Historical data analysis and future prediction of lung cancer in Zhejiang province, China

Abstract

Environmental factors have known to cause lung cancer. However, the relationship between lung cancer and the environmental factors are not well explored or quantified. Using Zhejiang Province of China as a case study, the spatiotemporal distribution of lung cancer and associated environmental factors were investigated. The data of lung cancer were obtained from Zhejiang Cancer Registry Annual Report and China Cancer Registry Annual Report from 2000 to 2022. The environmental factors are derived from China County Statistical Yearbook and so on. The JoinPoint, SaTScan, spatial autocorrelation technologies and spatial regression models were used to investigate the spatiotemporal clustering, spatial distribution characteristics analysis and correlation of environmental factors of lung cancer. The incidences of lung cancer in Zhejiang Province increased over time, most incidences occurred in the north and became less so in the south of the province, showing a spatial aggregation distribution. The increase of urbanization rate has increased the incidence of lung cancer, and the proportion of environment-related lung adenocarcinoma has increased year by year. The spatial regression model showed that the incidence of lung cancer was negatively correlated with vegetation index, and positively correlated with annual average per capita GDP. There was spatial heterogeneity in the distribution of lung cancer incidence in Zhejiang Province, China, and the areas with high incidence were located in areas with high degree of industrialization and urbanization. Vegetation and urbanization rate have an impact on the incidence of lung cancer.

Supplementary Information

The online version contains supplementary material available at 10.1038/s41598-025-07200-5.

Introduction

Lung cancer is one of the malignant tumors with the highest morbidity and mortality rates in the world, seriously endangering human health¹. Global Cancer Epidemiology Statistics (GLOBOCAN) showed that in 2020, there were about 2.2 million new cases of lung cancer worldwide and about 1.8 million cases of lung cancer deaths, accounting for 11.4% and 18.0% of all new cases of malignant tumors and cancer-related deaths, respectively^2,3. It is predicted that the annual number of lung cancer deaths worldwide will reach 3 million by 2035⁴. As the country with the second largest population in the world, China faces unprecedented challenges in the prevention and treatment of lung cancer. Over the past 10 years, with the aging of the population and the popularization of lung cancer screening across the country, the incidence of lung cancer in the Chinese population has been on a slow rise^3,5,6. GLOBOCAN showed that in 2022, there were about 871,000 new cases of lung cancer in China and about 767,000 cases of lung cancer death, accounting for 18.1% and 23.9% of all malignant tumor cases and cancer-related deaths, respectively⁷.

A large number of studies have confirmed that smoking is the main risk factor for lung cancer^8,9. A study from the UK Biobank found that both intrauterine tobacco exposure and post-birth smoking increased the risk of lung cancer¹⁰. Chronic comorbidity is also a risk factor for lung cancer. He et al.¹¹ and Choi et al.¹² reported that patients with immune-related diseases had an increased risk of developing lung cancer. Wei et al.¹³found that a high neuroticism score was associated with an increased risk of lung cancer. Cowdery et al.¹⁴confirmed that depressive symptoms are a risk factor for lung cancer. Moreover, outdoor air pollution causes millions of deaths worldwide each year, of which more than 500,000 are due to lung cancer¹⁵. Indoor solid fuel¹⁶. and outdoor PM2.5 exposure¹⁷have also been shown to increase the risk of death from lung cancer. In addition, lack of physical activity and poor sleep¹⁸and an unhealthy diet¹⁹are also risk factors for lung cancer. Ji et al.²⁰found that the inflammatory factor C-reactive protein was associated with an increased risk of lung cancer. The World Health Organization (WHO) estimates that non-smoking lung cancer patients account for 25% of all cases of lung cancer, ranking seventh among the causes of cancer-related death in the world²¹. Lung cancer is mainly divided into small cell lung cancer and non-small cell lung cancer; the latter mainly includes squamous cell carcinoma and adenocarcinoma. The occurrence of squamous cell carcinoma is closely related to smoking, and adenocarcinoma is mostly related to external factors such as gene mutation and environmental factors^22–24. In the beginning of the 2000s, the proportion of adenocarcinoma in lung cancer in China was about 30%²⁵. However, with the continuous advancement of urbanization and industrialization, the problem of environmental pollution has become increasingly prominent, and the impact of air pollution on the occurrence and development of lung cancer has become increasingly obvious, the incidence of adenocarcinoma has shown a significant upward trend²⁶.

With the rapid advancement of urbanization, industrialization, and aging, the incidence of lung cancer remains high in Zhejiang Province. Since 2020, the incidence and mortality rate of lung cancer have been the first cancer spectrum in Zhejiang Province for 20 consecutive years²⁷. However, Current research on the relationship between environmental factors and lung cancer is very limited. Identifying high-risk areas and potential causative factors for lung cancer is important to provide a basis for rational interventions and health service policies. Therefore, we collected data on the incidence and mortality of lung cancer in Zhejiang Province from 2000 to 2019, using spatial epidemiological methods, in order to provide scientific evidence for formulating rational intervention measures and health service policies.

Materials and methods

Data collection

The data of lung cancer in Zhejiang Province from 2000 to 2019 are derived from the Annual Report of Zhejiang Cancer Registry from 2000 to 2022 and the Annual report of China Cancer Registry.The PM2.5 data comes from 0.01°x 0.01° resolution PM2.5 data shared by the Atmospheric Composition Analysis Group at the University of Washington.

The vegetation index is from NASA

MOD13A3 set (https://search.earthdata.nasa.gov/search).Social and economic data come from Zhejiang Statistical Yearbook and China County Statistical Yearbook.

The three-year average values of GDP, per capita GDP, economic density, population density, total grain production in each county, number of industrial enterprises per km2,used as explanatory variables.

Statistical analysis

JoinPoint analysis

Modeling was performed by the grid search method (GSM), and the model was optimized by the Monte Carlo permutation test. The annual percent change (APC) and average annual percent change (AAPC) for lung cancer incidence and mortality were calculated^28,29.

Spatiotemporal scan statistics

The incidence and mortality of lung cancer were detected by spatiotemporal clustering using spatiotemporal scanning statistics. Given that the minimum temporal resolution of the data was yearly, the scanning parameters were set as follows: the time range was 2014–2016, with the time window constrained to a 1-year interval in the analysis; the retrospective analysis adopted spatiotemporal scanning as the index; the statistical selection of discrete scanning was based on a Poisson model; the maximum spatial clustering size was limited to 50% of the potential risk population; and the Monte Carlo simulation was restricted to 999 times.

where n is the number of lung cancer and death cases in the scan window, N is the total number of lung cancer and death cases in the population, and E is the expected number of lung cancer and death cases under the null hypothesis. If the null hypothesis is true, I = 1 when the scan window has more lung cancer and death cases than expected; otherwise, I = 0.

Spatial autocorrelation analysis

Anselin’s local Moran’s I coefficient was used to detect the “hotspots” of tumor monitoring sites, so that the type of spatial clusters with high or low incidence of monitoring sites could be determined: The global Moran’s I coefficient (1) and the local Moran’s I coefficient (2) were calculated as follows:

1

2

3

where n is the number of geographical units, is the spatial weight coefficient of regions i and j, the conceptualization of the spatial relationship was based on inverse distance weights, is the distance between regions i and j, is the bandwidth, and are the incidence or mortality of lung cancer in regions i and j, respectively, and is the average incidence or mortality of lung cancer in all geographical units.

Spatial regression analysis

Spatial regression analysis can fully consider the autocorrelation and heterogeneity of spatial data and can better utilize and analyze its spatial characteristics. The general form of the global regression model is as follows :³⁰

Based on whether the residuals were independent and whether the Lagrange multiplier (LM) and the robust Lagrange multiplier (R-LM) were statistically significant and based on other indicators such as R²we evaluated the submodel’s Akaike information criterion (AIC), Schwarz criterion (SC), and maximum log-likelihood (LIK) to determine the type of submodel^31,32.

Sensitive analyses

To further investigate the impact of different weighting methods on the spatial autocorrelation and spatial regression results of lung cancer incidence and mortality rates in this study, the following sensitivity analyses were conducted using R language, and the specific formulas are shown in the attachment:

1. Adjust the power parameter (Power = 1–3) and bandwidth (bandwidth = 100 km, 150 km) of the inverse distance weighting;

2. Construct a spatial weight matrix using the K-Nearest Neighbors (KNN) method. Based on the 14 study sites in this research, the approximate number of nearest neighbors (K) was calculated using the empirical formula, yielding K ≈ 3 or 4. Thus, results for K = 3 and K = 4 were compared;

3. Construct a spatial weight matrix using the Gaussian kernel density method, with bandwidths consistent with those of IDW (bandwidth = 100 km, 150 km).

4. Replace the aforementioned spatial weight matrices in the spatial regression analysis to evaluate their effects.

Results

Descriptive statistics of lung cancer

From 2000 to 2019, the total incidence of lung cancer in men, women, and both sexes in Zhejiang Province, China, showed an increasing trend; the total incidence of lung cancer increased 168% from 40.59 per 100,000 in 2000 to 108.75 per 100,000 in 2019. Among them, the incidence of lung cancer in males increased 110% from 59.00 per 100,000 in 2000 to 123.72 per 100,000 in 2019, and in females, it increased 340% from 21.34 per 100,000 in 2000 to 93.89 per 100,000 in 2019. The incidence in males was significantly higher than that in females. At the same time, from 2000 to 2019, the total lung cancer mortality rate of men, women, and both sexes in Zhejiang Province, China, also showed an increasing trend, with the total lung cancer mortality rate of men and women rising from 35.11 per 100,000 in 2000 to 52.31 per 100,000 in 2019. Among them, the total lung cancer mortality rate of men increased from 49.86 per 100,000 in 2000 to 75.68 per 100,000 in 2019, and in women, it increased from 19.67 per 100,000 in 2000 to 29.10 per 100,000 in 2019. The mortality rate was significantly higher for males than for females.

The main types of lung cancer in Zhejiang Province are adenocarcinoma, squamous cell carcinoma, and large cell lung carcinoma. Among them, lung adenocarcinoma ranks first, with its proportion rising year by year, from 43.52% in 2012 to 73.75% in 2019.

Time trends in lung cancer

JoinPoint analysis showed that during 2000–2019, the incidence of lung cancer in Zhejiang Province has an inflection point in 2014, and the APC (95%CI) corresponding to the two intervals was APC_{2000 − 2014} = 3.72 (95%CI 2.95–4.50, P < 0.001) and APC_{2014 − 2019} = 10.16 (95%CI 6.31–14.16, P < 0.001), and global AAPC_{2000 − 2019} =5.38 (95%CI 4.33–6.44, P < 0.001) (Table 1). There was no inflection point in the incidence of lung cancer in males in Zhejiang Province from 2000 to 2019 (AAPC_{2000 − 2019} = 3.76, 95%CI 3.20–4.32, P < 0.001), with an average annual increase of 3.76%. From 2000 to 2019, the incidence of lung cancer in females in Zhejiang Province showed an inflection point in 2014, and the APC (95%CI) corresponding to the two intervals was APC_{2000 − 2014} = 5.45 (95%CI 4.34–6.57, P < 0.001) and APC_{2014 − 2019} = 16.76 (95%CI 11.01–22.81, P < 0.001), and global AAPC_{2000 − 2019} = 8.32 (95%CI 6.79–9.86, P < 0.001), with an average annual increase of 8.32% .

Table 1.

Trends in incidence and mortality of lung cancer in Zhejiang Province from 2000 to 2019.

Type	Joinpoints	Years	APC	95%CI	P	Years	AAPC	95%CI	P
Incidence (total)	1	2000–2014	3.72	2.95–4.50	< 0.001	2000–2019	5.38	4.33–6.44	< 0.001
		2014–2019	10.16	6.31–14.16	< 0.001
Incidence (male)	0	2000–2019	3.76	3.29–4.23	< 0.001	2000–2019	3.76	3.20–4.32	< 0.001
Incidence (female)	1	2000–2014	5.45	4.34–6.57	< 0.001	2000–2019	8.32	6.79–9.86	< 0.001
		2014–2019	16.76	11.01–22.81	< 0.001
Mortality (total)	3	2000–2003	0.69	− 2.77 to 4.27	0.667	2000–2019	2.15	0.96–3.36	< 0.001
		2003–2006	7.58	0.32–15.37	0.042
		2006–2014	3.01	2.05–3.97	< 0.001
		2014–2019	− 1.45	− 2.97 to 0.11	0.064
Mortality (male)	1	2000–2010	4.30	3.42–5.20	< 0.001	2000–2019	3.11	2.49–3.73	< 0.001
		2010–2019	1.80	0.78–2.82	0.002
Mortality (female)	3	2000–2002	−4.41	− 8.92 to 0.32	0.064	2000–2019	2.36	1.41–3.31	< 0.001
		2002–2006	7.23	4.67–9.85	0.001
		2006–2012	4.48	3.36–5.61	0.002
		2012–2019	−1.15	− 1.78 to 0.51	0.002

APC annual percentage change, AAPC average annual percent change, CI confidence interval.

During 2000–2019, the mortality rate of lung cancer in Zhejiang Province showed three inflection point in 2003, 2006 and 2014, and the APC values corresponding to the three intervals were APC_{2000 − 2003} = 0.69 (95%CI −2.77 to 4.27, P = 0.667), APC_{2003 − 2006} = 7.58 (95%CI 0.32–15.37, P = 0.042), APC_{2006 − 2014} = 3.01 (95%CI 2.05–3.97, P < 0.001) and APC_{2014 − 2019} = − 1.45 (95%CI −2.97 to 0.11, P = 0.064). The global AAPC (AAPC_{2000 − 2019} = 2.15, 95%CI 0.96–3.36, P < 0.001) indicates that the average annual increase of lung cancer mortality in Zhejiang Province during 2000–2019 was 2.15%. The mortality rate of male lung cancer in Zhejiang Province from 2000 to 2019 showed an inflection point in 2010, with an average annual increase of 3.11%. The mortality rate of female lung cancer in Zhejiang Province from 2000 to 2019 showed three inflection point in 2002, 2006, and 2012, with an average annual increase of 2.36%.

Spatial scanning of lung cancer at tumor monitoring sites in Zhejiang Province

The SaTScan spatial analysis of the incidence of lung cancer in Zhejiang Province from 2014 to 2016 showed that one cluster with high incidence and one cluster with low incidence were found in the study area (Fig. 1). The primary cluster area with high incidence (LLR = 45.40, RR = 1.30, P < 0.001) was located in northern Zhejiang, covering Daishan, Cixi, Yinzhou, Shangyu, and Jiashan with a radius of 138.38 km. Among them, 1795 observed cases and 1457 expected cases were reported. The primary cluster area with low incidence (LLR = 33.58, RR = 0.70, P < 0.001) was located in southern Zhejiang, covering Lucheng and Xianju, with 503 observed cases and 700 expected cases (Table 2).

Fig. 1.

Spatial clustering map of lung cancer incidence in Zhejiang Province, China, from 2014 to 2016.

Table 2.

Spatial analysis of lung cancer incidence in Zhejiang province, china, from 2014 to 2016.

Statistic	High-incidence cluster	Low-incidence cluster
Cities	Daishan, Yinzhou, Cixi, Shangyu, Jiashan	Lucheng, Xianju
Radius (km)	138.38	92.72
Observed cases	1795	503
Expected cases	1457	700
RR	1.30	0.70
LLR	45.40	33.58
P-value	< 0.001	< 0.001

Spatial autocorrelation analysis

Spatial autocorrelation analysis of morbidity

Table 3 Showed the results of Moran’s I global spatial autocorrelation analysis. The average incidence of lung cancer in monitoring sites in Zhejiang Province, China, during 2014–2016 showed spatial autocorrelation (Moran’s I = 0.373, P = 0.013), indicating that the average incidence of lung cancer in the monitoring area was spatially positive. It was statistically significant in both males and females (male Moran’s I = 0.294, P = 0.026; female Moran’s I = 0.355, P = 0.008) .

Table 3.

Moran’s I for lung cancer incidence in Zhejiang province, china, from 2014 to 2016.

Gender	Moran’s I	E(I)	Var(I)	P-value
Males	0.294	− 0.077	0.168	0.026
Females	0.355	− 0.077	0.154	0.008
Overall	0.373	− 0.077	0.166	0.013

The significance level is 0.05. Moran’s I indicates the Moran index, E(I) is the global correlation statistic, and Var(I) is the standard deviation of the global correlation statistic.

Univariate LISA local spatial autocorrelation analysis: Local hotspots of the average incidence of lung cancer were observed from the the LISA results in Zhejiang Province, China, from 2014 to 2016. The overall average incidence was statistically significant at monitoring sites. The LISA results of the average incidence of male lung cancer were statistically significant at monitoring sites, and the LISA results of the average incidence of female lung cancer were statistically significant at monitoring sites. The average incidence of lung cancer showed two clusters, which were high-high and low-low. The high-high cluster of the incidences was located in Cixi, Shangyu of Shaoxing, Ningbo, and Daishan of Zhoushan, and the low-low cluster of which was located in Yongkang and Lucheng of Wenzhou. The average incidence of male lung cancer occurred at both high-high and low-low clusters, and the high-high cluster area was located in Jiaxing, Cixi, Shangyu of Shaoxing, and Ningbo. Whereas the low-low cluster area is located in Yongkang and Lucheng of Wenzhou. The average incidence of female lung cancer also showed high-high and low-low clusters. The high-high cluster of female lung cancer was located in Cixi, Ningbo, and Daishan of Zhoushan; and the low-low cluster of female lung cancer was located in Yongkang and Lucheng of Wenzhou.

Spatial autocorrelation analysis of mortality

Moran’s I global spatial autocorrelation analysis. showed that there was no spatial autocorrelation in the three-year overall mean lung cancer mortality at the monitoring sites during 2014–2016 (Moran’s I = 0.175, P = 0.091). After gender stratification, there was no spatial autocorrelation in the mean lung cancer mortality of males (Moran’s I = 0.133, P = 0.108). The mean mortality rate of female lung cancer was spatially positive (Moran’s I = 0.292, P = 0.026) (Table 4).

Table 4.

Moran’s I for lung cancer mortality rate in Zhejiang province, china, from 2014 to 2016.

Gender	Moran’s I	E(I)	Var(I)	P-value
Males	0.133	− 0.077	0.171	0.108
Females	0.292	− 0.077	0.162	0.026
Overall	0.175	− 0.077	0.182	0.091

The significance level is 0.05. Moran’s I indicates the Moran index, E(I) is the global correlation statistic, and Var(I) is the standard deviation of the global correlation statistic.

The LISA results of the average mortality of lung cancer in Zhejiang Province from 2014 to 2016 displayed local hotspots and clusters. The overall average mortality rate of lung cancer was statistically significant at monitoring points; the LISA result of the average mortality rate of male lung cancer was statistically significant at monitoring sites and was consistent with the distribution of the overall average incidence; the LISA result of the average mortality rate of female lung cancer was statistically significant at monitoring sites. The average mortality rate of lung cancer showed three clusters, which were high-high, high-low, and low-high. The high-high cluster was located in Cixi and Shangyu of Shaoxing, high-low cluster was located in Yongkang, and low-high cluster was located in Ningbo. The average mortality rate of male lung cancer showed three clusters, which were high-high, high-low, and low-high. The high-high cluster was located in Cixi and Shangyu of Shaoxing, high-low cluster was located in Yongkang, and low-high cluster was located in Ningbo. The average mortality rate of female lung cancer showed three clusters, high-high, low-high, and low-low. The high-high cluster was located in Cixi and Shangyu of Shaoxing, low-high cluster was located in Ningbo, and low-low cluster was located in Yongkang.

Spatial regression analysis

The spatial regression analysis showed that the average incidence of male lung cancer was negatively correlated with the number of beds in medical institutions per 10,000 people (β = −0.347, P = 0.021) and the number of doctors per 10,000 people (β = − 0.865, P = 0.018), with no correlation with the other variables (Table 5). The average mortality rate of male lung cancer was negatively correlated with the number of beds in medical institutions per 10,000 people (β = −0.238, P = 0.030) and the number of doctors per 10,000 people (β = − 0.596, P = 0.025) but was not correlated with the other variables (Table 6). The average incidence of female lung cancer was significantly correlated with NDVI (β = −163.485, P = 0.037) and GDP per capita (β = 4.002, P = 0.035), indicating that the average incidence of female lung cancer was negatively correlated with NDVI and positively correlated with annual average GDP per capita. The mortality of female lung cancer was significantly correlated with NDVI (β = − 58.979, P = 0.006), indicating that the mean mortality of female lung cancer was negatively correlated with NDVI. Total lung cancer incidence and mortality were not significantly associated with all included variables at the 0.05 level.

Table 5.

Model screening and fitting results of regression analysis of environmental factors and lung cancer incidence.

Variable	Males					Females					Overall
	Model	Coefficient	Std. error	Statistic	P	Model	Coefficient	Std. error	Statistic	P	Model	Coefficient	Std. error	Statistic	P
PM2.5	OLS	0.988	0.744	1.329	0.208	OLS	0.248	0.796	0.312	0.760	OLS	0.576	0.719	0.802	0.438
NDVI	OLS	− 130.675	75.132	−  1.739	0.108	OLS	− 163.485	69.858	− 2.340	0.037	OLS	− 142.135	66.179	− 2.148	0.053
Population density	OLS	− 110.862	107.905	− 1.027	0.325	OLS	11.114	112.820	0.099	0.923	SEM	− 56.665	71.633	− 0.791	0.429
Economic density	SEM	− 10.711	6.897	− 1.553	0.120	OLS	3.049	9.820	0.311	0.762	SLM	− 3.694	6.766	− 0.546	0.585
GDPPC	OLS	2.345	1.922	1.221	0.246	OLS	4.002	1.685	2.375	0.035	OLS	3.066	1.667	1.839	0.091
Number of medical beds	OLS	− 0.347	0.131	− 2.642	0.021	OLS	− 0.141	0.161	− 0.878	0.397	SLM	− 0.199	0.103	− 1.936	0.053
Number of Doctors	OLS	− 0.865	0.315	− 2.744	0.018	OLS	− 0.253	0.397	− 0.640	0.534	SLM	− 0.452	0.255	− 1.770	0.077
Urbanization rate	OLS	− 155.652	145.183	− 1.072	0.305	SEM	− 181.985	103.840	− 1.753	0.080	SEM	− 173.908	95.356	− 1.824	0.068
Total grain production	OLS	− 0.024	0.431	− 0.055	0.957	SEM	− 0.313	0.353	− 0.888	0.374	SEM	− 0.412	0.312	− 1.322	0.186
Number of industrial enterprises	OLS	7.052	12.703	0.555	0.589	OLS	8.878	12.644	0.702	0.496	OLS	7.290	11.729	0.622	0.546

Significance value bold.

Table 6.

Model screening and fitting results of regression analysis of environmental factors and lung cancer mortality.

Variable	Males					Females					Overall
	Model	Coefficient	Std. error	Statistic	P	Model	Coefficient	Std. error	Statistic	P	Model	Coefficient	Std. error	Statistic	P
PM2.5	OLS	0.820	0.521	1.574	0.141	OLS	0.117	0.230	0.508	0.621	OLS	0.431	0.349	1.234	0.241
NDVI	OLS	− 88.172	54.761	− 1.610	0.133	OLS	− 58.979	17.665	− 3.339	0.006	OLS	− 68.603	33.767	− 2.032	0.065
Population density	OLS	− 74.544	77.929	− 0.957	0.358	OLS	10.560	32.704	0.323	0.752	OLS	− 35.550	51.393	− 0.692	0.502
Economic density	OLS	− 6.969	6.769	− 1.029	0.324	OLS	1.175	2.849	0.412	0.687	OLS	− 3.234	4.482	− 0.722	0.484
GDPPC	OLS	1.372	1.409	0.974	0.349	OLS	1.018	0.517	1.970	0.072	OLS	1.083	0.896	1.210	0.250
Number of medical beds	OLS	− 0.238	0.097	− 2.461	0.030	OLS	− 0.040	0.047	− 0.849	0.412	OLS	− 0.140	0.065	− 2.143	0.053
Number of Doctors	OLS	− 0.596	0.232	− 2.567	0.025	OLS	− 0.056	0.116	− 0.487	0.635	OLS	− 0.328	0.162	− 2.029	0.065
Urbanization rate	OLS	− 69.960	107.280	− 0.652	0.527	OLS	− 30.739	43.455	− 0.707	0.493	OLS	54.012	69.026	− 0.782	0.449
Total grain production	OLS	0.020	0.309	0.066	0.949	OLS	− 0.053	0.125	− 0.423	0.680	OLS	− 0.022	0.201	− 0.109	0.914
Number of industrial enterprises	OLS	4.125	9.164	0.450	0.661	OLS	2.164	3.702	0.585	0.570	OLS	2.480	5.948	0.417	0.684

Significance value bold.

Sensitive analyses

The sensitivity analysis results of spatial autocorrelation show that. Among the results for men, When the IDW method has Power = 1 and Bandwidth = 100 km (Moran’s I = 0.439, P = 0.006), The KNN method shows stronger spatial autocorrelation when K = 4 (Moran’s I = 0.388, P < 0.001) and the Kernel method shows stronger spatial autocorrelation when Bandwidth = 100 km (Moran’s I = 0.456, P = 0.004). Among the outcomes for women, When the IDW method has Power = 3 and Bandwidth = 150 km (Moran’s I = 0.421, P = 0.004), The KNN method shows stronger spatial autocorrelation when K = 3 (Moran’s I = 0.379, P = 0.001) and the Kernel method shows stronger spatial autocorrelation when Bandwidth = 128.5835 km (Moran’s I = 0.360, P = 0.002). In the results of the total population, when the IDW method had Power = 3 and Bandwidth = 150 km (Moran’s I = 0.466, P = 0.003), The KNN method shows stronger spatial autocorrelation when K = 3 (Moran’s I = 0.435, P = 0.001) and the Kernel method shows stronger spatial autocorrelation when Bandwidth = 100 km (Moran’s I = 0.447, P = 0.003).

The same analysis was conducted on the mortality rate. It is observed that the construction methods of different spatial weight matrices have a certain influence on the spatial autocorrelation results of the incidence rate. In the results of men and the total population, The significance of the results of spatial autocorrelation is inconsistent. This might be because as the power increases, the distance attenuation effect intensifies, and the weights of distant neighbors are significantly weakened. Introducing more long-distance regions as neighbors may dilute the local spatial correlation. However, as the bandwidth further expands, some medium-distance similar regions may be included to restore some spatialautocorrelation. However, the results for women were relatively uniform, possibly because the spatial heterogeneity of lung cancer mortality rate among women was stronger, which was consistent with the original conclusion.

Discussion

At present, it is recognized that lung cancer is the number one malignant tumor threatening human health, with the highest incidence and mortality rates in the world, accounting for 21.3% of the global incidence of cancer³³. The incidence of lung cancer in Zhejiang Province, China, showed a significant upward trend from 2000 to 2019^34–36, increasing from 40.59 per 100,000 in 2000 to 108.75 per 100,000 in 2019, and the incidence in males was significantly higher than that in females^37,38. At the same time, the mortality rate of lung cancer in Zhejiang Province, China, rising from 35.11 per 100,000 in 2000 to 52.31 per 100,000 in 2019. The age-specific incidence and mortality of lung cancer are at a low level before the age of 40 and then increase rapidly. Morbidity and mortality peaked in patients aged over 80 years in both urban and rural areas³⁹ and were higher in males than in females⁴⁰. The effect of air pollution on the occurrence and development of lung cancer is becoming increasingly obvious. Studies have shown that the proportion of lung adenocarcinoma has increased year by year, while the proportion of lung squamous cell carcinoma has not changed significantly. According to WHO statistics, adenocarcinoma has been the most common histological type of lung cancer worldwide since 2004⁴¹.

Although smoking is one of the main risk factors for lung cancer^2,42,43, in-depth studies have found that exposure to environmental tobacco smoke has gradually increased among non-smoking people (also known as passive smoking). The number of lung cancer cases in this group is increasing year by year¹⁵. According to estimates of the WHO, non-smoking lung cancer patients account for 25% of all lung cancer cases, ranking seventh in the cause of cancer-related death in the world²¹. In 2013, the WHO for the first time identified PM2.5 as carcinogenic and considered it as a common and major environmental carcinogen⁴⁴. The damage to the lungs caused by air pollution cannot be underestimated, and any burning matter will produce particulate matter, including PM2.5. PM2.5 in the atmosphere can directly enter the lungs, causing damage⁴⁵. The lungs play an important role in blood circulation. The blood obtains oxygen and expels carbon dioxide through the lungs, and PM2.5 is disseminated throughout the body along with the blood. The harmful substances attached to the surface of PM2.5, including some polycyclic aromatic hydrocarbons, are strong carcinogens. A cohort study by the American Cancer Society showed that the risk of death begins to increase if atmospheric concentrations of PM2.5 are greater than 10 µg/m³ for a long period of time. For every 10 µg/m³ increase in concentration, the overall risk of death increases by 4%, and the risk of death from lung cancer increases by 8%⁴⁶ The Canadian Census Health and Environment Cohort (CanCHEC) study, in which subjects were followed-up with for 20 years and assigned to regions across Canada based on space weather classification types, found that exposure to PM2.5 was associated with an increased risk of lung cancer death⁴⁷.

The spatial and temporal distribution characteristics and spatial variability of lung cancer incidence and mortality at the county (district) level in Zhejiang Province, China illustrated that the environmental factors such as smoking, PM2.5 and urban green covers affect the trend of lung cancer. Therefore, change of these environmental factors could potentially reduce the incidences of lung cancer although the effects of the disease can be complex. In these clusters, Hangzhou, Huzhou, Jiaxing, and Ningbo belong to the eastern region of Zhejiang Province, China. Research shows that the incidence and mortality of lung cancer in the eastern region are higher than those in the central and western regions⁴⁸, and the economic development of the eastern region of Zhejiang Province is better than that of the central and western regions. The study found that the incidence and mortality of lung cancer were associated with socioeconomic status. Population aging, family income, and access to treatment and diagnosis are also reasons for the significant increase in lung cancer incidence²⁶. Similarly, the imbalance in the level of economic development in different regions of Zhejiang Province may also lead to spatial differences in the incidence of lung cancer, and the eastern region is relatively developed economically. However, western Zhejiang is a low-incidence cluster, while eastern Zhejiang is a high-incidence cluster, which may be due to the stronger industrial foundation, higher urbanization level, and higher economic development in eastern Zhejiang. It may also be related to the high industrial infrastructure area near southern Jiangsu Province, China, and the frequent sandstorms. This indicates that the level of economic development and exposure to other risk factors also vary greatly within the same region. It may also be related to the large size of the elderly population, the high degree of aging, and the accelerated aging process³⁷. While economically developed regions have better medical resources, better medical technology, and higher health literacy, they are also associated with more industrial pollution and exposure to risk factors. Lung cancer screening and the promotion of westernized diets may also play a role. Therefore, the areas with high lung cancer incidence and mortality rates are relatively developed.

Clusters with high lung cancer mortality rates basically coincide with clusters with high incidence rates, suggesting that China’s high incidence of lung cancer may be the biggest reason for its high mortality rates. The prognosis of lung cancer patients in China is poor, and the survival rate is lower than that in developed regions such as Europe and the United States. The main reason may be that the number of patients in the early clinical stage is small, the later stage of clinical diagnosis and treatment is not standardized, and the overall medical level of China still lags behind that of developed countries such as Europe and the United States. The population with the highest incidence of lung cancer is the elderly population, and with increasing age, the functions of the body’s vital organs, immunity, and other bodily functions weaken, so the probability of detection of malignant lung cancer tumors is relatively high and the survival rate of patients is low. Therefore, the coverage of early screening of lung cancer should be expanded, and the standardization and homogenization of clinical diagnosis and treatment of lung cancer should be promoted to reduce the mortality rate of lung cancer in China.

In 2019, The General Office of the State Council promulgated the “Healthy China Action - Cancer Prevention and Control Implementation Plan (2019–2020).” In 2021, The General Office of the State Council issued the Key Tasks for Deepening the Reform of the Medical and Health System, and the National Health Commission issued the Notice on Issuing the Action Plan for Improving the Quality of Cancer Diagnosis and Treatment. The introduction of the above-mentioned policies, regulations, and documents aims to improve cancer prevention and control and enhance the prevention service system. We should continue to promote the formulation and improvement of other relevant policies and regulations and conduct timely risk communication to the public to reduce the risk of lung cancer. Some scholars have proposed that the concept of “One Health” and public health service with multisectoral participation should be advocated to slow down and deal with the spread of various diseases in the Chinese population by mobilizing the full cooperation of various disciplines and departments^49,50. In addition to health administrations and agencies, other sectors should also be involved in this action⁵¹.

In China, as the pressure of lung cancer prevention and follow-up is concentrated in community health service centers, the prevention and control of lung cancer are highly emphasized in the healthcare system. In terms of healthcare services, such as health education, different sectors can provide services for different groups. For example, social organizations can participate in publicity, actively carry out various forms of publicity and education, and popularize scientific knowledge of air pollution prevention and control. Moreover, multisectoral participation can promote intersectoral cooperation, thereby improving the efficiency of lung cancer prevention and control. The detection and screening of lung cancer patients require large human, financial, and material investments, which cannot be fully guaranteed in the short term if only relying on the actions of health administration departments and institutions. On the contrary, if the National Development and Reform Commission and the financial sector fully play their roles within their respective areas of responsibility and form a cooperative model, they can ensure the timeliness of the input of resources and enable healthcare institutions to carry out screening services more effectively.

Conclusions

The spatial distribution of lung cancer in Zhejiang Province, China, showed significant spatial heterogeneity, and the high-risk clusters of lung cancer incidence and mortality were located in the northeast of Zhejiang Province, China, with a high urbanization rate. Vegetation and urbanization rate have significant effects on the incidence of lung cancer. In the context of accelerated aging, industrialization, and urbanization, lung cancer will still be highly prevalent, and all localities need to formulate prevention and control programs according to local conditions, increase and expand lung cancer screening, and strengthen environmental governance.

Limitations

First, in the present study, we only obtained environmental factors and socioeconomic data at the macro level, which cannot fully represent individual exposure levels, which may affect the results of this study to some extent. Second, cancer is a chronic disease, caused by a variety of factors acting on the human body. There is an incubation period and there are various factors affecting the disease that have not been considered. Third, for the multicollinearity of variables, we analyzed the relationship between lung cancer and environmental factors one by one through a spatial regression model, emphasizing the random association between each factor and lung cancer. Finally, data collection and verification at the national data center are time consuming, and cancer reports usually have a certain delay. In subsequent studies, we will further expand the time and space to provide a more detailed theoretical basis for lung cancer research.

Electronic supplementary material

Below is the link to the electronic supplementary material.

Author contributions

H.X. conceptualized and designed this study. H.H. and G.L. developed methodologies for data collection and analysis. Y.G. and M.Z. participated in data collection and analysis. Y.G. and H.L. participated in writing and editing. All authors were subsequently involved in critically revising the manuscript. All authors have approved the final version of the manuscript to be published and agreed to be accountable for all aspects of the work.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported by Soft Science Project of Ministry of Science and Technology (2011GXQ4D050 and 2007GXS3D086), National Social Science Foundation (07BSH035).

Data availability

The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation. Data is provided within the manuscript or supplementary information files.

Declarations

Competing interests

Conflict of InterestThe authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.