Topographic metabolic map of head and neck squamous cell carcinoma

using 18F-FDG PET and CT image fusion

Denise Takehana Dos Santos, DDS,a Eduardo Nobrega Pereira Lima, MD,b Rubens Chojniak, MD, PhD,c

and Marcelo Gusmao Paraiso Cavalcanti, DDS, PhD,d Sao Paulo, BrazilUNIVERSITY OF SAO PAULO COLLEGE OF DENTISTRY AND CANCER HOSPITAL

Objective. To propose a methodologic approach to evaluate head and neck tumors in order to identify and distinguish areas

of higher metabolic activity inside the lesion.

Study design. The sample consisted of 17 patients with squamous cell carcinoma of the head and neck. Images were

simultaneously acquired using a nondedicated PET-CT device and an independent workstation with ENTEGRA 2 NT software to

generate the image fusion between PET and CT. Sites of higher metabolic activity inside the tumor were classified as centric or

eccentric according to their relative location to the lesion center.

Results. Seventy-seven percent (n = 13) of the patients presented the site of higher metabolic activity at the center of lesion,

and in 23% (n = 4) the uptake of the tracer was increased at the periphery of the lesion.

Conclusions. This technique gave a realistic view of the functional metabolism, locating the anatomical tumor area and

helping in future treatment planning.

(Oral Surg Oral Med Oral Pathol Oral Radiol Endod 2005;100:619-25)

619

Prediction and detection of cancer response to radio-therapy is a challenge in oncology. Most commonly, thetreatment planning is based on anatomical aspects suchas size and volume of the lesions. Tumor size andvolume have a major impact on both treatment deci-sion process and patient outcome.1-2 The detection ofa cancer with a small size or volume, that can beeffectively treated, generally results in a more favorableoutcome.3 The precision radiation therapy techniquesrequire accurate tumor identification and delineation.Inaccurate assessment of the target can lead to failureto meet the treatment goals, with a higher probabilityof tumor recurrence and an additional, unnecessary,radiation burden.4

Technologic advances have already led to funda-mental improvements used to visualize tumors and toassess their metabolic and physical effects on the body.5

Supported by grants from CNPq (Dr Marcelo Cavalcanti, #309331/

2003-0), Brasilia, Brazil, and funds from CAPES (Dr Denise Santos,

PhD fellowship), Brasilia, Brazil.aPhD student, Oral and Maxillofacial Radiology, University of Sao

Paulo College of Dentistry.bDirector, Nuclear Medicine Division, Department of Imaging,

Cancer Hospital.cDirector, Department of Imaging, Cancer Hospital.dProfessor, Oral and Maxillofacial Radiology, University of Sao

Paulo College of Dentistry. Adjunct Assistant Professor, Department

of Radiology, University of Iowa College of Medicine, Iowa City.

Received for publication Dec 16, 2004; returned for revision Feb 22,

2005; accepted for publication Feb 25, 2005.

1079-2104/$ - see front matter

� 2005 Mosby, Inc. All rights reserved.

doi:10.1016/j.tripleo.2005.02.076

A number of new developments in diagnostic imaginghave potential impact for the detection, diagnosis,staging, and follow-up of cancer patients. Some of thediagnostic imaging developments include computerizedtomography (CT), magnetic resonance imaging (MRI),Doppler and endoluminal ultrasound techniques, newcontrast and radiopharmaceutical agents, positronemission tomography (PET), single-photon-emissiontomography (SPECT), and MR spectroscopy or spec-troscopic imaging.3,5-8

PETwith [18F]2-fluoro-2-deoxy-D-glucose (18F-FDG)provides images that quantitatively portray glucoseconsumption and therefore lesion metabolic activity,though it presents low anatomical resolution.3-5,9

Recently, a new diagnostic approach that combinessimultaneously acquired x-ray CT and18F-FDG PEThas been proposed. It can bring an anatomic localiza-tion to metabolic lesion activity, improving diagnosticcapabilities.6-9

The aim of this work therefore was to propose amethodologic approach to evaluate head and neck squa-mous cell carcinoma (SCCA) in order to identify anddistinguish areas of higher metabolic activity insidethe tumor by combining the functional metabolic andmorphologic data simultaneously acquired in a non-dedicated PET-CT device.

MATERIAL AND METHODSThe study population consisted of 17 patients with

SCCA of the head and neck (confirmed by histopath-ologic examination), staged clinically according to the1988 American Joint Committee on Cancer10 criteria in

OOOOE

620 Takehana Dos Santos et al November 2005

T2 and T3, with and without lymph node involvement(N0 and N1), and with no metastasis (M0) in order tostandardize the sample. These patients were submittedto a nondedicated 18F-FDG PET imaging using a systemwith low dose CT and positron emission coincidenceacquisition capabilities (Millennium VG Hawkeye; GEMedical Systems, Milwaukee, Wisc). Initially the x-rayimages were made with maximum power of 140 kVand2.5 mA, and 80 cross sectional slices were obtained; theslice thickness is fixed at 10 mm. Each slice acquisition,which is termed a half scan, takes approximately 14seconds.

Positron emission coincidence imaging was initiated60 minutes after intravenous injection of 8-10 mCi(296-370 Becquerel) of 18F-FDG. The positron emitter18F-FDG was employed in our study because when it isadministered intravenously it is trapped as 2-deoxyglu-cose-6-phosphate within cells. The unstable proton-richfluorine nucleus rids itself of the excessive charge withemission of a positively charged electron, becoming amore stable nucleus. These positrons travel a fewmillimeters before interacting with an electron, in anannihilation reaction, which results in the release of 2photons at approximately 180 degrees to one another.Simultaneous detection of the photons by opposeddetectors is used to build up a 3-dimensional represen-tation of tissues. The data are displayed in a mannersimilar to conventional sectional imaging (matrix of 2563 256). The images produced reflect glucose uptakewithin various tissue, including malignant cells due tothe increased glucose metabolism of the tumor. Theemission data were acquired for 10 rotations of 360degrees during 30 minutes, and the imaging data werethen transferred by network DICOM 3.0 to anENTEGRA 2 NT workstation (GE Medical Systems)to generate groups of individual images (metabolic andanatomic data) and image fusion (CT 1 PET). The CTdata were used for x-rayebased attenuation correctionof the PET emission data. The pixel value in HU unitswas transformed in linear attenuation rate in units ofcm�1 to 511 keV by means of a bilinear function definedwith 3 coordinator points: �1.000 HU = 0 cm�1;0 HU = 0.093 cm�1; and 1.326 HU = 0.172 cm�1.

In those patients with anomalous concentrations of18F-FDG, the lesion was depicted on 3 planes (axial,coronal, and sagittal) in CT, PET, and the image fusionat the computer screen (Fig 1). The crosshair was used tofacilitate the analysis and localization of the site ofhigher metabolic activity. The findings were defined aspositive in the presence of a well defined focal area ofincreased uptake in regions unrelated to the normalbiodistribution of the tracer on visual inspection. Twoexaminers interpreted the images in different sessions inorder to get an agreement. Subsequently, the sites of

higher metabolic activity inside the tumor were identi-fied and classified as centric (Fig 2, A) or eccentric (Fig2, B), according to their relative location to the center ofthe lesion using a progressive uptake scale in order tocorrelate the site precisely with the intensity of activityof the tumor. The progressive uptake scale was localizedon the righthand screen. In addition, we established asequence to perform this analysis, observing the 3planes (axial, coronal, and sagittal) in order to get thesites separately. Then, we combined this informationdetermining if the site of higher uptake is centric oreccentric by its predominance on the 3 planes (eg, ifcentric in both axial and sagittal views, the site wasconsidered centric). On the image fusion we identifiedanatomic localization of the lesions combined withmetabolic activity, which permitted us to distinguishthe low and high activity of the recurrence or anothermetastasis.

RESULTSTable I summarizes the clinical and pathologic data

of 17 patients with primary SCCA of head and neck(nasal fossa, orbit region, base of tongue, floor mouth,hypopharynx, and larynx) in pretreatment. We found3 patients staged in T2N0M0, 3 in T2N1M0, 6 inT3N0M0, and 5 in T3N1M0.

From the analysis of groups of individual images(metabolic and anatomic data) and image fusion(CT 1 PET), we found anomalous concentrations of18F-FDG in all patients. Observing the axial, coronal,and sagittal views in association (Fig 1) and separately(Figs 3 and 4), we found 77% of the patients with the siteof higher activity at the center of lesion (Figs 1 and 2, A)according to their relative geometric location. Thosepatients were classified in T2N0M0 (3 patients),T2N1M0 (3 patients), T3N0M0 (6 patients), andT3N1M0 (1 patient) (Table I). The grades of tumordifferentiation were I (4 patients), II (7 patients), and III(1 patient). We detected in 23.00% of the patients adifferent behavior, with the tracer increased eccentri-cally to the lesion (Figs 2, B and 4). All those patientswere staged in T3N1M0 (Table I). Regarding thedifferentiation, we found 3 patients in grade III and1 in grade II.

Figure 3 illustrates a case where PET-CT addedvaluable information that permitted an accurate plan-ning and follow-up treatment of a patient with SCCAin the posterior-inferior portion of the nasal fossa. Thetumor was not clearly demonstrated on the CT images(Figs 3), especially because of the absence of intrave-nous contrast. However, the metabolic and the imagefusion allowed better visualization of the lesion. Theabsence of activity at the posterior-inferior portion ofthe nasal fossa was observed and the presence of another

OOOOE

Volume 100, Number 5 Takehana Dos Santos et al 621

Fig 1. SCCA of the floor of mouth is depicted in 3 planes (axial, coronal, and sagittal) and in 2 modalities of imaging (CT andmetabolic) as well as the image fusion at the computer screen, demonstrating anomalous concentration of 18F-FDG. The crosshairwas used to facilitate the analysis and localization of the site of higher metabolic activity. The site of higher activity is centric whencorrelated to the progressive uptake scale (right side).

lesion was well visualized due to the high uptake of theFDG, near to the oropharynx. Functional and morpho-logic data were combined, making possible to identifythe metastases of the primary lesion.

DISCUSSIONA clear understanding of the decision therapy process

must accompany knowledge of both the disease and theimaging systems. Furthermore, the tumor therapy and

prognosis depends on factors such as tumor grade,size, site, vascularity, lymphatic drainage, host immuneresponse, age, gender, nationality, and performancestatus.9 Computer-combined CT and PET with 18F-FDGhas become the reference imaging modality for treat-ment planning owing to its acceptable costs, wideavailability, absence of geometric distortion, and inher-ent ability to provide information on tissue density,which is useful for dose calculation.5 The limitations of

OOOOE

622 Takehana Dos Santos et al November 2005

Fig 2. Metabolic images of 2 lesions, showing areas of higher activity inside the tumor: centric (A) and eccentric (B) in aprogressive uptake scale (blue to white).

detection variables relate to equipment (spatial andcontrast resolution), patient and target organ consid-erations, and interobserver variations. In our opinion,the improved and more sophisticated imaging techni-ques and procedures have advanced cancer stagingthrough more accurate and specific characterization.

Another important factor to be considered is themetabolism of the lesion. The metabolic lesion behaviormight not be homogeneous in a volume distribution,and more active areas might not to be found in thegeometrical center but in some other region of thegrowing lesion.1,2,9 The rate and intensity of accumu-lation of radiolabeled metabolites or drugs are analyzedto evaluate tumor biology and therapeutic response or tomake predictions. In the present study we choose thetracer 18F- FDG, which was demonstrated in recentpublications to have a 96% sensitivity for the detectionof primary tumors.4,5 Also, for the detection of lymphnode metastasis, the sensitivity and specificity are,respectively, 88% and 93%5 and, for the differentiationof recurrent tumors from posttreatment changes, 94%and 82%.6 We found anomalous concentration of the18F- FDG in all the patients in areas unrelated to thenormal biodistribution of the tracer. It was correlated totheir histopathologic diagnosis (ie, SCCA), confirmingthe efficiency of this tracer in the detection of malignanttumors. Several processes determine FDG uptake in

tumor cells. Of major importance is the integrity of thevascular network that is necessary for supplyingnutrients to the cell. With an intact vascular supply,FDG enters the tumor cells by using the same facilitatedtransport mechanism as glucose (via cell surfaceproteins).11 Also, in vitro studies have shown thatFDG uptake was determined by the number of viabletumor cells within a lesion (tumor cell density).Nontumoral tissue, such as necrotic and fibrotic tissue,may reduce tracer uptake.12 The bottom line is thatFDG accumulation within a tumor is likely relatedto a complex interaction between the cellular energydemand and the tumoral microenvironment.13 In ourstudy we demonstrated the application of FDG PET tothe management of head and neck SCCA tumors inorder to identify and to distinguish areas of highermetabolic activity inside the lesion.

Hemingway et al12 stated that the presence of arecurrent tumor can be obscured by dense scar tissueresulting from previous surgery or radiotherapy, andPET imaging has proved useful in demonstrating thepresence of an otherwise masked disease. In agreement,we could emphasize that one of the advantages ofPET-CT over other imaging modalities is that it doesnot depend upon such changes and consequently maylead to earlier diagnosis of tumors. In addition, it candemarcate the tumor margin at sites where it was

OOOOE

Volume 100, Number 5 Takehana Dos Santos et al 623

Table I. Summary of 17 patients with SCCA head and neck tumors

Patient Base pathology/TNM stage PET/CT

High uptake

inside the tumor

1 SCCA glotte T2N0M0 grade II Anomalous concentration of FDG in the infra

glottic region

Centric

2 SCCA nasal fossa T3N0M0 grade III Multiples areas of anomalous concentrations of

FDG in the axial and apendicular skeleton and

in the projection of the left nasal fossa

Centric

3 SCCA nasal fossa T3N0M0 grade II Areas of anomalous concentration of FDG in the

posterior right nasal fossa (moderate grade). High

grade in the right cervical projection

Centric

4 SCCA orbit T3N0M0 grade I High grade of anomalous concentration of FDG in

the orbit region (right projection)

Centric

5 SCCA tongue T3N1M0 grade II High grade anomalous concentration of FDG in the

base of tongue

Eccentric

6 SCCA tongue T3N1M0 grade III Anomalous concentration of the projection in

posterior part of the base of tongue

Eccentric

7 SCCA larynx T3N1M0 grade II Anomalous concentration of the FDG in the lateral

left portion of laringopharynx

Centric

8 SCCA floor mouth T2N1M0 grade II Two areas of anomalous concentration of the FDG in

the projection left parafaringea

Centric

9 SCCA pyriform sinuses T2N0M0 grade I High grade of anomalous concentration of FDG in the

cervical projection D

Centric

10 SCCA glotte T2N1M0 grade II High anomalous concentration of FDG in the projection

of oropharynx and anterior to globe

Centric

11 SCCA mandible T3N1M0 grade III Anomalous concentration of FDG in the mandible Eccentric

12 SCCA larynx T3N1M0 grade III Anomalous concentration of FDG in the inferior region

of nasal fossa E

Eccentric

13 SCCA pyriform sinuses T3N0M0 grade I High grade of anomalous concentration of FDG in the

cervical projection D

Centric

14 SCCA nasal fossa T2N1M0 grade III Anomalous concentration of FDG in the projection of

nasal fossa

Centric

15 SCCA larynx T3N0M0 grade II Anomalous concentration of FDG in the projection of

larynx and hypopharynx

Centric

16 SCCA larynx T3N0M0 grade II Anomalous concentration of FDG in the projection of

larynx and hypopharynx

Centric

17 SCCA tongue T2N0M0 grade I Anomalous concentration of FDG in the lateral portion

of the base of tongue

Centric

difficult to evaluate the lesion clinically. In our study, theextent of tumors was clearly demarcated by PET-CT.

In the present study, we correlated our findings to theprogressive uptake scale distinguishing easily severeinflammation from malignant processes. Applying theparticular tools of the computer program, we adjustedthe windows in order to establish maximum andminimum values of uptake inside the tumor. Then weidentified and classified the sites of major activity insidethe tumor as centric or eccentric, according to the uptakeof the tracer. In the majority of the cases (77%) theanomalous concentration of the FDG was localized atthe center of the lesion, and in 23% the uptake of thetracer was increased eccentrically of the lesion. Thiscould be explained as a result of a wide variation in thebehavior of the tumors. According to the inherentgrowth of the tumors, it can make cells capable of livingfar from the blood supply (hypoxic cells) and thereby

favoring the progression of the tumor toward evergreater malignancy.13 The reverse pattern (a cord oftumor cells surrounded by capillaries) is also possible,and in both cases mitoses will occur causing differencesin uptake of the tracer.13 Also, the entire sample of ourwork consisted of patients in pretreatment (in radio-therapy planning), which excludes the possibility of theincreased uptake occurring in the presence of hypoxiccells that resisted radiotherapy.

Another finding to be considered is the presence oflymph node involvement and tumor differentiationpredominantly of grade III in patients with the highereccentric uptake. In our opinion, this may be anindication that the tumor is continuing to grow andspread thorough the tissue. In contrast, the centricbehavior could be related to the increased metabolism ofthe cells due to keratin synthesis, without mitosis orsignal of growth. In those patients the predominant

OOOOE

624 Takehana Dos Santos et al November 2005

tumor differentiation grade is I or II; only 1 patientpresented grade III. Therefore, other studies should bedone to clarify or to confirm these theories.

Some reports have demonstrated a diagnostic accu-racy variance from 64% to 100% and specificity variancefrom 70% to 100% using a nondedicated PET-CT device

Fig 3. Follow-up of SCCA at the posterior-inferior portion ofthe nasal fossa in axial (A) and coronal (B) views. The tumoris not clearly demonstrated on CT images. We observed 2areas of anomalous concentration of 18F-FDG in PET andPET/CT images. One showed an absence of activity at theposterior-inferior portion of the nasal fossa (white arrows),and the presence of another lesion was well visualized, owingto the high uptake of FDG, near to the oropharynx (yellowarrows).

in head and neck malignancies.4,14-17 According tothose authors, these results were in agreement withthose obtained in dedicated PET. In addition, with theintroduction of attenuation correction the diagnosticaccuracy and the specificity may be improved, justifyingthe clinical use of the nondedicated PET-CT device.Based on this literature review, we felt comfortableusing a similar dual-head gamma camara, a systemwith low-dose CT and positron emission coincidence

Fig 4. SCCA in the posterior part of the tongue andparapharynx is depicted in axial (A) and coronal (B) views,demonstrating anomalous concentration of 18F-FDG. Thesites of higher activity are eccentric (twice) when correlated tothe progressive uptake scale (right side).

OOOOE

Volume 100, Number 5 Takehana Dos Santos et al 625

acquisition capabilities, to obtain the topographicmetabolic map of the lesions.

Even with these technologic advances, it is necessaryto emphasize that imaging patterns do not allow estab-lishing a histologic diagnosis or pathologic grading.Several studies continue on the appropriate useful-ness of diagnostic imaging in cancer management,because the current applications of imaging techniquesin clinical oncology are still less than optimal. One areaof continuing concern is the type and appropriateperiodicity of imaging tests for following cancerpatients after treatment. The latest technologies havenot only improved cancer care and lessened patientsuffering, but they also may have reduced overallmedical expenditures by shortening and simplifyingdiagnostic procedures, lessening hospitalization re-quirements, and helping to tailor therapeutic approachesmore appropriately to individual patient needs.18-21

In conclusion, the methodologic approach that com-bined the functional metabolic and morphologic datawas fundamental to obtain precise information regard-ing the anatomical localization of the head and neckSCCA and to define areas of higher metabolic activityinside the tumor. Owing to the digital fusion of PET andCT images, accurate spatial localization could be done,allowing for better target delineation in surgery andradiotherapy planning.

We are grateful to Dr Margareth R Mintzberg, DDS, forgrammatical assistance.

REFERENCES1. Cavalcanti MGP, Vannier MW. Measurement of the volume of

oral tumors by three-dimensional spiral computed tomography.Dentomaxillofac Radiol 2000;29:35-40.

2. Beyer T, Townsend DW, Brun T, Kinahan PE, Charron M, RoddyR, et al. A combined PET/CT scaner for clinical oncology.J Nucl Med 2000;41:1369-79.

3. Lucignani G, Jereczek-Fossa BA, Orechia R. The role ofmolecular imaging in precision radiation therapy for targetdefinition, treatment planning optimisation and quality control.Eur J Nucl Med Mol Imaging 2004;31:1059-63.

4. Ak I, Blockland JAK, Pauwels EKJ, Stokkel MPM. The clinicalvalue of 18F-FDG detection with a dual-head coincidencecamera: a review. Eur J Nucl Med 2001;28:663-78.

5. Wong WL, Ussain K, Chevretton E, Hawkes DJ, Baddeley H,Maisey M, et al. Validation and clinical application of computer-combined computed tomography and positron emission tomog-raphy with 2-[18F]fluoro-2-deoxy-D-glucose head and neckimages. Am J Surg 1996;172:628-32.

6. Anzai Y, Carroll WR, Quint DJ, Bradford CR, Minoshima S,Wolf GT, et al. Recurrence of head and neck cancer after surgeryof irradiation: prospective comparison of 2-deoxy-2-[F-18]

fluoro-D-glucose PET and MR imaging diagnoses. Radiology1996;200:135-41.

7. Bar-Shalom R, Yefremov N, Haim N, Dann EJ, Epelbaum R,Keidar Z, et al. Camera-based FDG PET and 67Ga SPECT inevaluation of lymphoma: comparative study. Radiology 2003;227:353-60.

8. Wong FCL, Kim EE. Molecular imaging in oncology. Head andneck tumors. 1st ed. New York: Springer; 2000. p. 231-42.

9. Hannah A, Scott AM, Tochon-Danguy H, Chan JG, Akhurst T,Berlangieri S, et al. Evaluation of 18F-fluorodeoxyglucosepositron emission tomography and computed tomography withhistopathologic correlation in the initial staging of head andneck cancer. Annals Surg 2002;236:208-17.

10. Beahrs OH, Henson DE, Hutter RVP, Myers MH. Manual forstaging of cancer. American Joint Committee on Cancer. 3rd ed.Philadelphia: Lippincott; 1988.

11. Conti PS, Lilien DL, Hawley KPET. [F-18]-FDG in oncology.A clinical update. Nucl Med Biol 1996;23:717-35.

12. Hemingway R, Wong W, Chevretton EB, McGurk M. The use ofpositron emission tomography in the evaluation of orofacialmalignancy and disease. Br Dent J 1996;181:250-3.

13. Majno G, Joris I. Cells, tissues and disease. In: Kim EE,Jackson EF, eds. Principles of general pathology. New York, NY:Springer; 1996. p. 768-75.

14. Stokkel MP, Terhard CH, Hordijk GJ, van Rijk PP. The detectionof local recurrent head and neck cancer with fluorine-18fluorodeoxyglucose dual-head positron emission tomography.Eur J Nucl Med 1999;26:767-73.

15. Pai M, Park CH, Suh JH, Koh JH. Fluorine-18 fluorodeoxyglu-cose imaging using dual-head coincidence positron emissiontomography without attenuation correction in patients with headand neck cancer. Clin Nucl Med 1999;24:495-500.

16. Zimny M, Nowak B, Di Martino E, Jaenicke S, Cremerius U,et al. Hybrid PET vs. conventional imaging in the diagnosis ofhead and neck tumors. J Nucl Med 2000;41:292P.

17. Di Martino E, Novak B, Krombach GA, et al. Results ofpretherapeutic lymph node diagnosis in head and neck tumors.Clinical value of 18F-FDG positron emission tomography (PET).Laryngo Rhino Otol 2000;79:201-6.

18. Kim EE. Imaging strategies and perspectives in oncology. In:Kim EE, Jackson EF, eds. Molecular imaging in oncology. headand neck tumors. New York, NY: Springer; 2000. p. 13-5.

19. Gambhir SS, Czernin J, Schwimmer J, Silverman DHS, ColemanRE, Phelps ME. A tabulated summary of the FDG PETliterature. J Nucl Med 2001;42:1S-93S.

20. Ak I, Stockkel MP, Pauwels EK. Positron emission tomographywith 2-[18F]fluoro-2-deoxy-D-glucose in oncology: part II: theclinical value in detecting and staging primary tumours. J CancerRes Clin Oncol 2000;126:560-74.

21. Higashi T, Saga T, Nakamoto Y, Ishimori T, Mamede NH, WadaM, et al. Relationship between retention index in dual-phase18F-FDG PET and hexokinase-II and glucose transporter-1 ex-pression in pancreatic cancer. J Nucl Med 2002;43:173-80.

Reprint requests:

Dr Marcelo Cavalcanti, DDS, PhD

Faculty of Odontology

Department of Radiology

University of Sao Paulo

Av. Prof. Lineu Prestes, 2227

Sao Paulo, SP, 05508-900, Brazil

mgpcaval@usp.br