References
6O-i PATIENT
(/) DARZYNKIEWICZ Z, CARTER SP, MIKULSKI
SM, ET AL: Cytostatic and cytotoxic effects of Pannon (P-30 protein), a novel anticancer agent. Cell Tissue Kinet 21:169-182, 1988
50-
CD
PHYSICIAN
77771 MAMMOGRAM
(2) SCHULTZ R, R U E P, CHIRIGOS M, ET AL:
Establishment and characterization of a cell line derived from a spontaneous murine lung carcinoma (M109). In Vitro 13:223-231, 1977 (3) KRUSKAL WH, WAUJS WA: Use of ranks on
one-criterion variance analysis. J Am Stat Assoc 47:583-621, 1952 (4) GILL JL: Design and Analysis of Experiments in the Animal and Medical Sciences, chap 1. Ames, IA: Iowa State Univ Press, 1981, pp 82-83 (5) GILL JL: Design and Analysis of Experiments in the Animal and Medical Sciences, chap 3. Ames, IA: Iowa State Univ Press, 1981, pp 76-77
12 40-
o
a:
30-
u m 2
20-
MW/.
10-
(6) CYERT MS, KIRSCHNER MW: Regulation of
cdc2
20-29
30-39
40-49
50-59
60-69
70-79
80 +
of p34 protein kinase during mitosis. Cell 58:361-372, 1989 (8) MOSSMAN T: Rapid colorimetric assay for cellular growth and survival: Application to proliferation and cytotoxicity assays. J Immunol Methods 65:55-63, 1983
DECAOE OF AGE AT DIAGNOSIS Figure 1. Method of breast cancer discovery by decade of age at diagnosis.
(9) HINDLEY J . P H E A R G , STEIN M, ETAL:SUC1 +
encodes a predicted 13-kilodalton protein that is essential for cell viability and is directly involved in the division cycle of Schizosaccharomyces pombe. Mol Cell Biol., 7:504-511,1987 (10) DUNPHY WG, BRIZUELA L, BEACH D, ET AL:
The Xenopus cdc2 protein is a component of MPF, acytoplasmic regulator of mitosis. Cell 54:423-431, 1988 (//) DRAETTAG, BRIZUELA L, POTASHKINJ.ETAL:
Identification of p34 and p i 3 human homologs of the cell cycle regulators of fission yeast encoded by cdc2+ and sucl + . Cell 50:319-325, 1987 (12) RUSSELL P, NURSE P: Negative regulation of
mitosis by weel + , a gene encoding a protein kinase homolog. Cell 49:559-567, 1987 (13) KARN J, WATSON JV, LOWE AD, ET AL:
Regulation of cell cycle duration by c-myc levels. Oncogene 4:773-787, 1989 (14) COOPER JA, WHYTE P: RB and the cell cycle:
Entrance or exit?. Cell 58:1009-1011, 1989 (15) MINTZ B: Normal and abnormal modes of expression of developmentally regulated genes. In Theories and Models in Cellular Transformation (Santi L, Zardi L, eds). London: Academic Press, pp 45-54
Impact of Mammographic Screening on Breast Cancer Diagnosis1 The incidence of breast cancer in the United States has shown a striking apparent increase over the past decade (7), a trend paralleled in Utah. It has been speculated that this increase is due in part Vol. 82, No. 2, January 17, 1990
to the use of mammographic screening (7); that is, the increased detection of prevalent but asymptomatic disease by mammography may be resulting in an overestimation of true incidence. However, the only evidence supporting this hypothesis has been indirect. Although the effectiveness of screening mammography in detecting asymptomatic breast cancer has been demonstrated in trials involving specific populations (2,3), the fraction of newly diagnosed breast cancers in the general population that has been detected by means of mammographic screening alone is unknown. Such a study was recently undertaken in Utah. The Utah Cancer Registry is a statewide population-based registry with very complete coverage. Letters were sent to the physicians for all 244 breast cancer cases diagnosed from September 1 through December 31, 1987, asking whether the cancer was first discovered by: 1) the patient's reporting suspicious findings, 2) the physician's discovering suspicious findings on physical examination, 3) screening mammographic findings without prior clinical suspicion, or 4) unknown. Multiple inquiries were sometimes necessary to trace the physician who knew the flow of diagnostic information.
Information on diagnosis was collected on 200 of the 244 cases (82%). Lack of data was due to physician nonresponse (25 cases) or unknown details (19 cases). The cases on which information was known had similar median ages at diagnosis (63 yr) and stage distribution (in situ plus local, 71% vs. 73%) to those that were unknown. Of these 200 cases, 111 (55.5%) were first found by the patient, 30 (15%) were found by the physician, and 59 (29.5%) were discovered by screening mammography. The distribution of method of discovery by age is shown in fig. 1 and demonstrates: 1) The patient discovered the majority of cancers in all ages except the elderly. 2) In no age group did physical examination by physician find the majority of cancers. 3) No cancers were found by screening mammography in patients under the age of 40 (as expected, since the procedure is not recommended in this age group). 4) The fraction of cases discovered by mammographic
Downloaded from http://jnci.oxfordjournals.org/ at University of Victoria on June 6, 2015
MPF activity in vitro. Cell 53:185-195, 1988 (7) MORENO S, HAYLES J, NURSE P: Regulation
'Supported by Public Health Service contract NO1CN-55428 from the Division of Cancer Prevention and Control, National Cancer Institute, National Institutes of Health, Department of Health and Human Services.
LETTERS
153
154
In addition, large increases in incidence cannot be explained over many years merely by a modest increase in lead time due to mammographic screening. This postulated lead time of 2 years in premenopausal and 4 years in postmenopausal women (6), when applied to our study population, would have resulted in only a very small increase (2%) in the age-adjusted rates (calculated before and after adjusting the age at diagnosis of mammographically diagnosed cases). This implies that if increased mammographic screening is implicated in much of the recent increase in known incidence, this increase will be only temporary, and the rates should diminish within a few years. If this does not occur, one must consider other explanations, such as limited malignant potential of some mammographically discovered carcinomas. WILLIAM P. MCWHORTER HARMON J. EYRE
Division of Hematology-Oncology University of Utah School of Medicine Salt Lake City, UT 84132
References (/) NATIONAL CANCER INSTITUTE: 1987 Annual
Cancer Statistics Review. US Department of Health and Human Services, DHEW Publ No. (NIH) 88-2789. Washington, DC: US Govt Print Off, 1988 (2) MORRISON AS, BRISSON J, KHALID N: Breast
cancer incidence and mortality in the Breast Cancer Detection Demonstration Project. J Natl Cancer Inst 80:1540-1547, 1988 (3) CHU KC, SMART CR, TARONE RE: Analysis of
breast cancer mortality and stage distribution by age for the Health Insurance Plan clinical trial. J Natl Cancer Inst 80:1125-1132, 1988 (4) HOWARD J: Using mammography for cancer control: An unrealized potential. CA 37: 33-48,1987 (5) CENTERS FOR DISEASE CONTROL: Trends in
screening mammograms for women 50 years of age and older: Behavioral Risk Factor Surveillance System, 1987. MMWR 38:137-140, 1989 (6) MOSKOWITZ M: Breast cancer: Age-specific growth rates and screening strategies. Radiology 161:37-41, 1986
Colon Cancer Mortality Among Amputees Research in the etiology of colon cancer has identified physical activity as a risk factor, with colon transit time suggested as a mediating factor in that causal link (7,2). Studies of job activity and colon cancer have given credence to this hypothesis. An American study of 2,950 colon cancer cases from the Los Angeles County cancer registry (3) found a proportional incidence ratio of 1.6 when comparing cases with sedentary occupations to those with high activity occupations, and a recent 19-year follow-up of 1.1 million Swedish men (4) found a relative risk for colon cancer of 1.3 (1.2-1.5, 90% confidence interval) for men in sedentary occupations. In our earlier study (5), we found an increased cardiovascular mortality among a group of men with traumatic amputations; this letter presents data on colon cancer mortality in this same group of men, who can be considered, on the average, less physically active because of their disability. Our earlier study of men drafted into the U.S. Army and hospitalized in 19441945 for service-connected trauma to the extremities reported mortality from January 1946 to April 1977; these same particulars pertain to this report. The group at highest risk in the previous analysis, the traumatic amputation group, was composed of men who had an amputation at or above the knee or one at or above the elbow. Limiting mortality follow-up to 3,135 men with above-theknee amputations, we observed 11 colon cancer deaths versus only 7.55 deaths expected, calculated using U.S. male death rates. This yields a relative risk of 1.46, with 95% confidence interval 0.73-2.61. Of additional interest are the figures for bladder cancer, which show four cases observed versus 1.73 expected, for a relative risk of 2.32, with 95% confidence interval 0.62-5.93. Thus, our findings after nearly 30 years of mortality follow-up agree gener-
Journal of the National Cancer Institute
Downloaded from http://jnci.oxfordjournals.org/ at University of Victoria on June 6, 2015
screening was approximately one-third for ages 50-79 and approximately onequarter for ages 40-49 and 80+. The stage of disease at diagnosis correlated with the method of discovery; 64% of patient-detected cancers were in situ or local (confined to the breast), versus 70% of physician-detected cancers and 85% of the cancers discovered by screening mammography. This study shows that although most breast cancers are still being discovered by the patient, mammographic screening is now making a large contribution to breast cancer diagnosis in Utah. As mammographic screening becomes more prevalent, the fraction of cases discovered by this method will be expected to rise, and the shift toward earlier stage at diagnosis to continue. Recent increases in breast cancer incidence in Utah and the rest of the United States have come at a time of increased use of mammography. Before 1980, few women in the United States (4) were being screened by mammography. However, in a 1987 national survey (5), 29% of women age 50 or over reported having a screening mammogram in the previous year. The corresponding figure for the same survey in Utah was 34%. The Utah age-adjusted incidence rates (in situ cancers included) have risen from 79.3 per 100,000 women in 1979-81 to 105.7 in 1986-87, a 33% increase. Theoretically, this could be completely explained by the additional cases now found by mammographic screening (approximately 25%-30% of the 1987 cases), assuming that these cancers would not yet be evident clinically. However, between 1979-81 and 1986-87, a 30% increase in rates occurred in women aged 20-39, whereas mammographic screening did not find cases in this age group. In the ages 80+, the increase in incidence was twice the mammographic yield. If the rate increases in these age groups were due to increased detection, then improved patient and physician vigilance must be implicated.
6O-i PATIENT
(/) DARZYNKIEWICZ Z, CARTER SP, MIKULSKI
SM, ET AL: Cytostatic and cytotoxic effects of Pannon (P-30 protein), a novel anticancer agent. Cell Tissue Kinet 21:169-182, 1988
50-
CD
PHYSICIAN
77771 MAMMOGRAM
(2) SCHULTZ R, R U E P, CHIRIGOS M, ET AL:
Establishment and characterization of a cell line derived from a spontaneous murine lung carcinoma (M109). In Vitro 13:223-231, 1977 (3) KRUSKAL WH, WAUJS WA: Use of ranks on
one-criterion variance analysis. J Am Stat Assoc 47:583-621, 1952 (4) GILL JL: Design and Analysis of Experiments in the Animal and Medical Sciences, chap 1. Ames, IA: Iowa State Univ Press, 1981, pp 82-83 (5) GILL JL: Design and Analysis of Experiments in the Animal and Medical Sciences, chap 3. Ames, IA: Iowa State Univ Press, 1981, pp 76-77
12 40-
o
a:
30-
u m 2
20-
MW/.
10-
(6) CYERT MS, KIRSCHNER MW: Regulation of
cdc2
20-29
30-39
40-49
50-59
60-69
70-79
80 +
of p34 protein kinase during mitosis. Cell 58:361-372, 1989 (8) MOSSMAN T: Rapid colorimetric assay for cellular growth and survival: Application to proliferation and cytotoxicity assays. J Immunol Methods 65:55-63, 1983
DECAOE OF AGE AT DIAGNOSIS Figure 1. Method of breast cancer discovery by decade of age at diagnosis.
(9) HINDLEY J . P H E A R G , STEIN M, ETAL:SUC1 +
encodes a predicted 13-kilodalton protein that is essential for cell viability and is directly involved in the division cycle of Schizosaccharomyces pombe. Mol Cell Biol., 7:504-511,1987 (10) DUNPHY WG, BRIZUELA L, BEACH D, ET AL:
The Xenopus cdc2 protein is a component of MPF, acytoplasmic regulator of mitosis. Cell 54:423-431, 1988 (//) DRAETTAG, BRIZUELA L, POTASHKINJ.ETAL:
Identification of p34 and p i 3 human homologs of the cell cycle regulators of fission yeast encoded by cdc2+ and sucl + . Cell 50:319-325, 1987 (12) RUSSELL P, NURSE P: Negative regulation of
mitosis by weel + , a gene encoding a protein kinase homolog. Cell 49:559-567, 1987 (13) KARN J, WATSON JV, LOWE AD, ET AL:
Regulation of cell cycle duration by c-myc levels. Oncogene 4:773-787, 1989 (14) COOPER JA, WHYTE P: RB and the cell cycle:
Entrance or exit?. Cell 58:1009-1011, 1989 (15) MINTZ B: Normal and abnormal modes of expression of developmentally regulated genes. In Theories and Models in Cellular Transformation (Santi L, Zardi L, eds). London: Academic Press, pp 45-54
Impact of Mammographic Screening on Breast Cancer Diagnosis1 The incidence of breast cancer in the United States has shown a striking apparent increase over the past decade (7), a trend paralleled in Utah. It has been speculated that this increase is due in part Vol. 82, No. 2, January 17, 1990
to the use of mammographic screening (7); that is, the increased detection of prevalent but asymptomatic disease by mammography may be resulting in an overestimation of true incidence. However, the only evidence supporting this hypothesis has been indirect. Although the effectiveness of screening mammography in detecting asymptomatic breast cancer has been demonstrated in trials involving specific populations (2,3), the fraction of newly diagnosed breast cancers in the general population that has been detected by means of mammographic screening alone is unknown. Such a study was recently undertaken in Utah. The Utah Cancer Registry is a statewide population-based registry with very complete coverage. Letters were sent to the physicians for all 244 breast cancer cases diagnosed from September 1 through December 31, 1987, asking whether the cancer was first discovered by: 1) the patient's reporting suspicious findings, 2) the physician's discovering suspicious findings on physical examination, 3) screening mammographic findings without prior clinical suspicion, or 4) unknown. Multiple inquiries were sometimes necessary to trace the physician who knew the flow of diagnostic information.
Information on diagnosis was collected on 200 of the 244 cases (82%). Lack of data was due to physician nonresponse (25 cases) or unknown details (19 cases). The cases on which information was known had similar median ages at diagnosis (63 yr) and stage distribution (in situ plus local, 71% vs. 73%) to those that were unknown. Of these 200 cases, 111 (55.5%) were first found by the patient, 30 (15%) were found by the physician, and 59 (29.5%) were discovered by screening mammography. The distribution of method of discovery by age is shown in fig. 1 and demonstrates: 1) The patient discovered the majority of cancers in all ages except the elderly. 2) In no age group did physical examination by physician find the majority of cancers. 3) No cancers were found by screening mammography in patients under the age of 40 (as expected, since the procedure is not recommended in this age group). 4) The fraction of cases discovered by mammographic
Downloaded from http://jnci.oxfordjournals.org/ at University of Victoria on June 6, 2015
MPF activity in vitro. Cell 53:185-195, 1988 (7) MORENO S, HAYLES J, NURSE P: Regulation
'Supported by Public Health Service contract NO1CN-55428 from the Division of Cancer Prevention and Control, National Cancer Institute, National Institutes of Health, Department of Health and Human Services.
LETTERS
153
154
In addition, large increases in incidence cannot be explained over many years merely by a modest increase in lead time due to mammographic screening. This postulated lead time of 2 years in premenopausal and 4 years in postmenopausal women (6), when applied to our study population, would have resulted in only a very small increase (2%) in the age-adjusted rates (calculated before and after adjusting the age at diagnosis of mammographically diagnosed cases). This implies that if increased mammographic screening is implicated in much of the recent increase in known incidence, this increase will be only temporary, and the rates should diminish within a few years. If this does not occur, one must consider other explanations, such as limited malignant potential of some mammographically discovered carcinomas. WILLIAM P. MCWHORTER HARMON J. EYRE
Division of Hematology-Oncology University of Utah School of Medicine Salt Lake City, UT 84132
References (/) NATIONAL CANCER INSTITUTE: 1987 Annual
Cancer Statistics Review. US Department of Health and Human Services, DHEW Publ No. (NIH) 88-2789. Washington, DC: US Govt Print Off, 1988 (2) MORRISON AS, BRISSON J, KHALID N: Breast
cancer incidence and mortality in the Breast Cancer Detection Demonstration Project. J Natl Cancer Inst 80:1540-1547, 1988 (3) CHU KC, SMART CR, TARONE RE: Analysis of
breast cancer mortality and stage distribution by age for the Health Insurance Plan clinical trial. J Natl Cancer Inst 80:1125-1132, 1988 (4) HOWARD J: Using mammography for cancer control: An unrealized potential. CA 37: 33-48,1987 (5) CENTERS FOR DISEASE CONTROL: Trends in
screening mammograms for women 50 years of age and older: Behavioral Risk Factor Surveillance System, 1987. MMWR 38:137-140, 1989 (6) MOSKOWITZ M: Breast cancer: Age-specific growth rates and screening strategies. Radiology 161:37-41, 1986
Colon Cancer Mortality Among Amputees Research in the etiology of colon cancer has identified physical activity as a risk factor, with colon transit time suggested as a mediating factor in that causal link (7,2). Studies of job activity and colon cancer have given credence to this hypothesis. An American study of 2,950 colon cancer cases from the Los Angeles County cancer registry (3) found a proportional incidence ratio of 1.6 when comparing cases with sedentary occupations to those with high activity occupations, and a recent 19-year follow-up of 1.1 million Swedish men (4) found a relative risk for colon cancer of 1.3 (1.2-1.5, 90% confidence interval) for men in sedentary occupations. In our earlier study (5), we found an increased cardiovascular mortality among a group of men with traumatic amputations; this letter presents data on colon cancer mortality in this same group of men, who can be considered, on the average, less physically active because of their disability. Our earlier study of men drafted into the U.S. Army and hospitalized in 19441945 for service-connected trauma to the extremities reported mortality from January 1946 to April 1977; these same particulars pertain to this report. The group at highest risk in the previous analysis, the traumatic amputation group, was composed of men who had an amputation at or above the knee or one at or above the elbow. Limiting mortality follow-up to 3,135 men with above-theknee amputations, we observed 11 colon cancer deaths versus only 7.55 deaths expected, calculated using U.S. male death rates. This yields a relative risk of 1.46, with 95% confidence interval 0.73-2.61. Of additional interest are the figures for bladder cancer, which show four cases observed versus 1.73 expected, for a relative risk of 2.32, with 95% confidence interval 0.62-5.93. Thus, our findings after nearly 30 years of mortality follow-up agree gener-
Journal of the National Cancer Institute
Downloaded from http://jnci.oxfordjournals.org/ at University of Victoria on June 6, 2015
screening was approximately one-third for ages 50-79 and approximately onequarter for ages 40-49 and 80+. The stage of disease at diagnosis correlated with the method of discovery; 64% of patient-detected cancers were in situ or local (confined to the breast), versus 70% of physician-detected cancers and 85% of the cancers discovered by screening mammography. This study shows that although most breast cancers are still being discovered by the patient, mammographic screening is now making a large contribution to breast cancer diagnosis in Utah. As mammographic screening becomes more prevalent, the fraction of cases discovered by this method will be expected to rise, and the shift toward earlier stage at diagnosis to continue. Recent increases in breast cancer incidence in Utah and the rest of the United States have come at a time of increased use of mammography. Before 1980, few women in the United States (4) were being screened by mammography. However, in a 1987 national survey (5), 29% of women age 50 or over reported having a screening mammogram in the previous year. The corresponding figure for the same survey in Utah was 34%. The Utah age-adjusted incidence rates (in situ cancers included) have risen from 79.3 per 100,000 women in 1979-81 to 105.7 in 1986-87, a 33% increase. Theoretically, this could be completely explained by the additional cases now found by mammographic screening (approximately 25%-30% of the 1987 cases), assuming that these cancers would not yet be evident clinically. However, between 1979-81 and 1986-87, a 30% increase in rates occurred in women aged 20-39, whereas mammographic screening did not find cases in this age group. In the ages 80+, the increase in incidence was twice the mammographic yield. If the rate increases in these age groups were due to increased detection, then improved patient and physician vigilance must be implicated.
