- Email: [email protected]
THE HODGKIN MAZE
469 I believe that ambulance staff should be told to bring of carbon-monoxide poisoning straight to one of the above hospitals. On several occasions there has been more than a few hours delay while doctors less dogged, patient, and persistent than Dr. Snashall have tried to locate the whereabouts of the " closely guarded secret " he refers to. Westminster Hospital, London S.W.1. JOHN THURSTON. cases
THE HODGKIN MAZE SiR,ŇIn your editorial with this title1 you refer to the many uncertainties surrounding Hodgkin’s disease, and to MacMahon’s proposal 2,3 that it is not a single disease, but a combination of at least two and perhaps three aetiologically distinct entities. Arguing, like MacMahon,2,3 from an analysis of the sex and age distributions of death-rates, I concludedthat at least three distinctive groups are predisposed to the disease. However, the data then available to me were inadequate to define the sex and age pattern below the age of 15 years. Recently, Fraumeni and Lihave published the results of an epidemiologic studyof Hodgkin’s disease in childhood. The number of affected boys recorded in theirmulti-hospital survey, and the details of age at diagnosis, are sufficient to indicate the dependence of onset on age, during the early years (see accompanying figure). Allowing for a contribution (curve B) from a group with peak death-rates at about 30 to 34 years of age,3we see that age-specific initiationrates (dp/dt) in the predominantly childhood group (curve A) can be described by the stochastic equation:
made
(I have a latent-period correction of 1 year to allow for the average interval between the initiation of the disease process and diagnosis.) After correcting for latent period, death-rates in England and Wales, 1950-56for group B, with peak initiation at 28 years of age, are described by 4: For the late-onset group, with peak initiation in both sexes about 65 years of age, initiation-rates in England and Wales, 1950-56, are given by 4:e
at
Fraumeni and Li’s data 5 for childhood Hodgkin’s disease in multi-hospital series in the United States.
a
Age-specific diagnostic rates for boys in the hospital series are normalised to U.S. national age-specific death-rates, 1960-64, at the flat portion of the curve, from 5 to 9 years. Numbers of cases are shown against points. Ordinate: age-specific initiation-rates (dPjdt), allowing an average latent period of 1 year between initiation and diagnosis. Abscissa: estimated age (t) at initiation. (Log-log scales.) Curve A is based on the equation: dp/dt= 2k1 S1 t exp(-k lt2). Curve B is based on the equation: dp/dt= 3k2 Sz t2 exp(-k2t3)-see ref. 4. the different genotypes. Numerous instances of the familial aggregation of Hodgkin’s disease have been reported.22 Using Fraumeni and Li’s5 data for the U.S. population, 1960-64, I calculate that the value of sl,M-that is, the proportion of males predisposed at birth to the childhood form of Hodgkin’s disease-is about 1.3x10-5; it is similar in Whites and non-Whites. s1,F (for girls) is about 4 x 10-6. For England and Wales, the following values have been calculated4 from the Registrar General’ss statistics for 195056 : S2,M≃7.5x10—4; S2,F≃3.5x10—4; S3,M≃1.6x10—3; and Sg,F=8-7 x 10-4. From the epidemiological studies 2,3,5,6 we can conclude that the values of s differ quite markedly from country to country, and also from State to State 3 within the United States. The frequencies of predisposing genes are, perhaps, mainly responsible for the observed differences in the death-rates. Because the age-patterns give a good fit to autoaggressive statistics ", and because Hodgkin’s disease usually presents with multiple affected lymph-nodes,lO,l1 I have proposed4 that each form of the disease is indeed autoaggressive in character. The age-statistics suggest that a single mutant cell initiates the disease process in each of the three genotypes. If the mutant stem-cell is located centrally, and if the mutant products of the daughter cells in the forbidden clone attack peripheral target-cells, a multicentric origin can be readily explained.4,12 The location of lymph-nodes containing target-cells that are complementary to the mutant products of the forbidden clone is likely to be determined by genetic factors.4 The proliferation of lymphoid cells in Hodgkin’s disease could result from an indirect mechanism, and the primary autoaggressive attack might be directed against reticuloendothelial cells that normally regulate the release of lymphocytes from the lymph-nodes.4,12 Probably, Reed-Sternberg cells represent the damaged target-cells of the primary autoaggressive attack. Clinical evidence agrees well with "
equations derive from a simple model of spontaneous autoaggressive disease, 4,7 which states that such disorders are initiated by random somatic mutations in stemcells of the central system that controls growth. Equation (1) assumes that two specific somatic mutations have to occur in a single stem-cell; equation (2) assumes that three, and equation (3) assumes that six such events have to occur, in a genetically-predisposed person, to initiate the growth of a All the
" forbidden clone " a>9 of cells from the mutant stem-cell. The values of the kinetic constants that describe the formation of forbidden clones in males and females are: k1 1.1 x 10-2 yr -2 ; k2 - 3 x lO-5yr 3j and k3 10-llyr-6. Constants, with appropriate subscripts, define the proportion of the population, by sex, predisposed to each form of the disease. sl, S2, and Sg (by sex) represent genetically distinctive groups: in each group predisposition is probably polygenic, involving sex-linked as well as autosomal factors. Certain predisposing alleles might or might not be shared by 1. 2. 3. 4.
Lancet, 1969, ii, 728. MacMahon, B. Cancer Res. 1966, 26, 1189. Cole, P., MacMahon, B., Aisenberg, A. Lancet, 1968, ii, 1371. Burch, P. R. J. An Inquiry Concerning Growth, Disease and Ageing. Edinburgh, 1968. 5. Fraumeni, J. F., Li, F. P. J. natn. Cancer Inst. 1969, 42, 681. 6. Registrar General’s Statistical Review of England and Wales for the Year 1956; part III. H.M. Stationery Office, 1958. 7. Burch, P. R. J., Burwell, R. G. Q. Rev. Biol. 1965, 40, 252. 8. Burnet, F. M. The Clonal Selection Theory of Acquired Immunity. London, 1959. 9. Burnet, F. M. Cellular Immunology. London, 1969.
this general theory.4 Dr. Sirtori (Jan. 10, p. 92) suggests that viruses might be implicated in the pathogenesis of Hodgkin’s disease, and that their mode of action might be subtle and complex ". I have put forward a new theory of the mode of action of "
10. 11. 12.
Rosenberg, S. A., Kaplan, H. S. Cancer Res. 1966, 26, 1225. Smithers, D. W. Br. med. J. 1967, ii, 263. Burch, P. R. J. Nature, Lond. 1970, 225, 512.
470
viruses in classic infectious diseases: the same theory could well apply to certain malignant diseases.4,12 From their age-distributions, I conclude that various infectious diseases are basically autoaggressive in character, but that they lie at one end of a spectrum of types of disease. At this end, the forbidden clone only manifests itself when an invading micro-organism competes specifically for endogenous defence resources-e.g., immunoglobulin antibodieswhich are directed against the mutant, and therefore " notself ", forbidden clone. However, epidemiological evidence 2,5 offers little support for the idea that viruses have a causal role in Hodgkin’s disease. The General Infirmary, P. R. J. BURGH. Leeds 1.
CHROMOSOME AND D.N.A. ABNORMALITIES IN OVARIAN CYSTADENOMAS SIR,-Two recent chromosome studies have suggested the presence of aneuploid cell-lines in cystadenomas of the ovary.13.H We have measured the D.N.A. contents of epithelial cells in smears from a cystadenoma: the distribution of values shows a mode which corresponds to a chromosome number of about 70. The lesion was a lobulated mucinous multicystic tumour 15 cm. in diameter in a patient aged 70. Although histologically there was no evidence of malignancy, it was noted that most of the epithelial nuclei were large and contained two sexchromatin bodies (clearly seen in orcein-stained squash
preparations). We feel that D.N.A. estimations on interphase cells may be of great value in demonstrating aneuploidy where there are few mitoses or, as in carcmoma-in-situ of the cervix uteri/5 where only small amounts of material are available for study. A distinction cannot of course be made, by this technique, between an aneuploid cell-line showing only small changes in D.N.A. content and cells having a normal (euploid) karyotype. Where, however, the majority of cells are aggregated around a modal value which is significantly different from the diploid (or tetraploid) value, as in many carcinomas-in-situ of the cervix,15 aneuploidy is clearly present and the abnormal cells probably constitute a single clone. We have now studied the D.N.A. contents of 7 cystadenomas. (The work will be reported elsewhere together with D.N.A. and chromosome findings on a series of ovarian carcinomas.) In polyps of the colon and rectum, chromosome and D.N.A. studies have shown that aneuploidy is frequently present, with small deviations from diploidy or, less often, high (hypertriploid or hypotetraploid) modes 16; we suggest that aneuploidy is also commonly present in ovarian cystadenomas, the chromosome changes again usually being small in extent (the other 6 cystadenomas in our series had modal D.N.A. values in the diploid region). The D.N.A. contents of epithelial cells in histological sections of9 mucinous cystadenomas have been measured by Weiss et al. 17 Their distributions do not show very clear modes, presumably for technical reasons, and it appears that sections may generally not be as suitable as smears for demonstrating the presence of aneuploidy 15 ; confirmatory evidence of polyploidisation, which, as can be seen in the accompanying histogram, was very low in Fracarro, M., Mannini, A., Tiepolo, L., Gerli, M., Zara, C. Lancet, 1968, 1, 613. 14. Benedict, W. F., Rosen, W. C., Brown, C. D., Porter, I. H. ibid. 1969, ii, 640. 15. Atkin, N. B. Obstetl gynec. Surv. 1969, 24. 793. 16. Atkin, N. B. Proceedings of the 23rd Annual Symposium on Fundamental Cancer Research, University of Texas M.D. Andersen Hospital and Tumor Institute, March 5-7, 1969 (in 13.
17.
the press). R R., Richart, R. M., 1969. 103, 409.
Weiss,
Okagaki,
T.
Am. J. Obstet. Gynec.
Distribution of D.N.A. values of interphase epithelial cells in smear preparation from the ovarian cystadenoma. D =
diploid,
T
=
tetraploid. and
o
=
a
octonloid levels.
the lesion studied by us, may however be obtainable. Since the ovarian carcinomas that we have studied fall into two discrete groups having near-diploid and high modal D.N.A. values (or chromosome numbers) respectively, with a considerable difference in malignancy as shown by the poorer average degree of differentiation and worse prognosis of those in the high-value group, it is possible that cystadenomas having high modes also carry different clinical implications from those which are diploid or
near-diploid. Department of Cancer Research, Mount Vernon Hospital, Northwood, Middlesex.
N. B. ATKIN M. C. BAKER.
COMBINED SURGERY AND CHEMOTHERAPY FOR CARCINOMA OF THE BRONCHUS SiR,—Iwas interested to read of Mr. Dolton’s experience with oncolytic drugs combined with surgery (Jan. 3, p. 40), for we have, for some years, used a technique not dissimilar to his. However, at this stage the results in our cases show a considerable improvement over those for patients treated with surgery alone. We too have been using cyclophosphamide, in the following doses: on the precperative day 100 mg.; on the day of operation 1 g., followed by 500 mgms. ; and on the first postoperative day and thereafter, 200 mg. daily for ten days. (Unlike Mr. Dolton, we always used the intravenous route.) We used a scalp tourniquet in all our 100 cases, and with this regimen have had no cases of alopecia. The total dosage is thus 50% higher than that used at Wolverhampton. White-cell counts have been carried out in all cases, and although some low readings worried us initially, this caused no
problem. We have been fortunate in this area in having a population that moves little, and we were able, without untoward difficulty, to follow up 100 consecutive cases treated in this manner and to compare them with a control series of 100 patients treated without cyclophosphamide. We are waiting till we have a 5-year follow-up of both groups, but, since we are preparing a preliminary communication on the subject, many of the figures are available now.
THE HODGKIN MAZE SiR,ŇIn your editorial with this title1 you refer to the many uncertainties surrounding Hodgkin’s disease, and to MacMahon’s proposal 2,3 that it is not a single disease, but a combination of at least two and perhaps three aetiologically distinct entities. Arguing, like MacMahon,2,3 from an analysis of the sex and age distributions of death-rates, I concludedthat at least three distinctive groups are predisposed to the disease. However, the data then available to me were inadequate to define the sex and age pattern below the age of 15 years. Recently, Fraumeni and Lihave published the results of an epidemiologic studyof Hodgkin’s disease in childhood. The number of affected boys recorded in theirmulti-hospital survey, and the details of age at diagnosis, are sufficient to indicate the dependence of onset on age, during the early years (see accompanying figure). Allowing for a contribution (curve B) from a group with peak death-rates at about 30 to 34 years of age,3we see that age-specific initiationrates (dp/dt) in the predominantly childhood group (curve A) can be described by the stochastic equation:
made
(I have a latent-period correction of 1 year to allow for the average interval between the initiation of the disease process and diagnosis.) After correcting for latent period, death-rates in England and Wales, 1950-56for group B, with peak initiation at 28 years of age, are described by 4: For the late-onset group, with peak initiation in both sexes about 65 years of age, initiation-rates in England and Wales, 1950-56, are given by 4:e
at
Fraumeni and Li’s data 5 for childhood Hodgkin’s disease in multi-hospital series in the United States.
a
Age-specific diagnostic rates for boys in the hospital series are normalised to U.S. national age-specific death-rates, 1960-64, at the flat portion of the curve, from 5 to 9 years. Numbers of cases are shown against points. Ordinate: age-specific initiation-rates (dPjdt), allowing an average latent period of 1 year between initiation and diagnosis. Abscissa: estimated age (t) at initiation. (Log-log scales.) Curve A is based on the equation: dp/dt= 2k1 S1 t exp(-k lt2). Curve B is based on the equation: dp/dt= 3k2 Sz t2 exp(-k2t3)-see ref. 4. the different genotypes. Numerous instances of the familial aggregation of Hodgkin’s disease have been reported.22 Using Fraumeni and Li’s5 data for the U.S. population, 1960-64, I calculate that the value of sl,M-that is, the proportion of males predisposed at birth to the childhood form of Hodgkin’s disease-is about 1.3x10-5; it is similar in Whites and non-Whites. s1,F (for girls) is about 4 x 10-6. For England and Wales, the following values have been calculated4 from the Registrar General’ss statistics for 195056 : S2,M≃7.5x10—4; S2,F≃3.5x10—4; S3,M≃1.6x10—3; and Sg,F=8-7 x 10-4. From the epidemiological studies 2,3,5,6 we can conclude that the values of s differ quite markedly from country to country, and also from State to State 3 within the United States. The frequencies of predisposing genes are, perhaps, mainly responsible for the observed differences in the death-rates. Because the age-patterns give a good fit to autoaggressive statistics ", and because Hodgkin’s disease usually presents with multiple affected lymph-nodes,lO,l1 I have proposed4 that each form of the disease is indeed autoaggressive in character. The age-statistics suggest that a single mutant cell initiates the disease process in each of the three genotypes. If the mutant stem-cell is located centrally, and if the mutant products of the daughter cells in the forbidden clone attack peripheral target-cells, a multicentric origin can be readily explained.4,12 The location of lymph-nodes containing target-cells that are complementary to the mutant products of the forbidden clone is likely to be determined by genetic factors.4 The proliferation of lymphoid cells in Hodgkin’s disease could result from an indirect mechanism, and the primary autoaggressive attack might be directed against reticuloendothelial cells that normally regulate the release of lymphocytes from the lymph-nodes.4,12 Probably, Reed-Sternberg cells represent the damaged target-cells of the primary autoaggressive attack. Clinical evidence agrees well with "
equations derive from a simple model of spontaneous autoaggressive disease, 4,7 which states that such disorders are initiated by random somatic mutations in stemcells of the central system that controls growth. Equation (1) assumes that two specific somatic mutations have to occur in a single stem-cell; equation (2) assumes that three, and equation (3) assumes that six such events have to occur, in a genetically-predisposed person, to initiate the growth of a All the
" forbidden clone " a>9 of cells from the mutant stem-cell. The values of the kinetic constants that describe the formation of forbidden clones in males and females are: k1 1.1 x 10-2 yr -2 ; k2 - 3 x lO-5yr 3j and k3 10-llyr-6. Constants, with appropriate subscripts, define the proportion of the population, by sex, predisposed to each form of the disease. sl, S2, and Sg (by sex) represent genetically distinctive groups: in each group predisposition is probably polygenic, involving sex-linked as well as autosomal factors. Certain predisposing alleles might or might not be shared by 1. 2. 3. 4.
Lancet, 1969, ii, 728. MacMahon, B. Cancer Res. 1966, 26, 1189. Cole, P., MacMahon, B., Aisenberg, A. Lancet, 1968, ii, 1371. Burch, P. R. J. An Inquiry Concerning Growth, Disease and Ageing. Edinburgh, 1968. 5. Fraumeni, J. F., Li, F. P. J. natn. Cancer Inst. 1969, 42, 681. 6. Registrar General’s Statistical Review of England and Wales for the Year 1956; part III. H.M. Stationery Office, 1958. 7. Burch, P. R. J., Burwell, R. G. Q. Rev. Biol. 1965, 40, 252. 8. Burnet, F. M. The Clonal Selection Theory of Acquired Immunity. London, 1959. 9. Burnet, F. M. Cellular Immunology. London, 1969.
this general theory.4 Dr. Sirtori (Jan. 10, p. 92) suggests that viruses might be implicated in the pathogenesis of Hodgkin’s disease, and that their mode of action might be subtle and complex ". I have put forward a new theory of the mode of action of "
10. 11. 12.
Rosenberg, S. A., Kaplan, H. S. Cancer Res. 1966, 26, 1225. Smithers, D. W. Br. med. J. 1967, ii, 263. Burch, P. R. J. Nature, Lond. 1970, 225, 512.
470
viruses in classic infectious diseases: the same theory could well apply to certain malignant diseases.4,12 From their age-distributions, I conclude that various infectious diseases are basically autoaggressive in character, but that they lie at one end of a spectrum of types of disease. At this end, the forbidden clone only manifests itself when an invading micro-organism competes specifically for endogenous defence resources-e.g., immunoglobulin antibodieswhich are directed against the mutant, and therefore " notself ", forbidden clone. However, epidemiological evidence 2,5 offers little support for the idea that viruses have a causal role in Hodgkin’s disease. The General Infirmary, P. R. J. BURGH. Leeds 1.
CHROMOSOME AND D.N.A. ABNORMALITIES IN OVARIAN CYSTADENOMAS SIR,-Two recent chromosome studies have suggested the presence of aneuploid cell-lines in cystadenomas of the ovary.13.H We have measured the D.N.A. contents of epithelial cells in smears from a cystadenoma: the distribution of values shows a mode which corresponds to a chromosome number of about 70. The lesion was a lobulated mucinous multicystic tumour 15 cm. in diameter in a patient aged 70. Although histologically there was no evidence of malignancy, it was noted that most of the epithelial nuclei were large and contained two sexchromatin bodies (clearly seen in orcein-stained squash
preparations). We feel that D.N.A. estimations on interphase cells may be of great value in demonstrating aneuploidy where there are few mitoses or, as in carcmoma-in-situ of the cervix uteri/5 where only small amounts of material are available for study. A distinction cannot of course be made, by this technique, between an aneuploid cell-line showing only small changes in D.N.A. content and cells having a normal (euploid) karyotype. Where, however, the majority of cells are aggregated around a modal value which is significantly different from the diploid (or tetraploid) value, as in many carcinomas-in-situ of the cervix,15 aneuploidy is clearly present and the abnormal cells probably constitute a single clone. We have now studied the D.N.A. contents of 7 cystadenomas. (The work will be reported elsewhere together with D.N.A. and chromosome findings on a series of ovarian carcinomas.) In polyps of the colon and rectum, chromosome and D.N.A. studies have shown that aneuploidy is frequently present, with small deviations from diploidy or, less often, high (hypertriploid or hypotetraploid) modes 16; we suggest that aneuploidy is also commonly present in ovarian cystadenomas, the chromosome changes again usually being small in extent (the other 6 cystadenomas in our series had modal D.N.A. values in the diploid region). The D.N.A. contents of epithelial cells in histological sections of9 mucinous cystadenomas have been measured by Weiss et al. 17 Their distributions do not show very clear modes, presumably for technical reasons, and it appears that sections may generally not be as suitable as smears for demonstrating the presence of aneuploidy 15 ; confirmatory evidence of polyploidisation, which, as can be seen in the accompanying histogram, was very low in Fracarro, M., Mannini, A., Tiepolo, L., Gerli, M., Zara, C. Lancet, 1968, 1, 613. 14. Benedict, W. F., Rosen, W. C., Brown, C. D., Porter, I. H. ibid. 1969, ii, 640. 15. Atkin, N. B. Obstetl gynec. Surv. 1969, 24. 793. 16. Atkin, N. B. Proceedings of the 23rd Annual Symposium on Fundamental Cancer Research, University of Texas M.D. Andersen Hospital and Tumor Institute, March 5-7, 1969 (in 13.
17.
the press). R R., Richart, R. M., 1969. 103, 409.
Weiss,
Okagaki,
T.
Am. J. Obstet. Gynec.
Distribution of D.N.A. values of interphase epithelial cells in smear preparation from the ovarian cystadenoma. D =
diploid,
T
=
tetraploid. and
o
=
a
octonloid levels.
the lesion studied by us, may however be obtainable. Since the ovarian carcinomas that we have studied fall into two discrete groups having near-diploid and high modal D.N.A. values (or chromosome numbers) respectively, with a considerable difference in malignancy as shown by the poorer average degree of differentiation and worse prognosis of those in the high-value group, it is possible that cystadenomas having high modes also carry different clinical implications from those which are diploid or
near-diploid. Department of Cancer Research, Mount Vernon Hospital, Northwood, Middlesex.
N. B. ATKIN M. C. BAKER.
COMBINED SURGERY AND CHEMOTHERAPY FOR CARCINOMA OF THE BRONCHUS SiR,—Iwas interested to read of Mr. Dolton’s experience with oncolytic drugs combined with surgery (Jan. 3, p. 40), for we have, for some years, used a technique not dissimilar to his. However, at this stage the results in our cases show a considerable improvement over those for patients treated with surgery alone. We too have been using cyclophosphamide, in the following doses: on the precperative day 100 mg.; on the day of operation 1 g., followed by 500 mgms. ; and on the first postoperative day and thereafter, 200 mg. daily for ten days. (Unlike Mr. Dolton, we always used the intravenous route.) We used a scalp tourniquet in all our 100 cases, and with this regimen have had no cases of alopecia. The total dosage is thus 50% higher than that used at Wolverhampton. White-cell counts have been carried out in all cases, and although some low readings worried us initially, this caused no
problem. We have been fortunate in this area in having a population that moves little, and we were able, without untoward difficulty, to follow up 100 consecutive cases treated in this manner and to compare them with a control series of 100 patients treated without cyclophosphamide. We are waiting till we have a 5-year follow-up of both groups, but, since we are preparing a preliminary communication on the subject, many of the figures are available now.