ANTICANCER RESEARCH 24: 3115-3120 (2004)

Hypertension and Risk of from Small Cell Lung : A Retrospective Follow-up Study

ABE E. SAHMOUN1, L. DOUGLAS CASE2, SUDHIR CHAVOUR1, SOHAIL KAREEM1 and GARY G. SCHWARTZ2

1Department of Medicine, University of North Dakota School of Medicine, Fargo, ND 58102; 1Department of Public Health Sciences and Cancer Biology, Wake Forest University School of Medicine, Winston-Salem, NC 27157, U.S.A.

Abstract. Background: Although metastatic brain cancer is one Brain metastases (BM) are the most common intracranial of the most common forms of cancer, little is known about the tumor, significantly outnumbering primary brain tumors (1- factors associated with its development. We examined the 3). In men, primary tumors of the lung are the most hypothesis that hypertension may increase the risk for brain common cause of brain metastasis and account for metastasis (BM) in patients with primary small cell approximately 30% to 60% of all brain metastases (4-13). (SCLC). Materials and Methods: A retrospective review of Patients with small cell lung cancer (SCLC) develop brain medical charts of patients diagnosed with SCLC between June, metastases at an extremely high rate (14). At the time of 1986 and June, 2003 at MeritCare in Fargo, ND, USA, was done diagnosis of SCLC, 10% of patients present with brain to determine which of these patients subsequently developed brain metastases. This figure rises to 20% during therapy (15-20). metastases. The effects of hypertension, age, gender, body mass The actuarial probability of developing brain metastases index and the site of SCLC on the risk of developing BM were from SCLC increases with length of survival and reaches examined using both univariate and multivariable Cox 50% to 80% at two years from diagnosis (21, 22). proportional-hazards regression models. Two-way interactions Regardless of treatment, BM are associated with a poor between hypertension and other covariates were also included in prognosis. Untreated patients have a median survival of only the analyses. Results: Two hundred and thirty-two patients were about 1 month (23), and nearly all untreated patients die as identified with SCLC and 185 patients were eligible for this study. a direct result of the . Treatment of BM usually Eighty-five (45.9%) patients developed BM. Over 54% of SCLC involves whole brain radiotherapy. A recent phase III study occurred in the right lobe and more than 70% of the patients by Postmus et al. (24) found that the combination of whole with BM had them in multiple locations. The risk of BM is brain radiotherapy and teniposide results in a median significantly higher in younger patients (p-value < 0.03). survival of 3.5 months. Univariate analysis showed a hazard ratio (HR) for hypertension Metastases spread to the brain via arterial microemboli (25). of 1.01 (95%Confidence Interval (CI) 0.6-1.6) for BM from This suggests that mechanical factors such as increased blood SCLC. The multivariable Cox model showed an adjusted HR for flow or blood pressure may cause more microemboli to hypertension of 1.06 (95%CI 0.7-1.6) for BM from SCLC. dislodge and eventually be trapped within small cerebral Conclusion: As has been consistently observed for other lung vessels. According to the hemodynamic or mechanical theory , SCLC is more common in the right lung. The higher of metastasis (26), the major factor determining the frequency incidence of BM in younger patients suggests that more aggressive of organ involvement in metastasis is the number of cancer therapy is needed in these patients. Hypertension does not appear cells delivered to target organs in their arterial blood supply. to increase the risk of BM from SCLC. Thus, metastatic frequency should correlate with some parameter of target organ arterial blood flow. In a recent test of this theory using human autopsy data (27), a significant correlation was demonstrated between documented arterial Correspondence to: Abe E. Sahmoun, Ph.D., M.S., University of blood flow (ml/min/g target organ) and metastatic frequency. North Dakota School of Medicine, Department of Medicine, 1919 Elm Street North, Fargo, ND 58102, U.S.A. Tel: (701) 293-4131, Virdis et al. (28) noted an apparent association between Fax: (701) 293-4145, e-mail: [email protected] hypertension and primary cerebral cancers. We examined whether the risk for BM in patients with SCLC is increased Key Words: Metastasis, brain, small cell lung cancer, hypertension. in hypertensive patients. The rationale for this hypothesis is

0250-7005/2004 $2.00+.40 3115 ANTICANCER RESEARCH 24: 3115-3120 (2004) that cerebral metastases are found along the paths of Table I. Small cell lung cancer by demographic characteristics. arterial supply to the brain. Thus, factors which increase Total (N=185) cerebral blood flow or velocity, or damage the arterial wall may increase the incidence of BM. Age 67 (44-89)

Materials and Methods Gender Female 55 (30%) A retrospective analysis of charts of patients diagnosed with SCLC Male 130 (70%) between June 1986 and June 2003 were identified from the cancer registry of MeritCare, Fargo, ND, USA. MeritCare is a private Small cell lung cancer site institution that serves Fargo-Moorhead and surrounding areas. The Right lobe 101 (55%) study was approved by the Institutional Review Board. Left lobe 65 (35%) Lingula 4 (2%) Unknown 15 (8%) Study design. Data on age, gender, race, tumor site, body mass index, treatments, cigarette and alcohol use, blood pressure and Diastolic blood pressure 71 (46-110) brain metastases from patients with SCLC were abstracted using Systolic blood pressure 127 (78-190) electronic records and medical charts. Inclusion and exclusion criteria for this study are presented below: Blood pressure status Inclusion criteria: Normotensive 98 (53%) 1. Diagnosis of SCLC as a primary site using a pathology report Hypertensive 85 (46%) present in the medical records. Unknown 2 (1%) Exclusion criteria: 1. Diagnosis of any cancer other than primary SCLC; 2. Diagnosis of brain metastases as a secondary site from cancers Yes 99 (54%) other than SCLC; No 83 (45%) 3. Date of brain metastases is equal to or before the date of SCLC Unknown 3 (1%) diagnosis; 4. Patients who received a prophylactic cranial irradiation (29, 30). Non-steroidal anti-inflammatory Yes 19 (10%) Exposure definition. Hypertension, "high blood pressure", is defined No 162 (88%) as systolic blood pressure higher than 140 or diastolic blood Unknown 4 (2%) pressure over 90 (31). The review of patients’ medical charts showed that blood pressure was measured at each medical exam. Chronic kidney disease Three successive measures were read. We took the average of these Yes 5 (3%) measures at the time of SCLC diagnosis. If these average values No 177 (96%) were higher than 140 or 90, then the patient was considered Unknown 3 (1%) hypertensive. In addition, if the patient was on antihypertensive medication, he or she was also considered hypertensive. Body mass index at diagnosis Demographic covariates analyzed included age, gender, Normal (<25 Kg/m2) 91 (49%) hypertension and the site of SCLC. Factors such as obesity, Overweight (25+ Kg/m2) 77 (42%) corticosteroids use, non-steroidal anti-inflammatory use and Unknown 17 (9%) chronic kidney disease contribute to elevated blood pressure and may confound the association between hypertension and BM. Brain metastases status 85 (46%) Therefore, these factors were also included in our analyses. Most Single 13 (15%) patients were smokers (99%) and alcohol drinkers (97%). The Multiple 60 (71%) majority (96%) of the population in the Fargo-Moorhead area are Unknown 12 (14%) Caucasians, thus we limited our analysis to this group. Age was a continuous variable but since the relationship with BM was not Brain metastases site linear, we categorized age into the following groups (<64, and Pons 7 (10%) 65+). Patients discharged with BM (198.x) and SCLC (162.x) using Medulla 2 (3%) the International Classification of Diseases, Ninth Revision (32) Cerebrum 60 (82%) codes among any diagnosis field were considered to have 22 (30%) metastases.

Statistical analysis. The outcome variable was the time from the diagnosis of SCLC to the development of BM or last date of contact. Univariate and multivariable analysis were performed significant. All two-way interactions involving hypertension were using Cox proportional-hazards regression models (33). All assessed. Analyses were performed using SAS Software V8.0 (SAS p-values are two-sided, and p-values<0.05 are considered Institute, Cary, NC 25513,USA).

3116 Sahmoun et al: Hypertension and Brain Metastasis

Table II. Univariate and multivariable Cox proportional-hazards regression models of brain metastases from small cell lung cancer.

Variables N Events Median(range) HR* (95% CI) HR** (95% CI)

Age ≤ 64 78 49 48(30-56)*** 1.60(1.03-2.5) 1.59(1.03-2.5) 65+ 107 36 79(45-212) Ref Ref

Gender Male 130 54 49(37-86) Ref Ref Female 55 31 52(33-72) 1.07(0.7-1.7) 1.01(0.6-1.6)

Small cell lung cancer site Left lobe 65 27 51(34-112) Ref Ref Right lobe 101 52 48(34-72) 1.21(0.8-1.9) 1.11(0.7-1.8)

Blood pressure status Normotensive 98 49 50(36-90) Ref Ref Hypertensive 85 36 53(34-83) 1.01(0.6-1.6) 1.06(0.7-1.6)

Body mass index at diagnosis Normal (< 25 Kg/m2) 91 46 48(35-56) Ref Ref Overweight (25+ Kg/m2) 77 36 83(31-212) 0.67(0.4-1.05) 0.66(0.4-1.1)

* Univariate analysis. ** Multivariable analysis. *** Survival in weeks.

Results Table III. Distribution of blood pressure in patients diagnosed with primary SCLC (31).

We identified 232 patients with primary small cell lung Category N=185 cancer. Twelve patients (5.2%) received prophylactic cranial (%) irradiation and were excluded from this study. Of the 220 remaining patients, 24 (11%) presented with confirmed or Optimal suspected synchronous BM. Eleven patients had missing Systolic < 120 and diastolic < 80 30 (16) Pre-hypertension values and therefore these 35 patients were excluded. A total Systolic 120-139 or diastolic 80-89 68 (37) of 185 patients with SCLC were included in the analysis Stage 1 hypertension (Table I). The mean (±SD) age at diagnosis for these Systolic 140-159 or diastolic 90-99 70 (38) patients was 65.7±10.4. Women were diagnosed at a slightly Stage 2 hypertension earlier age than men (64.5±10.4 vs. 66.4±10.4). Most of the Systolic ≥ 160 or diastolic ≥ 100 15 (8) Missing 2 (1) patients (70.3%) were male, and the majority (52.9%) were normotensive. Eight-five (45.9%) patients developed BM. Forty-nine (57.6%) of these patients were in the younger age group and 36 (42.3%) patients in the older group (p- value<0.0001). The estimated median time to BM was 12 (9- 13) months. Over 70% of patients with BM had multiple were taking antihypertensive medication. Fifteen of these metastases and the cerebrum was the most (82.2%) patients were at stage 2 hypertension and 5 patients at stage metastatic site. The majority (54.6%) of SCLC occurred in 1 hypertension (Table III). There were no patients with the right lobe. normal blood pressure who were taking antihypertensive Fifty-eight (55%) patients with hypertension were in the medication. The duration between the diagnosis of primary younger age group (<65) and 27 (35%) patients in the older SCLC in patients with hypertension and BM was 10 months. age group. The distribution of patients with blood pressure Univariate analysis (unadjusted HRs) (Table II) showed at various ranges is presented in Table III. Twenty patients that patients with hypertension had a HR of 1.01 (95%CI

3117 ANTICANCER RESEARCH 24: 3115-3120 (2004)

0.6-1.6) for developing BM metastases from SCLC, Primary lung cancers were significantly more common in indicating that hypertension is unlikely to be a risk factor the right than in the left lobe of the lung. This finding has for BM. After adjusting for all the covariates, hypertension been consistently observed yet is poorly understood. For had a HR of 1.06 (95%CI 0.7-1.6) for developing BM example, in 1977, Parkash (43) described a marked excess metastases from SCLC, indicating that there is no of right vs. left lung cancers. Although the right lung is association between hypertension and BM (Table II). The slightly larger than the left, the excess in cancers on this side analysis of the interactions between the age by hypertension, is far greater than the small difference in lung mass. Parkash gender by hypertension, and primary site by hypertension speculated that the difference might reflect the easier access showed no statistical significance for these terms. of carcinogens to the right than the left lung, though About 3% of hypertensive patients may have high blood empirical support for this suggestion is scarce. pressure induced by corticosteroids (34). In our study, the The clinical and experimental literature on hypertension majority (98%) of patients received corticosteroids after the and metastases is contradictory. For example, Weiss and diagnosis of BM, thus we did not include this variable in our colleagues (27) have stressed the role of vascular factors in model. Non-steroidal anti-inflammatory drugs (NSAIDs) are favoring metastases to the brain and have cited both known to raise mean blood pressure by approximately 5 mm experimental and clinical observations to support this view. Hg (35) but only 20 patients were taking NSAIDs and only 4 Conversely, in an experimental study of the relationship patients showed this effect therefore it was also not included between hypertension and tumor growth and metastases in the model. Adjusting for body mass index, which was using transplanted Walker tumors in rats, Fisher et al. (44) measured at baseline, did not alter the HR for hypertension. showed that the incidence or size of metastases did not increase in tumor-bearing rats with hypertension, and tumor Discussion growth had no effect on blood pressure. The fact that hypertension does not seem to be a risk factor for metastases Small cell lung cancer is known to disseminate to the brain may be explained by the death of many of the cancer cells in a fairly large proportion of patients. Why this is the case before they reach the host site. For instance, Lang et al. (45) is not yet fully understood. The present investigation was showed that an important factor responsible for the death of undertaken to test the hypothesis that hypertension is a risk circulating cancer cells is their mechanical deformation while factor for brain metastases from SCLC. Our study found they are being squeezed through small blood vessels. that hypertension is unlikely to be a risk factor for brain Clinical observations of cancer patients and studies with metastases. experimental tumor systems have concluded that tumors In SCLC, the risk of developing brain metastatic disease is metastasize to specific organs independent of the rate of approximately 50% at 2 years, being as high as 80% in one blood flow, vascularity and number of tumor cells that reach series (36). In this study, the incidence of BM that occurred an organ. Tumor cells can reach the microvasculature of in patients was 48.6%. Our results confirm other findings on many organs, but extravasation into the organ parenchyma the relation between age and metastases frequency. King et and growth occur only in some organs (46-48). The al. (37), based on 100 autopsies of patients with metastatic potential and other properties of human tumors bronchogenic carcinoma, and Halpert et al. (38), based on depend on whether they are growing in orthotopic or 92 autopsies of similar patients reported an inverse ectopic sites in nude mice (49, 50), in part because of association between age and rate of nervous system differences in responses to organ-specific growth factors (51, metastasis. Galluzzi et al. (39) noted that the rate of 52), gradients of pH, oxygen, nutrients, and cell waste metastasis from bronchogenic carcinoma decreased with age products that develop within tumors (53, 54). and offered a number of possible explanations: 1) the This was a retrospective follow-up study conducted in a prevalence of squamous cell carcinoma increases with age private institution. It is likely that selection bias may be (although their own data indicate a declining rate of present as well as misclassification of the codification of the metastasis with age for all histological types of bronchogenic disease. It seems unlikely that such misclassification would cancer); and 2) older patients may die before there is be directional (i.e. patients who developed BM would be opportunity for the development of metastasis. A number of equally likely to have blood pressure accurately measured published series have a high percentage of bronchogenic as those without). However, nondirectional misclassification carcinomas as the cause of brain metastases with a could bias our results toward the null. It is possible that the predominance of males (40, 41). Ries (42) found that age is presence of occult metastases might have lowered the blood the strongest predictor of survival in lung cancer of the pressure in some patients with BM, leading to a bias away demographic variables. Our study found that women are from the null. Although this is possible in theory, our study diagnosed with SCLC at an earlier age than men, which is in gives little empirical support to this view. It is possible that accordance with other studies (55, 56). patients with controlled hypertension after its diagnosis are

3118 Sahmoun et al: Hypertension and Brain Metastasis no longer hypertensive. We re-analyzed our data excluding 15 Newman SJ and Hansen HH: Frequency, diagnosis, and all patients who were on antihypertensive medication. There treatment of brain metastases in 247 consecutive patients with was no association found between hypertension and BM. bronchogenic carcinoma. Cancer 33: 492-496, 1974. The information on the duration of the hypertension prior 16 Auerbach O et al: Histologic type of lung cancer in relation to smoking habits, year of diagnosis and sites of metastases. Chest to the time of the diagnosis of SCLC was not available in 67: 382-387, 1975. the patients’ medical chart. 17 Martini N et al: Oat cell carcinoma of the lung. Clin Bull 5: 144- In summary, this study confirmed a right-sided excess of 148, 1975. lung cancer and did not find evidence in support of the 18 Matthews MJ: Problems in morphology and behavior of hypothesis that hypertension is associated with an increased bronchopulmonary malignant disease. In: Lung Cancer, Natural risk of brain metastases from primary SCLC. The History, Prognosis and Therapy, Isreal L and Chahinian AP, prevalence of BM is significantly higher in younger patients, New York, Academic Press. 1976, pp 28-62. 19 Takita H et al: Small cell carcinoma of the lung: Clinicopathological which may suggest that more aggressive therapy might be studies. J Thor Cardiovasc Surg 66: 472-477, 1973. appropriate in these patients. 20 Watson WL and Berg JW: Oat cell lung cancer. Cancer 15: 759- 767, 1962. Acknowledgements 21 Nugent JL, Bunn PA Jr, Matthews MJ et al: CNS metastases in small cell bronchogenic carcinoma: Increasing frequency This research was supported by an intramural grant from the and changing pattern with lengthening survival. Cancer 44: University of North Dakota School of Medicine, USA. 1885-93, 1979. 22 Rosen S, Makuch R, Lichter A, Ihde D, Matthews M, Minna J, References Glatstein E and Bunn P: Role of prophylactic cranial irradiation in prevention of CNS metastases in SCLC. Am J 1 Posner JB and Chernik NL: Intracranial metastases from Med 74: 615-623, 1983. cancer. Adv Neurol 19: 579-592, 1978. 23 Markesbery WR, Brooks WH, Gupta GD et al: Treatment for 2 Walker AE, Robins M and Weinfeld FD: of patients with cerebral metastases. Arch Neurol 35: 754-756, 1978. brain tumors: The national survey of intracranial . 24 Postmus PE, Haaxma-Reiche H, Smit EF et al: Treatment of Neurology 35: 219-226, 1985. brain metastases of small-cell lung cancer: comparing 3 Wingo PA, Tong T and Bolden S: Cancer statistics. CA Cancer teniposide and teniposide with whole-brain radiotherapy--a J Clin 45: 8-30, 1995. phase III study of the European Organization for the Research 4 Takakura K, Sano K, Hoho S et al: Metastatic Tumors of the and Treatment of Cancer Lung Cancer Cooperative Group. J Central System Nervous. Tokyo: Igaku-Shoin, 1982. Clin Oncol 18(19): 3400-8, 2000. 5 Baker AB: Metastatic tumors of the nervous system. Arch 25 Delattre JY, Krol G, Thaler HT and Posner JB: Distribution of Pathol Lab Med 34: 495-537, 1942. brain metastases. Arch Neurol 45: 741-744, 1988. 6 Baker GS, Kernohan JW and Kiefer EJ: Metastatic tumors of 26 Erwing J: Metastasis in Neoplastic Disease. 3rd ed., chapt. 4, the brain. Surg Clin North Am 31: 1143-1145, 1951. Saunders, Philadelphia, 1928. 7 Chang DB, Yang PC, Luh KT et al: Late survival of non-small 27 Weiss L, Haydock D, Pickren JW and Lane WW: Organ vascularity cell lung cancer patients with brain metastases: Influence of and metastatic frequency. Am J Path 101: 101-113, 1980. treatment. Chest 101: 1293-1297, 1992. 28 Virdis R, Balestrazzi P, Zampolli M et al: Hypertension in 8 Lang EF and Slater J: Metastatic brain tumors. Results of children with . Human Hypertension 8: 395- surgical and nonsurgical treatment. Surg Clin North Am 44: 397, 1994. 865-872, 1964. 29 Glantz MJ, Choy H and Yee L: Prophylactic cranial irradiation 9 Le Chevalier T, Smith FP, Caille P et al: Sites of primary in small cell lung cancer: rationale, results, and malignancies in patients presenting with cerebral metastases: A recommendations. Semin Oncol 24(4): 477-483, 1997. review of 120 cases. Cancer 56: 880-882, 1985. 30 Cranial irradiation for preventing brain metastases of small cell 10 MacGee EE: Surgical treatment of cerebral metastases from lung cancer in patients in complete remission. Cochrane lung cancer: The effect on quality and duration of survival. J Database Syst. Rev 4: CD002805, 2000. Neurosurg 35: 416-420, 1971. 31 Seventh Joint National Committee (JNC7) on Prevention, 11 Markesbery WR, Brooks WH, Gupta GD et al: Treatment Detection, Evaluation, and Treatment of High Blood Pressure. for patients with cerebral metastases. Arch Neurol 35: 754- National Institutes of Health, 2003. 756, 1978. 32 World Health Organization. International Classification of 12 Sundaresan N and Galicich JH: Surgical treatment of brain Diseases. Manual of the International Statistical Classification metastases: Clinical and computerized tomography evaluation of Diseases, Injuries, and Causes of Death. Ninth Revision. Vol. of the results of treatment. Cancer 55: 1382-88, 1985. 1 Geneva: World Health Organization; 1977. 13 Zimm S, Wampler GL, Stablein D et al: Intracerebral 33 Cox DR: Regression models and life-tables. J R Stat Soc 34B: metastases in solid-tumor patients: Natural history and results 187-220, 1986. of treatment. Cancer 48: 384-394, 1981. 34 Nussberger J: Investigating mineralocorticoid hypertension. J 14 Sheehan RG: Southwestern internal medicine conference: Hypertens Suppl 21(2): S25-30, 2003. Small cell lung cancer: a problem of tumor heterogeneity. Am J 35 Chou CM: Evaluation and treatment of hypertension. Rheum Med Sci 292(4): 241-256, 1986. Dis Clin North Am 25(3): 521-37, 1999.

3119 ANTICANCER RESEARCH 24: 3115-3120 (2004)

36 Nugent MJ, Bunn PA, Matthews MJ et al: CNS metastases in small 48 Hart IR: Seed and Soil revisited: mechanisms of site-specific cell bronchogenic carcinoma: Increasing frequency and changing metastasis. Cancer Metastasis Rev 1: 5, 1982. pattern with lengthening survival. Cancer 44: 1885-93, 1979. 49 Fidler IJ, Naito S and Pathak S: Orthotopic implantation is 37 King AB and Ford FR: Clinical and anatomical study of essential for the selection, growth and metastasis of human renal neurological conditions resulting from metastases in CNS due cell cancer in nude mice. Cancer Metastasis Rev 9: 149, 1990. to carcinoma of lung: review of 100 cases. Bull Johns Hopkins 50 Fidler IJ: Orthotopic implantation of human colon carcinomas Hosp 70: 124-156, 1942. into nude mice provides a valuable model for the biology and 38 Halpert B, Fields WS and De Bakey ME: Intracranial therapy of cancer metastasis. Cancer metastasis Rev 10: 229, 1991. metastasis from carcinoma of lung. Surgery 35: 346-349, 1954. 51 Radinsky R: Paracrine growth regulation of human colon 39 Galluzzi S and Payne PM: Bronchial carcinoma: statistical study carcinoma organ-specific metastases. Cancer Metastasis Rev 12: of 741 necropsies with special reference to distribution of 345, 1993. blood-borne metastases. Br J Cancer 9: 511-527, 1955. 52 Radinsky R, Risin S, Fan et al: Level and function of epidermal 40 Simionescu MD: Metastatic tumors of the brain. A follow-up growth factor receptor predict the metastatic potential of study of 195 patients with neurosurgical considerations. J human colon carcinoma cells. Clin Cancer Res 1: 19, 1995. Neurosurg 17: 361-373, 1960. 53 Price JE, Natio S and Fidler IJ: Growth in an organ 41 Haar F and Patterson RH: Surgery for metastatic intracranial microenvironment as a selective process in metastasis. Clin Exp . Cancer 30: 1241-1245, 1972. Metastasis 6: 91, 1988. 42 Ries LA: Influence of extent of disease, histology, and 54 Sutherland RM: Cell and environment interactions in tumor demographic factors on lung cancer survival in the SEER microregions: the multicell spheroid model. Science 240: 177, 1988. population-based data. Semin Surg Oncol 10(1): 21-30, 1994. 55 Scheinok PA, Engler PE, Robertson PD, Hutter RVP and 43 Parkash O: Lung cancer: A statistical study based on autopsy Henson DE: Lung cancer: Difference in age at diagnosis data from 1928 to 1972. Respiration 34: 295-304, 1977. between men and women. N J Med 86(12): 951-7, 1989. 44 Fisher ER, Hellstrom HR and Fisher B: Experimental study of 56 Greenberg ER, Korson R, Baker J, Barrett J, Baron JA and Yates relationship between hypertension and tumor growth and J: Incidence of lung cancer by cell type: A population based study metastases. Br J Cancer 22(2): 342-345, 1968. in New Hampshire and Vermont. JNCI 72(3): 599-603, 1984. 45 Lang EF and Slater J: Metastatic brain tumors. Results of surgical and nonsurgical treatment. Surg Clin North Am 44: 865-872, 1964. 46 Fidler IJ, Gersten DM and Hart IR: The biology of cancer invasion and metastasis. Adv Cancer Res 28: 149, 1978. 47 Fidler IJ: The relationship of embolic homogeneity, number, size and viability to the incidence of experimental metastasis. Received July 12, 2004 Eur J Cancer 9: 223, 1973. Accepted July 26, 2004

3120