Figure 1 shows trends in age-standardized lung cancer incidence rates from 1997 to 2017 in Taiwan. In men, the overall incidence rate of lung cancer increased, but this increase slowed toward the end of the study period (AAPC = 0.8, 95% CI [CI]: 0.6, 1.1). In women, the incidence increased rapidly without interruption (AAPC = 3.0, 95% CI 2.8, 3.3). The incidence of adenocarcinoma is increasing most rapidly among histological types, particularly in women; the rate in men increased from 10.8 to 25.5 per 100,000 from 1997 to 2017 (AAPC = 4.0, 95% CI 3.7, 4.3), and in women it increased from 8.9 to 28.5 per 100,000 individuals from 1997 to 2017 (AAPC = 5.5, 95% CI 5.2, 5.8). The incidence of squamous cell carcinoma is decreasing, especially in women; the incidence in men decreased from 11.2 to 9.1 per 100,000 from 1997 to 2017 (AAPC = − 1.1, 95% CI − 1.5, − 0.8), and in women decreased from 2.1 to 1.2 per 100,000 persons (AAPC = − 3.5 , 95% CI – 4.0, – 3.0). The incidence of small cell carcinoma in men first increased and then slightly decreased, but the overall trend was increasing (AAPC = 0.7, 95% CI 0.0, 1.4); in women, the rate is low, approximately 0.6 per 100,000 individuals, and decreasing (AAPC = − 1.2, 95% CI − 2.0, − 0.4). The incidence of large cell carcinoma in men was also low, approximately 0.5 per 100,000 individuals, but increased (AAPC = 1.3, 95% CI 0.2, 2.4); in women, the rate is even lower at approximately 0.2 per 100,000 individuals, but has increased significantly in recent years (AAPC = 4.0, 95% CI 2.6, 5.4). Appendix 3 presents the proportions of histological types of lung cancer from 1997 to 2016: Adenocarcinoma was the most common histological type, with the proportion increasing to almost 90% in women and > 50% in men. Squamous cell carcinoma is another major type in men, but is less common in women and its proportions are decreasing every year in both sexes. The proportions of small cell carcinoma, large cell carcinoma, and other specified or unspecified carcinomas in men and women were sparse and decreased.
Figure 2 presents lung cancer incidence rates by age, period and cohort. The incidence rate increases with age, but this increase slows with age. In men, the time trends in incidence rates were different for the age groups; rates increased but became flat or fell after 2009 for age groups 70–74, 75–79, and 80–84, were flat for age groups 60–64 and 65–69, increased monotonically for younger (35– 39 , 40–44, 45–49, 50–54 and 55–59 years) and increased but slowed after 2009 among the youngest (30–34 years). In women, incidence in the oldest age group (80–84 years) increased in earlier periods before falling in later periods. In contrast, the percentages in the other age groups continue to grow. Trends in incidence rates for birth cohorts were more consistent across ages for both sexes. For men, the rates for earlier births are increasing. Rates in the birth cohorts from 1932 to 1957 remained stable, while those from 1957 to 1967 increased rapidly. The rate increase slowed in the 1967 to 1977 cohorts; in later cohorts, levels plateaued. For women, rates in earlier birth cohorts also increased monotonically. Rates for birth cohorts from 1932 to 1967 increased rapidly, those for birth cohorts from 1967 to 1977 increased slowly, and those for cohorts after 1977 increased rapidly. Appendices 4 and 5 present incidence rates by age, period and cohort for lung adenocarcinoma and lung squamous cell carcinoma, respectively. Appendix 6 presents AAPC incidence rates in different age groups for lung cancer, lung adenocarcinoma, and lung squamous cell carcinoma, respectively.
Figure 3 shows the age, period, and cohort effects of our analysis. The risk of lung cancer increases with age for both sexes, but these increases are slower in older age groups (and in the middle age groups for lung squamous cell carcinoma in women). The periodic effects of lung cancer were the same for both sexes. For the cohort effects of lung cancer, rates increased in earlier cohorts for both sexes, while they decreased slightly before increasing in men, but increased consistently in women in the 1932 to 1967 birth cohorts .Rates increased slowly for both sexes in the 1967 to 1977 birth cohorts; finally, levels plateaued in men but increased in women in later cohorts. The age effects of lung adenocarcinoma for both sexes are similar but more pronounced than for lung cancer overall. Lung adenocarcinoma increased slightly in both sexes over time. Cohort effects of lung adenocarcinoma increase in both sexes; they are rising rapidly for cohorts born after 1977 in women. The age effect of lung squamous cell carcinoma increases with age in men, but increases slowly in older age groups, and that in women also increases, but less so, and the increase begins to slow in the middle age groups. The periodic effects of squamous cell carcinoma of the lung are almost equal for both sexes, but decrease slightly in later periods in women. Regarding the cohort effects of lung squamous cell carcinoma, rates increased in early birth cohorts for both sexes, but decreased in males beginning with the 1932 birth cohort and in females beginning with the birth cohort in 1922
Figure 4 presents the SIR and AAPC maps of lung cancer in men. It is challenging to interpret the general trend from the noisy SIR (Fig. 4A) and AAPC (Fig. 4C) maps without stabilized kriging. However, after stabilized kriging, we identified several incidence hotspots (Fig. 4B) in the northern, western, and southwestern coastal cities; northeastern coastal cities; southeastern cities; and the mountainous regions of Taiwan. We also identified several hotspots showing rapid rate increases in northwestern and southern coastal cities, southern cities, and mountainous regions (Fig. 4D).
Figure 5 depicts the SIR and AAPC maps of lung cancer in women. Again, it is challenging to interpret the general trends from the noisy SIR (Fig. 5A) and AAPC (Fig. 5C) maps without stabilized kriging. After stabilized kriging, we identified several incidence hotspots (Fig. 5B) in northern, western and southwestern cities, northeastern coastal cities and mountainous areas, and southeastern mountainous areas of Taiwan. We also identified several hotspots in northern, western, and southern coastal cities and towns and rural towns in central Taiwan where levels increased rapidly (Fig. 5D).
Figure 6 presents the SIR and AAPC maps of lung adenocarcinoma in men. Hotspots are similar to those for overall lung cancer in men. Incidence incidence hotspots (after stabilized kriging; Fig. 6B) were found in northern, western, and southwestern coastal cities, northeastern coastal cities, southeastern cities, and mountainous areas of Taiwan. Hotspots with increasing rates (after stabilized kriging; Fig. 6D) were found in western and eastern coastal cities and towns and southern cities and mountainous areas of Taiwan.
Figure 7 shows the SIR and AAPC maps of lung adenocarcinoma in women. Hotspots are similar to those for overall lung cancer in women. Incidence incidence hotspots (after stabilized kriging; Fig. 7B) were found in northern, western and southwestern coastal cities, northeastern coastal cities, mountainous areas, and southern mountainous areas of Taiwan. Hotspots with increasing rates (after stabilized kriging; Fig. 7D) are observed in northern, western, southern, and southeastern coastal cities and towns; the southern mountainous regions; and rural towns in central Taiwan.
Figure 8 shows the SIR and AAPC maps of lung squamous cell carcinoma in men. After stabilized kriging, incidence rate hotspots (Fig. 8B) were identified in southwestern and northeastern coastal towns and cities. Elevated rates (after stabilized kriging; Fig. 8D) were found in northwestern, southwestern, southern, and eastern coastal cities and rural towns in central Taiwan.
We could not identify any hotspot for squamous cell lung carcinoma in women because the number of cases was insufficient (Appendix 7).