Short-term, medium-term, long-term, and lifetime risks of developing and dying of breast carcinoma in a Westernized Chinese population

Evidence from Hong Kong between 1976 and 2000

Abstract

BACKGROUND

Estimates of lifetime and shorter-term risks of developing and dying of breast carcinoma are useful translational statistics for risk communication; however, these statistics are unknown for Chinese women, who account for one-fifth of the world's female population. In the current study, the authors examined age-specific 5-year, 10-year, 20-year, and lifetime risks of breast carcinoma incidence and disease-related mortality among Hong Kong women.

METHODS

Multiple-decrement life tables were constructed by applying age-specific incidence and mortality rates obtained from cross-sectional data to a hypothetic birth cohort. Incidence, mortality, and population data from 1976 through 2000 were used.

RESULTS

For Hong Kong women in the period 1996–2000, the overall lifetime risk of developing invasive breast carcinoma was 5.7% (1 in 17), compared with 10.6% (1 in 9) for U.S. Asian/Pacific Islander women and 14.5% (1 in 7) for U.S. white women. The lifetime breast carcinoma–related mortality risk for Hong Kong women in that same period was 1.7% (1 in 58). For women age 50 years, the 5-, 10-, and 20-year risks of developing breast carcinoma were 0.6% (1 in 172), 1.1% (1 in 87), and 2.1% (1 in 47), respectively. The absolute lifetime probability of developing breast carcinoma increased from 3.81% (1 in 26) during the period 1976–1980 to 5.73% (1 in 17) during the period 1996–2000; 44.5% of this change can be attributed to the increased longevity of women in the latter time period, and the remaining 55.5% can be ascribed to a secular rise in breast carcinoma incidence.

CONCLUSIONS

As China becomes more Westernized, the authors expect that the risks of developing and dying of breast carcinoma for Chinese women will become more similar to the corresponding risks for Western women. The current data from Hong Kong, the most Westernized Chinese community, appear to support this conclusion. Cancer 2005. © 2004 American Cancer Society.

The lifetime (i.e., up to age 85 years) risk of developing breast carcinoma, 1 in 9 for Caucasian populations,1 is a commonly quoted statistic in the scientific and lay media. Nonetheless, misconceptions about this frequently cited number often generate excessive fear and anxiety among the general public. For instance, there have been numerous surveys showing that women in all age groups and from various communities overestimate their risk of developing breast carcinoma by as much as an order of magnitude.2–5 Therefore, it is useful to project risk estimates over a shorter time horizon and for specific ages, as such projections tend to be more accurate and more meaningful for the purposes of risk communication.

Whereas there have been numerous reports on the short-term and lifetime risks of breast carcinoma in Caucasian communities,1, 6, 7 data on the majority of the world's female population outside of Europe and North America (particularly in China and the rest of Asia) are scarce. In many developing countries, including mainland China, incomplete data capture underlies this knowledge gap. Specifically, what remains to be characterized is local women's absolute risk of developing breast carcinoma in the short (5 years), medium (10 years), and long term (20 years), as well as over a lifetime, as opposed to incidence, which is the likelihood that breast carcinoma will develop over some shorter fixed interval (usually 1 year). Moreover, studying changes in Hong Kong women's risk estimates over time may reveal what is likely to come for the rest of China. Hong Kong is currently in the midst of the epidemiologic transition that the West has nearly completed and that mainland China is only beginning to experience.8

The current article describes a longitudinal analysis of the age-specific risks of developing and dying of breast carcinoma within the next 5, 10, and 20 years, as well as the corresponding lifetime risks for Hong Kong women between 1976 and 2000.

MATERIALS AND METHODS

Sources of Data

Data on breast carcinoma incidence were obtained from the Hong Kong Cancer Registry. Details regarding the registry's history, objectives, management, and registration coverage are documented elsewhere.9 In brief, the Hong Kong Cancer Registry is a population-based registry covering the entire resident population of Hong Kong. Information on breast carcinoma cases was collected from both the private and public service sectors (primarily through clinical oncology, radiation oncology, and histopathology departments) and from the government's Births, Deaths, and Marriages Registry, as well as through voluntary notification by medical practitioners. The completeness and quality of the data have been reported to be good,10 and the Hong Kong Cancer Registry is an accredited member of the International Association of Cancer Registries.1 Midyear population statistics and mortality data were obtained from the Hong Kong Census and Statistics Department.

The current analysis was based on 26,807 cases (out of a total of 26,870, as age at diagnosis was unknown in 63 cases) of invasive breast carcinoma (International Classification of Diseases [8th and 9th editions] code 174) reported by healthcare institutions in Hong Kong between January 1976 and December 2000. In addition, to view Hong Kong women's breast carcinoma risks in an international context and to compare these risks across geoethnic boundaries, we obtained corresponding incidence data from the National Cancer Institute Surveillance, Epidemiology, and End Results (SEER) Program (which covers 12 areas—San Francisco, CA; Connecticut; Detroit, MI; Hawaii; Iowa; New Mexico; Seattle, WA; Utah; Atlanta, GA; San Jose–Monterey, CA; Los Angeles, CA; and Alaska [Alaska Native Registry]), along with mortality data and demographic statistics for all 12 SEER areas during the period 1996–2000 from the U.S. National Center for Health Statistics. Racial and ethnic classifications were made, primarily on the basis of information from medical records or death certificates, by SEER or by the National Center for Health Statistics. The SEER data set accounted for a total of 126,280 breast carcinoma cases (of which 121,349 were first primary cases) in the current analysis.

Statistical Analysis

We calculated age-specific incidence, breast carcinoma–related mortality, and all-causes (excluding breast carcinoma) mortality rates within 5-year age groups (ranging from birth–5 years to 85 years or older) for each of five 5-year calendar periods—1976–1980, 1981–1985, 1986–1990, 1991–1995, and 1996–2000. Incidence rates take into account only the first occurrence of breast carcinoma in each affected individual. We used 5-year analysis periods to achieve stability in the estimation of rates, particularly for those age groups that were limited in size within certain single years.

Estimates of the risk of developing (i.e., being diagnosed with) breast carcinoma were computed on the basis of breast carcinoma incidence and all-causes (excluding breast carcinoma) mortality rates, using the methodology of Feuer and colleagues.1, 11–13 Specifically, we constructed standard multiple-decrement life tables starting from birth for a hypothetic cohort of 10 million females. In this model, each individual was subject to one of two mutually exclusive events: 1) the development of invasive breast carcinoma or 2) death due to causes other than breast carcinoma, with the individual never having developed an invasive breast malignancy. Probabilities derived from the calculated incidence and mortality rates were applied to the cohort to yield the number of women expected to develop breast carcinoma and the number of women expected to die of other causes without having developed breast disease within each 5-year age interval. Those who survived and were free of breast carcinoma within a given age interval proceeded to the next interval as the at-risk population. The risk of developing breast carcinoma within a particular age interval was calculated as the number of incident malignancies expected during that interval divided by the total number of at-risk (i.e., breast carcinoma–free) individuals at the beginning of that period. The lifetime risk of developing breast carcinoma (i.e., the probability of developing breast carcinoma from birth onward) was estimated by dividing the sum of all expected breast carcinomas in the life table by the number of individuals in the initial birth cohort.

In addition to these short-term and lifetime risks of developing breast carcinoma, corresponding risk estimates for breast carcinoma–related death were computed. These calculations were performed using a methodology similar to the one described above—specifically, we constructed a standard multiple-decrement life table in which each person in a hypothetic cohort of 10 million female newborns was exposed to the risk of dying of breast carcinoma and the risk of dying of any other cause based on existing mortality data.

To assess the sensitivity of our model, we calculated cumulative (from birth) and age-conditional probabilities of developing or dying of breast carcinoma in the absence of other competing causes of death. Probabilities calculated in this way are better for comparing populations with different underlying rates of other-cause mortality; in fact, these probabilities are similar to the cumulative rates (which approximate cumulative risk) that have been used to compare countries with dramatically different rates of other-cause mortality.14 We also computed the risk of developing or dying of breast carcinoma in the presence of other-cause mortality risk in U.S. Asian/Pacific Islander women and applied the results to the other two populations analyzed (Hong Kong women and U.S. white women), again to test whether differences in other-cause mortality among these populations would influence our findings.

Finally, we plotted the secular trend in lifetime breast carcinoma risk for each 5-year period between 1976 and 2000. Two important factors can account for temporal changes in longitudinal breast carcinoma incidence—namely, enhanced detection via mammographic screening and general increases in longevity. To date, mammography uptake rates in Hong Kong have been very low. According to a population-based general household survey conducted in 2002 by the Hong Kong Census and Statistics Department, only 0.84% of all women age ≥ 50 years reported undergoing regular mammographic screening (unpublished data). Thus, it is unlikely that a detection bias would have any appreciable impact on the 25-year trend in breast carcinoma risk in Hong Kong; however, as rates of mortality due to causes other than breast carcinoma grow smaller, more women are living longer and thus have a higher lifetime risk of developing breast carcinoma. To estimate the effect of this overall decrease in mortality on breast carcinoma risk, we repeated our life table calculations of lifetime risk for each 5-year time interval using actual rates of breast carcinoma incidence while holding the mortality rate constant at the 1976–1980 level.

All analyses were conducted using DEVCAN (Version 5.1)13 and S-PLUS (Version 3.4; Insightful Corp., Seattle, WA) software.

RESULTS

The life table for invasive breast carcinoma risk (Table 1) summarizes results for Hong Kong women as calculated using incidence and mortality data for the period 1996–2000. The at-risk population decreased successively from one interval to the next; the cohort initially consisted of 10 million newly born individuals. The number of individuals who developed breast carcinoma and the number of individuals who died of other causes (i.e., not breast carcinoma) within each age interval are also presented. The number of women who developed breast carcinoma initially was low but increased significantly from age 30 years onward and peaked in the sixth decade of life. The lifetime cumulative probability of developing breast carcinoma was 5.73% (1 in 17), or 4.86% (1 in 21) from birth to age 85 years. It is important to emphasize that the lifetime figure is a cohort probability and does not represent personal risk. In other words, it does not mean that if a woman lives beyond age 85 years, she will have had a 5.73% chance of developing breast carcinoma at some point in her life, since this risk would depend on her survival to age 85 years and beyond; rather, it suggests that if we follow 10,000 women from birth to at least age 85 years, 573 of them will be diagnosed with breast carcinoma during that interval. Table 2 is the corresponding life table for breast carcinoma– related mortality risk. Unlike incidence, breast carcinoma–related mortality increases almost monotonically from one age interval to the next. The lifetime risk of dying of breast carcinoma was calculated to be 1.72% (1 in 58).

Table 1. Life Tables Detailing the Probability of Developing Invasive Breast Carcinoma for Hong Kong Women, 1996–2000a

Age (yrs)	Total no. alive and free of breast carcinoma at beginning of interval	No. developing invasive breast carcinoma during interval	No. who died of other causes during interval	Cumulative probability (from birth) of developing breast carcinoma
a Numbers in parentheses indicate calculations in the absence of other competing causes of death.
0–4	10,000,000 (10,000,000)	0 (0)	32,849 (0)	0.000000 (0.000000)
5–9	9,967,151 (10,000,000)	0 (0)	10,207 (0)	0.000000 (0.000000)
10–14	9,956,943 (10,000,000)	7 (7)	5,511 (0)	0.000001 (0.000001)
15–19	9,951,425 (9,999,993)	115 (116)	8,879 (0)	0.000012 (0.000012)
20–24	9,942,430 (9,999,877)	757 (762)	12,383 (0)	0.000088 (0.000089)
25–29	9,929,290 (9,999,115)	3,572 (3,600)	14,030 (0)	0.000445 (0.000449)
30–34	9,911,688 (9,995,514)	10,299 (10,396)	16,779 (0)	0.001475 (0.001488)
35–39	9,884,611 (9,985,119)	22,537 (22,795)	23,458 (0)	0.003729 (0.003768)
40–44	9,838,616 (9,962,324)	40,827 (41,420)	37,676 (0)	0.007812 (0.007910)
45–49	9,760,113 (9,920,904)	53,612 (54,652)	58,563 (0)	0.013173 (0.013375)
50–54	9,647,938 (9,866,253)	56,011 (57,525)	89,125 (0)	0.018774 (0.019127)
55–59	9,502,802 (9,808,728)	54,905 (57,079)	148,154 (0)	0.024264 (0.024835)
60–64	9,299,743 (9,751,649)	47,722 (50,625)	237,211 (0)	0.029037 (0.029898)
65–69	9,014,809 (9,701,024)	46,934 (51,590)	408,857 (0)	0.033730 (0.035057)
70–74	8,559,018 (9,649,434)	49,354 (57,866)	697,941 (0)	0.038665 (0.040843)
75–79	7,811,722 (9,591,568)	49,950 (65,721)	1,082,758 (0)	0.043660 (0.047415)
80–84	6,679,014 (9,525,847)	49,503 (79,651)	1,570,444 (0)	0.048611 (0.055380)
85+	5,059,066 (—)	87,082 (—)	4,971,984 (—)	0.057319 (—)

Table 2. Life Tables Detailing the Probability of Dying of Breast Carcinoma for Hong Kong Women, 1996–2000a

Age (yrs)	Total no. alive at beginning of interval	No. who died of breast carcinoma during interval	No. who died of other causes during interval	Cumulative probability (from birth) of dying of breast carcinoma
a Numbers in parentheses indicate calculations in the absence of other competing causes of death.
0–4	10,000,000 (10,000,000)	0 (0)	32,849 (0)	0.000000 (0.000000)
5–9	9,967,151 (10,000,000)	0 (0)	10,207 (0)	0.000000 (0.000000)
10–14	9,956,943 (10,000,000)	0 (0)	5511 (0)	0.000000 (0.000000)
15–19	9,951,432 (10,000,000)	7 (7)	8879 (0)	0.000001 (0.000001)
20–24	9,942,546 (9,999,993)	83 (84)	12,383 (0)	0.000009 (0.000009)
25–29	9,930,080 (9,999,910)	292 (294)	14,033 (0)	0.000038 (0.000038)
30–34	9,915,755 (9,999,616)	1,017 (1,026)	16,793 (0)	0.000140 (0.000141)
35–39	9,897,946 (9,998,589)	3,275 (3,313)	23,511 (0)	0.000467 (0.000472)
40–44	9,871,159 (9,995,276)	5,875 (5,960)	37,868 (0)	0.001055 (0.001068)
45–49	9,827,416 (9,989,316)	8,673 (8,842)	59,110 (0)	0.001922 (0.001953)
50–54	9,759,632 (9,980,474)	11,331 (11,639)	90,386 (0)	0.003055 (0.003117)
55–59	9,657,915 (9,968,835)	13,689 (14,235)	150,937 (0)	0.004424 (0.004540)
60–64	9,493,289 (9,954,600)	13,817 (14,660)	242,646 (0)	0.005806 (0.006006)
65–69	9,236,826 (9,939,940)	13,050 (14,340)	419,800 (0)	0.007111 (0.007440)
70–74	8,803,976 (9,925,599)	13,455 (15,784)	719,606 (0)	0.008456 (0.009018)
75–79	8,070,915 (9,909,815)	16,340 (21,552)	1,121,454 (0)	0.010090 (0.011174)
80–84	6,933,122 (9,888,263)	19,826 (32,100)	1,634,664 (0)	0.012073 (0.014384)
85+	5,278,631 (—)	51,183 (—)	5,227,448 (—)	0.017191 (—)

Figure 1 plots cumulative probabilities of developing and dying of breast carcinoma for Hong Kong, U.S. Asian/Pacific Islander, and U.S. white women during the period 1996–2000. U.S. white women had the highest risk estimates, followed by U.S. Asian/Pacific Islander and Hong Kong women, in that order. The divergence in risk among these three ethnic groups increased with increasing age and was more pronounced with regard to incidence (as opposed to mortality). It is noteworthy that although the cumulative incidence curves for U.S. Asian/Pacific Islander and Hong Kong women were clearly divergent, the corresponding mortality curves were much less so and even converged in the ninth decade of life. Similarly, the lifetime breast carcinoma risk estimates for U.S. white and Asian/Pacific Islander women were 14.3% (1 in 7) and 10.5% (1 in 9), respectively, whereas the corresponding mortality rates were 3.2% (1 in 31) and 1.9% (1 in 54), respectively.

Figure 1. Cumulative probabilities from birth of (A) developing and (B) dying of invasive breast carcinoma for Hong Kong, U.S. Asian/Pacific Islander, and white women, 1996–2000.

Tables 3 and 4 show incidence and mortality risk estimates, respectively, within the next 5, 10, and 20 years and for the remainder of life for breast carcinoma–free women ages 20, 30, 40, 50, 60, and 70 years. For both Hong Kong and U.S. women, these results clearly indicate the positive dependence of breast carcinoma incidence and mortality risk on age. Of note was the finding that for a woman of any given age, the risk over the next x years was always higher than the risk over the next y years for a woman who was z years older, provided that x = y + z. For instance, the estimated risk of developing breast carcinoma within the next 20 years for a Hong Kong woman whose current age was 40 years was 2.1% (1 in 48), whereas the risk of developing breast carcinoma within the next 10 years for a 50-year-old woman was found to be 1.1% (1 in 87). This is because the 50-year-old woman had already survived without breast carcinoma for 10 additional years relative to the 40-year-old woman, who had yet to assume the breast carcinoma and all-causes mortality risks associated with the coming decade.

Table 3. Estimated Risk of Developing Invasive Breast Carcinoma within the Next 5, 10, and 20 Years and in Patients' Remaining Lifetimes for Hong Kong and U.S. Women, 1996–2000

Current age (yrs)	Hong Kong								U.S. SEER
	All women								Asian/Pacific Islander								White
	5 yrs		10 yrs		20 yrs		Remaining lifetime		5 yrs		10 yrs		20 yrs		Remaining lifetime		5 yrs		10 yrs		20 yrs		Remaining lifetime
	Risk (%)	One in:	Risk (%)	One in:	Risk (%)	One in:	Risk (%)	One in:	Risk (%)	One in:	Risk (%)	One in:	Risk (%)	One in:	Risk (%)	One in:	Risk (%)	One in:	Risk (%)	One in:	Risk (%)	One in:	Risk (%)	One in:
SEER: Surveillance, Epidemiology, and End Results Program. a Estimates were truncated at age 85 years; therefore, 20-year risk estimates are not available for this age group.
20	0.0	13,132	0.0	2,297	0.4	268	5.8	17	0.0	13,793	0.0	2,594	0.4	252	10.6	9	0.0	14,660	0.0	2,338	0.4	232	14.5	7
30	0.1	962	0.3	302	1.3	78	5.7	17	0.1	976	0.4	278	1.7	60	10.6	9	0.1	880	0.4	257	1.9	54	14.5	7
40	0.4	241	1.0	104	2.1	48	5.4	18	0.5	195	1.3	77	3.5	29	10.3	10	0.6	178	1.5	67	4.3	23	14.2	7
50	0.6	172	1.1	87	2.1	47	4.6	22	1.0	97	2.2	45	4.7	21	9.2	11	1.3	78	2.9	35	6.6	15	13.1	8
60	0.5	195	1.0	98	2.1	48	3.6	28	1.3	76	2.6	38	5.0	20	7.3	14	1.9	52	4.0	25	8.0	13	11.0	9
70a	0.6	173	1.2	86	—	—	2.8	36	1.4	73	2.6	38	—	—	5.2	19	2.4	42	4.6	22	—	—	8.1	12

Table 4. Estimated Risk of Dying of Breast Carcinoma within the Next 5, 10, and 20 Years and in Patients' Remaining Lifetimes for Hong Kong and U.S. Women, 1996–2000

Current age (yrs)	Hong Kong								U.S. SEER
	All women								Asian/Pacific Islander								White
	5 yrs		10 yrs		20 yrs		Remaining lifetime		5 yrs		10 yrs		20 yrs		Remaining lifetime		5 yrs		10 yrs		20 yrs		Remaining lifetime
	Risk (%)	One in:	Risk (%)	One in:	Risk (%)	One in:	Risk (%)	One in:	Risk (%)	One in:	Risk (%)	One in:	Risk (%)	One in:	Risk (%)	One in:	Risk (%)	One in:	Risk (%)	One in:	Risk (%)	One in:	Risk (%)	One in:
SEER: Surveillance, Epidemiology, and End Results Program. a Estimates were truncated at age 85 years; therefore, 20-year risk estimates are not available for this age group.
20	0.0	119,711	0.0	26,510	0.0	2,130	1.7	58	0.0	472,157	0.0	31,696	0.0	2,119	1.9	54	0.0	142,204	0.0	20,863	0.1	1,759	3.3	31
30	0.0	9,754	0.0	2,310	0.2	526	1.7	58	0.0	8,124	0.0	2,264	0.2	504	1.9	53	0.0	6,692	0.1	1,913	0.2	403	3.3	31
40	0.1	1,680	0.1	679	0.4	249	1.7	59	0.1	1,726	0.2	646	0.5	208	1.8	55	0.1	1,390	0.2	507	0.6	155	3.2	31
50	0.1	861	0.3	390	0.5	188	1.6	64	0.1	684	0.3	304	0.7	140	1.7	59	0.2	507	0.5	219	1.2	87	3.1	32
60	0.1	687	0.3	353	0.6	168	1.3	74	0.2	500	0.4	253	0.8	131	1.4	71	0.3	297	0.7	137	1.6	62	2.7	36
70a	0.2	654	0.3	295	—	—	1.1	87	0.2	519	0.4	255	—	—	1.1	92	0.5	203	1.0	100	—	—	2.3	44

Note that the number of women who develop breast carcinoma peaks in the sixth decade of life (Table 1) primarily because women begin to die of other causes during this time, and as a result, proportionally fewer women survive to be exposed to breast carcinoma risk in subsequent decades. Despite this phenomenon, 5- and 10-year breast carcinoma risk estimates remain fairly steady in the sixth through eighth decades of life (Table 3).

Comparison of Hong Kong and American women revealed that both the risk of developing breast carcinoma and the risk of dying of breast carcinoma were consistently smaller (by a factor of two to three on average) in the former population. This relation held true across all current age categories as well as all time horizons. When ethnicity is taken into consideration, there is a clear dose-response gradient, with the risk for U.S. Asian/Pacific Islander women falling between that of Hong Kong women (95% of whom are ethnically Chinese) and that of U.S. white women. In all three population groups, the finding that the breast carcinoma–related mortality risk was much smaller than the risk of developing breast carcinoma suggests that many who are diagnosed with this malignancy eventually die of other causes.

Our sensitivity analyses (Fig. 1; Tables 1, 2), which assumed the absence of competing causes of death, demonstrate that our findings are robust and were not influenced by potential differences in all-causes mortality across the population groups studied. Results similar to those presented in Tables 3 and 4 were obtained when all-causes mortality risk estimates for U.S. Asian/Pacific Islander women were applied to Hong Kong and U.S. white women (data not shown).

Figure 2 depicts the trend in lifetime probability of developing invasive breast carcinoma from 1976 through 2000. The absolute risk increased from 3.81% (1 in 26) at the beginning of the interval to 5.73% (1 in 17) at the end of the interval. Risk calculations performed with mortality rates held constant over time yielded lower overall risk estimates (4.88% [1 in 20] in 2000) and a slower rate of increase. Assuming independence between breast carcinoma incidence and mortality due to other causes, we can partition the observed increase in lifetime breast carcinoma risk into two parts. The absolute increase of 1.92 breast carcinoma cases per 100 women between 1976 and 2000 can be disaggregated into 1) the contribution of an increase in longevity for women during the interval ([0.0573 − 0.0488]/0.0192 = 44.5%) and 2) the contribution of a secular increase in breast carcinoma incidence over that time ([0.0488 − 0.0381]/0.0192 = 55.5%).

Figure 2. Lifetime probability of developing invasive breast carcinoma for Hong Kong women, 1976–2000.

DISCUSSION

We have reported both lifetime and shorter-term risk estimates for the development of and death due to invasive breast carcinoma in a Westernized Chinese population, with these estimates being one-half to two-thirds as large as the corresponding SEER estimates. Based on data from 1996–2000, the current lifetime risk (age 85 years and beyond) of developing invasive breast carcinoma for Hong Kong women is 1 in 17. The most recent Hong Kong Cancer Registry figure of 1 in 24 relied on single-year incidence and mortality data, and the estimate was truncated at age 75 years.9 In the current study, the probability of being diagnosed with breast carcinoma between birth and age 75 years (Table 1) was found to be 1 in 23.

Many have previously misinterpreted the truncation at age 85 years (which is close to the life expectancy of Hong Kong women [84.5 years in 2002]) to mean that all women are assumed to reach this age in the life table calculation. This is incorrect, because each age interval is assigned a weight based on the probability of surviving to that interval; thus, the reduced probability of reaching older ages, according to a standard mortality distribution, is considered in the lifetime risk measure.

Lifetime risk estimates, although useful for monitoring secular trends and making comparisons across geoethnic boundaries, are valid for a newborn only if incidence and mortality patterns remain stationary over his or her entire lifetime. It is clear that we cannot assume this to be the case, and our current findings (Fig. 2) do in fact demonstrate that it is not. Given the current state of understanding of the underlying etiology and pathogenesis of breast carcinoma and our inability to predict changes in population risk factors, it would be difficult to accurately project long-term incidence. Moreover, the effects of mammography and other screening technologies (effects that have yet to materialize locally) may include inflated incidence rates (due to detection bias), whereas more effective preventive measures may lead to a reduction in future cases; factors such as these make long-term forecasting an unreliable exercise.

Shorter-term risk estimates, however, are less sensitive to changes in future incidence and mortality rates and are therefore more useful for women who are currently alive without malignant disease. For instance, our findings indicate that a woman who is now 40 years old and has no history of breast carcinoma has a 1 in 241 chance of developing breast carcinoma by age 45 years and a 1 in 104 risk of developing breast carcinoma by age 50 years. In contrast, a 50-year-old woman currently has a 1 in 172 probability and a 1 in 87 probability of being diagnosed with breast carcinoma by age 55 years and age 60 years, respectively. Similarly, the subsequent lifetime risks of developing breast carcinoma for 40-, 50-, 60-, and 70-year-olds are 1 in 18, 1 in 22, 1 in 28, and 1 in 36, respectively. These estimates are strikingly diverse and differ from the 1 in 17 lifetime risk estimate for a newborn. Because the public often interprets lifetime risk to represent a woman's risk in the immediate future or even within the next year,1, 15 calculating and disseminating shorter-term ‘translational statistics’15 regarding risk, stratified by current age and based on prevailing incidence and mortality distributions, can help to put an individual's susceptibility to breast carcinoma in perspective and facilitate understanding on the part of the public.

Placing these risk estimates within a comparative framework of geoethnic heterogeneities is also important. As the results show, even in the highly Westernized Hong Kong Chinese society, breast carcinoma incidence is one-third to one-half less than in the U.S., which has incidence rates similar to those found in predominantly Caucasian populations in Western Europe, Canada, and Australia.10 The subsequent lifetime risks of being diagnosed with a malignant breast tumor for Hong Kong women ages 50, 60, and 70 years are similar to the 10-year breast carcinoma risks for these women's white American counterparts. In this age of Internet-savvy health information consumers, in which overseas cancer risks (mostly from developed Western countries) are widely reported online, it is particularly important to calculate and disseminate locally specific and relevant risk statistics that take into account geoethnic heterogeneities. Otherwise, inapplicable information may generate excessive fear and anxiety among women from different sociocultural backgrounds, and these feelings may in turn lead to inappropriate health-related behavior. One such example is the increasing use of regular mammographic examination as a screening practice in many Asian countries, such as Singapore.16 We have previously questioned the utility of regular screening regimens for women whose baseline breast carcinoma risk is much lower than that of Caucasian women, who have been involved in all randomized trials demonstrating the effectiveness of screening mammography. In fact, we concluded that mass screening via regular mammography may cause more harm than good in such low-risk communities.17 There may well be a significant proportion of low-risk women who attend these examinations at least in part because of their awareness of much higher overseas risk estimates, such as the commonly quoted 1 in 8 lifetime risk for U.S. women.15, 18

One caveat bears mentioning when relaying these risk estimates to women—our findings reflect risks for the general population, but women's individual risk profiles may differ according to their history of exposure to various risk factors (e.g., history of menstruation, fertility, and breastfeeding, as well as other hormonal and lifestyle-related factors). Risk prediction models, such as the Gail risk equations,19 exist for non-U.S. populations, but no such model has yet been developed for Chinese or Asian populations. A second limitation of the current study is the lack of information regarding in situ diagnoses (e.g., ductal carcinoma in situ). At present, the Hong Kong Cancer Registry (as well as most similar agencies in east Asia) does not routinely collect such data.

Our secular trend calculations and projections with mortality adjustments confirm that declining mortality has had a small but evident effect on the lifetime risk of developing breast carcinoma in Hong Kong, as has been observed in the U.S.1 Unlike in other countries, however, in Hong Kong, detection bias due to enhanced mammographic screening surveillance was not a significant contributor to the observed increase in risk over time. The residual secular increase in breast carcinoma incidence is primarily a cohort effect and can be understood as the direct and indirect consequence of Westernization and accompanying socioeconomic changes leading to low fertility, younger age at menarche, and older age at marriage and childbirth, as well as dietary excesses and reduced physical activity.

Local population-derived breast carcinoma risk estimates specific to non-Caucasian women in various age groups can be used by care providers to increase awareness and put breast carcinoma risk into perspective, thereby reducing general anxiety and confusion caused by statistics from overseas.20 The figures presented in the current study are informative with regard to corresponding risks in other similarly developed Asian countries (e.g., Taiwan, Singapore, Japan, Korea) and foreshadow the risks that mainland Chinese women may face in the coming decades as China continues through its epidemiologic transition.

Acknowledgements

The authors thank Marie Chi for her expert secretarial assistance in the preparation of the current article.