Hypoxia occurs to a variable extent in a vast majority of rodent and human solid tumors. It results from an inadequate and disorganized tumor vasculature, and hence an impaired oxygen delivery. A probe for the non-invasive detection of tumor hypoxia could find important utility in the selection of patients for therapy with bioreductive agents, anti-angiogenic/anti-vascular therapies and hypoxia-targeted gene therapy. In addition, tumor hypoxia has been shown to predict for treatment outcome following radio- or chemotherapy in human cancers, the underlying mechanism for which may involve hypoxia driving genetic instability and resulting tumor progression. Beyond oncology, utility can also be envisaged in stroke, ischemic heart disease, peripheral vascular disease, arthritis and other disorders. Design, validation, preclinical development and current status of a fluorinated 2-nitroimidazole, N-(2-hydroxy-3,3,3-trifluoropropyl)-2-(2-nitro-l-imidazolyl) acetamide (SR 4554, CRC 94/17), which has been rationally designed for the measurement of tumor hypoxia by magnetic resonance spectroscopy (MRS) and imaging (MRI), are reviewed. Application in positron emission tomography (PET) detection is also proposed. Design goals were: (i) a nitro group with appropriate redox potential for selective reduction and binding in hypoxic tumor cells; (ii) hydrophilic/hydrogen bonding character in the side chain to limit nervous tissue penetration and prevent neurotoxicity; and (iii) three equivalent fluorine atoms to enhance MRS/MRI detection, located in a metabolically stable position. Reduction of SR 4554 by mouse liver microsomes was dependent on oxygen content, with a half-maximal inhibition at 0.48 +/- 0.06%. SR 4554 underwent nitroreduction by hypoxic but not oxic tumor cells in vitro and electron energy loss spectroscopic analysis showed selective retention in the hypoxic regions of multicellular tumor spheroids. Pharmacokinetic design goals were met. In particular, low brain tissue concentrations were seen in contrast to excellent tumor levels, as measured by high performance liquid chromatography. The extent of this restricted entry to brain tumor was surprising given the overall octanol/water partition coefficient and was attributed to the hydrophilic/hydrogen bonding character of the side chain. Quantitative MRS was used to assess the retention of 19F signal in murine tumors and human tumor xenografts. The 19F retention index (FRI; ratio of 19F signal levels at 6 h relative to that at 45 min) ranged from 0.5 to 1.0 and 0.2 to 0.9 for murine tumors and human xenografts respectively. The correlation between SR 4554 retention and pO2 was not a linear one, but when FRI was > 0.5, the % pO2 < or = 5 mmHg was always > 60%, indicating that high FRI was associated with low levels of oxygenation. Finally, whole body 19F-MRI in mice demonstrated that SR 4554 and related metabolites localized mainly in tumor, liver and bladder regions. A selective MRS signal was readily detectable in tumors at doses at least 7-fold lower than those likely to cause toxicity in mice. We conclude that proof of principle is established for the use of SR 4554 as a non-invasive MRS/MRI probe for the detection of tumor hypoxia. Based on these promising studies, SR 4554 has been selected for clinical development.