Spatiotemporal and trend analysis of common cancers in men in Central Serbia (1999–2021)

1 Introduction

Cancer represents a significant public health issue in both Serbia and globally. It is a disease with an unknown aetiology, making it impossible to identify a single factor responsible for its occurrence [1]. Numerous factors, the most common being smoking, poor dietary habits, insufficient physical activity, stress, and environmental pollution, contribute to the development of non-communicable diseases [2].

A total of 19,976,499 cases of cancer were diagnosed in 2022 [3]. Figure 1 shows global cancer incidence in 2022 in men including non-melanoma skin cancer. The highest cancer incidence for men was in Australia at 514.3 people per 100,000 when all cancers were included [3]. There were 32 out of 185 countries for which the incidence rate was over 300 per 100,000, including the most developed countries, e.g., the United States (401.7), France (386.4), the United Kingdom (327.7), Italy (312.1), Japan (309.8), Germany (306.5), and some neighbouring countries – Hungary (395.9) and Croatia (367.3) [3].

Figure 1

World cancer incidence in 2022 in men including non-melanoma skin cancer [3, processed].

According to the World Cancer Research Fund, Serbia is among the countries with the highest cancer rates [3]. In the “Cancer Mortality in Serbia 1991–2015” study, the overall cancer rate for this period was 294.7 per 100,000 [1,4]. In 2022. Serbia was in the 37th place (289.9 per 100,000) [3]. According to the Cancer Register data, men were mostly diagnosed with and died of bronchus and lung, colon, rectum, and prostate cancer [5,6,7,8,9,10].

Mapping and spatiotemporal analysis are the most widely used spatiotemporal visualisation methods [11]. In epidemiology, “hotspot” is frequently used to refer to areas of elevated disease burden or high transmission efficiency [12]. Numerous studies have described the geospatial distribution of diseases using emerging hot spot analysis [2]. Dominguez et al. [13] used the Getis-Ord- G i * method for geospatial analyses to identify regional hot spots of diffuse gastric cancer in rural Central America, and Ruktanonchai et al. [14] examine where cancer-related “hot spots” exist among respondents in Alachua County, Florida, and the United States. Nawaz et al. [15] applied hotspot analysis (Getis–Ord G i * ) to identify statistically significant hotspots of cancer and immune cells, which are defined as tumour regions with a significantly high number of cancer cells or immune cells, respectively. Zhang and Tripathi [16] used multidisciplinary techniques and tools to identify hot spots using GIS and others, including global Moran’s I statistics and Getis–Ord G i * statistics. Shariati et al. [17] used hot spot analysis to investigate the spatiotemporal analysis and hotspot detection of COVID-19 globally.

Wan et al. [18] used GIS spatial analysis and scanning statistical analysis to study the major gynaecological cancers from 2016 to 2018. Campbell et al. [19] used hot spot analysis to investigate geographic cancer disparities in Oklahoma, and Shah et al. [20] used hot spot analysis to investigate colorectal cancer cases in Kuala Lumpur.

Nearchou et al. [21] used Getis-Ord G i * hotspot analysis to quantify the cold and hotspots, defined as regions with a significantly low or high number of each feature of interest, respectively. Olfatifar et al. [22] explored the spatial autocorrelation and the estimation of incidence rate variance among Iranian provinces. Bahri et al. [23] used the same method to compare diabetes type 2 patients’ data in both global and local scales in Iskandar Malaysia neighbourhoods. Mahara et al. [24] identified the hot spots of tuberculosis incidence in urban districts in Beijing, China.

In Central Serbia, most cancer research has been focused on medical statistics [25,26,27,28,29,30,31]. Markovic-Denic et al. [25] investigated cancer mortality in this region, while in another study [26], they analysed time trends in cancer mortality between 1985 and 2002. Ignjatović et al. [27] conducted a study titled “Cancer of Unknown Primary: Incidence, Mortality Trend, and Mortality-to-Incidence Ratio Associated with the Human Development Index in Central Serbia, 1999–2018,” using data from the national cancer registry. Šipetić Grujičić et al. [28] analysed the trends in colorectal cancer mortality rates in Central Serbia between 1999 and 2014. More recently, Nikolić et al. [29] examined trends in colorectal cancer incidence and mortality among men and women in Central Serbia from 1999 to 2020. Other studies include Cavic et al. [30] who analysed lung cancer in Serbia and Jovanović et al. [31] who investigated the incidence and mortality of adrenocortical carcinoma in Central Serbia.

Relatively, few studies in both Central Serbia and the Republic of Serbia have utilised a geographic-medical approach to disease research. Micić [32] conducted a geospatial analysis of cancer incidence and mortality in Central Serbia, as well as a geospatial analysis of lung cancer incidence and mortality in the male population from 1999 to 2013. In addition, Micić et al. [33] performed a “Medical–Geographic Analysis of Incidence and Mortality from Brain Tumors in the Population of Central Serbia for the Period 1999–2010.” Obradović Lović et al. [34] examined municipalities in Serbia vulnerable to COVID-19, considering age, as well as mortality data related to respiratory diseases, cancer, and circulatory system diseases in 2020. Kričković et al. [2] analysed non-communicable diseases, such as cardiovascular diseases and diabetes, in AP Vojvodina (Northern Serbia) using Getis–Ord G i * statistics. In another study, Kričković et al. [1] applied the DPSEEA framework to explore the relationship between arsenic concentrations in water and cancer rates in AP Vojvodina.

The Institute of Public Health of Serbia “Dr Milan Jovanović Batut” oversees the management of the Cancer Registry for the Republic of Serbia. The Serbian Cancer Register contains data on personal characteristics of new cases and deceased individuals, the potential occurrence of multiple primary cancers, the date of diagnosis, diagnostic methods used, and cancer characteristics such as primary and secondary anatomical location, histological type, and stage, disease outcome, and information about the healthcare institution reporting the malignancy [5]. Although the Cancer Registry was established in 1970, the epidemiological monitoring of malignant tumours in Serbia was, until recently, based solely on mortality data [5]. Inaccurate guidelines, inadequate training of medical personnel, and the lack of informatics support contributed to the underreporting of new cancer cases, as well as the relatively low quality of data in reporting forms [5]. From 1986 to 1995, the number of reported cancer cases was roughly equal to or even lower than the number of deaths from cancer [5]. In 1996, the Cancer Registry underwent reorganisation in line with recommendations from the International Agency for Research on Cancer (IARC) and the European Network of Cancer Registries (ENCR) [5]. The reorganisation included decentralisation, with regional registers established at the regional level, while the main database for the Republic of Serbia remains at the Institute of Public Health of Serbia [5]. It introduced active data collection alongside passive methods, expanded the range of information sources, enhanced the education and training of medical personnel, strengthened informatics support, and implemented a structured system for data feedback [5].

This research will not attempt to identify the specific causes of cancer, given the complex and multifactorial nature of the disease. Instead, it will focus on analysing the most prevalent cancers among men in Central Serbia from both geographic and medical perspectives, to identify cancer hotspots and examine temporal trends. A spatial analysis of cancer distribution is crucial from biomedical, economic, and behavioural viewpoints, particularly in Serbia, where the integration of health geography remains relatively underdeveloped compared to other EU countries [1,2].

2 Materials and methods

2.1 Study area

The Republic of Serbia is a medium-sized country situated in Southeast Europe, in the central part of the Balkan Peninsula, and on the southern edge of the Pannonian Basin [35]. Serbia is a parliamentary republic, covering an area of 88,499 km², with Belgrade as its capital [36]. The country consists of Central Serbia and two autonomous provinces: Vojvodina, and Kosovo and Metohija [36].

Central Serbia’s administrative region spans approximately 55,967 km² [37]. It is divided into three statistical NUTS 2 regions (the City of Belgrade, Šumadija and Western Serbia, and South and East Serbia) and 18 counties (NUTS 3) [37]. The region contains 4,252 settlements, of which 127 are classified as urban [37]. Figure 2 shows the study area position in Serbia and Europe.

Figure 2

Position of the study area in Balkan Peninsula and Serbia.

Seventy-three percent of Serbia’s total population lives in Central Serbia [37]. According to the 2022 census, the population of Central Serbia was 4,906,773 [38]. In the Belgrade region, 1,681,405 people were recorded, while the southern part of Serbia had 3,225,368 inhabitants [38]. This included 1,819,318 in the Šumadija and Western Serbia region and 1,406,050 in the Southern and Eastern Serbia region [38]. The average age of the population in 2022 was 42.7 years in the Belgrade region and 44.5 years in the southern regions (44.3 years in the Šumadija and Western Serbia region, and 44.8 years in the Southern and Eastern Serbia region) [38]. Serbia ranks among the most demographically aged countries in the world [39]. The aging of the total population in Serbia has been an ongoing process for more than 50 years, since the end of the 1960s when the population was demographically the youngest [39]. This aging process has intensified, as indicated by the data from the last four censuses (1991, 2002, 2011, and 2022) [39]. The average age has been increasing, the number of young people (younger than 15 years) has declined, and the population aged 85 and older has nearly doubled [39]. If we compare the 2011 and 2022 censuses, it is noticeable that the most significantly affected are the Southern and Eastern Serbia region, where the average age was 43.3 and 44.8 years, respectively [39]. In contrast, the most favourable demographic situation was observed in the Belgrade region, with an average age of 41.8 and 44.5 years, respectively [39]. In 2002, the average age in Central Serbia was 40.4 years [40]. In 2011, it was 43 years, and in 2022, it was 44.6 years. Clearly, the average age is increasing.

2.2 Data set

Data were collected from publicly available reports, then cleaned, pre-processed, and prepared for use in the ArcGIS Pro 3.0 geodatabase with Python scripts. The administrative boundaries used in the database were downloaded in the *.GPKG format from the GEOSrbija portal (https://opendata.geosrbija.rs/) [41], an integrated geospatial data system managed by the National Infrastructure of Geospatial Data (NIGP) under the Republic Geodetic Authority (RGA). The data were imported into the geodatabase and stored in the WGS 1984 UTM Zone 34N projected coordinate system.

Cancer data were sourced from publicly available publications by the Institute of Public Health of Serbia “Dr. Milan Jovanović Batut” including “Cancer Incidence and Mortality in Central Serbia from 1999–2015,” “Malignant Cancers in the Republic of Serbia 2016–2021,” and the “Health and Statistical Yearbook of the Republic of Serbia 2016–2021” [5,6,7,8,9,10,42,43,44,45,46,47,48,49,50,51,52,53,54,55,56,57,58,59,60,61,62,63,64]. These publications provided data on the most frequent cancers and their primary localisation at the county level, which was used in this research. Standardised male incidence and mortality rates were applied for analysis. Malignant tumours were coded according to the International Classification of Diseases – Tenth Revision (ICD-10, codes C00-C96), and the International Classification of Diseases for Oncology – Third Edition (codes 8000/3-9941/3), World Health Organization, 2000, Geneva. The registry excludes “in situ” tumours (codes D00-D09) [5].

Publications covering malignant tumours for the entire territory of the Republic of Serbia at the county level (together with the territory of AP Vojvodina) have been available only since 2020 [5,6,7,8,9,10]. However, analysis for this period at the county level for the entire territory of the Republic of Serbia was not conducted due to data inconsistencies and lack of access. In addition, data on incidence and mortality rates at the settlement level are not available in these publications and were therefore not considered. Consequently, this research focused on Central Serbia. The Statistical Office of the Republic of Serbia does not have data for the Autonomous Province of Kosovo and Metohija, where data have been unavailable since 1998 [4].

This study aims to analyse the spatiotemporal patterns of seven cancers – lung and bronchus, colorectal, stomach, bladder, laryngeal, pancreatic, and prostate – in the male population, using 23 years of data (1999 to 2021) across 18 counties in Central Serbia.

Figure 3 illustrates the research framework, outlining the workflow from data collection and processing to the application of geospatial analysis software and the presentation of results. The analysis began with an examination of cancer trends, followed by hotspot analysis (Getis-Ord G i * ) using Spatial Statistics on open data related to cancer mortality and incidence rates, with the goal of identifying hotspots for the most prevalent cancers.

Figure 3

Schematic representation of the research process.

3 Results

3.1 Results of trend analysis

The first analysis focused on trends in the incidence and mortality rates of the seven aforementioned types of cancer. The results are presented through parameters derived from the MK tests. The original Python test in the pymannkendall module provides nine parameters, but only three – p, z, and s are included in the tables as the most significant. These represent the p-value, the z-score, and Sen’s slope (or the Theil-Sen estimator).

Table A1 reveals increasing trends in lung and bronchus cancer incidence rates in five counties – Borski (p = 3.10 × 10⁻³, z = 2.96), Braničevski (p = 2.10 × 10⁻⁵, z = 4.25), Jablanički (p = 1.53 × 10⁻³, z = 3.17), Moravički (p = 4.34 × 10⁻³, z = 2.8), and Zaječarski (p = 1.95 × 10⁻⁴, z = 3.73). These values are bolded in Table A1. Similarly, in all other tables, values with significant trends are also bolded. The same manner is used for all other presented results. In contrast, the remaining counties did not exhibit any notable trends. The graphical representation of this analysis is shown in Figure A1.

Figure 4

Hot spot analysis of incidence rates for: (a) colorectal cancer, (b) bladder cancer, (c) pancreatic cancer, and (d) prostate cancer.

When analysing mortality rate trends for this cancer, an increasing trend was observed in four counties: Braničevski (p = 0.04, z = 2.01), Jablanički (p = 1.27 × 10⁻³, z = 3.22), Rasinski (p = 0.02, z = 2.35), and Zaječarski (p = 0.03, z = 2.17). In contrast, a decreasing trend was recorded in Beogradski County (p = 0.03, z = −2.12). The graphical representation of this analysis is shown in Figure A2.

Figure 5

Hot spot analysis of mortality rates for (a) lung and bronchus cancer, (b) stomach cancer, and (c) laryngeal cancer.

From Table A2, it is visible that the counties showed an increasing trend in colorectal cancer incidence rates: Borski (p = 1.27 × 10⁻³, z = 3.22), Braničevski (p = 3.98 × 10⁻³, z = 2.88), Jablanički (p = 2.00 × 10⁻⁶, z = 4.75), Kolubarski (p = 1.06 × 10⁻³, z = 3.27), Mačvanski (p = 3.79 × 10⁻⁵, z = 4.12), Moravički (p = 0.01, z = 2.56), Pčinjski (p = 1.80 × 10⁻³, z = 3.12), Pirotski (p = 1.26 × 10⁻³, z = 3.22), Pomoravski (p = 1.10 × 10⁻⁵, z = 4.40), Rasinski (p = 3.35 × 10⁻⁶, z = 4.65), Raški (p = 6.02 × 10⁻³, z = 2.75), Šumadijski (p = 0.03, z = 2.14), and Zlatiborski (p = 8.27 × 10⁻³, z = 2.64). In contrast, Beogradski County (p = 0.03, z = −2.17) recorded a decreasing trend in colorectal cancer incidence rates. The graphical representation of this analysis is shown in Figure A3.

Figure 6

Map of the Mann–Kendall (MK) analysis for cancer incidence: (a) lung and bronchus, (b) colorectal, (c) stomach, (d) bladder, (e) laryngeal, (f) pancreatic, and (g) prostate cancer.

Regarding mortality rates, the following counties showed an increasing trend in the number of deaths: Braničevski (p = 3.43 × 10⁻³, z = 2.93), Jablanički (p = 0.01, z = 2.56), Mačvanski (p = 0.02, z = 2.33), and Zlatiborski (p = 6.47 × 10⁻⁴, z = 3.41). No significant trends were observed in the other counties. The graphical representation of this analysis is shown in Figure A4.

Figure 7

Map of the Mann–Kendall (MK) analysis for cancer mortality: (a) lung and bronchus, (b) colorectal, (c) stomach, (d) bladder, (e) laryngeal, (f) pancreatic, and (g) prostate cancer.

The results of the MK analysis for stomach cancer are presented in Table A3. A decreasing trend in stomach cancer incidence rates was identified in four counties: Beogradski (p = 0.03, z = −2.17), Kolubarski (p = 0.02, z = −2.27), Nišavski (p = 2.17 × 10⁻³, z = −3.07), and Toplički (p = 1.06 × 10⁻³, z = −3.27). In contrast, no significant trends were detected in the other counties. The graphical representation of this analysis is shown in Figure A5.

Regarding mortality rates for this cancer, a decreasing trend was observed in 15 counties: Beogradski (p = 2.22 × 10⁻⁷, z = −5.18), Borski (p = 0.02, z = −2.35), Braničevski (p = 2.40 × 10⁻⁴, z = −3.67), Jablanički (p = 1.98 × 10⁻³, z = −3.09), Kolubarski (p = 3.58 × 10⁻⁴, z = −3.57), Mačvanski (p = 1.26 × 10⁻³, z = −3.22), Moravički (p = 2.35 × 10⁻⁵, z = −4.23), Nišavski (p = 1.26 × 10⁻³, z = −3.22), Pčinjski (p = 1.67 × 10⁻³, z = −3.14), Pirotski (p = 0.01, z = −2.62), Podunavski (p = 5.38 × 10⁻⁴, z = −3.46), Pomoravski (p = 3.25 × 10⁻⁴, z = −3.59), Rasinski (p = 0.01, z = −2.54), Toplički (p = 5.96 × 10⁻⁵, z = −4.01), and Zlatiborski (p = 2.59 × 10⁻³, z = −3.01). The graphical representation of this analysis is shown in Figure A6.

Table A4 presents the results of this analysis for bladder cancer. An increasing trend in incidence rates was observed in nine counties: Jablanički (p = 5.23 × 10⁻⁵, z = 4.05), Mačvanski (p = 2.95 × 10⁻⁴, z = 3.62), Nišavski (p = 5.96 × 10⁻⁴, z = 3.43), Pčinjski (p = 0.03, z = 2.17), Pirotski (p = 3.36 × 10⁻³, z = 2.93), Pomoravski (p = 4.22 × 10⁻⁵, z = 4.10), Rasinski (p = 2.65 × 10⁻⁴, z = 3.65), Šumadijski (p = 0.01, z = 2.67), and Toplički (p = 3.98 × 10⁻³, z = 2.88). No significant trend was detected in the other counties. The graphical representation of this analysis is shown in Figure A7.

The results of MK analysis of laryngeal cancer are presented through Table A5. It revealed a decreasing trend in incidence rates in two counties – Beogradski (p = 5.94 × 10⁻⁷, z = −4.99) and Moravički (p = 0.02, z = −2.25), with no significant trend detected in the other counties. The graphical representation of this analysis is shown in Figure A9.

Counties that recorded a decreasing trend in laryngeal cancer mortality rates include nine counties: Beogradski (p = 4.69 × 10⁻⁴, z = −3.50), Braničevski (p = 0.02, z = −2.38), Jablanički (p = 0.01, z = −2.64), Kolubarski (p = 2.81 × 10⁻³, z = −2.99), Mačvanski (p = 0.01, z = −2.49), Moravički (p = 0.02, z = −2.38), Pomoravski (p = 0.03, z = − 2.11), Rasinski (p = 2.99 × 10⁻³, z = −2.97), and Zlatiborski (p = 1.97 × 10⁻³, z = −3.10). The graphical representation of this analysis is shown in Figure A10.

Table A6 shows the analysis of prostate cancer trends with an increasing incidence rate in the following counties: Beogradski (p = 0.01, z = 2.78), Borski (p = 1.28 × 10⁻⁴, z = 3.83), Braničevski (p = 4.63 × 10⁻³, z = 2.83), Jablanički (p = 1.24 × 10⁻³, z = 3.23), Mačvanski (p = 1.67 × 10⁻³, z = 3.14), Moravički (p = 4.79 × 10⁻⁶, z = 4.57), Nišavski (p = 7.86 × 10⁻⁴, z = 3.36), Pčinjski (p = 0.02, z = 2.33), Podunavski (p = 0.03, z = 2.22), Rasinski (p = 2.97 × 10⁻⁵, z = 4.18), Šumadijski (p = 2.83 × 10⁻³, z = 2.99), and Zlatiborski (p = 7.17 × 10⁻⁴, z = 3.38). No significant trend was recorded in the other counties for incidence rates. The graphical representation of this analysis is shown in Figure A11.

In terms of mortality rate trends for pancreatic cancer, Beogradski (p = 5.86 × 10⁻⁴, z = 3.44), Kolubarski (p = 0.04, z = 2.06), and Moravički (p = 2.15 × 10⁻³, z = 3.07) counties stand out, showing an increase in the number of deaths. The graphical representation of this analysis is shown in Figure A12.

The analysis of prostate cancer incidence rates revealed an increasing trend in four counties: Braničevski (p = 3.34 × 10⁻³, z = 2.93), Jablanički (p = 0.01, z = 2.56), Mačvanski (p = 0.02, z = 2.33), and Zlatiborski (p = 6.47 × 10⁻⁴, z = 3.41). No trend was observed in the other counties. These results are presented through parameters of the MK test in Table A7. The graphical representation of this analysis is shown in Figure A13.

Additionally, an increase in prostate cancer mortality was observed in Jablanički (p = 3.98 × 10⁻⁴, z = 3.54), Mačvanski (p = 0.01, z = 2.46), and Nišavski (p = 1.52 × 10⁻³, z = 3.17) counties. No significant trends were recorded in the other counties. The graphical representation of this analysis is shown in Figure A14.

Results of hot spot analysis

The second analysis focused on identifying hot spots in the incidence and mortality rates of the seven previously mentioned cancer types. The results are presented using parameters derived from the Hot Spot Analysis tool in ArcGIS Pro – z-scores and p-values, which describe the analysed patterns.

Table A8 presents the parameters from the hot spot analysis for colorectal cancer. In terms of incidence rates, three counties were identified as new hot spots: Mačvanski (z = 3.54, p = 4.02 × 10⁻⁴), Moravički (z = 5.44, p = 5.26 × 10⁻⁸), and Zlatiborski (z = 5.47, p = 4.53 × 10⁻⁸). In addition, oscillating hot spots for colorectal cancer incidence rates were detected in Beogradski (z = 2.54, p = 0.01), Braničevski (z = 4.54, p = 5.56 × 10⁻⁶), Kolubarski (z = 3.22, p = 1.27 × 10⁻³), Podunavski (z = 2.48, p = 0.01), and Pomoravski (z = 5.12, p = 3.00 × 10⁻⁷) counties. No spatial pattern was detected in any county regarding mortality rates for this cancer. The graphical representation of this analysis for incidence rates is shown in Figure 4a.

The hot spot analysis for lung and bronchus cancer is presented through parameters in Table A9. It is evident that no spatial patterns in incidence were detected in any county. In contrast, regarding mortality for lung and bronchus cancer, new cold spots were identified in four counties: Kolubarski (z = 0.42, p = 0.67), Mačvanski (z = 0, p = 1), Moravički (z = 0, p = 1), and Zlatiborski (z = 0, p = 1). In addition, seven counties were identified as sporadic cold spots: Jablanički (z = 2.06, p = 0.04), Pčinjski (z = 2.06, p = 0.04), Pirotski (z = 2.64, p = 0.01), Rasinski (z = 2.01, p = 0.04), Raški (z = 0.53, p = 0.60), Zaječarski (z = 1.37, p = 0.17), and Zlatiborski (z = 0, p = 1.00). In any other of eight counties, no pattern was detected. The graphical representation of this analysis for mortality rate is shown in Figure 5a.

The hot spot analysis of the bladder cancer is presented in Table A10. The parameters for incidence rates showed that two counties were defined as consecutive hotspots: Podunavski (z = 4.91, p = 9.00 × 10⁻⁷) and Šumadijski (z = 4.75, p = 2.00 × 10⁻⁶). Oscillating hotspots for this cancer incidence were detected in Braničevski (z = 2.32, p = 0.02), Jablanički (z = 5.49, p = 3.94 × 10⁻⁸), Pčinjski (z = 5.49, p = 3.94 × 10⁻⁸), Pirotski (z = 5.07, p = 3.96 × 10⁻⁷), Pomoravski (z = 3.62, p = 2.95 × 10⁻⁴), Rasinski (z = 4.60, p = 4.32 × 10⁻⁶), and Zaječarski (z = 3.78, p = 1.58 × 10⁻⁴) counties. Sporadic hotspots for bladder cancer incidence were detected in Beogradski (z = 3.49, p = 4.88 × 10⁻⁴) and Kolubarski (z = 3.43, p = 5.96 × 10⁻⁴) counties. In addition, no spatial patterns were identified in any county for bladder cancer mortality. The graphical representation of this analysis for incidence rates is shown in Figure 4b.

Table A11 presents the parameters from the hot spot analysis for pancreatic cancer. In terms of incidence rates, an oscillating hot spot was detected in Borski county (z = 3.73, p = 1.95 × 10⁻⁴), while no patterns were detected in any other counties. Similarly, no spatial patterns were identified in any county for pancreatic cancer mortality. The graphical representation of this analysis for incidence rates is shown in Figure 4c.

The hot spot analysis for prostate cancer is presented in Table A12. The parameters for incidence rates indicate that eight counties were identified as oscillating hot spots: Beogradski (z = 2.91, p = 3.67 × 10⁻³), Borski (z = 4.97, p = 6.86 × 10⁻⁷), Braničevski (z = 5.12, p = 3.00 × 10⁻⁷), Kolubarski (z = 5.12, p = 3.00 × 10⁻⁷), Podunavski (z = 4.65, p = 3.35 × 10⁻⁶), Pomoravski (z = 4.75, p = 2.00 × 10⁻⁶), Šumadijski (z = 4.97, p = 6.86 × 10⁻⁷), and Zaječarski (z = 5.12, p = 3.00 × 10⁻⁷). In contrast, no spatial patterns were identified in any county for prostate cancer mortality. The graphical representation of this analysis for incidence rates is shown in Figure 4d.

Table A13 presents the parameters from the hot spot analysis for stomach cancer. In terms of incidence rates, no spatial patterns were identified in any county. In contrast, for mortality, a new cold spot was identified in two counties: Beogradski (z = −5.12, p = 3.00 × 10⁻⁷) and Kolubarski (z = −5.07, p = 3.92 × 10⁻⁷), while consecutive cold spots were identified in three counties: Moravički (z = −4.65, p = 3.35 × 10⁻⁶), Zaječarski (z = − 4.60, p = 4.32 × 10⁻⁶), and Zlatiborski (z = −4.65, p = 3.35 × 10⁻⁶). In addition, oscillating cold spots were detected in thirteen other counties: Borski (z = −4.38, p = 1.16 × 10⁻⁵), Braničevski (z = −4.54, p = 5.56 × 10⁻⁶), Jablanički, (z = −4.54, p = 5.56 × 10⁻⁶), Mačvanski (z = −5.12, p = 3.00 × 10⁻⁷), Nišavski (z = −4.86, p = 1.17 × 10⁻⁶), Pčinjski (z = −4.54, p = 5.56 × 10⁻⁶), Pirotski (z = −4.54, p = 5.51 × 10⁻⁶), Podunavski (z = −4.97, p = 6.86 × 10⁻⁷), Pomoravski (z = −4.86, p = 1.17 × 10⁻⁶), Rasinski (z = −4.81, p = 1.53 × 10⁻⁶), Raški (z = −5.28, p = 1.28 × 10⁻⁷), Šumadijski (z = −5.76, p = 8.44 × 10⁻⁹), and Toplički (z = −4.54, p = 5.56 × 10⁻⁶). The graphical representation of this analysis for mortality rates is shown in Figure 5b.

Table A14 presents the parameters from the hot spot analysis for stomach cancer mortality. As in the previous cancer analysis, no spatial patterns were identified in any county for incidence rates. In contrast, for mortality, a new cold spot was identified in Mačvanski county (z = −5.34, p = 9.46 × 10⁻⁸), while consecutive cold spots were identified in three counties: Moravički (z = −4.94, p = 7.79 × 10⁻⁷), Raški (z = −4.12, p = 3.79 × 10⁻⁵), and Zlatiborski (z = −4.94, p = 7.79 × 10⁻⁷). No spatial patterns were detected in the remaining fourteen counties. The graphical representation of this analysis for mortality rates is shown in Figure 5c.

4 Discussion

Malignant tumours, after diseases of the heart and blood vessels, are the most common cause of illness and death in the territory of central Serbia [74]. This study has not dealt with the causes of the occurrence of carcinomas but rather with identifying the most and least burdened areas [2].

New hotspots for colorectal cancer incidence were identified in three counties – Mačvanski, Moravički, and Zlatiborski. In addition, consecutive hotspots for bladder cancer emerged in two counties: Podunavski and Šumadijski. Conversely, new cold spots in mortality rates were identified for lung and bronchus cancer in four counties – Kolubarski, Mačvanski, Moravički, and Zlatiborski; for stomach cancer in two counties – Beogradski, and Kolubarski; and for laryngeal cancer in one county – Mačvanski. The study also identified an increasing trend in both incidence and mortality rates for lung and bronchus cancers and colorectal cancers in three counties. In addition, both incidence and mortality rates for prostate and pancreatic cancers were found to be rising in two counties. Conversely, a decreasing trend in both incidence and mortality rates was observed for stomach cancer in four counties and for laryngeal cancer in two counties.

Figure 6a displays the lung and bronchus cancer incidence trends, revealing a prominent increasing trend primarily in the north-eastern part of Central Serbia, along with two other counties – one in the western and another in the southern region. A similar pattern is observed for colorectal cancer incidence trends in Figure 6b, consistent with both the Mann–Kendall (MK) and hot spot analyses (Figure 4a). An increasing trend in colorectal cancer incidence is mainly found in the north-western part of the study area, except for Beogradski County, where the MK analysis indicates a decreasing trend, while the hot spot analysis categorizes it as an “oscillating hot spot.” In addition, the MK analysis shows an increasing trend in three south-eastern counties.

The analysis of stomach cancer incidence (Figure 6c) shows a decreasing trend in two counties in the north-western and two in the southern regions. Similarly, MK analysis of laryngeal cancer incidence (Figure 6e) identifies a decreasing trend in two counties – one in the north and another in the west of the study area.

Regarding bladder cancer, both the MK analysis (Figure 6d) and hot spot analysis (Figure 4b) reveal similar trends, with some form of increasing trend noted in 6 of the 18 counties. The MK analysis indicates an increasing trend, while the hot spot analysis categorizes five of these six counties as “oscillating hot spots” and one as a “new hot spot.”

Observing pancreatic cancer incidence (Figure 6f), the MK analysis identifies an increasing trend across 12 counties – eight in the northern and western regions and four in the southern region. However, the hot spot analysis only categorizes one of these counties as an “oscillating hot spot” (Figure 4c).

For prostate cancer incidence, the two analyses converge in only one county, where the MK analysis (Figure 6f) shows an increasing trend, and the hot spot analysis categorizes it as an “oscillating hot spot” (Figure 4d).

Figure 7a shows the MK analysis results for lung and bronchus cancer mortality rates, indicating an increasing trend in four counties and a decreasing trend in one county. In addition, the hot spot analysis in Figure 5a categorizes six counties in the southern, eastern, and central parts of the study area as “sporadic cold spots” and four counties in the western part as “consecutive cold spots.”

The MK analysis of stomach cancer mortality rates shown in Figure 7c aligns with the hot spot analysis shown in Figure 5b. In only three counties, the MK analysis detected no trend; in the remaining 15 counties, it identified a decreasing trend. Correspondingly, the hot spot analysis classified all counties in the study area with some types of cold spot. Specifically, it categorised 13 counties as “oscillating cold spots,” three counties as “consecutive cold spots,” and two counties as “new cold spots.”

For laryngeal cancer mortality rates, both analyses converge in three counties in the western part of the study area. Figure 7e shows the MK analysis results, highlighting a decreasing trend in five counties in the western region, three counties in the central region, and one in the southern part of the study area. The hot spot analysis of laryngeal cancer mortality rates (Figure 5c) categorizes three counties in the south-western part as “consecutive cold spots” and one county in the western part as a “new cold spot.”

In Figure 7b, the MK analysis indicates an increasing trend in four counties – two in the western part of the study area, one in the northern part, and one in the southern part. Figure 7d, showing MK analysis results for bladder cancer mortality rates, reveals an increasing trend in two counties in the northern part of the study area. Similarly, Figure 7f presents the MK analysis for pancreatic cancer mortality rates, showing an increasing trend in three counties in the northern part.

Finally, Figure 7g displays the MK analysis results for prostate cancer mortality rates, identifying an increasing trend in three counties – two in the southern and one in the northern part of the study area.

The hotspot analysis revealed a decline in mortality rates, whereas incidence rates showed an increase in cancer cases. The analyses suggest that, despite an increase in newly diagnosed cancer cases among the male population, the number of deaths is decreasing. This indicates improvements in the healthcare system in Central Serbia, as well as the effectiveness of prevention and treatment measures. However, the number of deaths remains high, along with the incidence of new cases, highlighting the need for targeted protective measures, particularly in vulnerable areas. The reduction in mortality can be largely attributed to enacted legislation [75,76,77,78,79,80,81,82,83,84,85]. At the end of 2014, the Government of the Republic of Serbia adopted the Law on Health Documentation and Records in the Field of Health [86], along with the accompanying Rulebook on Medical Documentation and Records in Healthcare [87], which further defined the management of the Cancer Register. While the Republic of Serbia has introduced certain planning documents for the prevention and control of chronic non-communicable diseases [75,76,77,78,79,80,81,82,83,84,85], it is necessary to review existing programs and adopt new ones to further enhance the prevention and protection of the population from malignant diseases. Adopting new prevention programs would certainly have the effect of reducing the number of newly diagnosed and deceased persons, which would result in the improvement of the public health of the population of Central Serbia. Better implementation of nationally organised screening programmes in Serbia is of the greatest importance [28]. Hence, prevention and early diagnosis could improve survival and decrease cancer mortality in Serbia [88]. Of course, there is always a question of whether the cancer mortality trends were a consequence of variations in the process of registering the causes of death, as well as the reliability and validity of death certificates [4].

Mihalj et al. [89] highlighted the association between mining activities and the increased incidence of bronchial carcinoma in both male and female populations during the period 2010–2020. Their study identified the Borski County as having the highest incidence rates of lung and bronchial cancer [89]. When compared with the results of our study, it is evident that the MK analysis of lung and bronchial cancer incidence in the Borski County also recorded a rising trend in new cases among the male population.

Marković-Denić et al. [26] analysed mortality trends from cancer in Central Serbia between 1985 and 2002 and observed an increasing trend in both sexes. Mortality from lung cancer rose in both men and women, as did mortality from colorectal cancer [26]. On the other hand, stomach cancer mortality in men showed a moderate decline since the 1990s, while prostate cancer mortality remained relatively stable [26]. The results of that study cannot be fully compared with our study, as it examines an earlier time period. However, a comparison indicates that stomach cancer mortality is also decreasing in our study, which may be partly attributed to improvements in public health policies and health awareness.

A study by Mihajlović et al. [88] analysed cancer incidence and mortality in Serbia from 1999 to 2009 and also reported increasing trends in lung, colorectal, prostate, and bladder cancer among men.

Nikolić et al. [29] indicated that standardised colorectal cancer incidence rates in Central Serbia have been rising significantly, with a 1% annual increase in men. Similarly, a study conducted by Šimunović Gašpar and Pavić [90], which analysed colorectal cancer incidence and mortality in Croatia, also observed an upward trend. When compared with our study, which also recorded an increase in colorectal cancer cases, it is evident that this cancer represents a major public health concern in both Serbia and neighbouring countries.

Studies that have employed spatial analysis to identify cancer hotspots [13,14,16,18,19,20,22] have highlighted the importance and effectiveness of using GIS in spatial-temporal disease analyses, a finding that has also been confirmed by this study.

Air pollution in the Republic of Serbia poses a serious public health problem. In IQAir’s 2019 World Air Quality Report, Serbia was ranked as the 5th most polluted country in Europe among 37 European countries and regions, with an average PM2.5 concentration of 23.3 μg/m³, adjusted for population size [91]. The country’s air pollution is driven by several factors, primarily its dependence on lignite and coal-powered plants in the energy sector, along with the use of solid fuels like coal and wood for home heating [91,92]. In addition, pollution is significantly worsened by emissions from an aging transportation fleet, industrial operations, waste disposal sites, and agricultural activities [91,92]. In central Serbia, only a small number of studies have been conducted on the link between air pollution and cancer occurrence. In a large area of Belgrade, Perišić et al. [93] estimated the effects of PM10, as well as its components (the elements As, Cd, Cr, Mn, Ni, and Pb, as well as benzo[a]pyrene) on lung cancer, and found Cr to be the major contributor to carcinogenic health risk. Future research in central Serbia should focus on examining the link between air pollution and cancer occurrence, a connection already established by numerous studies worldwide [94,95,96,97,98,99].

Studies examining the incidence and mortality of stomach cancer reveal significant regional variability, largely influenced by population eating habits and cultural practices [100,101,102]. Factors such as high salt intake, consumption of products containing N-methyl-N-nitro-N-nitrosoguanidine, the use of additives and preservatives, and insufficient intake of fresh fruits and vegetables (leading to reduced levels of vitamins C, E, and beta-carotene) contribute to these differences [100,101,102]. Other risk factors include smoking, alcohol consumption, radiation exposure, toxic substances, Epstein-Barr virus infection, pernicious anaemia, gastroesophageal reflux, blood type A, low socioeconomic status, and genetic predisposition [100,101,102].

Since the 1986 IARC Monograph on “Tobacco Smoking,” studies have provided sufficient evidence to establish a causal relationship between cigarette smoking and cancers of the nasal cavities, paranasal sinuses, nasopharynx, stomach, liver, kidney (renal cell carcinoma), uterine cervix, adenocarcinoma of the oesophagus, and myeloid leukaemia [103]. Smoking habits, alcohol consumption, and their interaction are the main risk factors for cancers of the oesophagus, oral cavity, and pharynx [104]. Despite some measures by the Serbian healthcare system to control risk factors, such as the introduction of a legislative framework for tobacco control, more efforts are needed to reduce these risks [88,105]. The WHO has identified alcohol consumption as 1 of the top 10 risk factors contributing to the global burden of disease [106].

The 2019 population health survey of Serbia, conducted by the Republic Statistical Office (RZS) in cooperation with the Institute for Public Health of Serbia and the Ministry of Health of the Republic of Serbia, showed that in the territory of Central Serbia, a significantly higher percentage of obese people is recorded in the territory of Southern and Eastern Serbia (23.1%), in all age groups from 45 to 84, among the poor population (25.7%), the least educated (26.6%), as well as those living in suburban settlements (23.6%) [107]. In 2019, the residents of Southern and Eastern Serbia (93.0%), the poorest (89.2%), and those with the lowest level of education (89.9%) used bread daily in their diet [107]. The region of Šumadija and Western Serbia stands out for its smaller percentage of daily smokers (23.9%), while in Southern and Eastern Serbia, there is a higher percentage of smokers belonging to the age group of 15–19 years (21.4, and 17.5%, respectively) [107]. Residents of Southern and Eastern Serbia (4.0%) consume significantly more alcohol on a daily basis, in contrast to residents of Šumadija and Western Serbia, where this percentage is the lowest (2.1%) [107]. Working towards improving diet, physical activity, body weight, as well as alcohol and smoking cessation are the most important preventive measures [28]. Figure 5 shows that cold spots (both new and consecutive) occur in the Western Serbia region for mortality rates of lung and bronchus, stomach, and laryngeal cancer, which may be associated with healthier lifestyles and attitudes towards health. In contrast, oscillating hot spots were detected in some counties in the Eastern Serbia for the incidence rates of bladder, pancreatic, and prostate cancer, which could be linked to the previously mentioned “worse lifestyle.”

According to publications [5,6,7,8,9,10] on the territory of the Republic of Serbia, the Autonomous Province of Vojvodina has the largest number of newly diagnosed and deceased cancer patients. Data on cancers in the territory of AP Vojvodina at the county level are publicly available in the period 2016–2021 [5,6,7,8,9,10], and it was not possible to conduct a hot spot analysis and calculate trends in the specified period. To examine trends in the movement of a disease and determine the focus, it is necessary to analyse a period of at least 10 years, which was not possible in this case. Certainly, one of the reasons for the inconsistency of the data is the change in jurisdiction for recording newly diagnosed and deceased cancer patients on the territory of AP Vojvodina. The Cancer Registry of Vojvodina has data on patients and deaths as of 2012. By adopting certain legal regulations, the Institute for Public Health of Serbia took over, i.e., centralised the process of collecting, processing and analysing the aforementioned data.

In the publications of the Institute for Public Health of Serbia, Dr. Milan Jovanović Batut [5], it can be seen that in 2021, the age-specific incidence and mortality rates of all cancer localisations in men in the territory of the Republic of Serbia are the highest in the over 75 categories, and that it is increasing with years of age. In lung and bronchus cancer, the highest incidence rate is in the 70–74 age category and mortality in the 60–69 age category [5]. In the case of colon and rectal cancer in men during the same year, the largest number of newly diagnosed cases was recorded in the 70–74 age category, and the number of deaths in the age group over 75 years. The largest number of newly diagnosed and deceased from prostate cancer was recorded at the age of over 75 years [5]. Data on age-specific incidence and mortality rates of other cancers, as well as data presented at the level of districts and settlements in the territory of central Serbia, are not available in public publications. Hence, it is impossible to examine the age-specific incidence and mortality rate in the territory of central Serbia. In future research, age-specific rates, professional structure, as well as the socio-economic status and living conditions of patients and deceased in the “hot spots” of cancer in central Serbia should be examined to examine the causes of cancer, which is a very difficult and complex task.

As demonstrated in this research, the study relied exclusively on publicly available data, and the scope and direction of the research were dependent on it. The unavailability of data at the level of settlements and counties for certain indicators of health status is a major limiting factor. A suggestion of the study is to make statistical data on carcinoma available at the level of settlements and municipalities, so that one can carry out a more detailed geographical analysis of a disease [2]. As a result, the health conditions of the population of the Central Serbia would be significantly improved [2]. Publicly available data on the population health presented at a local level could have a financial impact [1]. Providing adequate financial resources would help ensure a positive economic impact by reducing the intensity of monitoring and the healthcare costs for the cancer-affected population.

The importance of spatial clusters, or “hotspots” in disease epidemiology has been increasingly recognised, and targeting hotspots is often seen as an important component of disease-control strategies [12]. The response to screening in Serbia remains low and needs to be improved, along with health education on the importance of early detection programs [108]. Access to cancer care innovations, including wider accessibility of clinical studies, needs to be intensified and guided by consensus actions to reduce further inequities in cancer outcomes currently observed between high and low/middle-income countries [108]. The opportunistic screening methods available usually do not target the general population, or at least the population at risk [88]. This type of study could assist in evaluating the effectiveness of policies and could facilitate prioritising resources [1,109].

According to the Program for Improving Cancer Control in the Republic of Serbia for the Period 2020–2022, the key measures for cancer protection include prevention, diagnosis and treatment, research on malignant diseases, psychosocial services, as well as rehabilitation, supportive oncology, and palliative care [75]. The diagnostics and therapy of malignant diseases are becoming increasingly complex due to numerous advancements in medical, biomedical, and biotechnological research [75].

In addition to a multidisciplinary approach to oncology patients, modern diagnostic and therapeutic procedures require the integration of health services, the establishment of centres of expertise, and national reference networks for conducting complex procedures [75]. These also involve centres dedicated to the application of innovative therapies, and the assessment of treatment quality. However, resources for implementing these measures are often limited, making it necessary to prioritise and select procedures that offer the greatest benefit to the population while enhancing treatment outcomes [75].

Many patients encounter slow and inadequate diagnostic processes, often enduring long wait times [75]. Significant limitations primarily affect pathological and radiological diagnostics. There is an insufficient number of specialists, and specific cytological, pathological, biochemical, and genetic analyses are not readily available across various centres, nor are they evenly distributed throughout the territory [75]. According to the publication Health and Statistical Yearbook of the Republic of Serbia 2021, published by the Institute for Public Health of Serbia “Dr. Milan Jovanović Batut,” the Nišavski County in central Serbia stands out for having the highest number of doctors and nurse-technicians per 100,000 inhabitants (443 doctors and 853 nurse-technicians per 100,000) [64]. In contrast, the Rasinski and Mačvanski counties have the lowest numbers of doctors (220 in Rasinski, and 223 in Mačvanski) and nurse-technicians (485 in Rasinski, and 521 in Mačvanski) [64]. The hot spot analysis also found that in the Nišavski County, an oscillating cold spot was detected for stomach cancer mortality, while no patterns were identified for other cancers. This may be partially attributed to the positive organisation of the healthcare system and its impact on public health. However, more comprehensive research is needed to further examine this connection.

In Central Serbia, there is a notable shortage of psychologists, and no psychiatrists are employed in health institutions specialising in oncology, including both secondary and tertiary care centres with oncology departments [75]. For instance, the Institute of Oncology and Radiology of Serbia, the University Children’s Clinic, the Institute for Mother and Child, KC Niš (Clinic for Children’s Internal Diseases), and KC Kragujevac (Clinic for Paediatrics), each employ has only one psychologist [75]. Several other institutions do not have any psychologists, which represents a significant gap in care [75]. The Program for Improving Cancer Control in the Republic of Serbia for the 2020–2022 period notes that while some institutions may have a psychologist, they are not specifically assigned to work with oncology patients [75].

Experts from various fields need to collaborate in the analysis of carcinoma, examining the causes of the disease collectively rather than in isolation [2]. While developed countries have made significant advancements in the medical and geographical approaches to disease studies, Serbia remains behind, with medical geography still underdeveloped [33].

Population health is one of the key indicators of societal development, aligning with Sustainable Development Goal 3 (Good Health and Well-Being) [110]. According to the United Nations [110], the goal is to ensure healthy lives and promote well-being for all people, at all ages, encompassing major health priorities such as reproductive, maternal, newborn, child, and adolescent health; communicable and non-communicable diseases; universal health coverage; and access to safe, effective, high-quality, and affordable medicines and vaccines for all. Therefore, achieving these objectives should be a priority in the development of the Republic of Serbia.

5 Concluding remarks

As presented in this study, lung, bronchus, colorectal, stomach, bladder, laryngeal, pancreatic, and prostate cancers are among the most prevalent diseases affecting the male population in Central Serbia. This research stands as one of the pioneering geographical medical studies focused on analysing cancer patterns in the region. Such studies are invaluable as they enable the tracking of disease trends within specific populations and offer insights into potential future developments. A comprehensive understanding of spatiotemporal patterns is critical not only for medical geography but also for effective public health management and policy-making. By mapping the distribution of cancer cases, this study helps identify high-risk areas, assess the effectiveness of existing health protection strategies, and inform the development of targeted interventions. Ultimately, the findings of this research can play a crucial role in guiding public health initiatives, optimising resource allocation, and enhancing preventive measures to improve cancer outcomes in Central Serbia. As previously noted, the primary limitation of this study is the lack of access to cancer statistical data at the settlement level. Therefore, it is recommended to make cancer data publicly available at both the settlement and municipal levels to enable a more comprehensive and precise geographical analysis of the disease. Future research could explore the underlying socioeconomic and environmental factors contributing to cancer hotspots identified in this study, offering a deeper understanding of the social determinants of health. Expanding the scope of medical geography research to include a wider range of health conditions could help create a more comprehensive framework for public health planning in Central Serbia. Finally, longitudinal studies tracking cancer trends over time would be essential to monitor the effectiveness of interventions and assess changes in disease patterns, enabling more dynamic and responsive health policies at the national level.