2 from the United Kingdom,l-s and 1 from Texas.16 When the data were plotted on a double loga-

1094 CANCER November 1973 VOl. 82

rithmic scale there was a straight line relation- ship between frequency and age, the fre- quency varying with the 3rd power of the age. When a similar plot was made for clinical can- cers a similar relationship was found but the slope was steeper, corresponding to the 7th power of the age. According to the “multiple hit” theory of carcinogenesis originally pro- posed by Armitage and Doll2 Ashley suggests that this supports the idea that latent cancer is the result of a smaller number of hits than clinical cancer.

The significance of age and racial in- cidence: The presence of latent cancers even in low incidence groups suggests that, accept- ing the multi-stage hypothesis of carcinogene- sis, the initiation stage may occur commonly. As Doll suggests,O the increasing frequency with age may be due to the increased exposure to a hypothetical carcinogen, or it may be a direct or indirect consequence of the process of aging. The further development of the neo- plastic process may then depend on promoting factors which may be environmental or ge- netic. The fact that other hormone-related cancers such as breast, ovary, and endome- trium are also low in Japan,30 does suggest that there may be a possible common genetic basis. This again may be due to differences in the endocrine environment, i.e. hormone se- cretion pattern, or to differences in cellular re- sponsiveness. Both are well known phenom- ena in different strains of experimental ani- mals. Although differences in endocrine pat- tern have been reported in the Japanese, there is little information about other low incidence groups.

LOCAL FACTORS AFFECTING TUMOR PRODUCTION

So far all the factors I have discussed have been general factors which influence all cells in the gland. But since cancer is a focal dis- turbance we must also consider cellular factors which influence the development of a tumor in one localized area although admittedly in cells which may have been altered by general factors. Again, we have no real knowledge of possible localizing factors, but there is a ~uggestion21~~ that a type of atrophy associated with focal fibrosis may be followed by a pre- cancerous hyperplasia. There is little evidence to incriminate other localizing factors, al- though many have been suggested.

THE PATHOLOGY OF PROSTATIC CANCER

There are many reviews on the pathology of prostatic cancer and if the references I give are to those of my own it is not necessarily be- cause they are better than the others but be- cause I am more familiar with them.9.12 The basic facts are well known and I shall make no attempt to go into detail but only indicate three areas of interest. I shall not discuss the effects of treatment since these will be consid- ered later.

Site of Origin and Its Consequences

There are two main groups of prostatic glands-inner and outer. Benign enlargement of the prostate arises from the inner glands. Prostatic cancer arises from the outer glands. There are two main consequences of this. The first is that the disease is very difficult- impossible perhaps-to diagnose in its early stages. I t may be present for a considerable time before it involves any structure which may cause symptoms. The second consequence is that invasion of the capsule is common and early and the tumor cells soon involve peri- neural lymphatics and blood vessels in the periprostatic tissues. The extent of local spread is almost invariably found to be greater than the clinical assessment suggests. Distant spread is equally common and almost invariable in patients dying of the disease. Perhaps one of the best indications is the fre- quency with which prostatic cancer cells can be found in bone marrow punctures especially in the late stages of the disease.22 Even in la- tent cancers vascular and lymphatic involve- ment is very common.

Biological Activity of Prostatic Cancer

The only point to be stressed here is that there are two types of prostatic cancer which are morphologically indistinguishable yet dif- fer in the biological behavior. One behaves in the same way as any other clinical cancer, the other remains inactive or latent. Some cases of prostatic cancer diagnosed clinically may be of this biologically latent type. As might be ex- pected these are often discovered accidentally, e.g. after subtotal prostatectomy for benign enlargement. The fact that cancers of this type have a long survival period must always be remembered, particularly when assessing the results of treatment.

No. 5 PROSTATIC CANCER - Frun ks 1095

Factors Influencing Prognosis that grading based on cytologic criteria might be more valuable than that based on histo- This remains an unsolved problem. We

have found10 that there are few reliable prog- logic appearances. The indisputable features nostic features. Grading was not of value in associated with a Poor pr%nosis are pnetra- our series, although others find it useful.26gm tion of the capsule,6.*8 and the presence of

Vickery and Kerr28 and Esposti7 suggest tumor cells in marrow bi0psies.1~

REFERENCES

1. Andrews, G. S.: Latent carcinoma of the prostate. J . Clin. Pathol. 2:197-208, 1949.

2. Armitage, P., and Doll, R.: A two-stage theory of carcinogenesis in relation to the age distribution of human cancer. Br. J . Cancer 11:161-169.1957.

3. Ashley, D. J. B.: On the incidence of carcinoma of the prostate. J. Pathol. 90:217-224, 1965.

4. Bulbrook, R. D., Franks, L. M., and Greenwood, F. C.: Hormone excretion in prostatic cancer-The early and late effects of endocrine treatment on urinary oestrogens, 17-ketosteroids and 17-ketogenic steroids. Acta Endocrinol. 31:481-499, 1959.

5. Byar, D. P., and Mostofi, F. K.: Carcinogen of the prostate-Prognostic features in 208 radical prostatec- tomies. Cancer 305-13, 1972.

6. Doll, R.: The age distribution of cancer in man. In Thule International Symposia-Cancer and Ageing. Stockholm, Nordiska Bokhandelns Forlag, 1968; p. 15.

7. Esposti, P. L.: Cytologic malignancy grading of prostatic carcinoma by transrectal aspiration biopsy. S c ~ n d . J. Urol. Nephrol. 5,199-209,1971.

8. Franks, L. M.: Atrophy and hyperplasia in the prostate proper. J. P ~ t h o l . 68:617-621,1954.

9. Franks, L. M.: Latent carcinoma of the prostate. 3. P ~ t h ~ l . 68:603-616, 1954.

10. Franks, L. M., Fergusson, J. D., and Murbaghan, G. F.: An assessment of factors influencing survival in prostatic cancer-The absence of reliable prognostic features. Br. J . Cancer 12:321-326, 1958.

11. Franks, L. M.: Some comments on the long-term results of endocrine treatment of prostatic cancer. Br. J . Urol. 30:383-388, 1958.

12. Franks, L. M.: Recent research on prostatic pa- thology. In Pathology Annual, s. C. Sommers, Ed. New York, Appleton-Century-Crofts, 1967.

13. Franks, L. M.: The incidence of carcinoma of prostate-An epidemiological survey. I n International Symposium on Current Problems in the Epidemiology of Cancer Lymphomas and Leukemias. Dusseldorf, Springer-Verlag, 1971.

14. Gaynor, E. P.: Zur Frage des Prostatakrebses. Virchows Arch. [Pathol. Anat.] 301:602-652.

15. Haenszel, W., and Kurihara, M.: Studies of Japa- nese migrant-I. Mortality from cancer and other dis- eases among Japanese in the United States. J . Natl. Cancer Inst. 4043-68, 1968.

16. Halpert, B., Sheenan, E. E., Schmalhorst, W. R., and Scott, R.: Carcinoma of the prostate-A survey of 5,000 autopsies. Cancer 16:737-742, 1963. €7. King, H., Diamond, E., and Lilienfeld, A. M.:

Some epidemiological aspects of cancer of the prostate. J . Chronic Dis. 16:117-153, 1963.

18. Marmorston, J., Lombardo, L. J.. Myers, S. M., Grierson, H., Stern, E., and Hopkins, C. E.: Urinary excretion of neutral 17-ketosteroids and pregnanediol by patients with prostatic cancer and benign prostatic hypertrophy. J. Urol. 93:276-295, 1965.

19. Mehan, D. J., Broun, G. O., Hoover, B., and Sto- rey, G.: Bone marrow findings in carcinoma of the prostate. J. Urol. 95241-244, 1966.

20. Moore, R. A.: In Edncodrinology of Neoplastic Disease. New York, Oxford University Press, 1947; p. 194.

21. Moore, R. A.: The evolution and involution of the prostate gland. A m . J . Pathol. 12599-624, 1956.

22. Nelson, C. M. K., Boatman, D. L., and Flocks, R. H.: Bone marrow examination in carcinama of the prostate. J. Urol. 1973 (in press). 23. Reynoso, G., and Murphy, G . P.: Adrenalectomy

and hypophvsectomy in advanced prostatic cancer. Cancer 29:94i-945, 1972.

24. Robinson, M. R. G., and Thomas, B. S.: Effect of hormonal therapy on plasma testosterone levels in prostatic carcinoma. Br. Med. J . 4:391-894,1971.

25. Ruokonen, A., Laatikainen, T., Laitinen, E. A., and Vihko, R.: Free and sulfate-conjugated neutral steroids in human testis tissue. Biochem. 11:1411-1416, 1972.

26. Shelley, H. S., Auerbach, S. H., Classen, K. L.. Marks, C. H., and Wiederanders, R. E.: Carcinoma of the prostate-A new system of classification. Arch. Surg. 77:751-756, 1958.

27. Steiner, P.: Cancer-Race and Geography. Balti- more, Williams and Wilkins, 1954.

28. Vickery, A. L., Jr.. and Kerr, W. S., Jr.: Cancer of prostate treated by radical prostatectomy-A clinico- pathological survey of 187 cases followed for 10 years. Cancer 16:1598-1608, 1963.

29. Wiederanders, R. E., Stuber, R. U., Mota, C., O’Connell, D., and Haslam, G. J.: Prognostic value of grading prostatic carcinoma. 3. Urol. 89881-888. 1963.

30. Wynder, E .L., Mabuchi, K., and Whitmore, W. F.: Epidemiology of cancer of the prostate. Cancer 28: 344-360, 1971.