Dose to the thyroid from helical chest CT can vary significantly due to the random tube start point, pitch factor, thyroid position relative to the isocenter, and beam width. We used optically stimulated luminescence dosimeters (OSLDs) and an adult anthropomorphic phantom to investigate the uncertainty of thyroid dose estimate. Maximum gap or overlap in the helical beam was estimated using the above factors. Using the maximum gap/overlap over the thyroid, different possible scenarios were simulated and the degree of missed thyroid tissue by the primary beam was estimated. Results showed a variation of >30% in the average thyroid dose, and >50% if a single dosimeter was used to determine dose to the thyroid. Furthermore, measured doses were compared to those calculated by Monte Carlo simulation software, which automatically matches the anatomy of the localizer radiograph with the stylized computational phantom used for dose calculation. The difference was significant: the dose given by the Monte Carlo software was ∼50% lower than the average dose measured with the phantom in all three chest protocols. In addition, the software does not take the effect of the random tube start angle into account.