Breast Disease 13 (2001) 41–48 41 IOS Press Breast Cancer Screening in Taiwan and China a,* a microcalcifications, along with mammography within a long Chiun-Sheng Huang King-Jen Chang , period, may provide more effective protection for Taiwanese * Chen-Yang Shenb, and Chinese women against breast cancer. a Department of Surgery, National Taiwan University Hospital, Taipei, 100, Taiwan BREAST CANCER IN CHINESE Tel.: +886 2 23123456 ext 5080; Fax: +886 POPULATIONS 2 2392 8856;E-mail: email@example.com In contrast to incidence rates prevailing in b Institute of Biomedical Sciences, Academia women in Western countries, the women in Tai- Sinica, Taipei, 115, Taiwan wan and China are considered to have the lowest Tel.: +886 2 789 9036; Fax: +886 2 2782 incidence of breast cancer in the world. The age- 3047; E-mail: firstname.lastname@example.org adjusted breast cancer incidence rate of 1996 in Taiwanese women was 26.65/100,000 . This ABSTRACT: In contrast to incidence rates prevailing in figure is comparable to that in women living in women in Western countries, Chinese women in Taiwan and urban areas in China. On the basis of a prospec- China are considered to have the lowest incidence of breast tive survey, in women in Shanghai, the largest cancer in the world. However, in the past 20 years, breast city on the east coast of China, the breast cancer cancer incidence in Chinese women has seen a dramatic increase of 50–100%, which strongly supports the need for incidence was 26.5/100,000 from 1988 to 1992 breast cancer prevention and screening programs. It is also [2,3], which is about only one-fourth to one-third important to indicate that breast cancer in Chinese women is the incidence in white women in Western popu- characterized by younger age at tumor onset. More than lations. Given the dominant contribution of es- 50% of the total breast cancer diagnosed annually is found trogen to breast cancer development, the low in- in premenopausal patients, creating the need to initiate breast cancer screening programs in this popula- cidence in Chinese women can be attributable to tion. Initially, the breast cancer screening program de- lower estrogen level. On average, Asian women pended on breast self-examination. Since Chinese women have 20% lower serum estradiol levels than have relatively small breasts, it was assumed that breast Western women , which could be related to cancer was easier to detect by self-examination. However, this strategy has failed. The dilemma of breast cancer late menarche, early menopause, less common screening can be summarized by the fact that Chinese have a postmenopausal obesity, low fat or cholesterol rapidly increasing incidence of premenopausal breast cancer, intake from Eastern diets, or a combination of while the overall incidence is still low. Therefore, since these factors. premenopausal women have denser breasts than postmeno- In addition to environmental and dietary fac- pausal women, and Chinese women have smaller breasts and a higher percentage of dense breasts, increased mammogra- tors, the low incidence of breast cancer in Chi- phy screening frequency may be not the sole solution to nese women also may be associated with differ- increase detection in this age group. In our experience in ent genetic background. For incidence, a high Taiwan, the addition of breast ultrasound may be helpful. frequency of high-activity alleles/genotypes of Nearly all the nonpalpable cancers detected by mammogra- phy in our women are due to microcalcifications, and ultra- COMT, a gene involved in inactivation of the sound is more sensitive in detecting nonpalpable cancers; reactive metabolites of estrogen, has been identi- Therefore, we suggest that a screening program, based on fied in our population (e.g. >95% of women have ultrasound to detect nonpalpable cancers not associated with low-risk genotypes versus 72–87% in Western women) . This would contribute to markedly * Corresponding authors: Chiun-Sheng Huang and Chen- lower exposure of breast epithelium to tumori- Yang Shen genic estrogen metabolites, which also might ex- 0888-6008/01/$8.00 © 2001 – IOS Press. All rights reserved 42 C.-S. Huang et al. / Breast Cancer Screening in Taiwan and China plain the reduced risk of breast cancer seen in gov/public/factrbk96/c10.htm). Asian society, in Chinese women. general, is conservative and Chinese women are In the past twenty years, breast cancer inci- shy about discussing their breasts with their phy- dence in Chinese women has increased dramati- sicians; therefore, the earlier ages of breast can- cally, with a total increase of 50–100%. Epide- cer cannot simply be explained by earlier detec- miological findings to explore risk factors sug- tion due to public health policies for early gest that breast cancer in Chinese women is re- screening or younger women requesting asymp- productive hormone dependent and is similar tomatic screening. Instead, the evidences of ear- with respect to hormonal-reproductive risk fac- lier tumor onset may be contributed by two non- tors to that in high-incidence areas [1,3]. There- mutually exclusive possibilities. The first possi- fore, given the similarity of patterns and the bility suggests a cohort effect, which indicates magnitude of association between major risk that the contribution due to recent adaptation to factors and breast cancer found in Chinese Western life-style leading to higher estrogen ex- women, the recent trend of earlier menarche and posure may be relatively more pronounced an increasing proportion of women with post- among younger women. This possibility gains menopausal obesity due to modern affluence will support from a more dramatic increase of breast provide a partial explanation for recent steep in- cancer incidence in younger women, being crease in breast cancer incidence. Furthermore, 87.1% and 85.4% in the age groups 35–44 and given the important role of breast feeding to pre- 45–54, respectively (as compared to an increase vent breast cancer, a dramatically decreasing of 50% in all women) . The second possibility trend (from 90% breast feeding in the 60’s to suggests that genetic predisposition may be more 30% in the 80’s in Taiwan) may be specifically important in Taiwanese and Chinese breast can- correlated to the increasing trend of breast can- cers than in others. Unique p53 mutation pat- cer. Interestingly, though the incidence is low, tern/site  and novel regions of genomic dele- breast cancer has remained one of the most tion  have been found in breast cancers in common cancers in women in Taiwan and China these population, suggesting Chinese breast can- and is among the leading causes of cancer mor- cer may genetically differ from Western breast tality [1,2]. Cervical cancer is the leading cancer cancer. On the other hand, a few studies have in women in these populations, but, consistently, attempted to seek for germ-line mutation in high- the incidence has begun to decrease recently in penetrant genes, including BRCA1 and BRCA2, Taiwan and in China. In the future, it is expected in Chinese women or breast cancers [8,9], but no that breast cancer will replace cervical cancer as conclusive findings are available to examine the the most common female cancer. These figures contribution of these genes to early tumor onset. strongly indicate that breast cancer is an issue of particular public health and clinical importance. It is intriguing that Taiwanese and Chinese BREAST CANCER SCREENING breast cancer is characterized by younger age at PROGRAM IN CHINESE POPULATIONS tumor onset. More than 50% of total breast can- cer diagnosed annually in Taiwan is composed of In Taiwan, in order to combat the increasing patients younger than 50 years of age, and this threat of breast cancer, the Department of Health proportion is higher than those observed in West- has recently launched a 3-year breast cancer ern populations . Moreover, the incidence screening campaign. During this period, a total of ratio of younger (less than 50 years) versus older 1,000,000 women aged 35 years or older will (50 years and older) breast cancer in Taiwan is receive a physical examination by public health 0.25, more than two-fold that (0.09) in the United nurses; in the event of suspicious findings, the States (calculated based on the publication of the examinee will be referred to a hospital for further National Cancer Institute, USA, shown in the examinations, including breast ultrasound and/or following Internet address:http://www.nci.nih. mammography. This is not the first action taken C.-S. Huang et al. / Breast Cancer Screening in Taiwan and China 43 by the Department of Health in Taiwan to combat method for the early detection of breast cancer. breast cancer. Between 1992 and 1997, a pro- Therefore, mammography and breast ultrasound gram was carried out to screen mothers, daugh- were added to the screening modalities in Taiwan ters, sisters, and grandmothers of breast cancer when the high-risk group was screened for breast patients identified at hospitals . The partici- cancer. Doubts about self-examination were pants underwent annual screening using a combi- supported by the results of a randomized trial of nation of physical examination, mammography, breast self-examination in Shanghai , in and breast ultrasound. Female relatives of breast which about 225,000 Chinese women were ran- cancer patients were selected for screening be- domized to either a self-examination instruction cause of concern about cost-effectiveness. Al- group or a control group. After 5 years of fol- though mammographic mass screening for breast low-up, the number of breast cancers detected in cancer has proven to be effective in lowering the two groups was equal, and the breast cancers mortality in countries with a high incidence of detected in the self-examination group were not breast cancer , the application of mass diagnosed at an earlier stage or at a smaller size screening to a low incidence area such as Taiwan than those in the control group. Cumulative may not be worthwhile in terms of cost- breast cancer mortality rates during the 5 years effectiveness. Screening high-risk women, there- from entry into the trial also were almost identi- fore, was considered a better approach. The re- cal in the two groups. An interesting, yet not sults showed that the mean sojourn time (duration surprising, finding in this trial was that more be- of the pre-clinical screen-detectable phase) in this nign breast lesions were detected in the self- high-risk group was shorter than that in females examination group than in the control from the general population. Annual screening group. Benign breast tumors are better defined of this high-risk group was estimated to result in and easier to detect than malignant lesions, which a significant 33% reduction in breast cancer further explains why self-examination or exami- mortality compared with a nonsignificant 20% nation by physicians fails to detect many cancers reduction using 2-yearly and 3-yearly screening at an early stage. Although the follow-up period regimes, respectively . The cost-per-year of was still short and the benefits of self- lives saved by mass screening is estimated to be examination have probably not yet emerged, the 15 times higher that involved in screening the results justify consideration of other screening high-risk group. Although screening of the high- modalities. risk group is cost effective, we must keep in mind Two population-based breast cancer screenings that it probably only covers 5% of the incidence using mammography, with or without clinical of breast cancer. examination, were conducted in Singapore and As Taiwan is an area of low breast cancer in- Hong Kong, with most of the participants being cidence, the Department of Health’s policy for Chinese women [13,14]. As in Taiwan, Singa- breast cancer prevention has been breast self- pore is experiencing an increasing incidence of examination, which does not involve a large fi- breast cancer. A randomized trial of screening nancial outlay. The theory was that women using mammography without clinical examina- might be able to detect subtle changes in their tion in women aged 50–64 years was started in own breasts that might be missed on clinical 1994 . For every 1,000 women screened, 4.8 breast examination by doctors. Since Taiwanese cancers were detected. The percentage of stage 0 women have relatively small breasts, it was as- and stage 1 cancers was 64% in screened women, sumed that self-examination would lead to earlier compared to 26% in non-screened women. detection of breast cancer. However, many When only invasive cancers were considered, breast cancers were diagnosed in the late stage. 65% of cancers detected in screened women were Although many factors contributed to this late node-negative, compared with 47% in non- discovery of cancer , many clinicians screened women. As this trial only screened doubted that self-examination was an effective women aged 50–64 years, it cannot give any in- 44 C.-S. Huang et al. / Breast Cancer Screening in Taiwan and China formation on whether, or how, younger women women in this age group [21–23]. Most people should be screened. who are against annual screening for women in Clinical breast examination and mammogra- their 40s believe that both the incidence and de- phy were used in another mass screening pro- tection rate of breast cancer in younger women is gram in Hong Kong between 1993 and 1995 . lower . A total of 13,033 women aged 40–70 years were Although the reasons for the discrepancy in screened, and 8,504 women underwent mammog- screening efficacy between different age groups raphy. In all, 42 cancers were detected, 16 of are not well understood, different tumor biology which were nonpalpable. Four palpable cancers and mammographic test characteristics in were not detected by mammography. The cancer younger women must be considered . The detection rate was 4.61/1,000 in the 40–49 year doubling time for breast cancer in women under age group and 6.46/1,000 in the over 50 year age the age of 50 is 80 days, while that in women group. When these two trials were compared, aged 50 to 70 is 157 days . The Swedish mammography screening plus clinical examina- two-county trial found that the tumor types tion achieved a better detection rate than mam- tended to be more aggressive in younger women mography alone in women over the age of 50. . These screening results suggest that these This finding is consistent with the concept that a women are less likely to benefit. The lower screening program using both mammography and benefit from screening women aged 50 years or clinical examination can achieve a higher sensi- less, compared to those over 50 years, is due to a tivity than either modality alone . The lower shorter sojourn time in the younger age group detection rate in younger women raises the ques- . It has also been suggested that screening tion whether the incidence in this age group is every 2 or 3 years might be sufficient in women lower (not the case in Taiwan, and probably not aged 50 years or older, while annual screenings in Hong Kong) or the sensitivity of mammogra- might be required in women aged 40 to 49 years phy in younger women is lower. Mammography . As a high incidence of interval cancers rep- alone failed to detect 15% of breast cancers iden- resents either rapid progression of breast tumors tified by clinical examination, and the sensitivity or poor sensitivity of the screening method, the was lower in younger women . Thus, com- shorter sojourn time indicates a rapid progression bined clinical examination and mammography is of breast tumors in younger patients, which ne- desirable. cessitates a shorter screening interval or a more sensitive screening modality. The ability of mammography to detect breast cancer is affected CONTROVERSY OF MAMMOGRAPHY by breast density [29,30], its sensitivity being SCREENING IN YOUNG WOMEN 80% in women with fatty breasts and only 30% in women with extremely dense breasts . Although it is well documented that mass After adjustment for age, menopausal status, use screening using mammography lowers the mor- of hormone replacement therapy, and body mass tality of breast cancer among women over the age index, the odds ratio for interval cancer in women of 50, women aged 40 to 49 do not appear to with extremely dense breasts compared with benefit to the same extent. The reduction in those with fatty breasts is about 6. In regard to breast cancer mortality varied from 18% to 35% age and breast density, more premenopausal than for women who began screening between the postmenopausal women have dense breasts . ages of 35 and 49 years [16–19]. Therefore, The chance of having biopsy after mammography mammographic screening is recommended every examination is also different between young [40– 1–2 years for all women in their 40s , while 49] and old (50 and older) age groups. One study the American College of Radiology, the Ameri- noted that the probability of abnormal mammo- can Cancer Society, and the American Medical grams at first screening is the same in different Association advise annual screening for all age groups, but the positive predictive value of C.-S. Huang et al. / Breast Cancer Screening in Taiwan and China 45 screening mammography declines from about are more likely to be found with sonogra- 18% in women older than 60 to 4% in women phy. One study demonstrated that 0.3% of aged 40 to 49 , resulting in more unnecessary 12,706 examinations performed by ultrasound, or intervention procedures in women younger than 2.8% of 1,575 lesions detected, were cancers that 50 years. were detected by ultrasound but not by mammog- raphy or physical examination . The EGBCS concludes that the high rates of false-positive CURRENT BREAST CANCER SCREENING outcomes would lead to unnecessary further in- POLICY IN TAIWAN: MAMMOGRAPHY vestigation. However, in this particular study, AND ULTRASOUND ultrasound was used to examine contralateral or ipsilateral breasts of cancer patients for multifoci As mentioned above, in contrast to Western lesions, in addition to primary cancer. The prob- countries, breast cancer in Taiwan is character- ability of finding another lesion is expected to be ized by an earlier onset, with peak incidence at low, and the intention of the examiner to biopsy age 40–49, and more than 50% of breast cancers the detected lesion in these high-risk patients will arising in premenopausal women. Therefore, the be high. Thus, the cancer detection rate was low dilemma of breast cancer screening in Taiwan and the false-negative rate high in this study. In can be summarized by Taiwan having a rapidly another prospective study, when breast ultra- increasing incidence of breast cancer, especially sound was used to screen 3,626 women with in premenopausal women, while the incidence is dense breasts and normal mammographic and still low compared to Western countries. Since physical examination findings, 11 (0.3%) were premenopausal women have denser breasts than found to have cancers . These cancers, iden- postmenopausal women, and Taiwanese women tified by ultrasound alone, did not differ in terms have smaller breasts and a higher percentage of of tumor size and stage from nonpalpable cancers dense breasts, increasing the frequency of mam- detected by mammography and were smaller and mography screening may not be the sole solution at a lower stage than palpable breast cancers. In to increasing the detection rate in this age group. women with dense breasts, overall cancer detec- The addition of other screening modalities, such tion increased by 17%, and the number of tumors as breast ultrasound, may be helpful, as breast detected only by imaging increased by 37%. ultrasound can detect some breast cancers in These results support the idea that ultrasound is dense breasts that are missed by mammography very helpful in detecting nonpalpable cancers [30,32]. missed by mammography of dense breasts, more A consensus statement by the European Group frequent in premenopausal women. for Breast Cancer Screening (EGBCS) is, how- One retrospective study in Japan reported the ever, against the use of ultrasound in population use of breast ultrasound in a non-randomized screening at any age due to the high rates of both mass screening . In one group of 15,935 false-positive and false-negative results associ- women, only physical examination was per- ated with mass screening using breast ultrasound formed, and 5 breast cancers were detected. In . Many of the false-negative results are due another group of 18,539 women, ultrasound plus to difficulty in microcalcification detection using physical examination was performed, and 22 breast ultrasound, although we found that it is not cancers were detected. Sixteen of these 22 impossible to detect microcalcification, not nec- women had early breast cancer and 13 (59.1%) of essarily associated with mass, in nonpalpable the 22 cancers were not palpable. Of the 22 breast lesions using this method . Others women, 50% were younger than 50 years. Al- have reported that they could use ultrasound to though the overall cancer detection rate was low, localize microcalcifications detected by mam- which might be due to the low incidence of breast mography . Since mammography is not usu- cancer in Japan or the low sensitivity of the ally requested in younger women, calcifications screening modalities, breast ultrasound is useful 46 C.-S. Huang et al. / Breast Cancer Screening in Taiwan and China in mass screening for the detection of early breast the sensitivity. The mean time for performing a cancers, many of which will be in women under bilateral ultrasound screening is about 4 minutes the age of 50 and will be missed by physical ex- (range 1.5–9.75 minutes), which is not too time- amination. One criterion for evaluating screening consuming for mass screening. Two additional efficacy is that more than 50% of screen-detected advantages of breast ultrasound are that no radia- cancers should be smaller than 15 mm , and tion is involved and no pain is felt during exami- breast ultrasound should be able to do this. nation. Although the radiation risk from mam- If the main limitation of ultrasound in mass mography screening is probably not too high to screening is the difficulty in detecting DCIS asso- prohibit women under the age of 50 from benefit- ciated with microcalcifications, the significance of ing from screening, ultrasound involves no radia- DCIS needs to be clearly understood. Although tion, allowing more frequent screening in a short many pathological and molecular biological stud- interval. In addition, in contrast to mammogra- ies suggest that many cases of DCIS will progress phy, since ultrasound examination does not in- to invasive carcinoma if undetected or untreated volve pain, compliance will be better, especially in [38,39], it is estimated that only 30-50% of DCIS Chinese women with smaller and denser breasts. will progress to invasive cancer, with the remain- In Taiwan, a randomized mass screening has der regressing or remaining indolent [40,41]. In been proposed recently to investigate the role of addition, it is not documented whether all invasive breast ultrasound in breast cancer screening. The carcinomas arise from in situ cancers. In the first results should be able to define whether breast ul- mammography screening conducted on 1,000 trasound is useful in reducing breast cancer mor- women aged 40 to 49 and first screened by mam- tality in women aged 40–49. mography, 1.5 cases were DCIS and 1.5 cases in- vasive cancer, compared with 2 DCIS and 7 inva- sive cancer for every 1,000 women aged 50 to 69 SUMMARY . In the Sickles series, in which most of the patients were Caucasians, microcalcifications sug- As the incidence and mortality of breast cancer gesting malignancy constituted 42% of nonpalpa- has increased in Chinese women, there is a need to ble breast cancers . In our series, nearly all the initiate a more effective breast cancer screening nonpalpable cancers detected by mammography in program. Unlike Western countries, a higher pro- Taiwanese women were due to the presence of portion of premenopausal breast cancer is found in microcalcifications. Ultrasound is probably more our population, and Chinese women have denser sensitive than mammography in detecting nonpal- breasts. To solve this particular problem, we be- pable cancers; however, it fails to detect microcal- lieve that increasing only the frequency of mam- cifications. Mammography is the best tool to de- mography screening may not be the ideal solu- tect DCIS associated with microcalcifications, tion. However, it is possible that the addition of which probably will not progress rapidly to inva- breast ultrasound will be helpful. A randomized sive cancer; therefore, it can be performed over a screening trial has been proposed recently in Tai- longer interval, with ultrasound being added to wan to address the role of breast ultrasound in detect nonpalpable cancers not associated with breast cancer screening. The results should be microcalcifications. able to determine whether it will be useful in re- Although the sensitivity and specificity of ducing breast cancer mortality in women aged 40– breast ultrasound is highly operator- and machine- 49. dependent, and screening the whole breast is time consuming, the examiner will gain experience in the same way that every physician learns to master REFERENCES the use of the stethoscope. As the examiner be- comes more experienced, biopsy need not be per-  Yang PS, Yang TL, Liu CL, Wu CW, Shen CY. A case-control study of breast cancer in Taiwan- a low formed on all solid masses  without affecting incidence area. Br J Cancer 1997; 75:752–756. C.-S. Huang et al. / Breast Cancer Screening in Taiwan and China 47  Jin F, Devesa S, Chow W, Zheng W, Ji B, Fraumeni  Hendrick RE, Smith RA, Rutledge JH III, Smart CR. JF, Jr., Gao, YT. Cancer incidence trends in urban Benefit of screening mammography for women aged Shanghai, 1972–1994: an update. Int J Cancer 1999; 40–49: a new meta-analysis of randomized controlled 83:435–440. trials. J Natl Cancer Inst Mono 1997; 22:87–92.  Gao YT, Shu XO, Dai Q, et al. Association of men-  Smart CR, Hendrick RE, Rutledge JH III, Smith RA. strual and reproductive factors with breast cancer risk: Benefit of mammography screening in women ages results from the Shanghai Breast Cancer Study. Int J 40–49 years: current evidence from randomized con- Cancer. 2000; 87:295–300. trolled trials. Cancer 1995; 75:1619–1626.  Bernstein L, Yuan JM, Ross RK, et al. Serum hor-  Falun Meeting Organizing Committee and Collabo- mone levels in pre-menopausal Chinese women in rators. Breast-cancer screening with mammography in Shanghai and white women in Los Angeles: results women aged 40–49 years: Swedish Cancer Society from two breast cancer case-control studies. Cancer and the Swedish National Board of Health and Wel- Causes Control 1990;1:51–58. fare — report of the organizing committee and col-  Huang CS, Chern HD, Chang KJ, Cheng CW, Hsu laborators, Falun Meeing, Falun, Sweden (21 and 22 SM, Shen CY. Breast cancer risk associated with March 1996). Int J Cancer 1996; 68:693–699. genotype polymorphism of the estrogen-metabolizing  Andersson I, Janzon L. Reduced breast cancer mor- genes, CYP17, CYP1A1 and COMT- a multigenic tality in women under age 50: updated results from study on cancer susceptibility. Cancer Res 1999;59: Malmö Mammographic Screening Program. J Natl 4870–4875. Cancer Inst Mono 1997; 22:63–67.  Lo YL, Yu JC, Huang CS, et al. Allelic loss of the  Eastman P. NCI adopts new mammography screening BRCA1 and BRCA2 genes and other regions on 17q guidelines for women. J Natl Cancer Inst 1997; and 13q in breast cancer among women from Taiwan- 89:538–540. the area of low incidence but earlier onset. Int J Can-  Feig SA, D’Oris CJ, Hendrick RE, et al. American cer 1998; 79:580–587. College of Radiology guidelines for breast cancer  Lou MA, Tseng SL, Chang SF, et al. Novel patterns screening. AJR Am J Roentgenol 1998; 171:29–33. of p53 abnormality in breast cancer from Taiwan: ex-  Leitch AM, Dodd GD, Costanza M, et al. American perience from a low incidence area. Br J Cancer 1997; Cancer Society guidelines for the early detection of 75:746–751. breast cancer: updated 1997. CA Cancer J Clin 1997;  Tang NL, Pang CP, Yeo W, et al. Prevalence of mu- 47:150–153. tations in the BRCA1 gene among Chinese patients  Stapleton S. House: annual mammograms at 40. with breast cancer. J Natl Cancer Inst. 1999; 91:882– American Medical News 1999; 42:27–29. 885.  Kerlikowske K, Grady D, Barclay J, Sickles EA, Ea-  Sng JH, Chang J, Feroze F, et al. The prevalence of ton A, Ernster V. Positive predictive value of screen- BRCA1 mutations in Chinese patients with early on- ing mammography by age and family history of breast set breast cancer and affected relatives. Br J Cancer cancer. JAMA 1993; 270:2444–2450. 2000; 82:538–542.  Sirovich BE, Sox HC Jr. Breast cancer screening.  Lai MS, Yen MF, Kuo HS, Koong SL, Chen TH, [Review] [92 refs] Surg Clin North Am 1999; 79:961– Duffy SW. Efficacy of breast-cancer screening for 990. female relatives of breast-cancer-index cases: Taiwan  Peer PGM, van Dijck JAAM, Hendriks JHCL. Age multicentre cancer screening (TAMCAS). Int J Can- dependent growth rate of primary breast cancer. Can- cer 1998; 78:21–26. cer 1993; 71:3547–3551.  Chie WC, Chang KJ. Factors related to tumor size of  Tabar L, Fagerberg G, Chen HH, et al. Efficacy of breast cancer at treatment in Taiwan. Prev Med 1994; breast cancer screening by age: New results from the 23:91–97. Swedish Two-County trial. Cancer 1995; 75:2507–  Thomas DB, Gao DL, Self SG, et al. Randomized trial 2517. of breast self-examination in Shanghai: methodology  Duffy SW, Chen HH, Tabar L, Fagerberg G, Paci E. and preliminary results. J Natl Cancer Inst 1997; Sojourn time, sensitivity and positive predictive value 89:355–365. of mammography screening for breast cancer in  Ng EH, Ng FC, Tan PH, et al. Results of intermediate women age 40-49. Int J Epid 1996; 25:1139–1145. measures from a population-based, randomized trial of  Mandelson MT, Oestreicher N, Porter PL, et al. Breast mammographic screening prevalence and detection of density as a predictor of mammographic detection: breast carcinoma among Asian women: the Singapore comparison of interval- and screen-detected cancers. J Breast Screening Project. Cancer 1998; 82:1521–1528. Natl Cancer Inst 2000; 92:1081–1087.  Chan LK, Lam HS, Chan ES, et al. Mammogram  Kolb TM, Lichy J, Newhouse JH. Occult cancer in screening of Chinese women in Kwong Wah Hospital, women with dense breasts: detection with screening Hong Kong. Australasian Radiol 1998; 42:6–9. US—diagnostic yield and tumor characteristics. Radi-  Baines CJ, Miller AB. Mammography versus clinical ology 1998; 207:191–199. examination of the breasts. J Natl Cancer Inst Mono  Kerlikowske K, Grady D, Barclay J, Sickles EA, Ea- 1997; 22:125–129. ton A, Ernster V. Positive predictive value of screen- 48 C.-S. Huang et al. / Breast Cancer Screening in Taiwan and China ing mammography by age and family history of breast  Smith RA. Screening fundamentals. J Natl Cancer Inst cancer. JAMA 1993; 270:2444–2450. Mono 1997; 22:15–19.  Gordon PB, Goldenberg SL. Malignant masses de-  Feig SA. Age-related accuracy of screening mammog- tected only by US. A retrospective review. Cancer raphy: how should it be measured? Radiology 2000; 1995; 76:626–630. 214:633–640.  Teh W, Wilson AR. The role of ultrasound in breast  Ernster VL, Barclay J. Increases in ductal carcinoma cancer screening. A consensus statement by the Euro- in situ (DCIS) of the breast in relation to mammogra- pean Group for Breast Cancer Screening. Eur J Can- phy: a dilemma. J Natl Cancer Inst Mono 1997; cer 1998; 34:449–450. 22:151–156.  Huang CS, Wu CY, Chu JS, Lin JH, Hsu SM, Chang  Page DL, Dupont WD, Rogers LW, Landenberger M. KJ. Microcalcifications of non-palpable breast lesions Intraductal carcinoma of the breast: Follow-up after detected by ultrasonography: correlation with mam- biopsy only. Cancer 1982; 49:751–758. mography and histopathology. Ultrasound Obstet Gy-  Rosen PP, Braun DW, Kinne DE. The clinical signifi- necol 1999; 13:431–436. cance of pre-invasive breast cancer. Cancer 1980;  Leucht WJ, Leucht D, Kiesel L. Sonographic demon- 46:919–925. stration and evaluation of microcalcifications in the  Sickles EA. Mammographic features of 300 consecu- breast. Breast Disease 1992; 5:105–123. tive non-palpable breast cancers. AJR Am J Roentge-  Okamoto H, Ogawara T, Arihara F, et al. Usefulness nol 1986; 146:661–663. of ultrasonography combined with conventional  Stavros AT, Thickman D, Rapp CL, Dennis MA, physical examination in mass screening for breast Parker SH, Sisney GA. Solid breast nodules: use of cancer: a retrospective study of Yamanashi Health sonography to distinguish between benign and malig- Care Center results from 1989 to 1994. Jap J Cancer nant lesions. Radiology 1995; 96:123–134. Res 1996; 87:317–323.