This randomized controlled trial will compare oral 5-fluorouracil derivatives, TS-1, with intravenous standard chemotherapy such as taxanes in women with metastatic or recurrent breast cancer. Patients with hormone-resistant breast cancer are assigned to either TS-1 (40–60 mg twice daily for 28 consecutive days, followed by a 14-day rest period) or standard chemotherapy (docetaxel 60–75 mg/m² at 3- or 4-week intervals, paclitaxel 175 mg/m² at 3- or 4-week intervals or paclitaxel 80–100 mg/m² weekly, followed by a 1-week rest period). Treatment will be repeated until tumor progression or ≥4 courses for TS-1 and ≥6 courses for taxanes. The primary endpoint is overall survival. Secondary endpoints are progression-free survival, time to treatment failure, adverse events, health-related quality of life and cost-effectiveness. A threshold hazard ratio of 1.333 will be used to determine whether overall survival in the TS-1 group is equivalent (not inferior) to that in the taxane group. The target number of registered patients is 600.

Cryo-electron microscopy of vitreous sections (CEMOVIS) and cryo-electron tomography (cryo-ET) of vitrified specimens are gradually gaining popularity. However, similar to the conventional methods, these techniques tend to produce different images of the same sample. In CEMOVIS, the mechanical stress caused by sectioning may cause inaccuracies smaller than those caused by crevasses. Therefore, we examined Escherichia coli cells by using CEMOVIS and cryo-ET to determine the differences in the computed sizes of the envelope layers, which are smaller than crevasses. We found that the width of the periplasmic space in vitreous sections and tomograms was 12 and 14 nm, respectively; furthermore, while the distance between the outer membrane (OM) and the peptidoglycan (PG) layer was almost equal (11 nm) in the two techniques, that between the plasma membrane (PM) and PG was clearly different. Thus, the observed size difference can be mainly attributed to the PM–PG distance. Since our data were obtained from images acquired using the same microscope in the same conditions, the size differences cannot be attributed to microscope-related factors. One possible factor is the angle of the cutting plane against the long axis of the cell body in CEMOVIS. However, the same PG–OM distance in both methods may exclude the variations caused by this factor. Furthermore, the mechanical stress caused by vitreous sectioning or high-pressure freezing may result in shrinkage. If this shrinkage is responsible for the nanometre-scale deformation in CEMOVIS, this factor will have to be considered in determining the molecular resolution obtained by CEMOVIS.