Developing novel imaging techniques and instruments has been a major effort in the molecular imaging field. OI is a novel technique with many good characteristics such as high sensitivity, low-cost, ease of use, relatively high-throughput, and short acquisition time 
. Recent advances in optical imaging instruments and molecular probes have made it an excellent tool for both small animal research and clinics. In this study, we demonstrate systematically that OI techniques can also be used to image visible and near-infrared light produced by radioactive materials. This would make OI a modality of choice for evaluation of bioluminescent, fluorescent and radioactive probes.
The success of using radionuclides for OI as demonstrated here is expected to have a major impact to the molecular imaging field. First, radioactive agents are traditionally studied by PET, SPECT or γ cameras, which are expensive, hard to maintain and not widely available to many researchers. Our results clearly show that the commercially available OI instrument can be used for studying radioactive probes possessing both β+ and β−-emitting radionuclides, as well as for bioluminescence and fluorescence probes. Considering much lower cost and wider accessibility of the OI instruments than that of PET and SPECT, radioactive probe development will likely be dramatically accelerated by using this approach. Optical imaging systems such as IVIS can image up to five animals simultaneously, while small animal PET and SPECT may only image one mouse at a time. The high throughput manner of OI equipments will also help to improve the speed of radioactive probe development.
Although OI is an important tool in animal research, its clinical applications have been severely hampered by very limited OI probes approved by the Food and Drug Administration (FDA). So far only iodocyanine green dye (IC-Green) has been approved for use in humans 
. On the other hand, radioactive imaging has a longer history in biomedical imaging and has been widely used in clinics for the past several decades. Many FDA approved SPECT and PET probes including [18
F]FDG have been developed for imaging different diseases and molecular targets. New applications for these radioactive probes may be developed in conjunction with radioactive OI techniques. The research presented here opens a new avenue for small animal imaging research, as well as for imaging patients in clinics.
In this research, we have evaluated three radionuclides (18
I, and 90
Y) for small animal radioactive OI because of their important roles in nuclear medicine. 18
F is the most often used PET radionuclide 
, and 131
I and 90
Y are two most widely used radionuclides for radiotherapy 
. In vivo
optical images with reasonable sensitivity can be quickly obtained for all three radionuclides. These encouraging results suggest that the radioactive OI can be a powerful tool for fast preliminary evaluation of 18
I, and 90
Y labeled compounds. This will be important for 90
Y based agents development, since it has been relatively difficult to obtain in vivo
information for a 90
Y agent through non-invasive imaging method.
Compared to conventional fluorescence and bioluminescence imaging, radioactive OI has some unique properties. It has wide emission spectrum as demonstrated here, so that a radioactive probe can be monitored at different wavelengths. More importantly, radioactive OI does not require excitation light, which is a significant advantage over traditional OI. The radioactive OI signal generated by a radioactive probe is constitutive, which is very different from fluorescence and bioluminescence probes. The radioactive OI can be performed by monitoring spectral windows that differ from typical FLuc spectrum. Therefore, it is possible to perform BLI and radioactive OI in the same animal with proper emission filters. Radionuclides also generally emit low levels of light and are not expected to interfere with BLI.
It should be noted that although only β+ and β− emitters were evaluated in this study, many radionuclides (α, β+, β−, electron capture, etc.) which emit charged particles are likely to be suitable for OI. The detection and imaging sensitivity depends on the physical properties of radionuclides, particularly their energy. As shown in this study, the radionuclide with higher energy generates stronger optical signals. Radioactive OI shares some common disadvantages as other OI techniques: limited tissue penetration and relatively poor quantification ability compared to PET and SPECT. However, subcutaneous tumor models and superficial disease models have been widely used in the medical research. For these models, our results indicate, radioactive OI is a suitable technique. Despite the stated disadvantages, OI techniques have advanced rapidly over the past couple of years. OI has evolved from a basic research tool to a modality which has great potential utility for patient imaging. For instance, fluorescence molecular tomography (FMT) and optical fiber based technology could potentially lead to deeper tissue imaging. These techniques may be very helpful for achievement of deeper tissue imaging with radioactive OI as well. Future research in radioactive OI would further accelerate the translation of OI into clinical application.
The optical signals detected could be originating from Bremsstrahlung or Cerenkov radiation. Bremsstrahlung radiation is a well-known physical phenomenon, which refers to the electromagnetic radiation produced by slowing down or deflection of charged particles (especially electrons) in the Coulomb fields of atomic nuclei 
In Cerenkov radiation, optical signal is emitted when a charged particle travels through an insulator at a speed greater than the speed of light in that medium 
. Both of them have continuous energy spectra in optical and NIR range and are generated by radioactive disintegration of a charged particle 
. The recent publication on radioactive OI by Robertson et al. strengthens our data. They propose Cerenkov radiation as the origin of the optical signals generated from positron emitting nuclides 
. Cerenkov radiation was supported by the following data: (1) The obtained radioactive optical spectrum and (2) the increased light output as the refractive index of medium increases. Initially, we assigned generation luminescence signal to Cerenkov radiation as well. We obtained similar results which optical signal strength was enhanced with increasing refractive indexes of medium (Fig. S1
). But the spectra obtained from IVIS system () were quite different from the calculated Cerenkov emission in literature 
. Radioluminescence intensities, measured with a Fluoro Max-3 spectrofluorometer (Jobin Yvon Inc., Edison, NJ), were consistent with the spectra obtained from the IVIS system (Fig. S2
). Moreover, water has a refractive index of 1.332. The lower energy threshold for the emission of Cerenkov radiation in water is thus calculated to be 263 KeV 
. However, radioactive optical signals could also be observed for 111
In, which emits particles with the energy below the theoretical threshold. All these data imply that other mechanisms such as Bremsstrahlung radiation may also contribute to the radioactive OI. It is unclear as for the role of Bremsstrahlung radiation compared to Cerenkov. The mechanism of radioactive OI still needs to be further investigated.