Background/aim The increasing burden of Lung Cancer (LC) in Crete, Greece has raised concern about potential association of environmental risk factors with LC. Main aim of this study was to assess Outdoor Air Pollution (OAP) and the risk for LC incidence, as well as the association of OAP with LC mortality and survival; for the first time in Crete using LC primary data.

Methods 5,000 LC cases (diagnosed from 1992 to 2013) were obtained and followed-up from the population-based Cancer Registry of Crete. The Age-Adjusted Incidence Rates (AAIR) and the Age-Standardised Mortality Rates (ASMR) were calculated. Data on outdoor air concentrations of particulate matter [PM2.5, between 2.5 µm and 10 µm (PM2.5–10) and PM10], PM2.5 absorbance (black carbon measure), nitrogen dioxide (NO2) and nitrogen oxides (NOx) were sampled using Harvard impactors (OAP time series1982–2014). Spatio-temporal statistics were used to explore associations between LC and OAP, and develop a multivariate predictive model of current and future risk. All tests were conducted at a=0.05 in STATA and ArcMap 10.3.1.

Results LC in Crete accounts for 40.2 new cases/100,000/year for both genders. Annual median estimates of environmental concentrations were: PM2.5=20.7 (±1.5) µg/m³, PM10=38.9 (±2.5) µg/m³, PM2.5–10=59.6 (±3.7) µg/m³, PM2.5 absorbance=1.2 (±0.3) × 10−5 per m, NO2=15.2 (±3.8) µg/m³ and NOx=20.1 (±4.9) µg/m³. A strong positive association was found between LC and the OAP estimates. The highest risk for LC was observed in the major urban centres, several south-east and north-west rural regions (RR=3.2, 95% CI: 1.638 to 4.765). All associations were significantly increased, while LC hot spots due to OAP were identified, during the financial crisis. Significant increase of the RRs is estimated for the next 10 years, especially in north-west rural regions (RRexpected=3.9, 95% CI: 1.372 to 6.428).

Conclusion OAP seems to be an important determinant of LC, especially during the financial crisis. The joint contribution of OAP and other risk factors on LC outcomes has major adverse effects on LC outcomes with significantly diverse geographical patterns. Targeted interventions have to be performed in the current and future risk areas; even among this genetically homogeneous population.