More recently, a complex MS inhalation study comparing inhalation and post-inhalation periods of various durations Cyclopamine concentration (Study 1) determined that long-term inhalation, i.e., for 18 months, was sufficient to demonstrate an MS concentration-dependent increase in lung tumors without the need of a post-inhalation period (18 + 0 schedule) (Stinn et al., 2012). The concentration–response relationship for 18 months of MS inhalation observed in Study 1 was reproduced and refined in the current Study 2. Thus, intra-laboratory reproducibility was achieved. The use of two different generations of a filtered reference cigarette (2R4F and 3R4F) for MS generation

had no influence on the tumor response, as would be expected based on the results of comparative chemical-analytical, in vitro, and in vivo studies showing no apparent differences between the two reference GDC-0199 in vitro cigarettes (Roemer et al., 2012). Long-term MS inhalation studies with the A/J mouse using a similar study design have not been reported by other laboratories. Thus, information on inter-laboratory reproducibility of using MS inhalation on this mouse strain could only be obtained by analyzing tumor response data obtained with the more common 5 + 4-month schedule. Of the published

studies, some had very low group sizes (D’Agostini et al., 2001), and some used nose-only exposure (Hamm et al., 2007) instead of whole-body exposure as in ifenprodil the current study; these studies were not included in the current assessment of reproducibility. A direct comparison of both whole-body and nose-only exposure modes in a 5 + 4-month schedule did not find a statistically significant MS effect after nose-only exposure while whole-body exposure was positive (Curtin et al., 2004). For an inter-laboratory comparison of the results of the 5 + 4-month schedule, therefore, only four whole-body MS inhalation studies qualified (Fig. 7): the first was the whole-body exposure part of the above comparative study (Curtin et al., 2004);

the second included the A/J mouse as one of several strains that were compared in terms of cancer susceptibility (Gordon and Bosland, 2009); the third was a study experimentally conducted at TNO Quality of Life, Zeist, the Netherlands (Stinn et al., 2010); and the fourth was part of the complex study design of Study 1 (Stinn et al., 2012). The tumor multiplicities obtained in the four available studies were reproducible. The correlation obtained by linear regression analysis of the combined dataset of the four studies was lower (R2 = 0.68) than that of the 18 + 0-month dataset obtained in one laboratory. In part, this may be due to the lower effect size after 5 + 4 months, which was two to three times smaller than after 18 months. This was apparently associated with a higher relative variability within studies and may have also contributed to a larger inter-laboratory variability.