The evaluation of tobacco products is complex due to a multitude of factors including product diversity, limited testing standards, and variability in user behavior. Alternative approaches in current testing paradigms have limitations that generally truncate their applicability beyond screening for hazard identification; this is also true for toxicological evaluations of tobacco products. In a regulatory context, results from tobacco product toxicity assessments are extrapolated to the in vivo condition to assess human health relevance at the individual and population level. A key limitation of alternative approaches is the difficulty and uncertainty in extrapolating results to adverse outcomes relevant to chronic tobacco exposures in humans. This difficulty and uncertainty are increased when comparing toxicological outcomes between tobacco products. Given that the interpretation and quantification of differences in assay results (e.g., mutagenicity) for tobacco product comparison may be inconclusive, the predictive value of these approaches for human risk of relevant downstream pathologies (e.g., carcinogenesis) can be limited. Development and validation of fit-for-purpose alternative approaches that are predictive of human toxicity and dose response assays with adequate sensitivity and specificity for product comparisons would help advance the field of predictive toxicology.