uptake, inverse correlation, HMPAO uptake, hypoxic, Michael Hoskinson, IAZA, brain tissue, clinical centers, oxygen utilization, squamous cell carcinoma, Histogram, perfusion, avidity, Imaging agent, hypoxic tumor, tissue perfusion, Tumor Hypoxia, Matthew B. Parliament, Fox Chase Cancer Center, Edmonton, Alberta, Canada, Rebecca Sief Government Hospita4 Safed, University of Alberta, Leonard I. Wiebe, Lon E. Golberg, David Groshar, Raul C. Urtasun, John R. Mercer, Alexander J.B. McEwan, Rezaul H. Mannan, J. Donald Chapman Cross Cancer Institute
Imaging Tumor Hypoxia and Tumor PerfusionDavid G
roshar, Alexander J.
B. McEwan, Matthew B. Parliament, Raul C. Urtasun, Lon E. Golberg, Michael Hoskinson, John R
. Mercer, Rezaul H. Mannan, Leonard I. Wiebe and J. Donald Chapman Cross Cancer Institute, Edmonton, Alberta, Canada
; Rebecca Sief Government Hospita4 Safed, Israel; University of Alberta
, Edmonton, Alberta
, Canada; and Fox Chase Cancer Center
, Philo4e4hia, Pennsylvania
Recently, an analogue of misomdazole labeled with в~8F
Tumor perfusion and oxygenation status have been suggested was shown to selectively bind to hypoxic cells within tu as factorswhichmayinfluencetreatmenot utcomein cancer mors and myocardium(11,12). However, imagingwith this patientsN. uclearmedicineassaysof tumorperfusion[@Fc agent is limited to those centers with positron emission[email protected]
@1enoaamminee(HMPAO)aJndtumorhypoxia tomography (PET) systems, which precludes its routine
(IAZA)Jhave recently been de use in most clinical centers. Our group has reported the
veloped and descnbed. We report on measurementsof perfu synthesis of a novel misomdazole analogue, iodoazomycin
sion and oxygenation status of 27 tumors in 22 patients using arabinoside(IAZA), which can be labeled with в[email protected]
these probes. An inverse correlation between tumor uptake of HMPAaOndIAZAwasmeasur(epd< 0.05w),@sieverpeer
The bindingof this compound to the acid insoluble fraction of EMT-6 cells has been shown to be strongly dependent,
fusion deficit usually associatedwith an increaseduptake of the at 37В°Co,n the degree of hypoxia (13). Iodine-123 can be
hypoxic marker.This trend was observedforlimited stage small cell lung carcinoma, squamous-cellcarcinoma of the head and neck,soft-tissuesarcomab, rainmetastasesfromsmall-cellung
detected by both planar scintigraphy and single-photon emission
computer tomography (SPECT) with nuclear
carcinoma and adenocarcinomaof the prostate as a group, but medicine equipment available in most imaging depart
not for glioblastoma muftiforme.Whereas each Imaging agent ments. Furthermore,the 13-hrhalf-lifeof в[email protected]
can yield informationabout the physiologicalstatusoftumor and the 100-mm of в~8aFllows for delayed imaging to be per
normaltissue,the [email protected]
formed. Animal and preliminary human data have mdi
couldbe importantin cancertherapy.
cated that optimum tumor-to-background ratios are ob
J Nuci Med 1993; 34:885-888
tained after two to three pharmacologic half-lives during which time clearance of unbound compound from oxic
Regional tissue perfusion can also be an indicator of
t hasbeenrecognizedfor over30yr thatthepresenceof hypoxic cells within some tumors may play a role in their resistance to radiotherapyand chemotherapy (1,2). Several groupshave attemptedto developusefultechniquesfor measuring tissue hypoxia, although no simple method has yet been established. Measurements by invasive oxygen electrodes (3,4), by nuclear magnetic resonance spectros copy of phosphorus signals associated with cellular energy metabolism (5,6) and by uptake of radiolabeledbioreduc tive drugs (e.g., misonidazole) into hypoxic tissues (7,8) have been proposed. This latter technique involves the systemic administration of a radiolabeled bioreductive drug which can be reduced within cells to reactive inter mediatesthat bind to cellular moleculesat rates inversely
tissueviability, and this property is commonly usedin the evaluation of ischemic heart disease
(15). Several groups have shown @9'c-hexamethylpropylenaminoexime (HM PAO) to be an effective indicatorof tissue and tumor per fusion in animals(16) and in humans(17). Studieswith the Dunning rat prostate carcinoma (P3327-AT) have demon strated an inverse relationship between HMPAO and IAZA uptake in untreatedtumorsand in those treatedwith photodynamictherapy(PDT) (18).To assistin understand ing THE RELATIONSHIP
between tumor perfusion and tumor hypoxia in man, we made a qualitative comparison of HM PAO and IAZA distribution in patients with a variety of tumors at differentsites in which significanthypoxic frac tions were expected.
proportional to intracellular oxygen concentration. Car bon-14 and tritium-labeled radiosensitizers have been shown to selectively bind to hypoxic regions of solid tu morsin bothrodentsandhumans(8в"10).
METhODS Patient population
Twenty-two patients (16 male and 6 female) with primary small-cell lung cancer
(SCLC, n = 4), glioblastoma multiforme
(n = 4), head and neck squamous-cell carcinoma (n = 6), brain
metastases from SCLC (n = 2), soft-tissue sarcoma (n = 5) and
CroFRsoesCrccaeonirvcreeeJdrsIunpls2oti0ntu,d1tee91,n91cr25er;6epo0vrUiInsntiIsoDva:nerc.rAscJietAy.pBvte.eMF.d,cEeEdbmw2. 2ao,nnD1t,9oe9nГA,3o§[email protected]
= 1) were consent.
entered into The protocol
the study after had previously
been approvedby the institutionalethics committee. The anatom
TABLE 1 Disthbutionof LAZAand HMPAOUptakeAccordingto TumorPathdogyand Numberof Sites uptakeDecreasedEquaiIncreNaos.eodf DecreasedEqualIncreasuepdtaSkCeLHCMPAO tumor sftesIAZA
ical locationof the tumorswas definedby CT imaging,radiogra 30в"9m0 m after intravenous injection of 20 mCi (740 MBq) of
Two independent observers compared the uptake pattern of each of the two radiopharmaceuticalsT. he degreeof uptake was
Iinagmg was performed with a General Electric
400 AC gamma graded by the following criteria:
camera system (General Electric, Milwaukee, WI
) interfaced to a
Picker @c:5s12 computer system
(Picker International,Bedford, 1: Increased (tumor uptake higher than background).
2: Nonavid(tumoruptakeequalto thatof background).
3: Decreased (tumor uptake less than background). Back
groundwas definedas uptakeincontralateranlormaltissue.
Unlabeled IAZA (1-(5'-iodo-5'deoxy-$-D-arabinofuranosyl)-2nitroimidazole)was prepared as described elsewhere (14). Iodine The comparativedatawere tabulatedand comparedusing the 123, 1,110 MBq (30 mCi), as Na! (Nordion International Ltd., chi-square test.
VancouverC, anadai)n 0.1ml solutionof 0.1M NaOHcontained
in a 3-mi V-vial was evaporated to dryness at 40В°Cwith a stream RESULTS
of nitrogen gas. The dry residue was treated with 1.3 mg of IAZA
Twenty-seven tumors were identified as known lesions
and 3.1 mg of pivalic acid as a solution in 100 @olf aqueous in the 22 patients. Each tumorwas analyzed independently
methanoalndanalyzedby high-pressulirqeuidchromatographyand tumor uptake of both imaging agents was scored and
(HPLC). The chemical purity of the crude product was 96% with a 4%impUrityidentifiedas 1-(@-D-arabinofuranosyl)-2-nitroimida
zole, the hydrolysis product of IAZA. The radiochemicalpurity tumors showed increased uptake, 13 tumors (48%) showed
was measured as 92.6% with a 4% impurity identified as
equal uptake and 1 tumor (4%)showed decreased uptake.
iodide. The sample was purified by HPLC and the solvent re For HMPAO images, 13 tumors (48%)showed decreased
movedin vacuo. The patientdosewas preparedby dissolvingthe uptake, 7 tumors (26%)showed equal uptake and 7 tumors
в~231-IAZAin 5.7 ml of sterile saline containing 10 mg of unlabeled (26%) showed increased uptake. Tumors displaying de
IAZA and filtering the solution into a sterile multidose vial. The creased HMPAO uptake tended to be IAZA-avid (Figs. 1
purifiedsamplehad nodetectablechemicalimpuritiesanda >99% and 2). In two of the tumors displayingdecreased HMPAO
radiochemicalpuritywhen analyzed by HPLC.
uptake, the SPEC1' images revealed a вoedoughnutвp·attern
Becausethiswasa novelcompoundandbecausethisclassof with enhanced perfusion at the tumorperiphery. In tumors
drughas previously shown neurotoxicity, в[email protected]
[3A.9-9.3 mCi displaying this вoedoughnuptaвtte· rn of HMPAO uptake,
(145в"34M3Bq)] was given by slow intravenous mm. In the later phases of the study, after no
observed, the radiopharmaceutical was given by bolus injection. corresponded to the photopenic center of the perfusion
Lugol'siodinewasadministereodrallyfor3 dayspriorto imaging image.
to block thyroid uptake of free radioiodine. As previously de
Eight of 13 tumors (62%)with decreased HMPAO up
scribed (14), anterior and posterior planar static images
were take showed increased IAZA uptake, whereas only 5 of 14
obtained of the area of interest at 1 and 16в"2h4r postinjection. tumors (36%) with normal or increased HMPAO uptake
SPECF imagingof the area of interest was performedbetween showed increasedIAZA uptake (Fig. 3). Consideringtu
16в"2h4r postinfusion. We have reported the results of IAZA mors from all sites, these proportions are not significantly
imaging, but not data from the PAO study, from four patients in different (x2 1.73, ns). We observed, however, that none
of the four glioblastomas displayed IAZA avidity, despite
all having striking perfusion defects. Since these tumors
HMPAO images of tumor perfusion were acquired within may represent a unique category, we examined the pat
1 wk of acquiring the IAZA image. Anterior and posterior planar ternsof HMPAO andIAZA uptakeexcludingglioblastoma
images and SPEC!' images of the area of interest were obtained multiforme. Eight of nine tumors (89%) with decreased
of Nudear MedicineвVўol. 34 вNўo. 6 вJўune 1993
FiGURE1. AxialSPECTimagesof a patientwithSCLCbrain metastasis.The HMPAOstudy (top row) showsdecreasedperfu sion at the tumor site (arrow)with increasedIAZA uptake(bottom row).
HMPAO uptake showed increased IAZA uptake, whereas only 5 out of 14 tumors (36%) with normal or increased HMPAO uptake showed increased IAZA uptake ([email protected]
= 6.03, p < 0.05). DISCUSSION The patterns of IAZA uptake reported in this study are concordant with human tumor hypoxic fraction data mea sured with 3H-misonidazole(10). In that study, 12 of 27 tumors (44%) showed evidence of significant hypoxic frac tions by levels of bound 3H-misonidazole significantly higher than those of background. The distribution of tumor sites investigated in the 3H-misomdazole and this study were different. Nevertheless, in both studies a large pro portionof the SCLCs investigated showed avidity for these hypoxic markers. The inverse correlation between human tumor perfusion and hypoxic marker uptake is consistent with our study with animal tumors (18). In that study, rat prostate carcinomas(R3327-AT)were investigatedbefore and after photodynamic therapy, a procedure known to induce perfusion shutdown in treated tissues (6,19). These studies demonstrated that the extent of perfusion deficit and tumor cell
hypoxia in PDT-treated tumors was larger than that found in nontreatedtumors. In Figure 1, there is IAZA avidity at the site of a cerebral metastasis in a patient with SCLC, associated with a dis cordant HMPAO image. It is unlikely that the IAZA up take reflects breakdown of the blood-brain barrier. The negligible IAZA uptake found in gliomas, and the absence
SO FIGURE 2. CoronalSPECTimagesofapatientwithabulkyneck node metastasisfrom a squamous-caicl arcinomaof unknownpn marysite.The HMPAOstudy(toprow)showsdecreasedperfusion inthelymphnodemetastasi(sarroww) ithcorrespondinAglA avid ity (bottomrow).Iodine-123uptakeinthe salivaryglandsis presum ablydue to partialdelodination(14). of HMPAO uptake, make it unlikely that such a nonspe cific uptake mechanism, normally demonstrated by gluco heptonate scintigraphy, is responsible for the patterns of uptake reported in this paper. Whereas an inverse correlation between tumor perfu sion and hypoxic marker avidity was found in this study, an absolute correspondence between perfusion shutdown and tumor hypoxic fraction was not expected.The extent of perfusion reduction in a specific tumor and/or tissue required to produce significant zones of cells at oxygen levels
low enough to promote the reduction and binding of thesemarkersis probably tissue-specific.Endogenous1ev els of nitroreductases, tissue rates of oxygen utilization, the вoequalityoвf ti·ssuevesselsand other biologicalparam eters will impact on such correlations. In four patients with glioblastoma multiforme, we noted that the tumors showed significantly reduced perfusion relative to surrounding nor mal brain tissue, yet displayed no significant avidity for IAZA. If this result is confirmed by additional studies, this
Incrsasвўd Non-Avid [email protected]
0. Dicreued 0 Decreased Non-Avid Increased
HumanTumorHypoxia Grosharet al.
IAZA HMPAO FiGURE 3. Histogram shows the corn paratWedegreeof IAZAand HMPAOup take observedin 27 tumorsites. 887
could suggest that malignant glioma cells are extremelysquamous cell carcinoma
metastases and its relationship in [email protected]
sensitive to ischemic cell death
with viable hypoxic cells being a rarity in these tumors. Additional information isradiation therapy
. Intl Radiat Oncol Bid Phys 1988;14:831-838. 5. RofstaEdK,DeMutPh,FentoBnM,CeciderTS,utherlaRndM.P-31NMR spectroscopy
and W'В°2ciyospectrophotomeuy in prediction of tumor ra
required before realistic modeling of tumor/tissue blood
dioresistance caused byhypoxia. JntlRadiatOncolBidPhys 1989;16:919-
flow and viable hypoxic fraction as measured by IAZA avidity can be performed. The 16-hrimagingtime for IAZA was utilized because of
923. 6. ChapmaJnD, McPheeMS. WalzN, et al. Nuclearmagnetircesonance spectroscopy and sensitizer-adduCT measurements
of photodynamic thera py-incluced ischemia in solid tumors. INcA Can Inst 1991;83:1650-1659.
a late afternoon injection. Studies are planned to evaluate the 8-hr to 16-hr postinjection imaging period. Reduced regional HMPAO uptake is unable to distin
7. chapman3D. Hypoxicsensitizersi:mplicationsfor radiationtherapy.N ESIgIJMe4 1979;301:1429-1432. 8. ChapmaJnD,FrankoAl, SharpliJn.A markefrorhypoxicellsintumors with potential clinical applicability. BrI Cancer 1981;43:546-550.
guish between necrotic or hypoxic tissue and is likely to underestimate the extent of viable tumor. The patterns of IAZA uptake shown in Figure 1 suggest that this marker can discriminateviable hypoxic tumor, a parameterwhich
9. UrtasunRC,ChapmaJnD, RaleighJA, FrankoAJ,KochCJ.Bindingof в~H-misonidazotolesolidhumantumorsas a measureoftumor hypoxia.hit IRadiat OncoIBioIPh)@c 1986;12:1263-1267. 10. ChapmaJnD,UrtasunRC,FrankoAi,RaleigJhA,MeekerBE, McKinnon 5k The measurement of oxygenation status of individual tumors. In: вoePre
could have significant prognostic potential. The inverse correlation of uptake of the two radiopharmaceuticals sug gests a complex relationshipbetween impairedtissue per
diction of Response in Radiation Therapy. The Physical and Biological Basis
в·.Am Assoc Phys Med [email protected]
No. 7 (Part 1); 1989:49в"60. 11. KohW-i, RaseyJS,EvansML, eta!.Imaginogfhypoxiainhumantumors
fusion and the presence of viable hypoxic cells. The data also suggest that both HMPAO and IAZA measurements of tumor physiologymay provide additionalinformation to
with [F-l8lfluommisonidazole. IntlRadiat OncolBiolPhys 1992;22:199- 212. 12. Shelton ME, Dence @SH, wangD-R, Welch MJ, Bergmann SR. Myocardial kinetics of fluorine-18-misonidazole: a marker of hypoxic myocardiuin. I
NuciMed 1989;30:351в"358. 13.MannanRH, Somaya\jвi ~VL,eeJ, Mci-caJrR,ChapmaJnD, WiebeU.
Radioiodinated 1-(5-iodo-5-decxy-$-D-arabinofuranosyl)-2-nitroimidazole (iodoazomycin arabinoside: IAZA): a novel marker of tissue hypoxia. I
ThisresearchwassupportedbygrantRI-14(3f)romtheAlberta Cancer Board research initiatives
Programme and by Amersham International plc, which provided the HMPAO. The authors thank Mrs. C. Loeffler for manuscript preparation. Dr. Groshar is a
NuciMed 1991;32:1764в"1770. 14. ParliamenMtB, ChapmaJnD,UrtasunRC,et al. Non-invasivaessessment ofhuman tumour hypoxia with в[email protected]
: preliniinaiy report of a CLINICAL STUDY
. BrI Cancer 199265:90-95. 15. Kloner RA, PrzyklenkK. Hibernationaland stunningof the myocardium
recipient of the American physician
[Editorial]. NEngilMed 1991;325:1877-1879. 16.SugaK, HonmaY, UchisakoH, et al. Assessmeonft @вoeTc-HMPAO
tumour scintigraphy using VX-2 tumours implanted in a lower limb muscle of rabbits. Nuci Med Commun 1991;12:611в"619.
1. GrayUI, CongeAr D, EbertM, HornseyS,ScottOCA.Concentrationf 17. PowellNP, MCCreadyVR, Tait D, et al. Technetium-99HmMPAOand
oxygendissolvedin tissuesat timeofirradiationas a factorin radiotherapy.
SPECrin the assessmentofblood flow
in human lung tumours. BrI Cancer
2. MoulderJE, ROckWeSll. Tumorhypoxia:its impacton cancertherapy. 18.MooreRB,ChapmaInD, MercerJR,etaLMeasuremeonftphotodynamic
therapy induced hypoxia in the Dunningprostatic tumor model by
3. CaterDB, SilverIA. Quantitativemeasuremenotsf oxygenationin normal
lodoazomycin arahinOside. INuci Med 199234:405-413.
tissues and in the tumors of patients before and after radiotherapy.Acta 19. MooreRB,Chapma3nD,MokrzanowsAkiD, ArnfieldMR, McPheeMS.
McEwan MB. Non-invasive monitoring of photodynamic therapy with
4. GatenbyRA, KesslerHB, RosenbiumJS,et al. Oxygendistributioin
The Journal of Nudear Medicine Vol. 34 No. 6 June 1993
MB Parliament, RC Urtasun, LE Golberg