Molecular Imaging: A Primer for Interventionalists and Imagers

Molecular Imaging: A Primer forInterventionalists and ImagersDavid S. Wang, MD, Michael D. Dake, MD, Jinha M. Park, MD, PhD, and Michael D. Kuo, MD

The characterization of human diseases by their underlying molecular and genomic aberrations has been the hallmarkof molecular medicine. From this, molecular imaging has emerged as a potentially revolutionary discipline that aimsto visually characterize normal and pathologic processes at the cellular and molecular levels within the milieu ofliving organisms. Molecular imaging holds promise to provide earlier and more precise disease diagnosis, improveddisease characterization, and timely assessment of therapeutic response. This primer is intended to provide a broadoverview of molecular imaging with specific focus on future clinical applications relevant to interventional radiology.

J Vasc Interv Radiol 2006; 17:1405–1423

Abbreviations: CEA � carcinoembryonic antigen, FDG � fluorodeoxyglucose, HSV1-tk � herpes simplex virus type 1 thymidine kinase, MMP � matrix met-alloproteinase, NIR � near-infrared, PET � positron emission tomography, SPECT � single photon emission computed tomography, VEGF � vascular endothe-lial growth factor

IN this era of molecular medicine, ourapproach to patient care is evolving asdisease is increasingly being definedby underlying molecular and genomicaberrations rather than by clinicalsigns and symptoms alone. Molecularimaging is an emerging diagnostic dis-cipline that aims to visually character-ize normal and pathologic processes atthe cellular and molecular levels in liv-ing organisms (1–4). Broadly multidis-ciplinary, molecular imaging incorpo-rates methods and concepts from

molecular and cell biology, imagingsciences, chemistry, high-throughputbiology (eg, genomics, proteomics),nanotechnology, pharmacology, andbioinformatics (1,5–7). It is throughmolecular imaging that radiology isexpected to play a critical role in ad-vancing molecular medicine and po-tentially revolutionize patient care andbiomedical research (8,9).

The potential benefits and applica-tions of molecular imaging stem fromtwo fundamental paradigm shifts, oneclinical and the other preclinical. Con-ventional clinical imaging, as prac-ticed by the modern radiologist, gen-erally relies on macroscopic anatomicand/or physiologic variations for dis-ease diagnosis and assessment. Suchmorphologic changes are often nonspe-cific and late phenotypic manifestationsof underlying molecular derangements(1,2). By contrast, molecular imaging ex-ploits the use of directed imagingprobes to sense the specific molecularalterations underlying diseases ratherthan downstream end effects at the tis-sue or organ level. This shift in focusfrom the nonspecific morphologic to themore specific molecular allows for ear-lier and more precise disease diagnosis,improved disease characterization, andmore meaningful monitoring of disease

progression. Moreover, imaging at themolecular level also confers patientspecificity; thus, molecular imagingstands to greatly facilitate the practice ofpersonalized medicine in predictingtherapeutic response and guiding treat-ment selection (10). Although the clini-cal potential of molecular imaging hasyet to be realized, several human trialsof molecular imaging are under way(11–13).

The potential impact of molecularimaging in biomedical research isequally promising. Although past de-cades have witnessed explosivegrowth in our understanding of thefundamental basis of physiology anddisease, it has become clear that cur-rent paradigms of biomedical investi-gation have inherent limitations. Tra-ditional in vitro research employs areductionist approach whereby eventsunder investigation are extracted andstudied in artificial environments (eg,cell culture studies). This is problem-atic because biologic processes rarelyoccur in isolation and are instead me-diated through a complex and dy-namic interplay of gene expression,signaling pathways, environmentalfactors, and inherent feedback mecha-nisms. Therefore, molecular imaging,in interrogating molecular phenomena

From the Department of Radiology and Center forTranslational Medical Systems (D.S.W., M.D.K.),University of California San Diego Medical Center,San Diego; Department of Radiology (D.S.W.), Stan-ford University School of Medicine, Stanford; De-partment of Radiology (J.M.P.), University of Cali-fornia Los Angeles Medical Center, Los Angeles,California; and Departments of Radiology, Medi-cine, and Surgery (M.D.D.), University of VirginiaHealth System, Charlottesville, Virginia. ReceivedJune 1, 2006; revision requested June 6; final revisionreceived June 15; and accepted June 19. Addresscorrespondence to M.D.K., Department of Radiol-ogy, University of California San Diego MedicalCenter, 200 W. Arbor Dr., San Diego, CA 92103;E-mail: [email protected].

None of the authors have identified a conflict ofinterest.

© SIR, 2006

DOI: 10.1097/01.RVI.0000235746.86332.DF

Emerging Technologies

1405

in living individuals, preserves thecontext of whole biologic systems andenables the transition from a reduc-tionist to an integrative and holisticapproach to research (14). In studyingdisease in physiologically intact envi-ronments, molecular imaging assaysare therefore more predictive and rel-evant. This technology also broadensthe capabilities of in vivo research. Be-cause molecular imaging interrogatesevents remotely and noninvasively,analyses can be performed in real timewith minimal disturbance to themodel system, allowing for continu-ous observation of dynamic processes.For example, whereas traditional invivo temporal profile studies requiredanimals to be serially killed at fixedtime points to obtain tissue for in vitroanalyses, molecular imaging enablesreal-time monitoring of molecularphenomena through repetitive imag-ing of a single animal. These distinctadvantages, combined with the com-mercial development of dedicatedsmall-animal imaging instrumenta-tion, have driven the application ofmolecular imaging tools in the preclin-ical arena. Imaging of experimentalsmall animal models, usually mice,has been rapidly adopted into the ba-sic science research of dynamic bio-logic processes such as hypoxia (15),inflammation (16), apoptosis (17), an-giogenesis (18), tumorigenesis (19),and gene expression (20,21). Moretranslational applications include stemcell trafficking (22) and monitoring ofthe distribution and efficacy of noveltherapeutic moieties (23). The poten-tial to streamline and accelerate drugdiscovery and development is perhapsone of the most promising applica-tions of molecular imaging and hasgarnered considerable interest fromacademia and industry (23–25).

Interdisciplinary collaborations formthe essential foundation for the contin-

ued advancement of molecular imag-ing. Naturally, active engagement byphysicians will be critical for success-ful clinical translation of this noveltechnology (26). The goals of this com-munication are to provide a basic in-troduction to molecular imaging andto stimulate discussion among inter-ventional radiologists about whetherand how it could be incorporated intofuture practice. This review beginswith an overview of the imaging mo-dalities used, followed by a discus-sion of basic molecular imaging ap-proaches. To conclude, specific appli-cations in cardiovascular diseases andoncology are detailed. Because theroots of molecular imaging are in mo-lecular biology, readers are referredelsewhere for a review of molecularbiology terminology and concepts (27–31). Several other in-depth reviews ofmolecular imaging are recommendedfor further reading (2,4,32–35).

IMAGING MODALITIES

Molecular imaging encompasses abroad set of technologies that coupleimaging modalities and contrastagents with molecular specificity.Analogous to stains used in histopa-thology, these agents, called molecularprobes or tracers, consist of a signalingcomponent that emits a detectable sig-nal and a targeting component thatconfers localization. This latter compo-nent can be a peptide, receptor ligand,enzyme substrate, oligonucleotide, orantibody. The imaging instrumenta-tion must then be able to remotely de-tect this signal with sufficient spatialresolution and sensitivity. Modalitiesused in molecular imaging includepositron emission tomography (PET),single photon emission computed to-mography (SPECT), magnetic reso-nance (MR) imaging, optical imaging,and ultrasound (US). These modalities

differ in terms of spatial resolution,temporal resolution, sensitivity inprobe detection, depth of signal pene-tration, availability of biocompatiblemolecular imaging agents, and, ofcourse, cost. Each has its unique advan-tages and disadvantages, and the choiceof imaging system ultimately dependson the question to be addressed. Table 1summarizes the characteristics of eachmodality.

Nuclear Imaging

Many argue that nuclear medicinespecialists have been practicing molec-ular imaging since 131I was first usedfor thyroid imaging in the 1940s. In-deed, PET and SPECT are routinelyused clinically and are the most prev-alent molecular imaging modalities todate. Hundreds of nuclear imagingprobes have been developed. They in-clude radiolabeled enzyme substrates,receptor ligands, antibodies, drugs,and oligonucleotides (36). Examples ofU.S. Food and Drug Administration–approved nuclear probes for SPECT andPET include the 111In–labeled octreotideanalogue Octreoscan (MallinckrodtMedical; Hazelwood, MO) (37) and the18F–labeled glucose analogue 18F-fluoro-deoxyglucose (FDG), respectively.

Nuclear imaging modalities re-motely sense molecular events by de-tecting radioactive emissions from tar-geted radionuclides. PET specificallydetects pairs of coincident �-rays thatresult from positron/electron colli-sions after positron emission. Com-mon positron-emitting isotopes in-clude 18F, 11C, 13N, 15O, and 124I (2,36).SPECT, by contrast, detects �-rays di-rectly from �-emitting isotopes such as99mTc, 111In, 123I, and 131I (2).

Nuclear medicine modalities havebeen at the forefront of molecular im-aging because of their high intrinsicsensitivity, their unlimited depth pen-

Table 1Attributes of Molecular Imaging Modalities (1,2)

Modality Sensitivity Spatial Resolution Temporal Resolution Penetration Depth Cost

SPECT Medium Low Low High MediumPET High Low Low High HighMR imaging Low High High High HighUS Medium Medium High Medium LowOptical imaging High Low High Low Low

1406 • Molecular Imaging: A Primer for Interventionalists and Imagers September 2006 JVIR

etration, and the relative ease of radio-labeling molecular probes (38). PET issuperior to SPECT for imaging molec-ular processes because it is 10–100times more sensitive and becausepositron-emitting isotopes can bereadily substituted for naturally occur-ring atoms (eg, substituting 18F for hy-drogen or 11C for elemental carbon)(2,6). In fact, the sensitivity of PET is inthe nanomolar to femtomolar range,and therefore radiolabeled probes areneeded in only tracer quantities (2).This enables better visualization oftargets of low concentration and min-imizes perturbation of the systemunder investigation. Similarly, theavailability of positron-emitting radio-nuclides of common elements allowsfor radiolabeling by direct isotopic sub-stitution, which minimizes alterations inthe biologic properties of the parentmolecule (39). This is particularly ad-vantageous in the use of imaging toevaluate novel drugs (25). Last, PET of-fers the additional benefit of providingquantitative measures of tracer uptake.

Although nuclear imaging modali-ties provide exceptional sensitivity,they lack the high anatomic detail andspatial resolution of MR imaging andUS. The limited spatial resolution ofcurrent clinical PET scanners rangesfrom 3 mm to 6 mm but is approach-ing 1 mm for small-animal PET (mi-croPET) scanners (38). However, thisconstraint is being addressed by theclinical introduction of integrated PET/CT systems that simultaneously achievehigh probe detection sensitivity andspatial resolution, allowing for directanatomic correlation and more precisetarget localization (40). CombinedPET/CT systems are expected to replacemost stand-alone PET systems in thenext decade (38). In addition, prototypePET/MR imaging scanners are on thedevelopmental horizon (41). Anotherdrawback of PET is the need for an on-site cyclotron to synthesize a broadrange of radioactive tracers because thehalf-lives of positron-emitting radionu-clides are short and the decay processcannot be controlled. Additional limita-tions include patient exposure to radia-tion and the high cost of associatedequipment.

MR Imaging

MR imaging is based on the detec-tion of a signal generated from proton

spin relaxation after the application ofa radiofrequency pulse. Paramagneticagents (eg, gadolinium) and super-paramagnetic agents (eg, iron oxide)alter relaxation times and are used asthe signaling component of MR molec-ular probes. Paramagnetic agentsshorten T1 relaxation times to producesignal enhancement on T1-weightedimages, whereas superparamagneticagents shorten T2 relaxation times toproduce negative contrast on T2-weighted images (42,43). The signal-ing component of gadolinium (Gd3�)–based imaging probes consists generallyof larger nanoparticles containingmultiple Gd3� ions in the form of li-posomes, dendrimers, or perfluorocar-bons (42,44). Similarly, T2 imagingprobes contain monocrystalline orpolycrystalline superparamagnetic ironoxide nanoparticles coated with dextranor other polysaccharides (42,43).

In direct contrast to nuclear imag-ing, MR imaging offers the principaladvantages of exceptional spatial res-olution (10–100 �m) and the ability toconcurrently provide anatomic andmolecular information. Further bene-fits include the absence of ionizing ra-diation and good depth penetration.The main disadvantage of MR imag-ing is its relatively low sensitivity inthe detection of imaging probes, ne-cessitating high probe concentrations(millimolar to micromolar range)and/or highly efficient signal amplifi-cation techniques (1). Either may leadto potential toxicities or complex am-plification steps that may inherentlylimit molecular imaging with MR im-aging. However, the development ofhigher field strengths and super-fastpulse sequences may potentially opennew avenues in the detection of futureMR imaging probes.

Ultrasound

Image generation by US is renderedby detection of differential reflectionof sound waves. US contrast agentsused in molecular imaging include en-capsulated microbubbles, liposomes,and perfluorocarbons emulsions (45,46).Signal enhancement is generated be-cause these agents oscillate strongly inresponse to acoustic pulses, makingthem more reflective than normal tis-sue (47). However, as a result of thesize of these agents (0.1–8 �m in di-ameter), they are restricted to the in-

travascular space; therefore, molecularimaging applications have been lim-ited to vascular processes such as an-giogenesis (48,49), inflammation (16),and thrombosis (50). Notably, effortsare under way to design hybrid probesthat can also carry therapeutic agents,providing simultaneous targeted de-livery of contrast and drug (51).

Although it is less sensitive thannuclear imaging and of lower spatialresolution than MR imaging, US offersa good balance between detection sen-sitivity and spatial resolution (50–500�m). Higher spatial resolution isachieved with higher frequencies, butat the expense of depth penetration(47). Like MR imaging, US can alsoextract anatomic and molecular infor-mation simultaneously. US further-more offers higher temporal resolu-tion, allowing for greater operatorcontrol and the ability to capture dy-namic processes. Additional advan-tages include portability, low cost, ab-sence of ionizing radiation, and thepotential for hybrid imaging/drug de-livery probes.

Optical Imaging

Although most radiologists areleast familiar with the modality of op-tical imaging, techniques such as fluo-rescence microscopy have been amainstay of in vitro molecular and cel-lular biology research for decades. Op-tical imaging detects light photonswith the use of sensitive CCD camerasthat use an absorptive process to con-vert light photons to an electric signal(52). In in vivo optical imaging, lightphotons emitted from imaging probesmust travel through intervening tissueto reach the detector. In its path, signalattenuation occurs as a result of scat-tering and absorption, mainly by he-moglobin (for visible light) and water(for near-infrared [NIR] light) (2,53).Therefore, the depth at which lightphotons can penetrate through opaquetissue has become the primary barrierin the translation of optical imagingfrom in vitro to in vivo application.

The two principal optical ap-proaches to molecular imaging are flu-orescence and bioluminescence imag-ing (53,54). In fluorescence reflectanceimaging, a fluorophore or fluorescentprotein is first illuminated and excitedby an external light source of onewavelength and then emits visible

Wang et al • 1407Volume 17 Number 9

light at a lower wavelength. Green flu-orescent protein from the jellyfish Ae-quorea victoria is perhaps the most wellcharacterized of fluorescent proteins,but has an emission wavelength (510nm) that overlaps with the autofluo-rescence of many tissues (55). Effortsto optimize signal-to-noise ratio haveled to the development of fluoro-chromes that emit light in the NIRrange (650–900 nm), in which there isminimal tissue absorption and auto-fluorescence and therefore improvedsignal penetration (53,56). Several NIRfluorochromes are now available (57),with one, indocyanine green, ap-proved by the Food and Drug Admin-istration for human use (58).

Bioluminescence imaging exploitsthe use of naturally occurring en-zymes that convert chemical energy tolight. The genes for these enzymeshave been cloned from many light-emitting organisms including firefly(Photinus pyralis), coral (Renilla renifor-mis), and jellyfish (A. aequorea) (59). Inthese reactions, the enzyme, generi-cally called luciferase, oxidizes its sub-strate, generically called luciferin, toproduce an electronically excited mol-ecule that emits light (54). Oxygen,adenosine triphosphate, and other co-factors are often required in such pro-cesses. The light emitted from these

bioluminescence reactions are lightblue to yellow-green in color (490–620nm) (59). However, as with fluores-cence, novel red-shifted luciferasesand luciferins (550–700 nm) have beenengineered to improve depth penetra-tion (59). In practice, the tissue beingstudied must first be genetically mod-ified to express the luciferase enzymewith use of established gene-transfertechniques (60). The subject then re-ceives systemic administration of theluciferin substrate before data acquisi-tion. An example of in vivo biolumi-nescence imaging is shown in Figure 1.

In comparing fluorescence and bi-oluminescence imaging, biolumines-cence has the significant advantage ofnot requiring external illumination tofluoresce; therefore, background noiseis minimal and detection sensitivity isincreased. As few as 1,000 luciferase-expressing tumor cells have been de-tected in mouse models (61). The mainlimitation of bioluminescence is theneed for genetic manipulation. How-ever, the recent development of novelrecombinant luciferase fusion reporterproteins composed of luciferase and atargeting moiety promises to expandthe application of bioluminescence im-aging. In a proof of concept study, re-combinant luciferase–vascular endo-thelial growth factor (VEGF) proteins

were produced in bacteria, purified,and then infused to target VEGF re-ceptors and image angiogenesis in amouse tumor model (62).

Because both optical imaging mo-dalities are limited by the depth atwhich probes can be visualized (�2cm) (2), applications are currently re-stricted to in vivo imaging of smallanimals and superficial structures.Clinical uses being explored includedermatologic (63), intraoperative (64),endoscopic (65), and intravascular im-aging (53). To overcome this depthlimitation, research in optical probesand detector technology are underway. Fluorescence-mediated tomogra-phy is a particularly promising newtechnology that is expected to improvespatial resolution and achieve depthpenetrations as great as 10 cm (66).Despite these limitations, small-ani-mal optical imaging has demonstratedtremendous growth in the laboratorysetting. Optical imaging is fast, easy toperform, and relatively inexpensivewhile providing high sensitivity andspatial resolution at the limited depthsrequired for small-animal imaging.Moreover, the plethora of optical re-porters and dyes available renders thetechnology highly versatile and en-ables the possibility for imaging mul-tiple processes simultaneously (34).

MOLECULAR IMAGINGSTRATEGIES

The four imaging systems and theirrespective contrast agents having beendiscussed, our focus turns to how mo-lecular imaging is specifically prac-ticed. The imaging strategies mostwidely used can be classified as directand indirect. Direct molecular imagingis characterized by the direct and spe-cific interaction of the molecular probewith a target, resulting in probe local-ization, accumulation, and/or activa-tion. As shown in Figure 2, the targetcan be a cell surface receptor, antigen,enzyme, transporter, channel, or nu-cleotide. Targets can be located on thecell surface or cell periphery, or in or-ganelles or the nucleus of the cell.Probes designed for direct imaging arehighly specific and can be used to in-terrogate only the intended moleculartarget. Hence, for each novel target, anew probe must be developed, charac-terized, and validated, all of which aretime-consuming and costly endeavors.

Figure 1. Bioluminescence imaging of mouse myocardium at 8 weeks after fireflyluciferase gene transfer and 8 minutes after luciferin infusion. (a) Control mice receivedinjections of saline solution. As expected, no luciferase activity was detected in these mice.(b) In experimental mice, the firefly luciferase gene was introduced into myocardial cellsby direct injection of recombinant viral vectors into the left anterior wall. Luciferaseactivity is demonstrated. Color scale indicates signal intensity. Images courtesy of K.M.R.Prasad and S.S. Berr of the University of Virginia Health System.


The second strategy embodies amore generalized and flexible ap-proach in which established imagingsystems can be linked to various mo-lecular processes of interest and indi-rectly report on their activity. As illus-trated in Figure 3, reporter geneimaging, the most common practice ofindirect imaging, entails expression ofa reporter gene encoding for a tar-getable receptor, enzyme, or trans-porter on the occurrence of a specificmolecular event (21,32,67). Comple-mentary reporter probes are then in-fused systemically and—instead of in-teracting with endogenous targets as indirect imaging—are bound, trapped, oractivated by the exogenous reportergene product to generate a detectablesignal. In some instances, the reportergene product itself provides the imag-ing signal (eg, fluorescent proteins), ob-viating reporter probes. Reporter geneproducts therefore act as surrogatemarkers for the level of expression of the

protein being studied. Although re-porter gene systems can be engineeredto interrogate a variety of biologicevents, they are constrained by the needfor genetic manipulation of the tissuebeing studied. In all cases, the reportergene construct must be first introducedinto the cells of living subjects.

Direct Imaging

The direct strategy is used across allfour imaging modalities and can befurther categorized by the specific tar-get type (Fig 2). Careful target selec-tion is critical for any direct imagingstudy to be informative. Naturally, thetarget should play an important role inthe initiation or progression of the dis-ease pathway being studied andshould also be available in sufficientlyhigh quantities to ensure generation ofa detectable signal. Imaging the endproducts of gene expression is there-fore preferred because transcription

and translation are intrinsic amplifica-tion mechanisms (68).

Imaging cell surface receptors andantigens is a common form of directimaging used in all imaging modali-ties (Table 2). In one example shownin Figure 4, colon carcinoma xeno-grafts in mice were selectively imagedby PET with use of 64Cu–labeled anti-body fragments specific for carcino-embryonic antigen (CEA) (69). Directimaging of enzymes entails enzyme-mediated trapping of a labeled sub-strate or activation of a silent probe.PET with FDG, a widely known exam-ple of the former, specifically imagesthe activity of hexokinase as an indi-cator of glucose use and metabolic ac-tivity. FDG is transported into cells byglucose transporters and then phos-phorylated by hexokinase to FDG-6-phosphate, a modification that pre-vents its diffusion out of cells (72).Relative to receptor imaging, directenzyme imaging provides inherentsignal amplification because enzymescan act on multiple probes, whereasreceptors generally interact with onlya single probe. By contrast, direct im-aging of messenger RNA has beenparticularly challenging, given the rel-atively low number of target messen-ger RNA available per cell (73).

Activatable Probes

The aforementioned approaches todirect imaging often suffer from highbackground noise. Because theseprobes emit signal constitutively, it isnot possible to differentiate probesthat have reached their target fromthose that have not. This necessitates atime delay between tracer injectionand imaging to allow washout of ex-cess unbound probes from the tissueand systemic circulation. Despite this,significant background noise can re-main as a result of nonspecific bind-ing. To circumvent this problem, acti-vatable probes (also called “smartprobes” or “molecular beacons”) havebeen developed for optical and MRimaging. Such agents are designed toemit signal only after activation by theintended target, usually an enzyme. Itis estimated that this strategy, throughbackground suppression, amplifiessignal generation as much as severalhundred fold (74,75).

As illustrated in Figure 5, opticalactivatable probes are composed of an

Figure 2. Targets for direct molecular imaging. Targets can be at the DNA, RNA, orprotein level. Proteins can be intracellular, on the cell surface, or in the cell periphery andfunction as receptors, antigens, enzymes, channels, or transporters. The number of targetsper cell increases as gene expression progresses to the final product.


NIR fluorescent molecule and aquenching molecule attached in closeproximity by a linker that can becleaved only by the target enzyme(74). If exposed to excitation light in itsnative state, the probe is spectrally si-

lenced by the quencher. When theprobe encounters its target, enzyme-mediated cleavage of the linker re-leases the inhibiting quencher, effec-tively “turning on” the probe.Activatable NIR probes have been used

to visualize the activity of several en-zymes, including cathepsin B and D(74), HIV protease, matrix metallopro-teinase (MMP)–2 (Fig 6) (76), protein ki-nase A (77), caspases (78), and thrombin(79).

Figure 3. (a) Concept of reporter gene imaging. Reporter DNA constructs are engineered to encode for a reporter gene (X) and a geneof interest (Y) that are linked such that the two genes are expressed together. In this illustration, genes X and Y are driven by the samepromoter. The construct is then delivered into the cell nucleus via a viral or nonviral vector. Because the expression of genes X and Yare linked, the level of reporter gene expression and activity are proportional to the degree of expression of gene Y. Imaging of thereporter gene product therefore provides a surrogate marker for expression of gene Y. (b) The reporter gene encodes for a detectableprotein product with use of established reporter gene/reporter probe combinations detailed in Table 3. Similarly to direct imaging,reporter probes interact with the specific cell surface receptor, enzyme, or transporter encoded by the reporter gene. Reporter receptorsselectively bind reporter probes, whereas transporters selectively internalize them. Reporter enzymes either activate reporter probes orentrap them by phosphorylation.

Table 2Examples of Direct Molecular Imaging

Molecular Probe

Modality Target ApplicationImaging

ComponentTargeting

Component Reference

SPECT Somatostatin receptorsubtype–2

Neuroendocrinetumor

111In Octreotide analogue* 37

PET CEA Colon carcinoma 64Cu Anti-CEA antibody fragment 69MR imaging HER-2/neu receptor Breast and other

cancersGadolinium Anti–HER2/neu antibody 70

US �v�3 integrin Angiogenesis Microbubble RGD-containing peptide† 49Optical imaging Epidermal growth factor

receptorMultiple cancers Cyanine 5.5 Epidermal growth factor 71

* Octreoscan, a Food and Drug Administration–approved agent for imaging neuroendocrine tumors in humans.† Peptides containing the amino acid sequence arginine-glycine-aspartate (RGD).


An MR imaging activatable probe,named EgadMe, has also been de-signed and consists of a caged Gd3�

ion that is undetectable in its nativestate because the cage prevents watermolecules from accessing the para-magnetic core, a requirement for T1signal enhancement (80). However,one wall of the cage is a galactopyr-anose ring that can be cleaved by theenzyme �-galactosidase. Probe activa-tion is thereby mediated by �-galacto-sidase through opening of the cage,allowing water molecules to enter andGd3� to exert its T1-lengthening effecton water molecule protons.

Indirect Imaging

Reporter gene imaging methodshave been developed for nuclear(21,81,82), optical (61,83), and MR im-aging (80,84). Classification of reportersystems can also be based on whetherthe reporter gene product is an en-zyme, receptor, or transporter (Table3, Fig 3b). The most common enzyme-based system uses herpes simplex vi-rus type 1 thymidine kinase (HSV1-tk)or its mutant derivative (HSV1-sr39tk)as the reporter gene product, whichenables imaging through phosphory-lation and subsequent intracellular

entrapment and accumulation of ra-diolabeled acycloguanosine (eg, ganci-clovir, penciclovir) or pyrimidine nucle-oside derivatives (82,89). The thymidinekinase of herpes simplex virus type 1 isused because it can phosphorylate acy-cloguanosine derivatives, whereas themammalian version, which is morespecific in its activity, cannot. In theexploitation of this difference in sub-strate specificity, background noise isminimized. HSV1-tk has the addedadvantage of also being a therapeuticgene in its ability to convert ganciclo-vir, a prodrug, into a cytotoxic com-pound that inhibits DNA synthesis.This strategy, called suicide gene ther-apy, has been actively pursued in on-cology (90).

Reporter gene imaging has beenused to study a variety of molecularand cellular processes such as cell traf-ficking, gene delivery methods, en-dogenous gene expression, signaltransduction pathways, and gene ex-pression regulation (32). For example,the p53 gene, the most commonly mu-tated gene known in human cancer,encodes for a key transcription factorthat halts cell cycle progression in re-sponse to DNA damage. Doubrovin etal (88) studied p53 regulation of geneexpression by placing a dual reporter

gene encoding for HSV1-tk and greenfluorescent protein under the controlof a p53-specific enhancer, and the re-sulting construct was subsequentlytransduced into tumor cells in rats.DNA damage–induced upregulationof p53 transcriptional activity wasthen imaged by PET through expres-sion of HSV1-tk and accumulation ofits reporter probe and confirmed byfluorescence microscopy of green flu-orescent protein expression.

Although clinical application of thistechnology is limited by the need tointroduce reporter gene constructsinto target tissue, reporter gene imag-ing promises to play a vital role in thedevelopment of human gene therapy(60,91). Gene therapy has been signif-icantly hampered by the inability toeffectively evaluate the delivery andexpression of therapeutic genes. Vari-ous genetic engineering strategieshave been developed to link the ex-pression of therapeutic genes with thatof reporter genes (92,93), allowing forin vivo reporter gene imaging to beused to monitor the location, magni-tude, and duration of therapeutic geneexpression (91,93). Indeed, reportergene imaging has been used exten-sively in animal models of gene ther-

Figure 4. Selective in vivo microPET imaging of colon cancer xenografts with 64Cu–anti-CEA antibody fragments. (a) MicroPET scanof a mouse administered 200 �Ci of FDG and scanned 1 hour after injection. The left shoulder was implanted with a CEA-negative C6glioma xenograft (arrowhead), and the right shoulder was implanted with a CEA-positive LS174T colon carcinoma xenograft (arrow).Both xenografts are imaged here. (b) MicroPET scan of the same mouse injected with 26 �Ci of 64Cu–anti-CEA antibody fragments andimaged at 5 hours with the highest retention shown in the LS174T tumor (arrow) and liver and lower retention in the control tumor(arrowhead). (c) The mouse was killed for whole-body autoradiography, and an anatomic photograph was taken at the time ofsectioning. (d) Digital autoradiograph of adjacent section. Color scale indicates signal intensity for 64Cu–anti-CEA antibody fragmentand FDG PET imaging. Reproduced with permission from Wu et al (69).


apy, and initial studies in humanshave also been completed (13).

CLINICAL APPLICATIONSIn its capacity to noninvasively in-

terrogate cellular, molecular, and ge-netic processes fundamental to dis-eases, molecular imaging has thepotential to affect all disciplines ofclinical medicine. A survey of the mo-lecular imaging literature reveals thatits application to human ailments isbeing explored in a range of disparatefields from neurology and psychiatryto infectious diseases and drug resis-tance. Fortuitously, cardiovasculardisease and cancer, the leading causesof mortality and areas most relevant tointerventional radiology, have been atthe forefront of translational molecu-lar imaging research.

Molecular Imaging ofCardiovascular Disease

Atherosclerosis and restenosis.—Ath-erosclerosis, the leading cause ofmorbidity and mortality in devel-oped countries, is in essence achronic inflammatory process (94,95).Macrophages are the principal in-flammatory cell mediator of ather-oma formation, progression, andeventual disruption (96). Because un-stable plaques are morphologicallycharacterized by a preponderance ofmacrophages (97,98), there has beenconsiderable interest in in vivo imag-ing of macrophage density in plaquesas a means to assess vascular inflam-mation and risk of rupture. Exploitingthe innate ability of macrophages tophagocytose foreign materials, two dis-

tinct human studies (99,100) infusedpatients with iron oxide nanoparticlesbefore carotid endarterectomy for di-rect macrophage imaging (99,100). En-suing MR imaging demonstrated pref-erential signal attenuation in rupture-prone plaques with histopathologicconfirmation of high macrophage den-sity and phagocytosis of iron oxidenanoparticles.

Other studies of atherosclerosishave focused on the underlyingmechanisms of macrophage-mediatedplaque rupture (101,102). Macro-phages secrete proteolytic enzymes,such as cathepsins and MMPs, thatweaken the structure of atheroscle-rotic plaques. As illustrated in Figure7, fluorescence-mediated tomographywith an activatable probe specific forcathepsin B in a mouse model ofatherosclerosis recorded fluorescencesignal that was spatially correlated toregions high in cathepsin B activity,macrophage infiltration, and lipidcontent (103). Investigative applica-tion of activatable protease-specificprobes could be extended to othervascular diseases, given that cathep-sins and MMPs have been implicatedin the pathogenesis of aneurysmsand restenosis (104–106).

The application of molecular im-aging in restenosis remains in itsearly stages. Postangioplasty and in-stent restenosis are maladaptive re-sponses to injury characterized bysmooth muscle cell proliferation andmigration, neointimal hyperplasia,and vascular remodeling (107). Manygroups have therefore used molecu-lar imaging tools to study the patho-logic activities of vascular smoothmuscle cells (108,109) An MR imag-ing study designed probes directedtoward tissue factor (109), a moleculeoverexpressed by arterial smoothmuscle cells after vascular injury(110). Directed probes were synthe-sized by conjugating perfluorocarbonemulsion Gd3� nanoparticles withtissue factor antibody fragments(109). Remarkably, these nanoprobeswere also loaded with antiprolifera-tive drugs, either paclitaxel or doxo-rubicin, to illustrate their potential asan imageable and targetable drug de-livery platform. After administrationof these therapeutic nanoparticles,MR imaging not only confirmedsmooth muscle cell–specific probedelivery but also provided a nonin-

Figure 5. Concept of activatable optical probes. In its native state, the fluorophore isstructurally linked in close proximity with a quenching molecule that prevents thefluorophore from emitting fluorescent signal on excitation. The linker, usually a shortpeptide, is designed to be specifically cleaved by an enzyme of interest. When the probeencounters this enzyme, enzyme-mediated cleavage of the linker releases the quencher,allowing the fluorophore to emit signal on excitation. The probe is therefore activated onencounter with the specific activating enzyme.


Figure 6. Selective in vivo NIR opticalimaging of tumor xenografts expressingMMP-2 with use of MMP-2–sensitiveactivatable probes. (a) Photograph ofmouse bearing tumor xenografts. The leftshoulder was implanted with an MMP-2–positive HT1080 fibrosarcoma xenograft,and the right shoulder was implantedwith an MMP-2–negative BT20 mammaryadenocarcinoma xenograft. (b) Raw NIRimaging of the same mouse 2 hours afterintravenous injection of theMMP-2–sensitive probe. The MMP-2–positive fibrosarcoma generatedsignificantly higher fluorescent signalintensity versus control. Signalattenuation was observed onpretreatment with MMP-2 inhibitors (notillustrated). Reproduced with permissionfrom Bremer et al (76).


vasive measure of local drug deliv-ery. Antiproliferative activity wassignificantly enhanced with the tar-geted probes relative to nonspecificcontrols.

Thrombosis.—Thrombosis is thepathologic hallmark of acute coro-nary syndromes, stroke, deep veinthrombosis, and pulmonary embo-lism. Several mediators of thrombosishave been visualized in attempt tofurther elucidate their role and func-tion and for in vivo thrombus detec-tion and characterization. Becausemany proteins of the coagulation cas-cade are proteases, they are par-ticularly amenable to imaging withactivatable probes (111). Indeed, acti-vatable NIR fluorescence probes se-lective for thrombin and factor XIIIhave been synthesized and tested inexperimental models of thrombosis(112,113). However, for the purposeof detecting thrombus clinically, fi-brin is the more appropriate targetbecause it is present in all thrombi,regardless of type and age. In a se-ries of animal studies of differentthrombotic pathologic processes(114–116), MR imaging detection ofacute, subacute, and chronic thrombiwas significantly improved with theuse of a fibrin-targeted contrast agent,EP-2104R (Epix Pharmaceuticals; Cam-bridge, MA), compared with conven-tional methods. EP-2104R is currentlyin clinical trials. Other fibrin-targetedmolecular probes have been tested fornuclear imaging (117), US (118), andMR imaging (119). Last, several nu-clear imaging tracers have been devel-

oped to label platelets by specificallytargeting glycoprotein IIb-IIIa receptors(117). One such agent has been ap-proved by the Food and Drug Admin-istration and was found to detect acutedeep vein thrombosis with sensitivityand specificity rates of 92% and 86%,respectively (120). Platelet-specific USand MR imaging probes have alsobeen reported (121,122).

Angiogenesis.—Angiogenesis, theformation of blood vessels, is ahighly dynamic process that occursin a variety of physiologic andpathologic states (123). Pathologic an-giogenesis plays a critical role in tu-mor progression, atherogenesis, pro-liferative retinopathies, and variousother inflammatory and ischemic dis-eases. Several imaging strategies havecentered on assessing new vessel for-mation. Because �v�3 integrins are se-lectively expressed in angiogenic butnot quiescent vessels, they represent apromising biomarker for distinguish-ing neovasculature from existing bloodvessels. Molecular probes specific for�v�3 integrins have been developed fornuclear imaging (124,125), MR imaging(126,127), US (49), and optical imaging(128).

Whether angiogenesis occurs de-pends on a fine balance between an-giogenic promoters and inhibitors.Because VEGF is a potent inducer ofangiogenesis, there has been consid-erable interest in imaging VEGF tostudy its regulation and biologic ac-tivities. For example, hypoxia hasbeen shown to induce expression ofVEGF and its receptors to mediate

hypoxia-induced compensatory an-giogenesis (129). In a rabbit hindlimb ischemia model, uptake of 111In-labeled VEGF121 was significantly en-hanced in ischemic tissue versus con-trol tissue, suggesting that VEGF re-ceptor imaging could be used to pre-cisely identify ischemic tissue (130).VEGF has also been used for thera-peutic angiogenesis via exogenousapplication and gene therapy (131).In an example of reporter gene moni-toring of gene therapy, Wu et al (132)linked the HSV1-sr39tk reporter genewith the VEGF121 therapeutic geneand then introduced this fusion geneinto ischemic myocardium in a ratmodel of myocardial infarction. Afteradministration of the reporter probe,noninvasive imaging by microPETwas found to correlate closely withVEGF121 and HSV1-sr39tk expressionas assessed by immunohistochemis-try and enzyme assays. As expected,VEGF121 gene therapy also induced asignificant increase in angiogenesis.

Cardiac stem cell therapy.—Correc-tive or regenerative cell therapy isthe use of cells as therapeutic agents.Headlined by the promising thera-peutic potential of stem cells, celltherapy is being investigated for therestoration of tissue and function in abroad range of selective or territorialcell-loss pathologic processes such astype I diabetes and myocardial in-farction, respectively (133). To date,more than 15 clinical trials have eval-uated the use of stem cells to regen-erate myocardial tissue in ischemicheart disease patients, with the ma-

Table 3Reporter Gene Systems

Reporter Gene Product Reporter Gene Product Category Reporter Probe Imaging Modality Reference

HSV1-tk, HSV1-sr39tk Enzyme 18F-labeled uracil oracycloguanosine analogues

PET 81,82

Luciferase Enzyme Luciferin Bioluminescence 61�-galactosidase Enzyme EGadMe MR imaging 80Dopamine-2-receptor Cell surface receptor 18F-fluoroethylspiperone PET 85Somatostatin receptor

subtype-2Cell surface receptor 111In–labeled

DTPA-octreotideSPECT 86

Transferrin receptor Cell surface receptor Transferrin-MION MR imaging 84Sodium iodide symporter Cell membrane transporter 131I SPECT 87Green fluorescent protein Fluorescent protein None Fluorescence

imaging88

Note.—DTPA � diethylenetriaminepentaacetic acid; MION � monocrystalline iron oxide nanoparticle.


jority demonstrating improved myo-cardial function (134). Central tostem cell therapy is the remarkableability of such cells to migrate fromthe site of transplantation to relevantfoci of disease and differentiate intothe appropriate cell type in responseto chemical cues in the microenviron-

ment. Reliable noninvasive means toassess the delivery, distribution, andfate of transplanted stem cells aretherefore imperative and are beingestablished with use of molecular im-aging techniques.

There are two distinct approachesto imaging cell-based therapies: di-

rect labeling and genetic integrationof reporter genes; both modificationsare rendered before transplantation(135). In the first approach, therapeu-tic cells are loaded intracellularlywith a biocompatible imaging agent.Although all imaging modalities cantheoretically be used for direct label-ing, MR imaging with iron oxidenanoparticles is by far the mostwidely used method (Fig 8). Thehigh spatial resolution of MR imag-ing is ideally suited for tracking stemcells in vivo, and iron oxide nanopar-ticles are less toxic, more stable, andmore sensitively detected than Gd3�.In fact, in vivo MR imaging of a sin-gle iron oxide-loaded cell with a 7-Tscanner has been reported (136). Ap-plied to a swine model of myocardialinfarction, labeled stem cells weresuccessfully imaged with a standard1.5-T clinical scanner for as long as21 days after percutaneous injectionunder fluoroscopy (137). Cell viabil-ity, proliferation, and differentiationcapacity were not adversely affectedby labeling in vitro. Despite thesepreliminary results, there are signifi-cant limitations to this approach. De-tection becomes more difficult overtime as contrast agents are progres-sively diluted with each cell division.In addition, imaging agents that per-sist after cell death can be engulfedby local macrophages and compro-mise imaging specificity.

In the second approach, reportergenes are engineered for constitutiveexpression and subsequently inte-grated into the genome of therapeuticcells. Longer-term imaging is thereforetheoretically possible because the re-porter genes are preserved and passeddown to progeny. All previously de-scribed reporter gene systems can bepotentially applied for cell tracking. Inan elegant animal study (22), embry-onic stem cells were transduced with afusion reporter gene encoding for fire-fly luciferase and a derivative ofHSV1-sr39tk and then injected into themyocardium of rats. Stem cell survival,proliferation, and migration were non-invasively assessed by biolumines-cence and PET imaging after infusionof luciferin and 9-(4-18F-fluoro-3-[hy-droxymethyl]butyl)guanine, respec-tively, over a period of 4 weeks. Bi-oluminescence and PET signals corre-lated closely, and both actually in-creased over time. Anatomic context,

Figure 7. NIR optical imaging of atherosclerosis in mice aortas with use of cathepsinB-sensitive activatable probes. (a) Sudan IV staining of longitudinally opened aorta, inwhich red areas represent lipid-rich areas stained with Sudan IV. (b) Corresponding NIRimage shows prominent cathepsin B signal from atherosclerotic lesions that matchesSudan IV staining. Native atherosclerotic lesions had NIR autofluorescence similar to thatin normal aorta. Reproduced with permission from Chen et al (103).


to determine the intracardiac locationof the transplanted cells, was pro-vided by FDG PET imaging of thehypermetabolic myocardial tissue.The main disadvantage of this ap-proach, again, is the need to geneti-cally manipulate cells—in this case,cells that hold the potential to trans-form and proliferate unchecked. Re-markably, this same study (22)showed preliminary data thatHSV1-tk and its derivatives, whichdouble as suicide genes, can be ex-ploited as a safety mechanismagainst tumor formation by adminis-tration of the prodrug ganciclovir.

Molecular imaging of cell therapyhas been demonstrated in experimen-tal models of several other disorders,including ischemic stroke (138), Par-kinson disease (139), type I diabetes(140), and muscular dystrophy (141).In human and large animal studies,transplantation is performed by di-rect injection or intravascular tech-niques under the guidance of con-ventional imaging. Interventional ra-

diologists are therefore projected toplay a large role in stem cell delivery(142,143).

Molecular Imaging of Cancer

Cancer diagnosis and staging.—Early diagnosis and accurate stagingremain the mainstays for successfulcancer therapy (144). Despite recentadvances in imaging technologies,the reliance of diagnostic imaging onmorphologic imaging leaves a con-siderable void in the accurate detec-tion of early-stage and small-volumetumors. Molecular imaging, in sens-ing molecular alterations in tumorexpression patterns, holds muchpromise to sensitively detect smallcancers, to better distinguish malig-nant from benign lesions, and to re-duce the delay between initial malig-nant transformation and diagnosis.

A common approach involves di-rect imaging of receptors or antigensthat are exclusively or highly ex-pressed by cancer cells (145). Numer-

ous examples exist, including thepreviously described Octreoscan and64Cu-labeled anti-CEA antibody frag-ments (Fig 4) (37,69), as well as ef-forts to image dopamine and folatereceptors overexpressed by pituitaryadenomas and ovarian carcinomas,respectively (146,147). An extensionof this is the use of NIR activatableprobes to image enzymes upregu-lated in cancer cells such as cathep-sins and MMPs (75). By imaging tu-mor-associated proteases, Weisslederet al (75) were able to detect tumorssmaller than 0.3 mm in diameter inanimal models. Central to these strat-egies is the identification and valida-tion of reliable and specific molecularmarkers of cancer (148). High-throughput screening technologiessuch as microarray analysis can com-prehensively analyze differences inexpression patterns between normaland malignant cells and have provedparticularly important in the discov-ery of novel tumor markers.

Instead of imaging specific tumor-associated markers, FDG PET detectsthe unusually high metabolic activitycommon to most tumors and hasrapidly emerged as a widely usedoncologic imaging tool. FDG PET hasbeen shown to be more accurate thanCT alone in the detection and stagingof several cancers, such as non–small-cell lung cancer, breast cancer,lymphoma, and melanoma (149).However, because of its spatial reso-lution limitations, PET alone is beingreplaced by combined PET/CT for itsimproved anatomic detail and moreprecise lesion localization (38). In aprospective study comparing CT,PET, and PET/CT for staging non–small-cell lung cancer, PET/CT pro-vided additional information in 41%of patients and demonstrated a sig-nificantly improved staging accuracyof 88% versus 40%–58% for PET orCT alone (150).

Because macrophage imaging hasbeen used to assess atherosclerosis,MR imaging of in vivo macrophagemonocrystalline uptake is being in-vestigated for diagnosis of lymphnode metastases. Because nodal tissuesinvaded by cancer cells are unable tophysically accommodate macrophages,the absence of T2 hypointensity oflymph nodes on MR imaging wouldprovide an imaging marker for nodalinvasion. In a human study of pros-

Figure 8. Longitudinal assessment of cardiac stem cell transplantation in a mouse modelof myocardial infarction by MR imaging with use of a 4.7-T scanner. Myocardial infarc-tion was surgically induced by coronary occlusion in experimental mice. Stem cellsloaded with iron oxide nanoparticles were directly injected into the infarcted region andtracked by MR imaging at day 1 (a), day 7 (b), and day 28 (c) after injection. Short-axis andcorresponding long-axis images are shown in the first and second columns, respectively.Note that iron-loaded stem cells, seen as negative contrast, were localized to the infractedanterolateral wall at each time point (arrows). Images courtesy of A.J. Katz and B.A.French of the University of Virginia Health System.


tate cancer (12), MR imaging after in-fusion of lymphotrophic monocrys-talline particles was able to detectnodal metastases smaller than 2 mmin diameter and yielded sensitivityand specificity rates greater than95%.

Tumor characterization.—Currentmethods of biologically characteriz-ing tumors require invasive tissue ac-quisition for analysis. Receptor- andenzyme-based molecular imagingmethods may also provide noninva-sive means of gaining real-time in-sights into tumor biology. For exam-ple, high expression levels of cathep-sins, MMPs, and epidermal growthfactor receptors have been linked tohighly aggressive and invasive tu-mors and poor prognosis (151–153).In a mouse study comparing aggres-sive and well-differentiated breastcancer tissue (154), differential levelsof cathepsin activity were depictedby NIR fluorescence imaging of a ca-thepsin B–sensitive activatable probe.Relative to the well-differentiatedcancer tissue, NIR fluorescence signalwas 1.5 times greater in the aggres-sive breast cancer tissue. This corre-lated well with increased cathepsin Bexpression and rapid tumor growth.Receptor imaging may also proveuseful for tailoring the use of tar-geted molecular therapeutics such astrastuzumab (Herceptin; Genentech;South San Francisco, CA), tamoxifen(AstraZeneca; Wilmington, DE), andgefitinib (Iressa; AstraZeneca). Meth-ods have been devised to imageHER-2/neu expression by MR imag-ing (70), estrogen and progesteronereceptor status by nuclear imaging(155), and epidermal growth factorreceptor expression by optical imag-ing (71).

Therapy assessment.—Decreased tu-mor size remains the traditional crite-rion to assess oncologic therapeuticresponse (156). As with disease de-tection, morphologic changes aftertherapy are late and nonspecific, pre-cluding timely adjustments in thera-peutic regimen. In addition, manynovel oncologic therapies aim to in-hibit cell proliferation and angiogene-sis and may therefore arrest tumorgrowth but not necessarily reduce tu-mor size (33). Surrogate imagingendpoints for monitoring therapeuticefficacy are therefore needed. Be-

cause molecular imaging allows fornoninvasive and repetitive probingof cellular and molecular processes,treatment response can be monitoredby performing serial imaging andnoting changes in the surrogate end-point.

These endpoints can be specific tothe drug’s therapeutic mechanism ora nonspecific change in tumor cellactivity and viability. For example,antiangiogenesis drugs and MMP in-hibitors can be monitored by imagingdecreases in angiogenesis markersand MMP enzyme activity, respec-tively (157,158). In a more general-ized approach, posttreatment de-creases in tumor cell metabolism asmeasured by FDG PET have beenshown to correlate well with a reduc-tion of viable tumor cells (159). Aspresented in Figure 9, serial FDGPET was used to assess the efficacyof imatinib (Gleevec; Novartis, EastHanover, NJ) in patients with gastro-intestinal stromal cancer (160). Imag-ing of apoptosis has been anotherproposed means to evaluate onco-logic treatment efficacy. Apoptoticcells express the membrane phospho-lipid phosphatidylserine on the outersurface of cell membranes early inthe apoptotic process (161). AnnexinV binds to phosphatidylserine selec-tively and with high affinity and hastherefore been a commonly used li-gand for apoptosis imaging (17). In amouse model of radiation therapy, 15Gy of radiation was given to one oftwo mouse hind legs inoculated withmurine mammary tumor cells. Ensu-ing fluorescence imaging with cya-nine 5.5–labeled annexin V demon-strated greater signal from the irradi-ated leg than from the contralateralcontrol, indicating an increase in ra-diation-induced apoptosis (unpub-lished data, H. Sun and D. Pan,2006).

CONCLUSIONS

The advent of molecular imaginghas brought about fundamental para-digm shifts in our approach to re-search and patient care. In the labora-tory setting, molecular imagingprovides innovative tools for dissec-tion of the cellular and molecular fea-tures of various biologic processeswithin the milieu of living organisms.

In clinical practice, molecular imagingholds promise to ultimately allow ear-lier and more precise disease diagno-sis, disease characterization, and as-sessment of therapeutic response. Thepractice of molecular medicine willconsist of prevention, diagnosis, andtreatment methods that directly targetthe molecular, cellular, or physiologicdefects responsible for disease. Molec-ular imaging sits at the interface of thistransition and offers a clear opportu-nity for interventional radiology tohave a strong presence in this age ofmolecular medicine (162).

The convergence of molecular im-aging and interventional radiologybenefits both fields. As with diagnosticradiology, interventional radiologyhas traditionally relied on morpho-logic imaging. In its promise to imagemolecular and cellular processes withhigh spatial and temporal resolution,molecular imaging tools can providesignificant improvements in target vi-sualization and characterization, ther-apy planning, procedural guidance,and treatment monitoring. In addition,the adaptation of MR, US, and opticalmolecular imaging methods for appli-cation in the interventional suite willallow interventional radiology to ad-vance from its reliance on ionizing ra-diation–based imaging. With the con-tinuing development of interventionalMR (163), intravascular US (164), andintravascular optical systems (53), theuse of these imaging modalities is notfar from reality. Conversely, the trans-lation of molecular imaging tech-nologies to clinical application may beaccelerated through the use of inter-ventional techniques. Molecular imag-ing probes are rendered useless if theycannot reach their target as a result ofinherent biologic barriers, degrada-tion, or nonspecific binding (1). Al-though there has been much researchto overcome these limitations, localdelivery may be necessary for certainclasses of molecular probes.

Historically, those who invest in theresearch and development of newtechnologies are able to assume own-ership of the innovation and im-plement it in practice. As molecularimaging matures, its first clinical ap-plications are likely to be in cardiovas-cular and oncologic imaging. Fortu-itously, cardiovascular disease andcancer are the likely principal clinical


domains of interventional radiology inthe intermediate future (162). Withmutual benefits in the integration

of molecular imaging technologieswith interventional techniques andthe arrival of molecular medicine,

we believe it is time for interventionalradiologists to assume key roles in thedevelopment of molecular imaging.

Figure 9. Longitudinal assessment of imatinib treatment in a patient with a pelvic gastrointestinal stromal tumor by FDG PET.Sequential PET scans were obtained in the same patient before treatment (a), 1 month after imatinib treatment (b), and after 16 monthsof continuous treatment (c). Images at each time point include a two-dimensional PET scan of the body (top row), an axial PET scanthrough the pelvic tumor (middle row), and a corresponding axial CT scan (bottom row). FDG uptake in the cardiac blood pool,the myocardium, the liver, the bowel, the bilateral renal collecting system, and the bladder were within physiologic limits in thispatient. The patient had similar blood glucose concentrations at each of these time points. Reproduced with permission fromDemetri et al (160).


APPENDIX I: GLOSSARY OFTERMS

Angiogenesis: formation of newblood vessels.

Antisense: sequence of DNA orRNA that is complementary to andbinds target DNA or RNA.

Apoptosis: programmed cell death.Autofluorescence: the inherent flu-

orescence of tissues.Cell surface receptor: protein

and/or polysaccharide structure onthe surface of a cell that selectivelybinds certain molecular messengers.

Clone: production of multiple cop-ies of a single gene.

Dendrimer: branching polymerused to transfer genetic material intocells.

Enhancer: nucleotide sequence thatenhances transcription of a particulargene.

Gene construct: an engineered se-quence of several genes spliced to-gether.

Gene expression: process by whichproteins are made from the instruc-tions encoded in DNA.

In vitro: in an artificial environ-ment outside the living organism.

In vivo: within a living organism.Ligand: molecule that binds to a

receptor.Liposome: closed lipid vesicle that

may be used to encapsulate materialsfor delivery into cells.

mRNA: messenger RNA; serves asthe template for protein synthesis.

Oligonucleotide: small chain ofDNA or RNA subunits, consisting of abase, a phosphate, and a sugar mole-cule.

Phagocytose: internalization of par-ticles by cells.

Phosphorylation: the addition ofphosphate to an organic compoundthrough the action of a phosphorylaseor kinase.

Progenitor cell: multipotential in-termediate stem cells that are directprecursors for tissue-specific maturecells.

Promoter: binding site in a DNAchain at which RNA polymerase bindsto initiate transcription of messengerRNA of one or more nearby genes.

Protease: enzymes that catalyze thebreakdown of proteins.

Reporter gene: gene that encodesfor a protein that is readily detectable,with or without the use of reporter

probes, and serves as a marker for ex-pression of the gene construct.

Stem cell: primitive precursor cellsthat possess the capability to prolifer-ate, self-renew, and differentiate into avariety of cell types.

Substrate: molecule acted upon byan enzyme.

Transcription: synthesis of messen-ger RNA by RNA polymerases usinginformation encoded in DNA.

Transduction: introduction and in-tegration of exogenous DNA into arecipient cell’s genome by use of a vi-ral system.

Transfection: introduction and in-tegration of exogenous DNA into arecipient cell’s genome by use of anonviral system.

Translation: synthesis of proteinusing the information encoded in mes-senger RNA.

Tumorigenesis: development of tu-mor cells.

Tumor markers: specific biomol-ecules, such as enzymes or antigens,that are associated with the presenceof cancer.

Vector: viral or nonviral vehicleused to deliver genes into a target cell.

APPENDIX II: WEB SITES

The Academy of Molecular Imag-ing: http://www.ami-imaging.org

The Society for Molecular Imaging:http://www.molecularimaging.org

Molecular Imaging Central: http://www.mi-central.org

References1. Weissleder R, Mahmood U. Molecu-

lar imaging. Radiology 2001; 219:316–333.

2. Massoud TF, Gambhir SS. Molecu-lar imaging in living subjects: seeingfundamental biological processes in anew light. Genes Dev 2003; 17:545–580.

3. Piwnica-Worms D. Introduction tomolecular imaging. J Am Coll Radiol2004; 1(suppl):S2–S3.

4. Jaffer FA, Weissleder R. Molecularimaging in the clinical arena. JAMA2005; 293:855–862.

5. Tempany CM, McNeil BJ. Advancesin biomedical imaging. JAMA 2001;285:562–567.

6. Hengerer A, Wunder A, WagenaarDJ, et al. From genomics to clinicalmolecular imaging. IEEE Eng MedBiol Mag 2005; 93:819–828.

7. Wickline SA, Lanza GM. Nanotech-nology for molecular imaging and tar-

geted therapy. Circulation 2003; 107:1092–1095.

8. Phelps ME. Nuclear medicine, mo-lecular imaging, and molecular med-icine. J Nucl Med 2002; 43:13N-14N.

9. Thrall JH. How molecular medicinewill impact radiology. Diagn Imaging(San Franc) 1997; 19:23,27.

10. Jain KK. Personalized medicine.Curr Opin Mol Ther 2002; 4:548–558.

11. Voges J, Reszka R, Gossmann A, et al.Imaging-guided convection-enhanceddelivery and gene therapy of glioblas-toma. Ann Neurol 2003; 54:479–487.

12. Harisinghani MG, Barentsz J, HahnPF, et al. Noninvasive detection ofclinically occult lymph-node metasta-ses in prostate cancer. N Engl J Med2003; 348:2491–2499.

13. Jacobs AH, Voges J, Kracht LW, et al.Imaging in gene therapy of patientswith glioma. J Neurooncol 2003; 65:291–305.

14. Kitano H. Systems biology: a briefoverview. Science 2002; 295:1662–1664.

15. Serganova I, Doubrovin M, Vider J, etal. Molecular imaging of temporaldynamics and spatial heterogeneity ofhypoxia-inducible factor-1 signaltransduction activity in tumors in liv-ing mice. Cancer Res 2004; 64:6101–6108.

16. Lindner JR. Assessment of inflam-mation with contrast ultrasound.Prog Cardiovasc Dis 2001; 44:111–120.

17. Zhao M, Beauregard DA, Loizou L, etal. Non-invasive detection of apo-ptosis using magnetic resonance im-aging and a targeted contrast agent.Nat Med 2001; 7:1241–1244.

18. Schirner M, Menrad A, Stephens A, etal. Molecular imaging of tumor an-giogenesis. Ann N Y Acad Sci 2004;1014:67–75.

19. Weissleder R. Scaling down imag-ing: molecular mapping of cancer inmice. Nat Rev Cancer 2002; 2:11–18.

20. Blasberg R. Imaging gene expres-sion and endogenous molecular pro-cesses: molecular imaging. J CerebBlood Flow Metab 2002; 22:1157–1164.

21. Gambhir SS, Herschman HR, CherrySR, et al. Imaging transgene expres-sion with radionuclide imaging tech-nologies. Neoplasia 2000; 2:118–138.

22. Cao F, Lin S, Xie X, et al. In vivovisualization of embryonic stem cellsurvival, proliferation, and migrationafter cardiac delivery. Circulation2006; 113:1005–1014.

23. Rudin M, Weissleder R. Molecularimaging in drug discovery and devel-opment. Nat Rev Drug Discov 2003;2:123–131.

24. Contag PR. Whole-animal cellularand molecular imaging to accelerate


drug development. Drug Discov To-day 2002; 7:555–562.

25. Wang J, Maurer L. Positron emis-sion tomography: applications indrug discovery and drug develop-ment. Curr Top Med Chem 2005;5:1053–1075.

26. Hillman BJ, Neiman HL. Translat-ing molecular imaging research intoradiologic practice: summary of theproceedings of the American Collegeof Radiology Colloquium, April 22–24, 2001. Radiology 2002; 222:19–24.

27. Wagenaar DJ, Weissleder R, HengererA. Glossary of molecular imagingterminology. Acad Radiol 2001; 8:409–420.

28. Pandit SD, Li KC. A primer on mo-lecular biology for imagers: IV. con-cepts and basic methods in molecularbiology. Acad Radiol 2004; 11:686–697.

29. Pandit SD, Li KC. A primer on mo-lecular biology for imagers: III. Pro-teins: structure and function. AcadRadiol 2004; 11:448–461.

30. Pandit SD, Bednarski M, Li KC. Aprimer on molecular biology for im-agers: I. DNA, how does it work?Acad Radiol 2003; 10:1215–1223.

31. Pandit SD, Li KC. A primer on mo-lecular biology for imagers: II. Tran-scription and gene expression. AcadRadiol 2004; 11:333–344.

32. Blasberg RG, Tjuvajev JG. Molecu-lar-genetic imaging: current and fu-ture perspectives. J Clin Invest 2003;111:1620–1629.

33. Herschman HR. Molecular imaging:Looking at problems, seeing solu-tions. Science 2003; 302:605–608.

34. Weissleder R. Molecular imaging:exploring the next frontier. Radiology1999; 212:609–614.

35. Miller JC, Thrall JH. Clinical molec-ular imaging. J Am Coll Radiol 2004;1(suppl):S4–S23.

36. Phelps ME. Inaugural article:positron emission tomography pro-vides molecular imaging of biologicalprocesses. Proc Natl Acad Sci U S A2000; 97:9226–9233.

37. Krenning EP, Kwekkeboom DJ, Bak-ker WH, et al. Somatostatin receptorscintigraphy with [111In-DTPA-D-Phe1]- and [123I-Tyr3]-octreotide: theRotterdam experience with more than1000 patients. Eur J Nucl Med 1993;20:716–731.

38. Gambhir SS. Molecular imaging ofcancer with positron emission tomog-raphy. Nat Rev Cancer 2002; 2:683–693.

39. Park JM, Gambhir SS. Multimodal-ity radionuclide, fluorescence, and bi-oluminescence small-animal imaging.IEEE Eng Med Biol Mag 2005; 93:771–783.

40. Townsend DW, Beyer T. A com-bined PET/CT scanner: the path totrue image fusion. Br J Radiol 2002;75:S24–S30.

41. Grazioso R, Ladebeck R, Schmand M,et al. APD-based PET for combinedMR-PET imaging [abstract]. Proc IntlSoc Mag Reson Med 2005; 13:408–.

42. Allen MJ, Meade TJ. Magnetic reso-nance contrast agents for medical andmolecular imaging. Met Ions Biol Syst2004; 42:1–38.

43. Basilion JP, Yeon S, Botnar R. Mag-netic resonance imaging: utility as amolecular imaging modality. CurrTop Dev Biol 2005; 70:1–33.

44. Artemov D. Molecular magneticresonance imaging with targeted con-trast agents. J Cell Biochem 2003; 90:518–524.

45. Christiansen JP, Lindner JR. Molec-ular and cellular imaging with tar-geted contrast ultrasound. IEEE EngMed Biol Mag 2005; 93:809–818.

46. Lindner JR. Molecular imaging withcontrast ultrasound and targeted mi-crobubbles. J Nucl Cardiol 2004; 11:215–221.

47. Liang HD, Blomley MJ. The role ofultrasound in molecular imaging. Br JRadiol 2003; 76(suppl):S140–S150.

48. Ellegala DB, Leong-Poi H, CarpenterJE, et al. Imaging tumor angiogene-sis with contrast ultrasound and mi-crobubbles targeted to alpha(v)beta3.Circulation 2003; 108:336–341.

49. Leong-Poi H, Christiansen J, KlibanovAL, et al. Noninvasive assessmentof angiogenesis by ultrasound andmicrobubbles targeted to alpha (v)-integrins. Circulation 2003; 107:455-460.

50. Lanza GM, Wallace KD, Fischer SE, etal. High-frequency ultrasonic detec-tion of thrombi with a targeted con-trast system. Ultrasound Med Biol1997; 23:863–870.

51. Unger EC, Hersh E, Vannan M, et al.Local drug and gene delivery throughmicrobubbles. Prog Cardiovasc Dis2001; 44:45–54.

52. Spibey CA, Jackson P, Herick K. Aunique charge-coupled device/xenonarc lamp based imaging system forthe accurate detection and quantita-tion of multicolour fluorescence. Elec-trophoresis 2001; 22:829–836.

53. Weissleder R, Ntziachristos V.Shedding light onto live moleculartargets. Nat Med 2003; 9:123–128.

54. Contag PR, Olomu IN, Stevenson DK,et al. Bioluminescent indicators inliving mammals. Nat Med 1998;4:245–247.

55. Tsien RY. Building and breedingmolecules to spy on cells and tumors.FEBS Lett 2005; 579:927–932.

56. Chance B. Near-infrared images us-

ing continuous, phase-modulated,and pulsed light with quantitation ofblood and blood oxygenation. Ann NY Acad Sci 1998; 838:29–45.

57. Lin Y, Weissleder R, Tung CH.Novel near-infrared cyanine fluoro-chromes: synthesis, properties, andbioconjugation. Bioconjug Chem2002; 13:605–610.

58. Slakter JS, Yannuzzi LA, Guyer DR, etal. Indocyanine-green angiography.Curr Opin Ophthalmol 1995; 6:25–32.

59. Hastings JW. Chemistries and colorsof bioluminescent reactions: a review.Gene 1996; 173:5–11.

60. Voss SD, Kruskal JB. Gene therapy:a primer for radiologists. Radiograph-ics 1998; 18:1343–1372.

61. Contag CH, Jenkins D, Contag PR, etal. Use of reporter genes for opticalmeasurements of neoplastic disease invivo. Neoplasia 2000; 2:41–52.

62. Park JM, Loening AM, Wu AM, et al.Bioluminescence imaging of angio-genesis with a novel renilla luciferase-VEGF121 fusion reporter gene. Ab-stract. Mol Imaging Biol 2005; 7:149.

63. Fischer F, Dickson EF, Pottier RH, etal. An affordable, portable fluores-cence imaging device for skin lesiondetection using a dual wavelength ap-proach for image contrast enhance-ment and aminolaevulinic acid-in-duced protoporphyrin IX. Part I:design, spectral and spatial character-istics. Lasers Med Sci 2001; 16:199–206.

64. Raabe A, Beck J, Seifert V. Tech-nique and image quality of intraoper-ative indocyanine green angiographyduring aneurysm surgery using sur-gical microscope integrated near-in-frared video technology. ZentralblNeurochir 2005; 66:1–6.

65. Funovics MA, Weissleder R, Mah-mood U. Catheter-based in vivo im-aging of enzyme activity and gene ex-pression: feasibility study in mice.Radiology 2004; 231:659–666.

66. Ntziachristos V, Ripoll J, WeisslederR. Would near-infrared fluorescencesignals propagate through large hu-man organs for clinical studies? OptLett 2002; 27:333–335.

67. Herschman HR. Noninvasive imag-ing of reporter gene expression in liv-ing subjects. Adv Cancer Res 2004;92:29–80.

68. Pomper MG. Molecular imaging: anoverview. Acad Radiol 2001; 8:1141–1153.

69. Wu AM, Yazaki PJ, Tsai S, et al.High-resolution microPET imaging ofcarcinoembryonic antigen-positivexenografts by using a copper-64-la-beled engineered antibody fragment.Proc Natl Acad Sci U S A 2000; 97:8495–8500.


70. Artemov D, Mori N, Ravi R, et al.Magnetic resonance molecular imag-ing of the HER-2/neu receptor. Can-cer Res 2003; 63:2723–2727.

71. Ke S, Wen X, Gurfinkel M, et al.Near-infrared optical imaging of epi-dermal growth factor receptor inbreast cancer xenografts. Cancer Res2003; 63:7870–7875.

72. Kelloff GJ, Hoffman JM, Johnson B, etal. Progress and promise of FDG-PET imaging for cancer patient man-agement and oncologic drug develop-ment. Clin Cancer Res 2005; 11:2785–2808.

73. Tavitian B, Terrazzino S, Kuhnast B,et al. In vivo imaging of oligonucle-otides with positron emission tomog-raphy. Nat Med 1998; 4:467–471.

74. Mahmood U, Weissleder R. Near-infrared optical imaging of proteasesin cancer. Mol Cancer Ther 2003;2:489–496.

75. Weissleder R, Tung CH, Mahmood U,et al. In vivo imaging of tumors withprotease-activated near-infrared fluo-rescent probes. Nat Biotechnol 1999;17:375–378.

76. Bremer C, Tung CH, Weissleder R.Molecular imaging of MMP expres-sion and therapeutic MMP inhibition.Acad Radiol 2002; 9:S314–S315.

77. Law B, Weissleder R, Tung CH.Mechanism-based fluorescent re-porter for protein kinase A detection.Chembiochem 2005; 6:1361–1367.

78. Bullok K, Piwnica-Worms D. Syn-thesis and characterization of a small,membrane-permeant, caspase-activat-able far-red fluorescent peptide forimaging apoptosis. J Med Chem 2005;48:5404–5407.

79. Tung CH, Gerszten RE, Jaffer FA, etal. A novel near-infrared fluores-cence sensor for detection of thrombinactivation in blood. Chembiochem2002; 3:207–211.

80. Louie AY, Huber MM, Ahrens ET, etal. In vivo visualization of gene ex-pression using magnetic resonanceimaging. Nat Biotechnol 2000; 18:321–325.

81. Gambhir SS, Barrio JR, Phelps ME, etal. Imaging adenoviral-directed re-porter gene expression in living ani-mals with positron emission tomogra-phy. Proc Natl Acad Sci U S A 1999;96:2333–2338.

82. Tjuvajev JG, Stockhammer G, Desai R,et al. Imaging the expression oftransfected genes in vivo. Cancer Res1995; 55:6126–6132.

83. Yang M, Baranov E, Jiang P, et al.Whole-body optical imaging of greenfluorescent protein-expressing tu-mors and metastases. Proc Natl AcadSci U S A 2000; 97:1206–1211.

84. Weissleder R, Moore A, Mahmood U,

et al. In vivo magnetic resonanceimaging of transgene expression. NatMed 2000; 6:351–355.

85. MacLaren DC, Gambhir SS, Saty-amurthy N, et al. Repetitive, non-invasive imaging of the dopamine D2receptor as a reporter gene in livinganimals. Gene Ther 1999; 6:785–791.

86. Rogers BE, McLean SF, Kirkman RL,et al. In vivo localization of[(111)In]-DTPA-D-Phe1-octreotide tohuman ovarian tumor xenografts in-duced to express the somatostatin re-ceptor subtype 2 using an adenoviralvector. Clin Cancer Res 1999; 5:383–393.

87. Haberkorn U. Gene therapy withsodium/iodide symporter in hepato-carcinoma. Exp Clin Endocrinol Dia-betes 2001; 109:60–62.

88. Doubrovin M, Ponomarev V, BerestenT, et al. Imaging transcriptional reg-ulation of p53-dependent genes withpositron emission tomography invivo. Proc Natl Acad Sci U S A 2001;98:9300–9305.

89. Gambhir SS, Bauer E, Black ME, et al.A mutant herpes simplex virus type 1thymidine kinase reporter geneshows improved sensitivity for imag-ing reporter gene expression withpositron emission tomography. ProcNatl Acad Sci U S A 2000; 97:2785–2790.

90. Perry MJ, Todryk SM, Dalgleish AG.The role of herpes simplex virus thy-midine kinase in the treatment ofsolid tumours. Expert Opin InvestigDrugs 1999; 8:777–785.

91. Min JJ, Gambhir SS. Gene therapyprogress and prospects: noninvasiveimaging of gene therapy in living sub-jects. Gene Ther 2004; 11:115–125.

92. Blasberg RG, Gelovani J. Molecular-genetic imaging: a nuclear medicine-based perspective. Mol Imaging 2002;1:280–300.

93. Ray P, Bauer E, Iyer M, et al. Moni-toring gene therapy with reportergene imaging. Semin Nucl Med 2001;31:312–320.

94. Libby P. Inflammation in atheroscle-rosis. Nature 2002; 420:868–874.

95. Ross R. Atherosclerosis: an inflam-matory disease. N Engl J Med 1999;340:115–126.

96. Li AC, Glass CK. The macrophagefoam cell as a target for therapeuticintervention. Nat Med 2002; 8:1235–1242.

97. Moreno PR, Falk E, Palacios IF, et al.Macrophage infiltration in acute cor-onary syndromes: implications forplaque rupture. Circulation 1994; 90:775–778.

98. Lendon CL, Davies MJ, Born GV, et al.Atherosclerotic plaque caps are lo-cally weakened when macrophages

density is increased. Atherosclerosis1991; 87:87–90.

99. Kooi ME, Cappendijk VC, CleutjensKB, et al. Accumulation of ultra-small superparamagnetic particles ofiron oxide in human atheroscleroticplaques can be detected by in vivomagnetic resonance imaging. Circula-tion 2003; 107:2453–2458.

100. Trivedi RA, JM UK-I, Graves MJ, et al.In vivo detection of macrophages inhuman carotid atheroma: temporaldependence of ultrasmall superpara-magnetic particles of iron oxide-en-hanced MRI. Stroke 2004; 35:1631–1635.

101. Sukhova GK, Shi GP, Simon DI, et al.Expression of the elastolytic cathep-sins S and K in human atheroma andregulation of their production insmooth muscle cells. J Clin Invest1998; 102:576–583.

102. Galis ZS, Khatri JJ. Matrix metallo-proteinases in vascular remodelingand atherogenesis: the good, the bad,and the ugly. Circulation Res 2002;90:251–262.

103. Chen J, Tung CH, Mahmood U, et al.In vivo imaging of proteolytic activityin atherosclerosis. Circulation 2002;105:2766–2771.

104. Liu J, Sukhova GK, Yang JT, et al.Cathepsin L expression and regula-tion in human abdominal aortic aneu-rysm, atherosclerosis, and vascularcells. Atherosclerosis 2006; 184:302–311.

105. Cheng XW, Kuzuya M, Sasaki T, et al.Increased expression of elastolyticcysteine proteases, cathepsins S andK, in the neointima of balloon-injuredrat carotid arteries. Am J Pathol 2004;164:243–251.

106. Wilson WR, Schwalbe EC, Jones JL, etal. Matrix metalloproteinase 8 (neu-trophil collagenase) in the pathogene-sis of abdominal aortic aneurysm. BrJ Surg 2005; 92:828–833.

107. Libby P, Tanaka H. The molecularbases of restenosis. Prog CardiovascDis 1997; 40:97–106.

108. Johnson LL, Schofield LM, Weber DK,et al. Uptake of 111In-Z2D3 onSPECT imaging in a swine model ofcoronary stent restenosis correlatedwith cell proliferation. J Nucl Med2004; 45:294–299.

109. Lanza GM, Yu X, Winter PM, et al.Targeted antiproliferative drug deliv-ery to vascular smooth muscle cellswith a magnetic resonance imagingnanoparticle contrast agent: implica-tions for rational therapy of resteno-sis. Circulation 2002; 106:2842–2847.

110. Marmur JD, Rossikhina M, Guha A, etal. Tissue factor is rapidly inducedin arterial smooth muscle after bal-


loon injury. J Clin Invest 1993;91:2253–2259.

111. Jaffer FA, Weissleder R. Seeingwithin: molecular imaging of the car-diovascular system. Circulation Res2004; 94:433–445.

112. Jaffer FA, Tung CH, Gerszten RE, etal. In vivo imaging of thrombin ac-tivity in experimental thrombi withthrombin-sensitive near-infrared mo-lecular probe. Arterioscler ThrombVasc Biol 2002; 22:1929–1935.

113. Jaffer FA, Tung CH, WykrzykowskaJJ, et al. Molecular imaging of factorXIIIa activity in thrombosis using anovel, near-infrared fluorescent con-trast agent that covalently links tothrombi. Circulation 2004;110:170–176.

114. Botnar RM, Perez AS, Witte S, et al.In vivo molecular imaging of acuteand subacute thrombosis using a fi-brin-binding magnetic resonance im-aging contrast agent. Circulation2004; 109:2023–2029.

115. Spuentrup E, Buecker A, Katoh M, etal. Molecular magnetic resonanceimaging of coronary thrombosis andpulmonary emboli with a novel fi-brin-targeted contrast agent. Circula-tion 2005; 111:1377–1382.

116. Sirol M, Fuster V, Badimon JJ, et al.Chronic thrombus detection with invivo magnetic resonance imaging anda fibrin-targeted contrast agent. Cir-culation 2005; 112:1594–1600.

117. Knight LC. Radiolabeled peptide li-gands for imaging thrombi and em-boli. Nucl Med Biol 2001; 28:515–526.

118. Lanza GM, Wallace KD, Scott MJ, etal. A novel site-targeted ultrasoniccontrast agent with broad biomedicalapplication. Circulation 1996; 94:3334–3340.

119. Flacke S, Fischer S, Scott MJ, et al.Novel MRI contrast agent for molec-ular imaging of fibrin: implicationsfor detecting vulnerable plaques. Cir-culation 2001; 104:1280–1285.

120. Bates SM, Lister-James J, Julian JA, etal. Imaging characteristics of a noveltechnetium Tc 99m-labeled plateletglycoprotein IIb/IIIa receptor antago-nist in patients With acute deep veinthrombosis or a history of deep veinthrombosis. Arch Intern Med 2003;163:452–456.

121. Takeuchi M, Ogunyankin K, PandianNG, et al. Enhanced visualization ofintravascular and left atrial append-age thrombus with the use of a throm-bus-targeting ultrasonographic con-trast agent (MRX-408A1): In vivoexperimental echocardiographic stud-ies. J Am Soc Echocardiogr 1999; 12:1015–1021.

122. Johansson LO, Bjornerud A, AhlstromHK, et al. A targeted contrast agent

for magnetic resonance imaging ofthrombus: implications of spatial res-olution. J Magn Reson Imaging 2001;13:615–618.

123. Carmeliet P, Jain RK. Angiogenesisin cancer and other diseases. Nature2000; 407:249–257.

124. Meoli DF, Sadeghi MM, KrassilnikovaS, et al. Noninvasive imaging ofmyocardial angiogenesis followingexperimental myocardial infarction.J Clin Invest 2004; 113:1684–1691.

125. Hua J, Dobrucki LW, Sadeghi MM, etal. Noninvasive imaging of angio-genesis with a 99mTc-labeled peptidetargeted at alphavbeta3 integrin aftermurine hindlimb ischemia. Circula-tion 2005; 111:3255–3260.

126. Sipkins DA, Cheresh DA, KazemiMR, et al. Detection of tumor angio-genesis in vivo by alphaVbeta3-tar-geted magnetic resonance imaging.Nat Med 1998; 4:623–626.

127. Winter PM, Morawski AM, CaruthersSD, et al. Molecular imaging of an-giogenesis in early-stage atherosclero-sis with alpha (v) beta3-integrin-tar-geted nanoparticles. Circulation 2003;108:2270-2274.

128. Lin PC. Optical imaging and tumorangiogenesis. J Cell Biochem 2003; 90:484–491.

129. Shweiki D, Itin A, Soffer D, et al.Vascular endothelial growth factor in-duced by hypoxia may mediate hy-poxia-initiated angiogenesis. Nature1992; 359:843–845.

130. Lu E, Wagner WR, Schellenberger U,et al. Targeted in vivo labeling ofreceptors for vascular endothelialgrowth factor: approach to identifica-tion of ischemic tissue. Circulation2003; 108:97–103.

131. Liddell RP, Weiss CR, Hofmann LV.Therapeutic angiogenesis: the nextfrontier for interventional radiology.Tech Vasc Interv Radiol 2004; 7:40–48.

132. Wu JC, Chen IY, Wang Y, et al. Mo-lecular imaging of the kinetics of vas-cular endothelial growth factor geneexpression in ischemic myocardium.Circulation 2004; 110:685–691.

133. Orlic D, Hill JM, Arai AE. Stem cellsfor myocardial regeneration. Circula-tion Res 2002; 91:1092–1102.

134. Sheikh AY, Wu JC. Molecular imag-ing of cardiac stem cell transplanta-tion. Curr Cardiol Rep 2006; 8:147–154.

135. Schellingerhout D, Josephson L.Molecular imaging of cell-based ther-apies. Neuroimaging Clin North Am2004; 14:331–342.

136. Shapiro EM, Sharer K, Skrtic S, et al.In vivo detection of single cells byMRI. Magn Reson Med 2006; 55:242–249.

137. Hill JM, Dick AJ, Raman VK, et al.

Serial cardiac magnetic resonance im-aging of injected mesenchymal stemcells. Circulation 2003; 108:1009–1014.

138. Kim DE, Schellingerhout D, Ishii K, etal. Imaging of stem cell recruitmentto ischemic infarcts in a murinemodel. Stroke 2004; 35:952–957.

139. Hakumaki JM, Savitt JM, Gearhart JD,et al. MRI detection of labeled neu-ral progenitor cells in a mouse modelof Parkinson’s disease. Dev Brain Res2001; 132:A43–A44.

140. Evgenov NV, Medarova Z, Dai G, etal. In vivo imaging of islet trans-plantation. Nat Med 2006; 12:144–148.

141. Walter G, Cordier L, Bloy D, et al.Noninvasive monitoring of gene cor-rection in dystrophic muscle. MagnReson Med 2005; 54:1369–1376.

142. Barbash IM, Leor J, Feinberg MS, et al.Interventional magnetic resonanceimaging for guiding gene and celltransfer in the heart. Heart 2004; 90:87–91.

143. Mahmood U. Can MR imaging beused to track delivery of intravascu-larly administered stem cells? Radiol-ogy 2004; 233:625–626.

144. Welch HG, Schwartz LM, Woloshin S.Are increasing 5-year survival ratesevidence of success against cancer?JAMA 2000; 283:2975–2978.

145. Achilefu S. Lighting up tumors withreceptor-specific optical molecularprobes. Technol Cancer Res Treat2004; 3:393–409.

146. Lucignani G, Losa M, Moresco RM, etal. Differentiation of clinically non-functioning pituitary adenomas frommeningiomas and craniopharyngio-mas by positron emission tomogra-phy with [18F]fluoro-ethyl-spiperone.Eur J Nucl Med 1997; 24:1149–1155.

147. Moon WK, Lin Y, O’Loughlin T, et al.Enhanced tumor detection using a fo-late receptor-targeted near-infraredfluorochrome conjugate. BioconjugChem 2003; 14:539–545.

148. Sidransky D. Emerging molecularmarkers of cancer. Nat Rev Cancer2002; 2:210–9.

149. Bar-Shalom R, Valdivia AY, BlaufoxMD. PET imaging in oncology. Se-min Nucl Med 2000; 30:150–185.

150. Lardinois D, Weder W, Hany TF, et al.Staging of non-small-cell lung cancerwith integrated positron-emission to-mography and computed tomogra-phy. N Engl J Med 2003; 348:2500–2507.

151. Maguire TM, Shering SG, DugganCM, et al. High levels of cathepsin Bpredict poor outcome in patients withbreast cancer. Int J Biol Markers 1998;13:139–144.

152. Ang KK, Berkey BA, Tu X, et al. Im-pact of epidermal growth factor re-ceptor expression on survival and


pattern of relapse in patients with ad-vanced head and neck carcinoma.Cancer Res 2002; 62:7350–7356.

153. Sato H, Takino T, Miyamori H.Roles of membrane-type matrix met-alloproteinase-1 in tumor invasionand metastasis. Cancer Sci 2005; 96:212–217.

154. Bremer C, Tung CH, Bogdanov Aet al.Imaging of differential protease ex-pression in breast cancers for detec-tion of aggressive tumor phenotypes.Radiology 2002; 222:814–818.

155. Jonson SD, Welch MJ. PET imagingof breast cancer with fluorine-18 ra-diolabeled estrogens and progestins.Q J Nucl Med 1998; 42:8–17.

156. Prasad SR, Jhaveri KS, Saini S, et al.CT tumor measurement for therapeu-

tic response assessment: comparisonof unidimensional, bidimensional,and volumetric techniques initial ob-servations. Radiology 2002; 225:416–419.

157. Rehman S, Jayson GC. Molecularimaging of antiangiogenic agents. On-cologist 2005; 10:92–103.

158. Bremer C, Tung CH, Weissleder R.In vivo molecular target assessmentof matrix metalloproteinase inhibi-tion. Nat Med 2001; 7:743–748.

159. Avril NE, Weber WA. Monitoringresponse to treatment in patients uti-lizing PET. Radiol Clin North Am2005; 43:189–204.

160. Demetri GD, von Mehren M, BlankeCD, et al. Efficacy and safety of ima-tinib mesylate in advanced gastroin-

testinal stromal tumors. N Engl J Med2002; 347:472–480.

161. Hakumaki JM, Liimatainen T. Mo-lecular imaging of apoptosis in can-cer. Eur J Radiol 2005; 56:143–153.

162. Partain CL, Chan HP, Gelovani JG, etal. Biomedical Imaging ResearchOpportunities Workshop II: reportand recommendations. Radiology2005; 236:389–403.

163. Elgort DR, Duerk JL. A review oftechnical advances in interventionalmagnetic resonance imaging. AcadRadiol 2005; 12:1089–1099.

164. Lee JT, White RA. Basics of intravas-cular ultrasound: an essential tool forthe endovascular surgeon. SeminVasc Surg 2004; 17:110–118.


Documents

Molecular Imaging: A Primer for Interventionalists and Imagers