8282134 Exposure-time-response analyses and quantitative cancer risk assessment for protracted exposures

<h3></h3> Background: Quantitative Risk Assessment (QRA) for protracted exposures and cancer risks often relies on cumulative exposure metrics and constant exposure scenarios. However, mounting evidence from epidemiological studies suggests that the exposure-dependent risk may be very different for subjects with the same cumulative exposure due to variations in the duration and intensity of exposure, as well as recency of exposure. This study aims to highlight the significance of incorporating time-related factors into QRA more transparently and demonstrates how recently developed analytical approaches can enhance risk assessment in such contexts. <h3>Methods</h3> We assessed the impact of exposure metrics that account for the recency of exposure on cancer risk, using the established association between benzene exposure and acute myeloid leukemia (AML) as an example. Standard life table methods were applied using cumulative and recency-weighted exposure metrics to estimate excess risks. Dynamic lag models (DMLs) were used to show how empirical data could be used to better inform risk assessment using data from a study on benzene poisoning (BP). <h3>Results</h3> For AML, excess risks varied by exposure metric and target age, with traditional cumulative exposure producing higher excess risk estimates compared to recency-weighted metrics, resulting in lifetime risks ranging from 52 to 98 per 1 million exposed workers. For BP, DMLs showed that risks were primarily driven by exposures within the last 5 to 10 years, and excess risks were similar when using an exposure metric consistent with the one used to derive the risk function. <h3>Conclusions</h3> Our findings highlight the potential for over- or under-estimating risks when ignoring the timing of exposure. Incorporating results from exposure-time-response (ETR) models into QRA may improve the accuracy of risk assessments and better inform regulatory decisions. Even without such data, risk assessments could benefit from a more transparent evaluation of time-related aspects of exposure.