Sri Lankan Patterns of Female Cancers: Incidence and Mortality Over 1995-2010

Background: Neoplasms are the second leading cause of deaths in Sri Lanka. Present study analysed the trends in incidence and mortality of all cancers, breast cancer, cervical, ovarian and uterine cancers among Sri Lankan females over 1995-2010. Methods: Cancer incidence was obtained from national hospital-based cancer registries. Cancer mortality was abstracted from World Health Organization database and Department of Census and Statistics Sri Lanka. Number of new cases and deaths were obtained by five-year age group for all cancers by sex and breast, cervical, ovarian and uterine cancers for females. Particular cancer specific incidence and mortality rates were directly age-standardised to the Segi-Doll world standard population. Age-standardised incidence and mortality for young adults (20-34 years), adults (35-64 years) and older adults (over 64 years) by the type of the female cancer over 1995-2010 were calculated. Results: Age-standardised rates for incidence for all cancers among females rose from 63.3 to 87.5 per 100 000 population during 1995-2010 and its morality increased from 44.5 to 53.5 per 100 000 population. In spite of having similar trends in both sexes, cancer incidence among females remained higher while mortality persisted lower than males. Breast cancer was the commonest cancer among females with its incidence and mortality increasing through-out. Cervical cancer incidence increased during 1995-2000, declined slightly in 2005 and remained stable over 2006-2010.Cervical cancer mortality remained stable over 1995-1999, declined in 1999-2003, increased slightly throughout 2003-2006 and remained stable during 2007-2010. Ovarian cancer incidence remained stable over 1995-2010. Its mortality remained stable over 1995-2000, declined slightly during 2000-2003 and increased in 2003-2010. Uterine cancer incidence and mortality increased steadily throughout 1995-2010. For all these cancers, incidence and mortality in 0-34 years remained low. Conclusions: Increasing trend of cancer incidence and mortality among females over 1995-2010, directs the need of revisiting female cancer control programmes.


Introduction
In global context, according to GILOBOCAN 2012, breast cancer is the second most commonest cancer overall with 1.7 million cases (12%) and ranking as the fifth leading cause of death accounting for 522 000 deaths (6.4%) while being the most frequent cause of death in less developed regions with 324 000 deaths, 14.3% of total [1]. Breast cancer is seen mainly among females resulting from cervical cancer, 87% of deaths occurs in less developed regions [1]. For 2012, 320 000 new cases of uterine cancer (4.8% of all female cancers) and 76 000 deaths (2.1% of female cancers) were estimated [1]. Ovarian cancer with an incidence of 239 000 new cases (3.6% of female cancers) and 152 000 deaths (4.3% of female cancers) in 2012 comes as seventh most common cancer and eighth commonest cause of cancer deaths among females. Highest incidence estimates of uterine and ovarian cancers are for developed world [1].
Sri Lanka with its population 20.4 million in 2012 has faced epidemiological transition from communicable diseases to non-communicable diseases within past two decades [3]. Among adults of 35-64 years, cancer mortality increased over 1950-2005 and had become the second leading cause of death among female aged 35-64 years in 2002-2006 [4]. Neoplasms were the second leading cause of hospital death in 2012 [5].
Population based cancer registry in the district of Colombo was established in 2012 with the technical support from the International Agency for Research on Cancer [7]. Still it is limited to Colombo district and only the first report which is for the single year of 2012, is still published. National Hospital-based Cancer registry reports provide large-scale data about cancer incidence by specific cancer type, age, gender and ethnicity providing potential for identifying important trends. However, being summary reports they fall short in fully addressing their potential in doing detail analysis of cancer trends over years. Mortality data comes from the Register General Department. The Registrar General's Department collects and compiles information on vital statistics. Department of Census and Statistics coordinates the dissemination and provides guidance for other government ministries. When a death is registered, a certificate is issued to the informant. A duplicate copy, with death declaration, is sent to the Registrar-General by divisional secretaries. There, all the returns are checked for possible entry errors. If any inconsistency is found it is referred back to the district office. The error free abstracts are used in the manual compilation of vital statistics and coded by a trained coder into International Classification of Diseases-10 (ICD-10) classification. ICD 10 classification has been used since 1997. Detailed information is processed using computers [8].
Well woman clinics (WWCs) was introduced in 1996 [9]. In WWCs, breast cancer early detection is conducted using clinical breast examination (CBE) and increasing awareness on breast self-examination (BSE). Women with high risk for breast cancer or having positive finding of CBE, are referred for mammography. Breast cancer early detection coverage was on only 2.2% among 35-59 year women in Gampaha district in Sri Lanka [10]. In WWCs, cervical cancer screening is conducted using Pap smear test [9]. There are no organised screening programmes for ovarian and uterine cancers in the country.
Females are bit hesitant to go for screening and to prioritize their health when it comes to female only cancers [11] in spite of having good literacy rate of 94.6% [5]. In order to evaluate effectiveness of present interventions and to plan for future, it is important to assess the recent trends in cancer incidence and mortality. Present study analyses incidence and mortality trends of female breast cancer, cervical, ovarian and uterine cancers in Sri Lanka over 1995-2010 and investigates the trends among young adults (20-34 years), adults (35-64 years) and older (over 64 years) age groups.

Materials and Methods
This study analyses female cancers consisting of breast, cervical, uterine and ovarian cancer trends among Sri Lankan females over 1995-2010.

Data sources
Sri Lankan cancer incidence among females by cancer type and five year age group were obtained published national hospital based cancer registry reports for 1995, 2000, 2005, 2006, 2007,2008,2009 and 2010 [6, 12-18] [8,19]. Data for both of sources are from Sri Lankan vital registration system. Mortality data used for present analysis were confined only to the data classified by ICD 9 and 10. For the said period, it was in ICD-9 and 10 formats. Denominator Sri Lankan populations by sex were obtained from the Department of Census and Statistics Sri Lanka [3].
As data were extracted from published sources, ethical clearance or any special permission were not obtained. Sources of data are referred as applicable.

Analysis
Descriptive data analysis was conducted using Microsoft Excel 10. Cancer-specific incidence and mortality by five-year age categories were grouped Cervical cancer was the second leading cancer in for females into all cancers, breast, cervical, ovarian and uterine cancers and for males only for all cancers. The rates were directly age-standardised per 100 000 population to the Segi and Doll world standard population [20][21]. Results were presented for all cancers contrasting the sex differentials. Age-standardised incidence and mortality for young adults (20-34 years), adults (35-64 years) and older adults (over 64 years) by the female cancer types over 1995-2010 were presented.

Results
From 1995-2010, total number of cancer incidence and mortality increased among Sri Lankan females. Incidence increased from 3875 of new cases in 1995 to 8970 in 2010. Cancer deaths rose from 2654 to 5398 over 1995-2010. Female population also expanded from 8.9 to 10.4 million during 1995-2010.
Female, age-standardised rates (ASR) for incidence for all cancers continued to rise over 1995-2010 from 63.3 to 87.5per 100 000 population with a slightly higher figure of 88.7 per 100 000 in 2000. Though trends are basically similar in both sexes, incidence among females was higher than males throughout 1995-2010 with sex

Discussion
Cancer incidence as well as mortality increased among females over 1995-2010. Cancer incidence was higher in females. Mortality was higher among males. Breast cancer was the leading cancer among females, with increasing trend in both incidence and mortality. Cervical cancer was second highest in list with an initial increase in incidence followed by a slight decline and thereafter remained stable. Its mortality had declined and thereafter increased. Ovarian cancer was the third in the incidence with a stable incidence and a slightly increasing mortality in latter period. Uterine cancer incidence and mortality increased over the period.
Increasing cancer incidence and mortality among Sri Lankan females as well as males over 1995-2010, reflects the epidemiological transition from communicable diseases to non-communicable disease occurring over recent decades. Sri Lanka has experienced  epidemiological transition reflected by profound decline in infectious diseases and pregnancy-related mortality in adults with an increase in circulatory diseases and external causes in males, cancers in females [4]. Females having higher incidence of cancers and a lower mortality compared to male, is consistent with GLOBOCAN 2012 estimates for South East Asia [1]. Commonest cancer among females is breast cancer and next commonest is cervical cancer in Sri Lanka. Globally, it is second cause of female cancer mortality in more developed region while first is lung cancer [1] while being commonest in South East Asia among females [22].
Breast cancer is early detectable and cervical cancer is preventable. Early treatment of these two cancers was proven to increase the survival and reduce the mortality [1]. In 1996, with the introduction of WWCs, breast cancer early detection and screening of cervical cancer using Pap smear test were initiated. This with parallel increased awareness might have accounted for having high incidence and low mortality due to cancer among females compared to males. However, this contrasts from global figure of male cancer incidence being 25% higher with 14.8% of male cancer incidence consisting of prostate cancer due high early detection of prostate cancer [1]. Female mortality being lower than males is consistent with the findings of more developed regions [1].
Breast cancer showed an increasing incidence through-out the period. It is highest in over 64 years in 1995, became highest in 35-64 years in 2000, equalised in these groups in 2007-2008 and became highest again in over 64 years in 2009-2010 while 20-34 years had a lower incidence. This increased incidence in 35-64 years could be attributed to the age of occurrence as well as to the early detection parallel to the provision CBE and awareness on BSE for 35-60-year group with initiation of WWC programme in1996.
Being commonest cause of cancer deaths among females, breast cancer mortality increased over 1996-2000 and 2003-2006 with over 64 years having highest mortality and 35-64 years having next highest mortality. This might be due to late detection of breast cancer being more in over 64 years than 35-64 years. Yet Sri Lankan findings are consistent with breast cancer being first commonest cause of cancer deaths in the South East Asia and in less developed countries in 2012 [1].
Breast cancer being commonest, reflects the need of reviewing the ongoing early detection programme. Recent studies highlighted the inadequacy of coverage of breast cancer detection and poor knowledge, attitudes and practices on breast cancer early detection among community as well as primary health care workers [23][24]. CBE coverage in WWCs was only 2.2% among 35-59 year women in Gampaha district in 2007. The overall procedure on CBE conducted by public health midwives at WWC was observed to be substandard in entire district in an assessment of conducted using Lot Quality Assurance Sampling in 2008 [11]. So, adopting evidence-based strategies is needed for the programme. A qualitative study conducted in the same district among public health midwives revealed non-availability of guidelines, inadequacy of training, lack of skills and material to provide health education, inability to provide privacy during CBE, lack of community awareness and motivation in 2008 [11].
Cervical cancer was the second leading cancer in the females over 1995-2010. Though cervical cancer is the second commonest female cancer in Sri Lanka, globally it is the fourth most common cancer in females [1]. The average risk of dying from cervical cancer before 75 years is three time higher in lesser developed than developed regions [1]. In the South East Asia, it is the second commonest cancer among females [1].
Cervical cancer incidence increased from 1995 to 2000, followed by a slight decline in 2005 and remained stable over 2006-2010. Initial increase may be parallel to introduction of Pap smear screening under WWC in 1996. Mortality which declined from 1995-2001, increased in a lesser extend throughout 2003-2010. This reflects that fact of lack of organised screening programme in past. However, not having decreasing trend in incidence and the raising trend in mortality directs towards the need of strengthening of the cervical cancer control programme.
sixth commonest cancer among females in South East Asia [1]. In present study, ovarian cancer was the third in the incidence of four female cancers and incidence remained stable throughout 1995-2010 in all ages. The highest incidence seen among 35-64 years shows need for giving more emphasis in diagnosing ovarian cancer as disease per say is having only vague symptoms from younger age. In spite of initial stable rates followed by a slight decline in mortality, its increase from 2003 to 2010 directs on need of early diagnosis and timely treatment.
Uterine cancer incidence and mortality increased steadily throughout the period. This reflects the need of more focused interventions with the intention of uterine cancer control. Increasing awareness, early diagnosing and timely treatment should be achieved. Globally it accounted for 4.8% of female cancers and in terms of mortality only 2.1% of cancer deaths among females reflecting the reasonable prognosis associated with the disease in developed world [1]. Therefore, there is need of strengthening early diagnosis and timely treatment of it in Sri Lankan set up. However, it is seventh commonest cancer among females in South East Asia [1].
Strengths of the present study are that it presents in recent trends of female cancer incidence and mortality using data best available. Incidence and mortality data were age-standardised to the to the world standard population as proposed by Segi (1960) and modified by Doll et al.(1966) to overcome confounding due to different age structures over the years [20]. Age-standardization to the world standard population allows comparison with similarly ASRs from the world and regions.
Incidence data were obtained from the published national hospital-based cancer registry reports. Not having a national population-based cancer registry data for the country for 1990-2010 was a barrier in getting 100% coverage of cancer incidence and mortality. There is a limitation in coverage of population in national hospitalbased cancer registry. For national hospital-based cancer registry, data for 1990-2000 were only from National Cancer institution and there-after, expanded to the data from newly establishing treatment units, pathological labs and OMF units. This has an effect of having data that are more complete in later years. However, national hospital-based cancer registry not having data of the patients taking treatment in private sector and abroad has result a certain underestimation of cancer incidence. Yet, histological confirmation was high with 92% in 2007 and over 94% in 2008-2010 [6,[16][17][18].
As data is obtained from published cancer registries, it included the codes used in those. In cancer registries before 2007, behaviour code 1 (uncertain whether benign or borderline malignancy) or 2 (carcinoma in situ) were included as cancers. From 2007 onwards, only behaviour code 3(Malignant primary site) was included as cancers. ICD 9 was used for coding incidence in 1995 and ICD O was used from 2000 onwards [6,[12][13][14][15][16][17][18]. These might have had certain effects on comparisons and decrease in incidence of breast, cervical, ovarian and all cancers from 2000 to 2005.
In spite of all the effort of getting best available mortality data, inherent issues of the under-registration of deaths could not be overcome particularly as over this period there was a civil conflict in the country. Since 1982 to 2000, there was an internal conflict was there and cease fire was there over 2001-2005. Internal conflict lasted up to 2009. Sudden increase of mortality in 2000 could be due to probable artefact resulted from restoration following decreased registration in previous years [4]. Even though trained coder does the coding at the central Registrar General Department [8,25], accuracy of documented cause of death leading to misclassification is unavoidable. One study conducted in quality and coverage of death registration in a district of Sri Lanka; found that 15.5% of medical officers misclassified the underlying cause of death (Fonseka, 1996). Another study conducted in a sample of deaths that had occurred in 2006-2008 shows that the majority of medical records was of average quality. Concordance between the underlying cause of death in the vital registration data and that from the 'gold standard' (medical records review), diagnosis was only 41.4% (n=249). The sensitivity of all leading causes of death and positive predictive value were below 67% [26]. Due to limitation of availability of data, present study does not show latest trends in incidence and mortality after 2010. Hence, effectiveness of very recent interventions such development of guidelines on early detection of breast cancer in 2011 and of gynaecological cancer in 2014 with revision of primary health care worker curriculums in 2015 cannot be assessed. Hence, further analysis would be needed to see the effectiveness of very recent interventions on female cancer control such development of guidelines on early detection of breast cancer in 2011 and of gynaecological cancer in 2014 with parallel revision of primary health care worker curriculums in 2015, when data are available. A study on cost-benefit analysis of using of different strategies in breast and cervical cancer control are present needs.
Increasing trend of breast cancer incidence as well as mortality directs the need of revisiting breast cancer control programme in the country. Cervical cancer, the second in list with an initial increase in incidence followed by a slight decline and thereafter remaining stable rather than declining, with a slightly increasing trend in mortality after initial decline, reflects the inadequacy of screening strategy with opportunistic screening using Pap smear test. Need of strengthening the cervical cancer programme exists. The stable incidence and a slightly increasing mortality in latter period for ovarian cancer and increasing incidence and mortality for uterine cancer draws the attention for strengthening awareness, early diagnosis and timely treatment for the gynaecological cancers.