Preventive Medicine 30, 174–177 (2000)doi:10.1006/pmed.1999.0595, available online at http://www.idealibrary.com on

Changing Patterns of Breast Cancer Stage at Diagnosis in SouthernItaly: Hospital Data as Indicators of Progressive Changes

Maurizio Montella, M.D.,*,1 Marina Buonanno, B.Sc.,* Edoardo Biondi, M.D.,† Anna Crispo, St.,*Mariarosaria De Marco, M.D.,* Mario Tamburini, M.D.,* Gerardo Botti, M.D.,* Gabriella Fabbrocini, M.D,‡

Immacolata Capasso, M.D.,* and Giuseppe D’Aiuto, M.D.*

*National Cancer Institute. Via Mariano Semmola, 80131 Naples

P

and Oncology and ‡Department of Biology and Cellular and Mole

“Federico II” University, Via

Background. In southern Italy diagnostic delay inbreast cancer patients has been demonstrated to berelated to the level of education and residency in ruralareas. In order to verify whether late breast cancerdiagnosis was actually in decline as a result of improv-ing socioeconomic conditions and ongoing preventionprograms, we evaluated clinical data from the tumorregistry of the National Cancer Institute, Naples.

Methods. Four thousand two hundred forty consecu-tive breast cancer patients admitted to our institutionfrom 1986 to 1997 were grouped into four 3-year periodsaccording to their admission date. Using multiple logis-tic regression, x 2 for trend and b-coefficient were cal-culated in each pT and pN categories in order to dis-cover the trend for the 1986–1997 period.

Results. A progressive, statistically significant de-crease in the number of patients with advanced cancerat the time of diagnosis was observed over the studyperiod. In particular, x 2 values for trend for each pTcategory, over the study period, were pT1 119.4 (P ,

0.001) with positive b-coefficient, pT2 13.4 (P 5 0.003)with negative b, and pT3–pT4 152.2 (P , 0.001) withthe strongest negative b.

Conclusions. Changing patterns of breast cancerstage at diagnosis have been demonstrated in womenliving in Southern Italy. They are consistent with anincreasing orientation toward prevention. Data fromhospital tumor registries are a useful source of infor-

mation on diagnostic delay. q 2000 American Health Foundation

and Academic Press

Key Words: breast cancer; staging; prevention; diag-nostic delay; tumor registry.

1 To whom reprint requests should be addressed. Fax: 0039-81-5466888. E-mail: legamont@tin.it.

1

, Italy; and †Department of Molecular and Clinical Endocrinologycular Pathology, Medical Statistical Division, Faculty of Medicine,ansini 5, 80131 Naples, Italy

INTRODUCTION

The comparative analysis of data from different tu-mor registries shows that the incidence rate of breastcancer in southern Italy is lower than in northern Italyand in northern Europe. In fact, in the period 1988–1992 the incidence rate ranged between 44.1/100,000(Ragusa) and 50.4/100,000 (Latina) in the south of thecountry and from 64.3/100,000 (Trieste) to 73.9/100,000(Parma) in northern Italy. Although less pronounced,southern mortality rates for the same period are evenlower than those in the north (Ragusa 35.7, Trieste41.3) [1]. Moreover, in 1990, the breast cancer incidencerate was 68.1/100,000 in the United Kingdom and 72.9/100,000 in Sweden [2].

Contrasting with such favorable epidemiological con-ditions, diagnostic delay in breast cancer patients hasbeen demonstrated in southern Italy as being relatedto low educational levels and residence in rural commu-nities [3]. However, over the past decades, and espe-cially in recent years, there has been a phenomenon ofprogressive sociocultural and economic homogenizationbetween urban and rural areas in southern Italy. Sucha reduction in the gap between urban and rural areascan be explained by the following factors: extension ofcompulsory schooling (an annual increase of 1.6% ofpeople with more than 5 years school attendance wasobserved), increased impact of mass media, and im-provements in economic conditions [4–7].

Moreover, in the past 15 years the European Commu-nity has carried out the “Europe Against Cancer” pro-

gram with the aim of reducing breast cancer mortalityby the year 2000 through health and prevention cam-paigns together with training programs for health oper-ators [8–10].

74 0091-7435/00 $35.00Copyright q 2000 by American Health Foundation and Academic Press

All rights of reproduction in any form reserved.

BREAST CANCER AND DIAGNOS

One would expect that the above mentioned mea-sures would lead to the progressive decline of the diag-nostic delay of breast cancer. The study of the changingstage patterns at admission has been previously demon-strated to be a useful tool in the verification of theactual decline in diagnostic delay for breast cancer [11].Such an investigation was carried out in our institution,

one of southern Italy’s two National Cancer Institutes.We considered all consecutive admissions for breastcancer from the period 1986–1997. The task was madeeasier by the fact that since 1986 we have had thehospital tumor registry data at our disposal.

MATERIALS AND METHODS

Four thousand two hundred forty consecutive femalepatients living in the Campania Region of southernItaly (surface area 13,595 km2, 551 municipalities, 20-to 89-year-old female population: 2,078,254) who hadreceived surgery for histologically confirmed breastcancer (ICD 174.0–174.9) [12] between January 1986and December 1997 were selected from the local hospi-tal tumor registry [13]. Data were gathered from hospi-tal files and by directly interviewing patients.

Patients were staged by pTNM (pathological classifi-cation of malignant tumors) [14], which remained con-stant over the time period considered. The specimenswere analyzed by two different pathologists, a compari-son of the results was made, and further considerationwas given to those cases that had an ambiguous inter-pretation.

Three groups of tumor size were considered: tumorsize # 2 cm [pT1], tumor size more than 2 cm butless than 5 cm [pT2], and tumors larger than 5 cm orinfiltrating adjacent structures [pT3–pT4] independentof regional lymph node status. A separate analysis wasmade for patients grouped as pN0 (no metastatic re-gional lymph nodes) versus pN1 (metastatic regionallymph nodes) for all T dimensions. The other variablesexamined were age (grouped as #40, 41–60, and .60),menopausal status (premenopausal, postmenopausal),education (illiterate, school #5 years, school .5 years,unknown), marital status (married, widowed, unmar-ried, divorced), residence (Naples, other municipalitiesin its province, other provinces in Campania), and his-tology type (ductal, lobular, mucinous, papillary, medul-lary, others).

The time of admission was stratified in 3-year periods(1986–1988, 1989–1991, 1992–1994, and 1995–1997).A x2 test for trends was calculated for each pT category,using multiple logistic regression, adjusted for terms

of age, menopausal status, and histology [15–17], andb-coefficient, which permitted a direct comparison be-tween standardized explanatory variables and theirpredictive capacity of the variable response value (theb-coefficient indicates the slope of the curve: for positive

TIC DELAY IN SOUTHERN ITALY 175

values the trend increases, and for negative values thetrend decreases).

RESULTS

Seven hundred forty-two women were admitted toour institution between 1986 and 1988, 841 between1989 and 1991, 1155 between 1992 and 1994, and 1502between 1995 and 1997.

The distribution of sociodemographic data for the ex-amined variables is reported in Table 1. Tables 2 and3 show the distribution of patients in groups of tumorsize and nodal involvement over the admission period.

A progressive, statistically significant decrease in pa-tients with advanced cancer at diagnosis was observedthrough the period 1986–1997 by analyzing both tumorsize (Table 2) and nodal involvement (Table 3). Thetumor-size group better reflects the possibility of earlydiagnosis through self-examination and thereforeyields results more relevant to diagnostic delay. Thephenomenon is particularly impressive when compar-ing the first with the last triennium.

In fact, in the first triennium pT1 cases were 48.9%,pT2 34.6%, and pT3–pT4 16.4%, while in the last trien-nium pT1 rose to 70.6%, pT2 decreased to 26.4%, and

pT3–pT4 decreased to 3.0%. The x2 values for trendwere calculated over the study period (1986–1997) inthe various pT categories: pT1 5 119.4 (P , 0.001),pT2 5 13.4 (P 5 0.003), and pT3–pT4 5 152.3 (P ,

TABLE 1

Total Patient Sociodemographics

No. %

Age at diagnosis (median: 55.0)#40 393 9.241–60 2084 49.2.60 1763 41.6

Menopausal statusPremenopausal 1519 35.8Postmenopausal 2721 64.2

EducationIlliterate 217 5.1School #5 years 2115 49.9School .5 years 1865 44.0Unknown 43 1.0

Marital statusMarried 3019 71.2Widowed 691 16.3Unmarried 428 10.1Divorced 102 2.4

ResidenceNaples (chief town) 1354 31.9Naples (province) 1699 40.1

Other provinces 1187 28.0

HistologyDuctal 3137 74.0Lobular 959 22.6Others 144 3.4

x2 (trend) 119.4 13.4 152.2

b

P value 0.0001 0.0003b-coefficient 0.31 20.11

a Statistics: the x2 value for trend and b-coefficient were estimatedfor age, menopausal status, and histology.

0.001), respectively (Table 2). Moreover, in the first 3-year period N0 cases were 51.9% and N1 cases were

176 MONTELLA ET AL.

TABLE 2

Distribution and Summary Statisticsa of pT Categories for Year of Diagnosis of Breast Cancer in Southern Italy

pT1 pT2 pT3 and pT4 Total

Years No. % No. % No. % No. %

1986–1988 363 48.9 257 34.6 122 16.4 742 100.01989–1991 463 55.1 248 29.5 130 15.5 841 100.01992–1994 736 63.7 334 28.9 85 7.4 1155 100.01995–1997 1061 70.6 396 26.4 45 3.0 1502 100.0Total 2623 61.9 1235 29.1 382 9.0 4240 100.0

48.1%, while in the last triennium N0 cases rose to62.9% and N1 cases decreased to 37.1%. The x2 valuefor trend was 35.9 (P , 0.001) in both N0 and N1; thevalue of the b-coefficient indicates an increase in theN0 category (b 5 0.17) (Table 3).

DISCUSSION

As our results show, the percentage of breast cancerpatients with large-size tumors (pT3–pT4) admitted toour institute has dramatically declined in recent years.A similar phenomenon was recorded for nodal involve-ment.

Italian data show that diagnostic delay is an im-portant determinant of stage at diagnosis in women

P value 0.0001b-coefficient 0.17

a Statistics: x2 value for trend and b-coefficient were estimated byunconditional multiple logistic regression equation including termsfor age, menopausal status, and histology.

0.000120.63

y unconditional multiple logistic regression equation including terms

campaigns and improved access to medical and surgicalconsultation, may have contributed to this.

Prevention Campaigns

Two main prevention campaigns have recently in-volved the female population of the Campania Region.They are the Europe Against Cancer and the Wvivaprograms. The former (“Europe Against Cancer”) pro-gram started in 1985 and supported the fight againstcancer, in particular against breast cancer, in memberstates of the European Community [19,20]. The Wvivaprogram was a public mammographic screening cam-paign designed to target the 50- to 69-year-old femalepopulation living in the Campania Region, recommend-ing and providing a biannual mammography [21].

Both the prevention programs and the resulting in-crease of mass media attention have contributed to thediffusion of a prevention culture in southern Italy.

Improved Access to Hospitals

The observed increase in admissions through theyears does not reflect an increased incidence, but ratherthe progressive increased availability of surgical beds inthe institute, the only specialized institution for cancertreatment in the region.

The series examined was taken from the hospitaltumor registry. Even though such data are not derivedfrom a population registry, one must take into accountthat our institution admits a large percentage (morethan 30%) of all breast cancer cases in the CampaniaRegion (female population: 2,881,327, ISTAT 1991; esti-mated new cases per year, 1300) [22,23]. The regionalmortality rate has passed from a standardized rate of

with breast cancer [18]. It is reasonable to argue thatthis progressive decline of stage at diagnosis arises froma progressive decline of diagnostic delay in southernItaly. Several factors, such as secondary prevention

TABLE 3

Distribution and Summary Statisticsa of pN Categories for Yearof Diagnosis of Breast Cancer in Southern Italy

pNO pN1 Total

Years No. % No. % No. %

1986–1988 385 51.9 357 48.1 742 100.01989–1991 447 53.2 394 46.8 841 100.01992–1994 734 63.5 421 36.5 1155 100.01995–1997 945 62.9 557 37.1 1502 100.0Total 2511 59.2 1729 40.8 4240 100.0

x2 (trend) 35.9

28.9 in 1980 to a rate of D1 31.2 in 1993 [24]. Theexamined sample should, therefore, be representativeof the entire population.

If this positive trend continues into the near futurethe rate of late diagnoses could become marginal, as it

BREAST CANCER AND DIAGNOST

is in the rest of Italy, bringing about a consequent in-crease in relative survival [25–27].

The role of ongoing breast cancer prevention pro-grams could, however, be enhanced by an increasedawareness and the involvement of patients and physi-cians [28].

CONCLUSIONS

Changing patterns of breast cancer stage at diagnosishave been demonstrated in women living in the southof Italy. They are consistent with an increasing orienta-tion toward prevention. Over the past years (1986–1997) the most significant data concern tumors oper-ated on while they were in their initial phase (T1passing from 363 in the 1986–1988 period to 1061 inthe 1995–1997 period). This was due to the informationcampaigns and secondary prevention measures re-ported in the study. The number of most anatomicallyadvanced tumors (T3–T4, which went from 122 in the1986–1988 period to 45 in the 1995–1997 period) ismarginal compared with the increase of T1. Because ofthe lack of recent data it is not possible to verify theimpact on mortality.

Tumor registry data, when reliable and analyzed in

Italian]. Rome: Ediesse, 1998.

a correct and appropriate manner, can be a usefulsource of information in the evaluation of epidemiologi-cal trends. Considering the low costs of such registriesand the possibilities that information technology offersin the implementation of data banks together with thehighly informative value of such studies, the authorsfirmly believe that there is a case for increasing invest-ment in the resources used in collecting data on routinerecoveries both in Italy and abroad.

REFERENCES

1. Zanetti R, Crosignani P, Rosso S. Il cancro in Italia [in Italian].Vol II, 1988–1992; Roma: Il Pensiero scientifico, 1997.

2. Parkin DM, Muir CS, Whelan SL, Ferlay J, Raymond L, YoungJ (editors). Cancer incidence in five continents, vol VII. Lyon:IARC Scientific Publication, 1997.

3. Montella M, Biondi E, De Marco MR, Botti G, Tatangelo F, Ca-passo I, et al. Sociodemographic factors associated with the diag-nostic staging of breast cancer in Southern Italy. Cancer 1995;76:1585–90.

4. Istituto Nazionale di Statistica. Annuario Statistico Italiano 1997[in Italian]. Roma: ISTAT, 1997.

5. Istituto Nazionale di Statistica. Rapporto Annuale ISTAT. Lasituazione del Paese nel 1997 [in Italian]. Roma: ISTAT, 1997.

6. Barbone F, Filiberti R, Franceschi S, Talamini R, Conti E, Mon-tella M, et al. Socioeconomic status, migration and risk of breastcancer in Italy. Int J Epidemiol 1996;25:479–87.

7. Toniolo P, Protta F, Capa APM. Risk of breast cancer, diet andinternal migrations in Northern Italy. Tumori 1989;75:406–9.

IC DELAY IN SOUTHERN ITALY 177

8. De Waard F, Kirkpatric A, Perri NM, Tornberg S, Tubiana M, deWolf C. Breast cancer screening in framework of the Europeagainst cancer programme. Eur J Cancer Prev 1994;3 (1Suppl): 3–5.

9. Dobrossy L. WHO initiative: cancer prevention in Europe. EurJ Cancer Prev 1993;2:251–3.

10. Weston R. Europe Against Cancer: cancer week UK. Eur J CancerCare 1994;3:105–10.

11. Mant D, Vessey MP, Neil A, McPherson K, Jones L. Breast selfexamination and breast cancer stage ad diagnosis. Br J Can-cer 1987;55:207–11.

12. Avtrandilov G, Berg JW, Carleschi R, Corbett HG, de Jong U,Delafresnaye J, et al. International classification of diseases foroncology. 2nd ed. Geneva: World Health Organization, 1990.

13. Montella M, Antignani F, Di Lucca G, Toglia G. In: ReichertzPL, Lindedberg DAB (editors). RTO-A management system forcomputerized cancer registry. Berlin: Medical Informatics Europe’85, Springer-Verlag, 1995;503–8.

14. International Union Against Cancer. TNM classification of malig-nant tumors. 4th ed. Berlin: Springer-Verlag, 1992.

15. Mantel N, Haenszel W. Statistical aspects of analysis of datafrom retrospective studies of disease. J Natl Cancer Inst1959;22:719–48.

16. Hosmer DW, Lemeshow S. Applied logistic regression. New York:Wiley, 1980.

17. Breslow NE, Day NE. Statistical method in cancer research. VolI. IARC Scientific Publication No. 32, Lyon: International Agencyfor Research on Cancer, 1980.

18. Rossi S, Cinini C, Di Pietro C, Lombardi CP, Crucitti A, Bellan-tone R, et al. Diagnostic delay in breast cancer: correlation withdisease stage and prognosis. Tumori 1990;76:559–62.

19. “Europe Against Cancer” Programme, GU C 50 2/26/87.20. “Europe Against Cancer” Programme [COM(89) 185].21. De Placido S, Nuzzo F, Perrone F, Noviello A, Delrio P, Di Palma

E, et al. The first breast cancer-screening program in southernItaly: preliminary results from three municipalities of Naplesprovince. Tumori 1995;81:7–12.

22. Assessorato Ricerca Scientifica ed Informatica. La Campania incifre [in Italian]. Regione Campania, 1993.

23. Assessorato Ricerca Scientifica ed Informatica. Annuario Statis-tico Campano [in Italian]. Regione Campania, 1995.

24. Montella M, Buonanno M, Iannuzzo MT, Celentano E, Castronu-ovo E, Capodanno G, De Marco MR, Tamburrini M, CreazzolaS, Ruffolo P, Fusco M, Palombino R. Trends of cancer mortality inCampania Region and in Italy, 1981–1993 & Incidence CampaniaCancer Register, 1996. Lega Italiana per la Lotta contro i Tumori,Napoli: Luglio, 1998.

25. Geddes M, Berlinguer G. La salute in Italia, rapporto 1998 [in

26. Forster RS Jr, Worden JK, Costanza MC, Solomon CJ. Clinicalbreast examination and breast self examination. Past and pres-ent effect on breast cancer survival. Cancer 1992;69 (7Suppl):1992–8.

27. Newcomb PA, Lantz PM. Recent trends in breast cancer inci-dence, mortality, and mammography. Breast Cancer Res Treat1993;28:97–106.

28. Brecheisen NL, Snyder TE. Breast cancer detection: improvingthe efficacy of screening mammography. Kans Med 1994;95:183.