Environmental Cancer Risk Scenarios for Lebanon: Quantitative Sensitivity Analysis and a Qualitative Genetic Susceptibility Framework

Abstract

Background: Environmental exposures and biological susceptibility contribute to the complexity in cancer vulnerability, particularly in settings where comprehensive environmental monitoring and longitudinal exposure data are limited. Scenario-based modeling can provide a structured approach for exploring potential environmental cancer risks under uncertainty and identifying priorities for future research and surveillance. Methods: We developed a scenario-based environmental cancer risk framework for Lebanon incorporating three exposure domains—perchlorate contamination, nitrate/nitrite exposure, and conflict-derived environmental contamination— and a qualitative Genetic Susceptibility Index (GSI). Monte Carlo simulations with 100,000 iterations were used to propagate uncertainty in literature-derived epidemiological effect estimates for colorectal and lung cancers under alternative hypothetical scaling assumptions. The GSI was used to organize qualitative susceptibility pathways but was not incorporated as a numerical modifier in the simulations. Thyroid cancer was evaluated separately through an illustrative hypothetical sensitivity analysis using author-defined relative-effect ranges because direct epidemiological estimates linking perchlorate exposure to thyroid cancer incidence were unavailable. Results: For colorectal cancer, the mean simulated relative effect increased from 3.79% (95% UI: 1.90–5.70%) under the low-exposure-contrast scenario to 7.72% (95% UI: 3.83–11.71%) under the moderate scenario and 16.09% (95% UI: 7.81–24.80%) under the high scenario, which applied the full published log-relative effect. For lung cancer, corresponding mean simulated relative effects were 2.18% (95% UI: 1.01–3.36%), 4.41% (95% UI: 2.02–6.83%), and 9.03% (95% UI: 4.09–14.13%). Under the full published exposure contrast, 92.34% of colorectal cancer simulations exceeded a 10% relative-effect threshold and 94.52% of lung cancer simulations exceeded a 5% threshold. The thyroid cancer analysis produced mean simulated relative effects of 1.03%, 3.50%, and 7.51% under the low-, moderate-, and high-scenario assumptions, respectively; these outputs were conditional on the prespecified hypothetical ranges. Conclusions: This framework demonstrates how alternative exposure-contrast assumptions can influence the distribution of literature-derived relative effects in a data-limited setting. By combining quantitative exposure-contrast sensitivity analysis with a qualitative genetic susceptibility framework, the approach may support environmental health planning, surveillance prioritization, and the design of future population-based studies in Lebanon. The thyroid cancer results illustrate model behavior under author-defined hypothetical assumptions and do not constitute epidemiological estimates of perchlorate-associated thyroid cancer risk. The findings are exploratory and should not be interpreted as estimates of cancer incidence, population-attributable risk, or causal effects in the Lebanese population.