The prognostic role of gender in survival of adult cancer patients. EUROCARE Working Group

Many observations indicate that women have a much longer expectancy of life than men. Some population-based studies on cancer patients support the idea of the role of gender in predicting survival. However, the data are somewhat contrasting and inconclusive. The purpose of this paper was to evaluate the prognostic role of gender for cancer patients, making use of the large set of survival data made available by the EUROCARE II project for the period 1985-1989. By applying a multivariate approach the major confounders such as age, geographical area and cancer site were considered in analysing survival data on more than 1 million cancer cases collected by 45 population-based cancer registries in 17 European countries. The results were consistent with the general observation that in the industrialised countries women tend to survive longer than men. The multivariate analysis showed better survival from cancer in women than in men, estimated as an overall 2% lower relative risk of dying. The female advantage was particularly evident in young cases, reduced in patients in middle age groups and in the oldest patients completely reversed so that at this age men had the better prognosis. Longer survival for women was not present immediately after diagnosis, but the major advantage was seen after 3 years of follow-up. The risk of death for women was significantly lower for cancer of the head and neck, oesophagus, stomach, liver and pancreas. For bladder cancer, the risk of death was significantly greater for women. These results can be explained by gender differences in sub-site distributions (head and neck and stomach) and by the differences in the stage at diagnosis (presumably bladder). However, the consistency of the data, evident only when a vast set of data is analysed, suggest that women may be intrinsically more robust than men in coping with cancer.

Introduction: the EUROCARE II Study

This introduction provides a general overview of the aims, methods and procedures used in the EUROCARE II study and the types of analyses presented in each article of this Special Issue of the European Journal of Cancer. The main aims of the EUROCARE II project are the updating of the survival database of the European Cancer Registries, the study of recent trends in relative survival rates and the interpretation of the survival differences observed both in time and across populations. Once having completed the preliminary stage of data checking, a total of 3,473,659 individual records from patients of all cancer sites, diagnosed between 1978 and 1989 and provided by 45 cancer registries in 17 European countries were accepted to build up the EUROCARE database. The quality of these data, in terms of the accuracy of the diagnosis and the validity of vital status assessment, was checked by indirect indicators, based on cross-validation analysis of consistency of the relevant variables. Statistical analysis was based on age-specific relative survival rates, computed for each cancer sites as the ratio of observed survival to the expected survival of the general population of the same area, gender and age, according to the Hakulinen method. An estimate of the European survival was computed as a weighted mean of the corresponding survival of the different countries, using as weights the expected yearly number of incident cases in each country. For comparison purposes, age-standardised survival was also calculated for Europe and for each country involved in the study.

Variation in survival of patients with lung cancer in Europe, 1985-1989. EUROCARE Working Group

In this study, we report on the variation in the prognosis for adult patients with lung cancer within Europe, by age, histology and country from 1985-1989. We considered trends in survival since 1978 for most countries. Survival analysis was carried out on 173,448 lung cancer cases diagnosed between 1985 and 1989 in 44 population-based cancer registries, participating in the EUROCARE study. Relative 1-year survival rates for patients with lung cancer varied from 24 to 40%, being highest in Finland, France, The Netherlands and Switzerland and lowest in Denmark, England, Poland and Scotland. Half of all patients under the age of 45 years died within 1 year of diagnosis, increasing to almost 80% for those aged 75 years or older. Whilst the prognosis for patients with non-small cell carcinoma remained more or less constant between 1978 and 1989 (25% in Denmark and 44% in Finland), that for patients with small cell carcinoma improved slightly, especially in The Netherlands (Eindhoven from 17 to 24%) and Switzerland (Geneva from 24 to 32%). In conclusion, a fairly large variation in lung cancer relative survival rates existed between European countries. The most likely explanation for the differences is the variation in access to specialised care. Except for a slight improvement in short-term survival for patients with small cell lung cancer, survival has remained poor since 1978.

Survival of European women with gynaecological tumours, during the period 1978-1989. EUROCARE Working Group

This study concerns the survival of European patients diagnosed between 1978 and 1989 with cancer of corpus and cervix uteri and ovary. Variations in survival in relation to age, country and period of diagnosis were examined. Data from the EUROCARE study were supplied by population-based cancer registries in 17 countries to a common protocol. Five years after diagnosis, relative survival rates were 75, 62 and 35% for cancers of the endometrium, cervix and ovary, respectively. Survival decreased markedly with age. The decrease was especially evident for ovarian cancer, which declined from 65% (15-45 years) to 18% (75+ years). In 1985-1989 there were important inter-country differences in survival for European women with gynaecological cancers: Eastern European countries were characterised by low 5-year relative survival whilst in Sweden, Austria, The Netherlands and Switzerland survival was generally higher than for other European countries. From 1978-1989, 5-year relative survival improved slightly for cervical cancer and improved more among the oldest patients. Prognosis also improved slightly for patients with ovarian tumours and this increase (around 20%) was concentrated among patients between 15 and 64 years of age. Intercountry differences in survival did not in general reduce over time, although for ovarian cancer survival differences narrowed probably in relation to the more widespread use of more effective chemotherapy. Intercountry and time differences in survival for cervical cancer are almost certainly related to variations in the effectiveness of cervical screening programmes. For corpus uteri cancer there was no improvement in survival over the period of this study and intercountry survival differences for this cancer are probably related to differences in patient management.

Small bowel enteroclysis in surgically treated obesity

PURPOSE

To define the indications, technical limitations and diagnostic yeld of small bowel transbuccal enema in the follow-up of surgical jejunoileal shunting in patients with complicated severe essential obesity.

MATERIAL AND METHODS

Three patients were submitted to surgical diversion: two of them underwent an intestinal bypass after Payne-De Wind (isoperistaltic end-to-side jejunoileostomy) and the other after Scott (end-to-end jejunoileostomy). The latter refers to intestinal recanalization and antiperistaltic lower end-to-side gastroenteric restoration.

RESULTS

Radiologic studies are the only means to depict the surgical small bowel. Radiographic follow-up needs barium sulfate administration and therefore cannot be performed any sooner than 30 days postoperatively. In the last three years the classic transbuccal enema has been performed with a Rollandi tube (with a terminal opening and a balloon). Both the anastomosis and the blind loop are difficult to demonstrate.

CONCLUSIONS

Jejunoileal bypass can be used to treat severe obsity uncontrollable otherwise, to reduce food absorption. Different severe complications may result and small bowel studies may permit to show late local complications. Small bowel enema is also indispensable in bypass reversal. There are no alternatives to this radiologic examination which is however very difficult to perform, because of the changes made by previous operation(s), and to interpret because the anastomosis, the sutured loop and wall changes are often poorly demonstrated.

Survival of women with breast cancer in Europe: variation with age, year of diagnosis and country. The EUROCARE Working Group

Breast cancer is the most frequent malignancy among women in developed countries. Prognosis is better than for other major cancers, and an improvement in survival has been reported for several populations in recent decades. Within the framework of EUROCARE, a population-based project concerned with the survival and care of cancer patients in Europe, we analysed data from 119,139 women diagnosed with breast cancer between 1978 and 1985 in 12 countries and followed for at least 6 years. Multiple regression models of relative survival, which take mortality from all other causes in each area into account, were used to estimate the effect of age, period of diagnosis and country on survival. For the comparison between countries, survival rates were age-standardised to the age structure of the entire study population. Women aged 40-49 years at diagnosis had the best prognosis in all countries and throughout the study period. Women younger than 30 years at diagnosis had a worse prognosis than those aged 30-39. The highest relative survival at 5 years was in Finland and Switzerland (about 74%), intermediate levels were found for Italy, France, The Netherlands, Denmark and Germany (about 70%) and the lowest rates were in Spain, the United Kingdom, Estonia and Poland (55-64%). During the 6 months following diagnosis, survival was highly dependent on age and was sharply lower in women older than 49 years. For women surviving more than 6 months after diagnosis, survival was similar for all ages, although women aged 40-49 still had the better prognosis. The average rate of death from breast cancer fell by about 2.5% for each year of diagnosis between 1978 and 1985. This improvement manifested mainly in younger and older women, for whom survival was initially less good. The largest improvement was seen in Poland (-15% death risk per year). We suggest that the better survival of women aged 40-49 at diagnosis is related to lower levels of circulating sex hormones, resulting in reduced stimulation of tumour cell growth. Early diagnosis may also be important in the peri-menopausal period due to increased diagnostic attention. Low survival in the United Kingdom may be due to inadequate adherence to consensus treatment guidelines and greater variation in treatment.

The cure for colon cancer: results from the EUROCARE study

The interpretation of time trends and geographical differences of population-based survival rates is generally not easy, due to the difficulty in disentangling the effects of observational biases, diagnostic and therapeutic procedures and their interactions. Whereas descriptive analysis of relative survival is generally based on survival levels estimated at fixed time since diagnosis, interpretation issues can take advantage from the analysis of the shape of the considered relative survival. Parametric survival models allowing the estimation of the fraction of cured patients are applied here to analyze and discuss the differences in colon cancer relative survival between European countries, according to age and period of diagnosis. The survival curves of colon cancer patients are described according to 2 parameters: the proportion of cured patients and the mean survival time of fatal cases. These parameters are estimated by least square nonlinear regression of relative survival values derived from the EUROCARE Project publication. Exponential and Weibull survival functions are used to model the relative survival curve for the fraction of fatal cases. The Weibull model gives generally a better fit with respect to the exponential model, thus indicating that the mortality rate for fatal cases is decreasing with time since diagnosis. For the youngest patients, however, the 2 survival functions give practically overlapping estimates. The overall proportion of colon cancer patients in Europe that are estimated to be cured was 38.6%. This proportion increased from 36% to 40% for patients diagnosed in 1978-1980 and in 1983-1985, respectively. Accordingly, mean survival time of fatal cases increased from 1.18 to 1.52 years. According to age, the proportion of cured patients present a marked decrease from young (48.4% at age 15-44 years) to middle-aged patients (38.6% at age 5564 years) and only a mild decrease from these to the oldest patients (34.4% at age 75 or more). The opposite effect was shown by survival time of fatal cases, i.e., 1.71, 1.75 and 0.77 years for the same age classes, respectively. Proportion of cured cases and mean survival time of fatal cases tended to be positively correlated with each other across countries. Our results are consistent with the hypothesis that a real improvement in colon cancer survival took place in Europe during the years 1978-1985 and also suggest that the well-known decrease of relative survival with age at diagnosis could be mostly due to a decreasing efficacy of early diagnosis for patients under 60 years old and to less effective therapies for older patients.

[Survival for digestive system tumors: Italian population-based data and international comparisons]

Survival for malignant tumour of the upper digestive tract, oesophagus, stomach and pancreas are analysed from the data of 5042 cases observed during the period 1976-1985 by the Italian cancer registries (CRs) of Varese, Latina, Ragusa and Florence. Three-year relative survival rates were 42% for patients with tumours of the upper digestive tract and 7% for those tumour of the oesophagus. Five-year relative survival for patients with stomach cancer was 20%. A worse prognosis was found for cases from the Latina and Ragusa CRs. One-year relative survival for patients with pancreatitic cancer was 17%. Younger age, histologic confirmation, period of diagnosis (for upper tract, stomach and pancreatitic cancers), and female gender (for tumours of the upper tract) resulted as positively associated to survival.

[Survival for tumors of the colon and the rectum in Italy]

In population based from four Italian cancer registries (Varese, Modena, Latina, Ragusa), five-years relative survival was 34% for colon cancer patients (no. = 1864) and 28% from rectum cancer patients (no. = 1117). However, rectum cancer patients showed higher survival rates in the first year of follow-up, which is possibly due to earlier diagnosis for sites more accessible to clinical examination. For colorectal cancer, age was an important prognostic factor, both in univariate and in multivariate analysis, adjusting for competitive mortality, sex, geographical area and year of diagnosis. In the same multivariate analysis, the prognosis was better for women, significantly for colon cancer and the excess death risk was reduced by 25-30% from 1978-79 to 1984-85. No significant differences were found among different geographical areas, possibly because of the low number of patients from some of the participating cancer registries.

Estimation and projections of colorectal cancer trends in Italy

BACKGROUND

Occurrence of and prognosis for tumours of the colon and rectum are thought to be changing rapidly due to simultaneous changes in risk factor prevalence, early diagnosis and treatment. In this paper time trends of morbidity, survival and mortality for colorectal cancer during the period 1970-1990 are estimated and analysed.

METHODS

Mortality trends were obtained from official death certificates. Relative survival rates were computed from population-based cancer registries. Incidence and prevalence rates were estimated from mortality and survival data.

RESULTS

Incidence rates were increasing during the period considered, with a lower rate of increase for the youngest birth cohorts. Relative survival rates of both colon and rectum cancers were higher for women, and for younger age groups, and were positively associated with period of diagnosis. No significant survival difference among the cancer registries used was found. A total of about 155,000 prevalent cases, 40% of which had been diagnosed > or = 7 years before, were estimated in the Italian population for the year 1990. Mortality rates were slightly increasing for men and stable for women. Projections of colorectal cancer trends to the year 2000 indicate major expected rises in both incidence and prevalence.

CONCLUSION

Colorectal cancer represents a problem of growing impact for health services in Italy. This conclusion can probably be extended to many developed countries.

Substantial variation in therapy for colorectal cancer across Europe: EUROCARE analysis of cancer registry data for 1987

To provide a quantitative description of the treatments applied to malignant colorectal cancer across Europe, we analysed all cases (11,333) of colorectal cancer registered in 1987 by 15 Cancer Registries in eight European countries. In a third of cancer registries, therapy was known for all cases, in the others 1-15% of registrations lacked treatment information. Eighty per cent of all patients received surgical resection, ranging from 58% (Estonia) to 92% (Tarn). The proportion of resections decreased with advancing age (85-73% for colon cancer; 85-70% for rectal cancer for < 65 years to > 74 years, respectively). Only 4% of colon cancer patients received adjuvant or palliative chemotherapy, range 1-12%. Sixteen per cent of rectal cancer patients received radiotherapy with great inter-registry variability (1-43%). Since the proportion of surgically resected patients correlated positively with the 5-year relative survival probability reported by the recently published EUROCARE study, this may be part of the explanation for the major differences in survival for these cancers among different European populations. The most likely determinant of this correlation is stage at diagnosis, but, quality of, and access to surgery, as well as access to endoscopy, may differ among countries and registry areas, and these may also contribute to inter-country survival differences.

Comparisons of colon-cancer survival among European countries: The Eurocare Study

Under the aegis of EUROCARE, a European Union project to assemble survival data from population-based cancer registries and analyze them according to standardized procedures, we have investigated and compared colon-cancer survival in 10 European countries. We analyzed 68,283 colon-cancer cases diagnosed between 1978 and 1985 and followed for at least 6 years. After calculating relative survival, putative factors prognostic for survival were investigated by univariate and multiple-regression analyses. Important intercountry colon-cancer survival differences exist within Europe, which are not explained by methodological differences, nor by demographic confounders. In patients aged 60 to 69, the mean European 5-year cumulative relative survival was 40%. Switzerland, Finland and The Netherlands had significantly higher 5-year relative survival, while one area in the UK and Cracow in Poland had significantly lower survival than this European estimate. Prognosis improved over time: from 1978 to 1985, the risk of death was reduced by about 4% per year in all countries studied. Age at diagnosis is inversely related to prognosis. Differences in health provision and hence in quality of care and stage at presentation seem largely responsible for the differences in colon-cancer survival found in the EUROCARE countries.

Estimation and projections of stomach cancer trends in Italy

Mortality data from official sources, and survival data from population-based cancer registries, are used for the estimation of incidence and prevalence of stomach cancer. Time trends of morbidity, survival, and mortality during the period 1970-90 are presented and analyzed. Incidence rates were decreasing during the considered period, but the rate of decrease was slowing down during the last decade. Almost stable rates, and even slightly increasing for women, were estimated for the youngest cohorts. Relative survival for stomach cancer was higher for women and for young ages; it was associated positively with period of diagnosis, and presented a significant South-North geographic gradient. Prevalence was estimated as decreasing during the period 1970-80, but increasing during the successive decade, due to both better survival and population aging. Projection of stomach cancer morbidity and mortality to the year 2000 showed that the disease should still be considered in Italy as a major public health problem.

Time trends of lung and larynx cancers in Italy

During the period 1970-1989, age-adjusted mortality rates for lung cancer in Italy increased by more than 50%, while rates for larynx cancer in males decreased by approximately 13%. This study aims to interpret this difference, which seems to contradict the finding that cigarette smoking is a common major risk factor for both lung and larynx cancer. To this end, we jointly analyzed the time trends of incidence, survival and mortality. We first examined survival data taken from the population-based Lombardy Cancer Registry (northern Italy). Based on data referring to 880 incident cases of larynx cancer, diagnosed during the period 1976-1987, we estimated a 3% annual increase in relative survival. By contrast, no significant period effect was observed for survival rates of 2,259 incident cases of lung cancer. National incidence rates were estimated using official mortality data and the above-described survival data. Age-adjusted estimated incidence rates increased, from 1970 to 1989, for both cancer sites: +55% for male lung, +56% for female lung, and +22% for male larynx. Moreover, the patterns of birth-cohort effect, which are diverging for mortality, are nearly parallel with regard to incidence. This analysis suggests that a substantial improvement in survival of larynx cancer patients may largely explain the differences in mortality trends for cancer of lung and larynx.

[Trends in cancer incidence as observed by Lombardy cancer registry, Varese Province]

Trends in cancer risk of the Varese province Cancer Registry from 1976 to 1987 have been considered. The analysis has been carried out subdividing the 12 years of observation in three quadriennia. In younger ages (0 to 54) an increase of melanomas, of testicular cancer and, only in females, of tobacco-related cancers (oral cavity, oesophagus, pancreas, larynx, lung, bladder and kidney) has been observed. In younger males, instead, a decrease of tobacco-related cancer risk is already observable. Below age 55 there is no reduction in gastric cancer risk. The increase of breast cancer for youngest women is only observed between the first and the second quadriennium. For older ages (from 55 to 74) tobacco-related cancers continue to increase, in both sexes, but mainly among females. Also increasing, for males, are colon and liver cancer risks. Among the same ages, gastric and cervix (for screened age groups) cancer risks are decreasing. These observations, most in accordance with patterns observed in other western countries, can be used for setting intervention and research priorities.

Survival for lung cancer in northern Italy

From the population covered by the Lombardy Cancer Registry (northern Italy), all 2,259 lung cancer patients diagnosed from 1976 to 1981 were followed through 30 November 1989. The length of follow-up ranged from eight to 13 years. A special investigation on long-term survivors showed that a negligible proportion (0.01 percent) of errors occur in determining life status when an active follow-up is adopted. Age, stage, and histotype were found to be statistically significant, independent, prognostic factors in multivariate analysis both by the Cox model and by a model considering the relative survival. Observed survival was 29 percent at one year after diagnosis, 13 percent at two years, eight percent at three years, five percent at five years and two percent at 10 years. Survival decreased with age, but the youngest patients of both sexes showed lower survival compared with immediately subsequent ages. Among morphologically confirmed tumors, epidermoid carcinomas and adenocarcinomas showed the highest survival (38 percent and 33 percent at one year, respectively); small cell carcinomas showed the poorest prognosis (one-year survival, 23 percent). Beyond the second year after diagnosis, differences between histotypes became slighter. Survival according to stage showed a decreasing pattern from limited to advanced tumors, one-year figures being 41 percent for localized tumors, 27 percent for regional metastasis, and three percent for distant metastasis. Relative survival in Varese was compared with that reported by other cancer registries in Western countries: the variability noted could be related to different modalities of registration and to different distribution of clinical and demographic factors.

Clinical and pathologic prognostic indicators in colorectal cancer. A population-based study

The institution of a colorectal Cancer Register in a health care district of Northern Italy gave the authors the opportunity to evaluate the prognostic relevance of several morphologic and clinical variables by univariate and multivariate analyses. Of the 134 patients registered in 1984, 132 were followed up until the end of 1989. Overall 5-year survival was 37%, but the figure increased to 43% when only colorectal cancer-related deaths were considered. Univariate analysis for clinical variables showed that TNM staging and age at diagnosis were significantly related to prognosis, whereas none of the other parameters were indicative of the clinical outcome. With a similar analysis, among the various morphologic variables, pattern of growth (infiltrating versus expanding) and extent of fibrosis (extensive versus little or absent) appeared to be indicators of prognosis. When the variables that were significant (stage, age, pattern of growth, and fibrosis) in the univariate analysis were entered into the Cox model of multivariate analysis, TNM staging was the only parameter that maintained an independent prognostic importance. The authors state that their results confirm the importance of stage in predicting survival for cancer of the large bowel and suggest that the possible prognostic value of clinical and morphologic variables should be investigated within each of the major TNM or Dukes' classes.

Clinical and pathologic prognostic indicators in colorectal cancer. A population-based study

The institution of a colorectal Cancer Register in a health care district of Northern Italy gave the authors the opportunity to evaluate the prognostic relevance of several morphologic and clinical variables by univariate and multivariate analyses. Of the 134 patients registered in 1984, 132 were followed up until the end of 1989. Overall 5-year survival was 37%, but the figure increased to 43% when only colorectal cancer-related deaths were considered. Univariate analysis for clinical variables showed that TNM staging and age at diagnosis were significantly related to prognosis, whereas none of the other parameters were indicative of the clinical outcome. With a similar analysis, among the various morphologic variables, pattern of growth (infiltrating versus expanding) and extent of fibrosis (extensive versus little or absent) appeared to be indicators of prognosis. When the variables that were significant (stage, age, pattern of growth, and fibrosis) in the univariate analysis were entered into the Cox model of multivariate analysis, TNM staging was the only parameter that maintained an independent prognostic importance. The authors state that their results confirm the importance of stage in predicting survival for cancer of the large bowel and suggest that the possible prognostic value of clinical and morphologic variables should be investigated within each of the major TNM or Dukes' classes.

Randomized trial of primary school education against smoking

Since 1974 an anti-smoking campaign consisting of a one-day educational intervention has been carried out in primary schools in Milan by the Italian League against Cancer (Milan Division). All but two of the 165 schools were randomized to evaluate the intervention effect. A total of 8549 children aged 9-10 were allocated to the intervention group and 8897 to the control group. Four years later a self-administered questionnaire was distributed in order to investigate the children's smoking habits. The proportion of smokers was 8.05% and 8.72% (p = 0.23) respectively for the intervention and control groups. It was concluded that sporadic educational intervention carried out during primary school years has little or no impact in preventing cigarette smoking in teenagers.

Survival and age at diagnosis of breast cancer in a population-based cancer registry

From the population covered by the Lombardy Cancer Registry, Italy, 1991 female breast cancer patients diagnosed from 1976 to 1981 were followed up until May 1987. Relative survival was 69% at 5 years and 58% at 10 years; median survival was 8.8 years. Ages 40-49 showed the best survival; ages 25-34 were 20% lower. From age 50 onwards, survival decreased progressively, with the exception of age group 65-74. We suggest that the best prognosis for ages 40-49, followed by the survival fall in subsequent ages, could be related to an anticipation of diagnosis in ages near menopause. The death hazard function showed a bimodal pattern, with a first peak in the first years after diagnosis, and a second one between the seventh and eighth years. The death hazard rate decreased by about 1% per year at each subsequent calendar year of diagnosis. When such an estimated calendar effect was taken in account, there were no considerable survival differences among Western countries covered by population-based cancer registries.

Breast cancer incidence and prevalence estimated from survival and mortality

Survival probability for female breast cancer patients was used to estimate incidence rates from breast cancer mortality data in Italy. The female breast cancer survival curve from the Lombardy Cancer Registry (LCR) was used to test the method on data from four local cancer registries, covering areas in different regions of Italy. In spite of the well known geographic variability of female breast cancer incidence and mortality, the results support the idea that survival probability does not change across the country and that the survival probability from the LCR is a good estimate of that in the country as a whole. Female breast cancer incidence and prevalence rates were then estimated for Italy, making use of a mathematical model specifically developed for chronic diseases. In 1985, crude incidence and prevalence rates of female breast cancer, for ages up to 74 years, were estimated as 71 and 701 per 100,000 women, respectively. Estimated incidence rates show a complex trend with age, increasing to a temporary pronounced peak at the age of 52. A marked cohort effect was found to increase significantly the risk of the disease from the 1886 to the 1930 birth cohorts by a factor of 2.9. After the 1930 cohort, risks have continued at a constant high level.