In a recent cohort study in aluminum foundries and remelting plants an unexpectedly high risk of lung cancer was found in workers in sand foundries. On the basis of present and historical measurement data, we developed a statistical model for exposure to total dust and crystalline quartz for different jobs and time periods. Cumulative dose estimates of total dust and crystalline quartz were calculated and used in a nested case-control study in the cohort. From the cohort of foundry workers (n = 5016), 46 cases of lung cancer were identified. The final analysis was performed on 31 cases and 233 controls with one year or more of employment. Historical measurement data from the 1960s and onward were collected, totaling 203 total dust and 103 crystalline quartz exposure observations. Regression models using the determinants of job title, time period, type of foundry, and size of production, were developed for assessing historical total dust and crystalline quartz air concentrations. These estimates were used to calculate individual cumulative exposure in the case-control study. In the multiple linear regression analysis, the determinants explained much of the variations in dust level (r2 = 0.58). The explained variation in crystalline quartz was much lower (r 2 = 0.13). The regression coefficients for the type of foundry, time period, and size of production were statistically significant for total dust. On the basis of the regression analysis, the final models were used to calculate individual cumulative exposures. The calculated cumulative dust and quartz exposures averaged 33 mg/m3 * year and 0.42 mg/m 3 * year, respectively. The odds ratios (ORs) were not significant, but showed dose-response trends for both dust and crystalline quartz.