Reliable information for cancer control in Cali, Colombia

Abstract Background: The Cali Population Cancer Registry (RPCC) has been in continuous operation since 1962 with the objective of producing valid statistics on the incidence of cancer, its patterns, trends and survival rates. Methods: During the period 2008-2012, 23,046 new cases were registered and during 2011-2015 there were 12,761 cancer deaths. The trend of the rates was described with the APC average annual change rate and with the Joinpoint analysis. We analyzed the individual data of 38,671 adults (15-99 years) diagnosed with cancer between 1995-2009, and we calculated the standardized net survival by age for the 14 most common cancer body sites, using the Pohar-Perme method. Results: Prostate and breast cancer were the first cause of cancer morbidity. The incidence rates in these were susceptible to early detection, tumors stabilized after decades of growth, while an increase in the incidence of colon cancer and papillary thyroid carcinoma was observed. The incidence rates of cervical and stomach cancer and conditions related to infectious agents decreased, although the number of absolute cases increased, due to the growth and aging of the population. Gastric cancer was responsible for the highest number of cancer related deaths. The types of cancer related to tobacco consumption (lung, oral cavity, esophagus, pancreas, urinary bladder) showed low numbers and a tendency to decrease. During the period 2000-2004, the 5-year net survival improved for cancers of the breast, cervix, prostate, melanoma and thyroid, although in the period 2005-2009 a stagnation was observed. In stomach, liver and lung cancer, the 5-year net survival was less than 15%. The 5-year overall survival in children was 51.0% (95% CI: 47.5, 54.3) and in adolescents 44.6% (95% CI: 36.0, 52.8). Comment: RPCC has been an advisor to the Colombian government in the evaluation of CPRs in the country and its data has contributed significantly to different aspects of cancer control in Colombia.


Introduction
Comprehensive cancer control is a strategic approach that brings together the main associations and organizations of a community to prevent or minimize its impact and to develop a plan to reduce the number of citizens who become ill or die from cancer. The plans are specific, based on an analysis of the cancer situation in each country 1 . It is essential that the information on incidence, mortality and survival be of high quality because it will help monitor and evaluate the programs. Population-based cancer registries (RCPobs) represent the gold standard for providing cancer incidence and survival figures in a region and are a key element in cancer control because they provide indicators for planning and evaluating cancer control activities and carrying out cancer research 2 . The information disclosed by these information systems in Colombia indicates that cancer is a public health problem that causes 63,000 new cases and 33,000 deaths each year 3 . To face this threat, the Colombian government has formulated a Ten-Year Plan for Cancer Control in Colombia (PDCC) 4 focusing on activities to control and reduce mortality from cancer of cervix, stomach, prostate, breast, colon and acute pediatric leukemias.
Colombia lacks an RCPobs with national coverage and for several decades the only source of valid cancer incidence information for the country was the Cali Population Registry of Cali (RPCC) 5 . Now it has three additional regional RCPobs that provide quality cancer incidence information in Pasto 6 , Manizales 7 and Bucaramanga 8 ; and two new ones in the process of consolidation in Barranquilla 9 and Medellín 10 . The coverage of these six regional RCBPs is less than 12.9% of the Colombian population. To overcome this limitation, health authorities use GLOBOCAN methods to make national and regional cancer incidence estimates based on mortality information 11 . The incidence/mortality ratio of the period of interest of each regional RCPob is incorporated into a mathematical model that uses this information and the mortality observed in each department as inputs to estimate the departmental and national incidence 3,11 . The validity of the estimates depends on the quality of the information and also on the accurate quality of the certification of general mortality and cancer in Colombia and the coverage of the certification is close to 100% 12 .
In this article, the Cali Population Registry discloses the most recent cancer statistics in Cali, Colombia, for incidence and mortality rates standardized by age (ASR) for all cancers for the periods 2008-2012 and 2008-2015, respectively; and the 5-year net survival estimates standardized by age for the 14 most common cancer sites from 1995 to 2009. Estimating the incidence of cancer in Colombia and creating some of the baseline indicators of the current PDCC in the city is a contribution made by the Universidad del Valle to the health authorities.

Materials and Methods
Population and registration area Cali is the third largest city in Colombia, capital of the Department of Valle del Cauca. According to the 2005 census and according to the projections of the DANE 13 , the estimated population for 2010 was 2.3 million inhabitants. 52% are females, and 26.2% self-identify as belonging to the black ethnic group 14 . The life expectancy at birth is 73.1 years for men, and 78.5 years for females 15 . The infrastructure for cancer care includes 165 oncology functioning services 16 , these services are in the urban area where 95% of the population resides in an area of approximately 110 km 2 that corresponds to 20% of the extension of the municipality of Cali (503 km 2 ).
Incidence and mortality information Information on the incidence of cancer was obtained from the database of the RPCC (2008-2012) and information on general mortality was obtained in the Municipal Public Health Secretary of Cali (2006)(2007)(2008)(2009)(2010)(2011)(2012)(2013)(2014)(2015). Details on the history, objectives, logistics and coverage of the RPCC have been previously described 5,17 . This same issue of Colombian Medical describes in detail the procedures and methods for estimating incidence, mortality and survival in adults 18 . In summary, the RPCC was established in 1962, it is a population-based cancer registry which provides continuous information on new cases of all types of cancer in permanent residents of Cali through active search and notification. Implementation of a childhood cancer outcomes surveillance system (VIGICANCER) within Cali's population-based cancer registry was carried-out in 2009; methodological details have been published recently 19,20 . Briefly, children and adolescents (<19 years of age) with new diagnosis of cancer and treated in a pediatric oncology unit of the city, are registered by the system and included in an active follow-up. VIGICANCER includes both children living in the city as well as children from other Colombian municipalities and provinces but treated in Cali. Vital status, relapse, treatment abandonment, and second neoplasms are the primary outcomes.

New cases of cancer (incidence)
In the quinquennium 2008-2012, 23,046 new cases of cancer were diagnosed among the permanent residents of Cali, for an average of 4,500 cases per year; 55% (12,613) occurred in females and the sex ratio was 1:2. The incidence rates standardized by age for all cancer sites per 100,000 person-years were 204.6 in men; and 185.1 in females. In the absence of other causes of death, the cumulative risk of developing cancer before reaching the age of 75 was 23.8% and 20.5% in males and females in Cali.
Cancer incidence rates per 100,000 person-year by sex and cancer location are shown in Cancer was the third cause of death in Cali after mortality due to cardiovascular diseases (26.0%) and unintentional or intentional injuries (20.1%).
In contrast to the number of deaths, standardized cancer mortality rates for all combined locations per 100,000 person-years were higher among males (107.0) than among females (85.9). Cancer of stomach, lung, colorectal, breast and prostate were the main causes of cancer-related death, together they represent approximately half of all cancer deaths (47.9%).   Tables 1 and 2 show the APC that represents the average percentage of annual increase or decrease in cancer incidence and mortality rates during the periods 1962-2012 and 1984-2015, respectively. In describing the change, three well-defined patterns were detected: increased or decreased when the APC was significantly different from zero (twotailed values p <0.05); otherwise the term stable or flat was used.

Changes in cancer morbidity and mortality
The incidence rates for all cancer body sites increased in male an annual average of 0.6% (95% CI: 0.4 -0.8) and remained stable in females. In contrast, mortality for all cancer body sites has been significantly decreasing at an annual average of 0.9% in male; (95% CI: -1.1; -0.6); and 1.3% in females, (95% CI: -1.4, -1.1).
Trend in cancer incidence rates (1962-2012) The incidence of cancer decreased in both males and females in the following sites: oral cavity and pharynx, esophagus, stomach, larynx, urinary bladder and leukemia of unspecified type. The decrease was only observed in male with pancreatic cancer and with Hodgkin's disease; and in females with cervical cancer.
In contrast, increased incidence rates of colorectal cancer, melanoma, non-Hodgkin's lymphoma and lymphoid leukemia were found in both males and females; breast and thyroid cancer increased in females only; and liver, prostate and testicular cancer in males only.
In females, there was no change in the risk of morbidity due to cancer of the liver, pancreas, lung, uterine body, ovary and Hodgkin's lymphoma and in males the incidence of thyroid cancer and myeloid leukemia remained stable.
Trend in cancer mortality rates  Mortality from cancer shows a favorable trend. There was only an increase in mortality rate from melanoma and colorectal cancer in men. In the rest of the neoplasms, there was evidence of a decrease in mortality rates for ten of the 17 main body locations. The decrease was observed in both males and females with cancer of the esophagus, stomach, lung, urinary bladder, lymphomas and multiple myeloma; only in males with cancer of the oral cavity and pharynx; and only in females with cancer of liver, pancreas, cervix and uterine body.
There were no changes in leukemia mortality in the entire population of Cali. Mortality rates for liver, pancreas, and prostate cancer remained stable in males; and females, there were no changes in mortality rates for breast, colorectal, ovarian and melanoma cancer.

Five-year net survival
For the analysis, a total of 38,671 patients diagnosed with cancer were included through 1995-2009. The distribution of the most frequent malignancies corresponded to breast (17.7%), prostate (17.3%), stomach (13.1%) and colorectal cancer (9.4%), while a smaller number of records were reported for liver cancer (2.2%), melanoma (1.8%), multiple myeloma (1.5%) and Hodgkin's lymphoma (1.0%). The median age at diagnosis for the period considered was 64 years. There has been an increase in the number of patients diagnosed for the last study period 2005-2009. The trend of net survival for certain types of cancer by sex and diagnosis period 1995-2009 is shown in Figure 1.    Five-year net survival (%) 6.6 (4.7 ; 8.6) 9.0 (6.8 ; 11.1) 9.6

Discussion
The RPCC of Universidad del Valle provides unique information of the statistics of cancer in Cali, during the 2008-2015 period. This information is necessary for health authorities to make estimates of cancer risk for other regions of Colombia that are lacking cancer registries. These statistics complement previous reports 5, 22 and provides uninterrupted continuous monitoring The offer of oncology care services in Cali corresponds to one sixth of the country's installed capacity 16  Pancreas rest are patients from the south-west, a region that represents 20% of the Colombian population 13 . Eighty five percent of the oncology services in Cali are private 16 , the care is not comprehensive and there are several barriers to accessing quality oncological care services. Government measures aimed at stabilizing the health system have been unsuccessful and there has been evidence of discriminatory behavior and risk selection of the oncological patients by the health care provider entities responsible for managing the risks related to the disease 23 . Therefore, the clinical outcomes remain unfavorable primarily because patients present late with in advanced stages of the disease and, thus, survival is low for most types of cancer compared to that observed in Europe and the United States 24,25. Coinciding with demographic changes there are significant variations in trends, patterns and differences in incidence rates and cancer mortality. The increase or decrease in the risk of morbidity and mortality due to this group of diseases is determined by different factors. So far, some are recognized and most are still to be identified. These changes may be the result of variations in the exposure of the population to different risk factors, better access to health services and improvement in diagnostic and treatment techniques 26,27 .
Although several threats persist, the available information shows evidence of advances in the control of some types of cancer in Cali.
Cancer related to infectious agents The incidence and mortality rates for stomach and cervix uteri cancer have decreased significantly over the last 55 years (Fig.  3). The descent is monotonic, continuous and began before knowing the prominent role in the processes of carcinogenesis of Helicobacter pylori 28 and the Human Papilloma Virus (HPV) 29 . These changes are not related to specific interventions against these infectious agents, they are the result of progress in the development and improvement of sanitary conditions. Economic development determined changes in lifestyles and modifications of the known risk factors for gastric cancer. Refrigeration facilitates the consumption of fresh foods and limits the use of chemicalbased food preservation methods (salting, desiccation, smoking, and acidification).
In the 21st century, gastric cancer still represents a great social burden in Cali and Colombia because it causes the highest number of deaths from cancer 3,5,22 . The disease is fatal when discovered clinically because the diagnosis is usually made in the advanced stages. From 1995 to 2009, the 5-year net survival of patients with gastric cancer in Cali was less than 20%, with a healing fraction of 15% and an average survival time of 6 months for uncured cases 22 ; similar results were observed in Chile and Costa Rica. But in other Latin American countries with equal or lesser development, the survival estimates were around 30%, which are comparable to those observed in the United States and Europe. It is also possible that estimates of gastric cancer survival in Ecuador and Cuba may be overestimated 24,25 .
Despite the continuous decline in the incidence and mortality of infection-related cancers, rates remain high and the number of cases continues to increase due to aging and population growth 30 . It is very likely that this downward trend will continue, even without additional interventions, in the years to come; however, under natural conditions, it will likely take many decades, if not centuries, before the incidence and mortality rates reach the values currently observed in the United States and Europe. It is a priority to implement additional measures to accelerate the decline, improve survival and achieve control 30 .
The perspectives for the control of gastric cancer are uncertain because therapeutic advances are insufficient, the pre-clinical results of efforts to develop vaccines against H. pylori have been disappointing 31 ; and the early detection of gastric cancer in Latin America has shown unreliable results 32,33 and low costeffectiveness 34 . During the first stages of tumor growth, cancer is clinically silent. Therefore, an alternative to control, is the implementation of primary prevention programs which would help eradicate H. pylori infection by reducing the risk of developing gastric cancer in people without precursor lesions 30,34,35 . Due to the above, it is necessary to develop 1) monotherapies to facilitate adherence to antibiotic treatment and 2) accurate non-invasive tests to identify premalignant gastric lesions and thus serve as a risk stratification tool of patients. The simultaneous detection of serum pepsinogens and antibodies against H. pylori has achieved this goal in Japan 35,36 . This strategy has not been adequately validated in Latin America and continues to be an option that requires exploration with a well-founded project of implementation.
The picture is different and more favorable for females with cervix uteri cancer (CUC). Mortality rates in Cali are close to the PDPCC goal 4 , but they are still three times higher compared to the United States and Europe; where the risk of cervix uteri cancer is half of that observed in Cali. The incidence and mortality rates have declined for many reasons, including declining fertility rates, improved socio-economic conditions and the establishment of a citywide program to prevent cervix uteri cancer via a widespread use of pap smear 29,37 .
The knowledge that certain genotypes of VPH infection are necessary to cause cervix uteri cancer has created new strategies for its prevention in the current PDPCC. As of 2012, the national guidelines for vaccination against VPH are established and coverage of 80% has been achieved, which unfortunately is now below 10% because of a mismanaged episode of massive psychogenic reaction in Carmen de Bolívar, a Colombian rural region. This reaction was supposedly associated with the VPH vaccine 38 . To increase the accuracy of cervix uteri screening, the Ministry of Health of Colombia incorporated HPV testing in cervical cancer screening programs. It is expected to achieve coverage of 80% of the target population in 2021 4 . These measures are essential to accelerate the control of this disease because the 5-year net survival of females with cervix uteri cancer in Cali was 57%, 10% points below that observed in affluent countries 24,25 .
Cancer related opportunity screening activities Prostate and breast cancer are the leading cause of cancer-related morbidity in males and females in Cali, respectively 5 . In Colombia there are no organized screening programs for either cancer and cancer control is based on specific opportunistic screening activities. Mammography, digital rectal examination and PSA allowed us to detect cases of disease that were previously unknown and contributed to increasing the incidence rates before the first quinquennium of the 21st century and since then, it has begun to decline. Most, but not all, of the increase may be due to earlier detection of the disease. Once the use of screening tests had been established the rates tended to stabilize as long as other factors causing the disease had not changed.
These changes were more evident in the population subject to screening, the group of 50-69 years of age, where there was also a turning point in the trend of incidence. Similar changes were documented in Costa Rica and Ecuador at the end of the first decade of the 21st century and were observed in Europe and the United States 20 years earlier.
Mortality from prostate cancer has decreased consistently since 1984 with an average annual percentage decrease of 2%; the decline occurred earlier than expected. This could not be attributed exclusively to the screening activities (Fig. 3). An influential and perhaps determining factor is the evolution of treatments with curative intent; it is likely that the use of PSA and digital rectal examination have contributed to maintaining and consolidating this trend 39 . However, mortality from breast cancer remained stable during the study period (Fig. 3).
The United States and Europe have made great advances in the control of prostate and breast cancer. Despite the high incidence rates (ASR: 119.8 and ASR: 124.9, respectively); the 5-year net survival is around 98.9% and 89.7%; and mortality rates around (ASR: 20.1, ASR: 21.2), respectively 11,24 . In Cali, 5-year net survival for the same neoplasms was 83.2% and 74.4%; and the mortality rates around (ASR: 17.4 and ASR: 13.8), respectively. The existence of a gap of 15 percentage points in 5-year net survival in a population where incidence rates are half of those observed in affluent countries 40 , suggests that the diagnosis of cases is made at more advanced stages and / or that the tumors are more aggressive. This will remain an area of future investigation.
The incidence and mortality due to colorectal cancer continues to rise in males and females in Cali. The reasons are that the screening activities are incipient, and the risk factors are difficult to control or are not clearly identified 41 ; it is a priority to promote an organized screening program to reverse the current trend. Until this intervention occurs, oncological care services must be oriented to the early diagnosis of suspected cases.
Cancer related to tobacco use The trend in the incidence of lung cancer correlates with the historical patterns of prevalence of cigarette smoking and there is sufficient evidence of a causal relationship between cigarette smoking and various types of cancer. The reduction in the number of cancer cases related to tobacco use in Cali, has been interpreted as a successful example of cancer control. This was due to the implementation of a very strong anti-smoking government campaign implemented in the seventies and that has been consolidated in the 21st century. The incidence rates of lung cancer for both sexes in Cali reflect the end of a tobacco-related epidemic that began in the 1970s and was interrupted around the 1980s 5,42,43 . Since then, there has been a significant decrease in tobacco-related cancer incidence and mortality: oral cavity and pharynx, esophagus, pancreas, lung and urinary bladder. The decrease was more consistent in the oral cavity and lung cancer in both males and females.
Net survival estimates Surveillance of cancer survival is important for health organizations, civil society and research agencies because it serves to formulate strategies and prioritize cancer control measures, and to evaluate effectiveness, as well as the cost effectiveness of these strategies 1 .
At the beginning of the 21st century, we began to monitor trends in cancer survival in Cali. Certification of cancer mortality Information on cancer mortality in liver, lung, brain and bones should be interpreted with caution, because in these sites, the occurrence of metastasis is frequent. In the Cali cancer registry, we found evidence that the primary site of some of these cancers came from locations different than these organs. It is important to understand that 45% of liver cancer cases corresponded to metastasis. It was also established that cancers of the bone (46%), lung (15%) and CNS (10%) corresponded to metastasis. There were 2,447 new cases and 450 deaths from cancer. The coding of the body locations made by the vital statistics office and the cancer registry were compared. The concordance (Landis criteria 47 for the coding of cases of liver, bone and lung cancer were considerable (Kappa = 0.64, 0.67 and 0.79, respectively) and the highest was for malignant tumors of the CNS (Kappa = 0.90).
Childhood cancer About 200,000 new childhood cancer cases per year are diagnosed in the world 48 , 84% occurring in low and middle-income countries 49, 50 . Taken into account that cancer in children is not amenable to primary or secondary prevention, survival is the most relevant metric to evaluate efforts aimed to control cancer burden in this population group. Cali's 5-year OS (51%) is 26% to 32% lower to outcomes reported in more affluent countries (77% to 83%) 51,52 . This implies that if in Colombia 1,500 to 1,600 children are treated for cancer per year then after 5 years 765 to 816 had died, and 390 to 512 would be preventable deaths. This survival gap persists in all cancer groups, except for Hodgkin disease (7% difference; 5-year OS: 88% vs 95%). Effectiveness decrease in cancer treatment is mainly related to intensity lost. Chemotherapy intensity is related both to dose and time interval among doses. Therefore, effectiveness of treatment is very dependent on the delays in treatment administration (adherence to treatment), being treatment abandonment the extreme example of this principle. Intensity lost has multifactorial causes involving the patient, their families, health providers, and the health system. Other path to reduce survival is due to treatment mortality and not because disease. This adverse outcome is both related to access to supportive care and advance disease at diagnosis. Access to timely and correct diagnosis and treatment is particularly related to poor outcomes in tumors that are dependent of the stage at diagnosis to achieve cure; retinoblastoma is the best example of this. Nevertheless, in the Latin-American context, the Argentinian hospital registries system reports a 3-year OS of 61.7% 53 , which compares favorably with Cali´s 3-year OS of 56.6% for this cohort.
Limitations RPCC does not actively monitor adults, and Cali lacks reliable statistics on population migration 13 . The RPCC has information about the cause of death through death certificates, but in some cases it can be difficult to determine if cancer is the basic cause of death.

Strengths
The RPCC has participated in many other collaborative studies and has been an advisor to the Colombian government in the evaluation of PBCR in the country and its data have contributed significantly to different aspects of cancer control in Colombia. The collaborative work with the SSPM of Cali facilitates access to information on general mortality and cancer; and allows an independent source of verification of new cases of cancer. Access to the information system of the Ministry of Health (SISPRO) and to the assurance databases provides a permanent update of the vital status and date of last contact.