at SciVerse ScienceDirect

Clinical Radiology 67 (2012) 966e975

Contents lists available

Clinical Radiology

journal homepage: www.cl in icalradiologyonl ine.net

Delayed response assessment with FDG-PET-CT following(chemo)radiotherapy for locally advanced head and necksquamous cell carcinomaR.J.D. Prestwich a,b,*, M. Subesinghe c, A. Gilbert b, F.U. Chowdhury a,c,M. Sen b, A.F. Scarsbrook a,c

aDepartment of Nuclear Medicine, St James’s Institute of Oncology, Leeds, UKbDepartment of Clinical Oncology, St James’s Institute of Oncology, Leeds, UKcDepartment of Clinical Radiology, St James’s Institute of Oncology, Leeds, UK

article information

Article history:Received 9 January 2012Received in revised form22 February 2012Accepted 26 February 2012

* Guarantor and correspondent: R.J.D. PrestwicJames’s Institute of Oncology, Bexley Wing, BeYorkshire, LS9 7TF, UK. Tel.: þ44 113 2067838; fa

E-mail addresses: Robin.Prestwich@leedsth.nhorg.uk (R.J.D. Prestwich).

0009-9260/$ e see front matter � 2012 The Royal Codoi:10.1016/j.crad.2012.02.016

AIMS: To analyse the diagnostic accuracy of delayed response assessment 2-[18F]-fluoro-2-deoxy-D-glucose (FDG) positron-emission tomographyecomputed tomography (PET-CT)following (chemo)radiation for locally advanced head and neck squamous cell carcinoma(HNSCC).MATERIAL AND METHODS: Forty-four consecutive patients who underwent a baseline and

response assessment using FDG PET-CT for HNSCC following (chemo)radiation betweenAugust 2008 and April 2011 were identified retrospectively. Clinicopathological findings andserial clinical follow-up provided the reference standard.RESULTS: Median follow-up was 14 months (range 5e43 months). Response assessment FDG

PET-CT was performed at 16.8 weeks (inter-quartile range 15.8e18.6 weeks). Thirty-one out of44 (70%) response assessment examinations showed a complete metabolic response. Seven outof 40 (18%) assessable primary tumours were positive. Eight out of 41 (20%) patients withpre-treatment nodal disease had equivocal or positive FDG uptake at response assessment. Thesensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) forprimary disease and nodal disease were 100, 89, 43, 100, and 100%, and 92, 63, and 100%,respectively. Seven patients had residual FDG-negative soft tissue detectable on the unen-hanced CT component of the response assessment images; all remained disease free afterclinical observation. Distant metastases were detected on response assessment FDG PET-CT infour out of the 44 patients (10%).CONCLUSION: The diagnostic accuracy of response assessment with FDG PET-CT performed

at approximately 16 weeks post-(chemo)radiotherapy is good. The very high NPV ofa complete metabolic response can be used to guide management decisions. Although the PPVis limited for local residual disease, FDG PET-CT is a powerful screening tool for the detection ofinterim metastatic disease.

� 2012 The Royal College of Radiologists. Published by Elsevier Ltd. All rights reserved.

h, Level 4, Bexley Wing, Stckett Street, Leeds, Westx: þ44 113 2067886.s.uk, rprestwich@doctors.

llege of Radiologists. Published by

Introduction

Radiotherapy remains a key treatment modality in themanagement of head and neck squamous cell carcinoma

Elsevier Ltd. All rights reserved.

R.J.D. Prestwich et al. / Clinical Radiology 67 (2012) 966e975 967

(HNSCC). Sixty percent of patients present with locallyadvanced disease.1 Concurrent chemoradiotherapy hasnow been established as a standard of care for unresectableHNSCC,2 and for organ-preservation strategies.3 Forpatients with persistent disease following non-surgicaltreatment, salvage surgery to the primary site and/or aneck dissection has curative potential. The optimalmanagement of the neck lymph nodes following (chemo)radiotherapy remains a controversial area4; some centreshave advocated the use of a routine neck dissection for N2/3disease regardless of treatment response.5,6 In the modernera many centres have adopted a selective policy of neckdissection only if there is evidence of residual nodaldisease.4,7 Surgery to the primary lesion and/or neckdissection post-(chemo)-radiotherapy are associated withsignificant morbidity.4,8 Accurate assessment of therapeuticresponse is essential to minimize unnecessary surgicalmorbidity in patients who have been successfully treatedand to optimize the chances of cure in those with persistentdisease.

Treatment response assessment has conventionallyutilized a combination of clinical examinationand anatomicalimaging with computed tomography (CT) and/or magneticresonance imaging (MRI). Clinical examination alone isinadequate, with very limited sensitivity in the order of50%.9,10 Theuseof anatomical imaging is inherently limitedbythe difficulty in detecting small residual deposits. Forexample, sub-centimetre residual lymph nodes fall below theconventional criteria for malignant lymph node involvementon anatomical imaging.11 Post-treatment changes can alsomake interpretation of images difficult. In addition, CTand/orMRI fail to differentiate residual masses with and withoutviable tumour.12 The use of functional imaging with 2-[18F]-fluoro-2-deoxy-D-glucose (FDG) positron-emissiontomography (PET)eCT offers the potential to overcomethese deficiencies by identifying residual disease based uponmetabolic activity rather than the presence or absence ofanatomical abnormalities. A number of studies have sug-gested that the use of FDG PET-CT may be superior toconventional imaging in response assessment.12,13

The optimal timing of response assessment remainsunclear. Late complete responses are well documenteddue to sub-clinical tumour cell damage following radio-therapy. For example, in a recent chemo-radiotherapystudy,14 response assessment was performed 8 weeks post-treatment; however, this is likely to have been too early asa higher proportion of patients were found to have hada complete response on imaging performed 8 months post-treatment. It is undesirable to allow an excessive delay inresponse assessmentwhen salvage options are available. Thetiming of imaging is likely to be particularly critical to theinterpretation of FDG PET-CT images, to allow accuratedifferentiation of residual disease from inflammatorychanges secondary to radiotherapy. In a recent meta-analysis, FDG PET-CT performed more than 12 weeks post-treatment had higher diagnostic accuracy.15

Prior to 2008, imaging response assessment was per-formed at approximately 3 months post-(chemo)radio-therapy with MRI or CT. In 2008 a new imaging algorithm

involving a baseline and delayed response assessmentFDG-PET-CT at approximately 4 months post-(chemo)radiotherapy for patients with locally advanced HNSCC wasintroduced at our tertiary referral centre was introduced atSt. James's Institute of Oncology. At the time of imple-mentation, data suggested that imaging early after treat-ment limited accuracy,29 although the optimal timing ofresponse assessment was uncertain. The aim of the presentstudy was to review the impact of this development and, inparticular, to assess the accuracy of a delayed FDG-PET-CTexamination for response assessment.

Materials and methods

Inclusion criteria

Consecutive patients who underwent FDG PET-CT forhead and neck cancer between August 2008 and April 2011were obtained from an institutional database. Electroniccase notes were used to identify patients who fulfilled theeligibility criteria for the study. Formal institutional reviewboard approval was waived for this retrospective study.

Eligible patients fulfilled all of the following criteria: (1)histologically confirmed squamous cell carcinoma of theoropharynx, oral cavity, hypopharynx, larynx, or paranasalsinuses; (2) reviewed by a specialist head and neck multi-disciplinary team (MDT) meeting; (3) TNM stage III or IV;(4) received radical non-surgical treatment (radiotherapyalone or chemo-radiotherapy); (5) FDG PET-CT performedas a baseline prior to treatment; (6) FDG PET-CT performedas response assessment post-(chemo)radiotherapy.

Exclusion criteria were (1) nasopharynx cancer; (2)previous therapeutic resection of primary or nodal disease;(3) previous history of radiotherapy; (4) FDG PET-CTperformed only following response assessment with CTand/or MRI.

Data collection

Demographic, clinical, and imaging data, including PET-CT findings, maximum standardized uptake value (SUV)at the primary site and lymph nodes at presentation,maximum SUV at the primary site and lymph nodes atresponse assessment, were obtained from review of insti-tutional electronic case note records and imaging review.

Staging

Conventional staging of locally advanced HNSCC wasroutinely performed by physical examination and neckpalpation, fibre-optic endoscopy, examination underanaesthetic with biopsy where indicated, MRI or contrast-enhanced CT of the head and neck region, and CT of thethorax. Staging FDG PET-CT was performed primarily toprovide a baseline for future response assessment. Resultswere routinely reviewed in a specialist head and neck MDTmeeting and a TNM classification, based on all availableclinical and radiological data, according to the AmericanJoint Committee on Cancer TNM staging was assigned.

R.J.D. Prestwich et al. / Clinical Radiology 67 (2012) 966e975968

Radiotherapy

During the early part of the study period, patientswere treated with a conformal technique involving aparallel opposed pair as previously described.16 Intensity-modulated radiotherapy (IMRT) was subsequently imple-mented into routine clinical practice. Institutional protocolswere followed with a radical treatment dose of 70 Gy in 35fractions over 7 weeks, with lower doses to prophylacticdose regions (54e63 Gy in 35 fractions over 7 weeks). Otherdose-fractionation regimens were occasionally used at thetreating clinician’s discretionwithin the period of the study.

Chemotherapy

Induction chemotherapy with docetaxel, cisplatin, and5-fluorouracil was delivered to a proportion of patients aspreviously described.17 Concurrent chemotherapy consistedof cisplatin 100mg/m2 days 1, 22, and 43. Carboplatin AUC 4was substituted for cisplatin if the patient’s creatinineclearance was <55 ml/min.

Response assessment and follow-up

Tumour response was routinely assessed 4 months afterthe completion of treatment by clinical examination, naso-endoscopy, and FDG PET-CT; examination under anaes-thetic (EUA) and biopsies were performed on clinicaldiscretion following response assessment. Patients withless than a complete response were evaluated for salvagesurgery. Subsequently, patients were followed up withphysical examination, and flexible endoscopy every 6e8weeks in the first year after treatment, every 3 months foran additional 2 years, and 6 monthly until discharge at 5years. Patients with biopsy-proven local and/or regionaldisease were considered for salvage surgery by the MDT.

FDG PET-CT protocol

FDG PET-CT examinations prior to June 2010 were per-formed on a 16-section Discovery STE PET-CT system (GEHealthcare, Amersham, UK), and from June 2010 on a64-section Philips Gemini TF64 system (Philips Healthcare,Best, the Netherlands). PET acquisition from skull vertex toupper thighs was performed 60min after a 400MBq dose ofintravenous FDG. A silence protocol was employed in theuptake period following tracer injection to minimizephysiological tracer activity within the head and neckregion. The CT component was performed according to astandardized protocol (without the use of iodinatedcontrast medium) with the following settings: 140 kV;80 mAs; tube rotation time 0.5 s per rotation; pitch 6;section thickness 3.75 mm (to match the PET sectionthickness). Patients maintained normal shallow respirationduring the CT acquisition. Images were reconstructed usinga standard OSEM algorithm with CT for attenuationcorrection. Both non-attenuation corrected and attenuationcorrected datasets were reconstructed.

Categorization of FDG PET-CT response assessment

FDG PET-CT images were reported qualitatively (byvisual assessment) and quantitatively (maximum SUV) bya dual-certified radiologist and nuclear medicine physician.Primary tumour and nodal maximal SUVs were docu-mented. Results of post-treatment FDG PET-CT were cate-gorized into “positive”, “equivocal”, and “negative” for theprimary site and nodal sites separately as previouslydescribed.12 Areas of FDG uptake were classified as positiveif uptake was focal, corresponding to structural abnormalityand being of greater intensity than background liveractivity. Uptake was classed as equivocal if focal FDG uptakewas below liver background but above that of surroundingnormal tissues. Uptake was classed as negative in theabsence of any abnormal focal FDG uptake. Diffuse FDGuptake in the absence of a corresponding anatomicalabnormality on the CT was considered to be radiotherapy-related and was also classed as negative. The presence orabsence of residual tissue on the CT component of the post-treatment FDG-PET-CT was recorded.

Analysis and statistics

Histopathology from either a biopsy or surgical proce-dure was defined as the reference standard for determiningthe presence of persistent or recurrent loco-regionaldisease. In patients who did not receive a biopsy, serialnegative physical examinations over the follow-up periodwere used as confirmation of disease-free status. Sensitivity,specificity, positive and negative predictive values werecalculated using Excel software (Microsoft, Redmond, USA).For the purpose of these calculations “equivocal” findingswere classified as positive.

Results

Forty-four patients were eligible for inclusion in theretrospective study. Median age was 55 years (range 29e75years). Thirty-one of the 44 (70%) were male. All patientswere World Health Organization (WHO) performancestatus 0 or 1. Tumour site, sub-site, TNM stage, histology,and treatment details are listed in Table 1. Median follow-upwas 14 months (range 5e43 months).

At baseline FDG PET-CT, performed prior to treatment,the median primary tumour maximum SUVwas 13.9 (range4e55). The median lymph node maximum SUV was 11.5(range 3e26). Response assessment FDG PET-CT was per-formed at a median of 16.8 weeks post-(chemo)radiation(interquartile range 15.8e18.6, range 9e24 weeks). Table 2summarizes the sensitivity, specificity, positive predictivevalues, negative predictive values for the primary tumours,nodes, and overall.

Assessment of primary tumour response by FDG PET-CTand correlation with outcome

Seven of 40 patients with a detectable primary at base-line imaging (18%) had persistent FDG uptake at the primary

Table 1Disease site, TNM stage, histology, and treatment.

Characteristic Number (%)

Primary tumour site:Oropharynx 30 (68%)Larynx 3 (7%)Hypopharynx 6 (14%)Paranasal sinuses 1 (2%)Unknown primary 4 (9%)

T stageTX 4 (9%)T1 10 (23%)T2 12 (27%)T3 9 (20%)T4 9 (20%)

Nodal (N) stageN0 3 (7%)N1 1 (2%)N2a 3 (7%)N2b 28 (64%)N2c 7 (16%)N3 2 (5%)

Metastases (M) stageM0 44 (100%)M1 0

StageIII 1 (2%)IV 43 (98%)

HistologySCC 44 (100%)

TreatmentRadical radiotherapy 7 (16%)Concurrent chemo-radiotherapy 24 (54%)Induction TPF (docetaxel, cisplatin,5-fluorouracil)þconcurrent

12 (27%)

ChemoradiotherapyConcurrent cetuximab and radiotherapy 1 (2%)

Radiotherapy dose schedule70 Gy in 35 fractions 43 (98%)65 Gy in 30 fractions 1 (2%)

Radiotherapy techniqueThree-dimensional conformal radiotherapy 23 (58%)Intensity-modulated radiotherapy 21 (42%)

Table 3Summary of positive positron-emissionecomputed tomography (PET-CT)response assessment in primary tumour site.

Primary site BaselineprimarySUV

ResponseassessmentSUV

True/falsepositive(þve)

Method ofdeterminingoutcome

Oropharynx 21.5 5.8 True þve Clinical progressionOropharynx 16.5 3.9 True þve Clinical and MRI

progressionOropharynx 10 6 True þve BiopsyOropharynx 20 3 False þve EUA and 12 months

clinical F/UOropharynx 14.9 3.2 False þve 15 months clinical F/UHypopharynx 14.4 8 False þve Negative biopsy and

8 months clinical F/UHypopharynx 15.1 4.1 False þve Negative biopsy and

9 months clinical F/U

SUV, standardized uptake value; MRI, magnetic resonance imaging; EUA,examination under anaesthesia; F/U, follow-up.

R.J.D. Prestwich et al. / Clinical Radiology 67 (2012) 966e975 969

tumour site on the response assessment, which was inter-preted as positive. Median primary site SUV on the follow-up assessment in this group was 4.1 (range 2.9e8). Detailsare shown in Table 3; based on clinico-pathological

Table 2Diagnostic performance of response assessment FDG PET-CT in 44 patientswith locally advanced head and neck cancer after (chemo)radiation.

Primarysite(n ¼ 40)

Necknodes(n ¼ 41)

Overall(primaryand neck)(n ¼ 44)

FDG-PET-CT positive 7 8 13FDG-PET-CT negative 33 33 31True positive 3 5 7True negative 33 33 31False positive 4 3 6False negative 0 0 0Sensitivity 100% 100% 100%Specificity 89% 92% 84%Positive predictive value 43% 63% 54%Negative predictive value 100% 100% 100%

correlation four out of seven were false positive. As shownin Table 3, the maximum SUV could not distinguish true andfalse positive findings.

All of the remaining (n¼ 33) assessments were classed asnegative. None of these patients with a negative responseassessment FDG PET-CT at the primary tumour site subse-quently developed local recurrence.

Assessment of cervical lymph node response FDG PET-CTand correlation with outcome

The details of the eight response assessment FDG PET-CTthat were classed as either positive or equivocal with regardto nodal disease are shown in Table 4. All of the fourpatients with a positive FDG PET-CT response assessment[median SUV of 5 (range 4e9.6)] were true positives. Theassessments in a further four patients were interpreted asequivocal with a median SUV of 2.6 (range 2.4e2.7). One ofthese patients subsequently progressed in the neck nodesand was classified as a true positive. The remainder werefalse positive; an example is shown in Fig 1. No patients(n¼ 33) with a negative FDG PET-CT response assessment inthe nodal region subsequently developed nodal recurrence.

Detection of distant metastases by responseassessment FDG PET-CT

Four of 44 (10%) patients had distant metastases detectedon response assessment FDG PET-CT imaging. Two patientshad new multiple lung nodules. One patient had FDG-avidmediastinal nodes, lung, and bone lesions. These threepatients all had subsequent radiologically confirmedprogressive metastatic disease and died during the follow-up period. The remaining patient had a solitary liverlesion, which was subsequently resected and found to bea metastasis from the base of tongue squamous cell carci-noma (Fig 2). All of these patients had had a baseline pre-treatment FDG-PET-CT examination with no evidence ofdistant metastases. Of the 40 patients with no evidence ofdistant metastases on response assessment FDG-PET-CT,

Table 4Summary of positive or equivocal positron-emissionecomputed tomography (PET-CT) response assessment in lymph nodes.

Primary site BaselineprimarySUV

ResponseassessmentSUV

Size of lymph nodeon CT at responseassessment (mm)

Positive orequivocalscan

Method of determiningoutcome

True or falsepositive (þve)

Hypopharynx 12.3 4.5 8 Positive Biopsy and þve neck dissection True þveSupraglottis 16.1 5.4 39 Positive Clinical progression True þveUnknown 12.9 9.6 33 Positive Clinical progression True þveOropharynx 16 4 12 Positive Clinical and MRI progression True þveOropharynx 8 2.7 15 Equivocal Negative neck dissection False þveOropharynx 16 2.4 7 Equivocal Clinical progression True þveOropharynx 12.6 2.1 7 Equivocal 14 months clinical F/U False þveUnknown unavailable 2.7 7 Equivocal Negative neck dissection and

36 months clinical F/UFalse þve

SUV, standardized uptake value; MRI, magnetic resonance imaging; F/U, follow-up.

Figure 1 Equivocal cervical lymphnode responseassessmentbyFDGPET-CT foraT3N2bpoorlydifferentiated squamous cell carcinomaof the left tonsilat baseline and response assessment 16 weeks post-chemoradiotherapy. (a) Baseline PET-CT. (b) Response assessment PET-CT. Black arrow ¼ primarytumour. Black arrowhead ¼ left level II lymph node; pre-treatment SUV of lymph node in (a) was 8.7 and post-treatment in (b) was 2.7. The primarytumour had a complete metabolic response to treatment. A left neck dissection was performed, which showed no evidence of residual disease.

R.J.D. Prestwich et al. / Clinical Radiology 67 (2012) 966e975970

Figure 2 Detection of interim distant metastases by response assessment PET-CT. FDG-PET-CT performed for a T2 N2b squamous cell carcinomaof the base of tongue at baseline and response assessment 16 weeks post-chemoradiotherapy. (a) Maximum intensity projection (MIP) pre-chemoradiotherapy, (b) MIP post-chemoradiotherapy, (c) baseline PET-CT (head and neck region), (d) baseline PET-CT (through liver),(e) response assessment PET-CT (through liver). Black arrow ¼ primary tumour, filled black arrowhead ¼ liver metastasis, emptyarrowhead ¼ regional lymph node metastasis. The patient underwent a liver metastectomy, which confirmed the diagnosis of metastaticsquamous cell carcinoma.

R.J.D. Prestwich et al. / Clinical Radiology 67 (2012) 966e975 971

one developed lung metastases 8 months later. The nega-tive predictive value of negative response assessment FDG-PET-CT for distant metastases was 98% (39/40).

Outcome of patients with FDG-negative residual tissue

One patient with a maxillary sinus tumour had FDG-negative residual disease in the primary tumour site onthe unenhanced CT component of the FDG PET-CT responseassessment (Fig 3); there has been no evidence of clinicalprogression 21 months following completion of treatment.Six patients had residual FDG-negative nodal tissuedetectable on the unenhanced CT component of theresponse assessment PET-CT; the sizes of residual lymphnodes ranged from 1e2 cm in short axis. An example isshown in Fig 4. All of these patients remain disease freeafter clinical observation of residual lymph node tissue aftera median of 16 months follow-up (range 13e36 months).

Discussion

Response to definitive (chemo)radiation is a powerfulpredictor of outcome.18 Accurate detection of treatmentfailure prior to subsequent disease progression is likely tofacilitate surgical salvage options and avoid unnecessaryintervention. Conventional response assessment caninvolve clinical examination and anatomical imaging withCT and/or MRI. Response criteria such as response evalu-ation criteria in solid tumors (RECIST)19 are based onmeasurement of lesions; a complete response requires the

complete disappearance of malignant lesions. However,residual tissue, particularly small lymph nodes, is commonfollowing (chemo)radiotherapy.12 In addition, radiation-induced changes are difficult to distinguish from residualdisease using conventional imaging. The limitations ofanatomical imaging techniques in identifying the presenceor absence of residual tumour in residual masses and indetecting small foci of disease has led to widespreadinterest in the use of functional imaging in this setting.

FDG PET-CT has an emerging role in the management ofHNSCC. Currently under investigation are the role of FDGPET-CT in staging, radiotherapy target definition, responseassessment, and recurrence detection.13 FDG PET-CT isa promising tool to evaluate response, as metabolic changesin tumours often occur prior to a reduction in tumour size.20

However, the interpretation of PET images can be chal-lenging post-radiotherapy, when asymmetry, inflamma-tion, and oedema are common causes of “benign” traceruptake. The optimal approach to the analysis of FDG uptakeremains uncertain. Commonly used approaches are a qual-itative visual assessment and semi-quantitative assessmentbased upon maximal SUV. Variable physiological FDGuptake in the head and neck region impedes the use ofa strict SUV threshold for defining abnormal areas. There islittle evidence to suggest that the use of SUV quantificationimproves diagnostic accuracy.13,21,22 Therefore, in this study,image interpretation was based upon visual interpretation;SUVs were documented where possible but were not inte-gral to the response assessment. The use of integratedPET-CT is more accurate than FDG-PET alone in HNSCC

Figure 3 FDG-negative primary tumour residual mass post-chemoradiotherapy. FDG PET-CT performed for a T4 N1 squamous cell carcinoma ofthe left paranasal sinus at baseline and response assessment 16 weeks post-chemoradiotherapy. (a) Baseline PET-CT. (b) Response assessmentPET-CT. Black arrowhead indicates paranasal tumour mass pre-treatment. Black arrow indicates residual FDG-negative residual tissue. Notebilateral physiological metabolic activity of extra-ocular muscles. The patient remains disease free after 21 months of clinical observation.

R.J.D. Prestwich et al. / Clinical Radiology 67 (2012) 966e975972

imaging.23e25 However, in the context of response assess-ment, no benefit has been shown for FDG PET-CT beyondFDG-PET.15 Despite these findings, the superior localiza-tion26 provided by integrated PET-CT is important forguiding biopsies and any surgical intervention.

A number of other prospective and retrospective studieshave reported the role of FDG-PET in response assessmentand have been the subject of meta-analyses.15,27 The largestand most recent meta-analysis15 was published in 2011 andincluded 51 studies involving 2335 patients. The includedstudies were very heterogeneous, involving differing headand neck tumour sites and stages, the use of standalone PETand integrated PET-CT, response assessment and surveil-lance follow-up imaging, assessment of primary and/orneck disease, and variable timing of PET ranging from 2weeks to more than 2 years post-treatment. For assessmentof the primary tumour site, 24 studies with 1122 patientswere included in the meta-analysis. The pooled sensitivity,specificity, PPV, and NPV was 79.9, 87.5, 58.6, and 95.1%.Thirty studies, comprising 1525 patients, were analysed forassessment of the response of neck disease; the same

measures were 72.7, 87.6, 52.1, and 94.5%, respectively. Thecombined accuracy for the primary site and lymph nodescould be assessed in 17 studies with 981 patients; theresults were 87.7, 87.8, 75.7, and 94.3%, respectively.

The purpose of the present retrospective study was toevaluate the diagnostic accuracy of FDG PET-CT in theassessment of response to (chemo)radiotherapy for locallyadvanced HNSCC at a single institution using a standardizedprotocol. Baseline FDG PET-CT was performed prior totreatment in all patients. Initial imaging is of value inresponse assessment by more accurately depicting theextent of disease prior to treatment and providing a base-line for comparison. This cannot completely replaceconventional staging, particularly as not all sites of diseasemay be FDG-avid prior to therapy, for example, necroticlymph nodes are recognized as commonly being FDGnegative.28 Delayed response assessment imaging wasperformed a median of 16 weeks post-(chemo)radiationbased on the concept that early inflammation post-radiotherapy would impede accurate detection of residualdisease, whilst the timing would also allow surgical

Figure 4 FDG negative lymph node residual mass post-chemoradiotherapy. FDG-PET-CT performed for a T2 N2b squamous cell carcinoma of theleft base of tongue at baseline and response assessment 16 weeks post-chemoradiotherapy. (a) Baseline PET-CT. (b) Response assessment PET-CT.Black arrowhead indicates pre-treatment malignant lymph node. Black arrow indicates residual FDG-negative residual lymph node tissue. Thepatient remains disease free after 36 months of clinical observation.

R.J.D. Prestwich et al. / Clinical Radiology 67 (2012) 966e975 973

intervention when necessary prior to tumour progression.Data available at the time of the implementation ofresponse assessment FDG-PET-CT suggested that FDG-PETperformed early post-treatment has limited accuracy,29

although the optimal timing was uncertain. In the meta-analysis by Gupta et al.15 assessments had a higher diag-nostic accuracy if performed more than 12 weeks followingthe completion of treatment. In this series FDG PET-CTresponse assessment for both primary tumour and lymphnode response had very high sensitivity, specificity, andNPV (100, 89, and 100% for primary disease and 100, 92, and100% for lymph node disease, respectively). The PPV forprimary tumour (43%) and lymph nodes (63%) was limited.

The limitations of the study include the retrospectivenature of the analysis, limited patient numbers and follow-up. Although the longest follow-up period was 43 months,most patients were treated in the latter part of the studyperiod once experience of PET-CT as a response assessmenttool was gained. However, these data are consistent withthe results of the meta-analyses.15,27 The excellent sensi-tivity and NPV of delayed response assessment FDG PET-CT

is sufficiently high to guide clinical management, avoidingunnecessary interventions, such as examination underanaesthetic, biopsy and neck dissection in the context ofa negative assessment. In addition, the high NPV can beused to make a case for a risk-adapted follow-up strategy.13

The PPV is limited for both the primary tumour and lymphnode response. In the analysis of diagnostic accuracy,assessments reported as equivocal were combined withthose reported as positive. In the assessment of lymph noderesponse, four assessments were reported as positive, ofwhich all four were true positives. In contrast, four assess-ments were reported as equivocal; only one of these wasa true positive and two underwent negative neck dissec-tions. Although based on limited numbers, these datasuggest that neck dissections may be unnecessary in manycases of equivocal imaging. These are consistent withprospective data from Porceddu et al.12 In this study of PET-directed management of the neck in 112 patients, PET wasperformed 12 weeks following completion of (chemo)radiotherapy. If nodal imaging was equivocal (defined asfocal FDG activity below liver background) PET imaging

R.J.D. Prestwich et al. / Clinical Radiology 67 (2012) 966e975974

was repeated after an interval of 4e6 weeks; a neckdissection was performed at this stage if imaging remainedequivocal. Of 11 equivocal PET assessments, 10 had negativerepeat assessments without subsequent nodal failure onfollow-up.

The absence of CT contrast medium with the PET-CTperformed in this study limited the ability to assess theincremental impact of response assessment PET-CT abovemorphological imaging with contrast-enhanced CT or MRI.However, the unenhanced CT did demonstrate residualmasses in seven patients both in primary and nodal sites,which were FDG negative. None of these patients relapsedon subsequent follow-up. Consistent with these data, in thestudy by Porceddu et al., 41 patients with PET-negativeresidual nodal masses were observed without subsequentnodal failure.12 These data therefore suggest that patientswith FDG-negative residual masses can be safely observedand spared surgical intervention.

The present data demonstrate the ability of responseassessment PET-CT to detect interimmetastases (four out of44 response assessment PET-CT examinations) with a highNPV. This is consistent with the high sensitivity of PET fordistant metastatic disease.30 These findings are similar tothose of Zundel et al. who reported the detection of fourcases of distant metastases in a series of 52 patientsundergoing response assessment PET-CT 4e6 months post-treatment.9 These data suggest that an additional advantageof response assessment with PET-CT is the detection ofinterim distant metastases.

In conclusion, the diagnostic accuracy of responseassessment with FDG PET-CT performed at approximately16 weeks post-(chemo)radiotherapy for locally advancedHNSCC is good. The very high NPV of a metabolic completeresponse can be used to guide management decisions.Although the PPV is limited for local residual disease, FDGPET-CT is a powerful screening tool for the detection ofinterim metastatic disease.

References

1. Seiwert TY, Cohen EE. State-of-the-art management of locally advancedhead and neck cancer. Br J Cancer 2005;92:1341e8.

2. Adelstein DJ, Li Y, Adams GL, et al. An intergroup phase III comparison ofstandard radiation therapy and two schedules of concurrent chemo-radiotherapy in patients with unresectable squamous cell head and neckcancer. J Clin Oncol 2003;21:92e8.

3. Forastiere AA, Goepfert H, Maor M, et al. Concurrent chemotherapy andradiotherapy for organ preservation in advanced laryngeal cancer. N EnglJ Med 2003;349:2091e8.

4. Ferlito A, Corry J, Silver CE, et al. Planned neck dissection for patientswith complete response to chemoradiotherapy: a concept approachingobsolescence. Head Neck 2010;32:253e61.

5. Brizel DM, Prosnitz RG, Hunter S, et al. Necessity for adjuvant neckdissection in setting of concurrent chemoradiation for advanced head-and-neck cancer. Int J Radiat Oncol Biol Phys. 2004;58:1418e23.

6. Stenson KM, Haraf DJ, Pelzer H, et al. The role of cervical lymphade-nectomy after aggressive concomitant chemoradiotherapy: the feasi-bility of selective neck dissection. Arch Otolaryngol Head Neck Surg2000;126:950e6.

7. Grabenbauer GG, Rodel C, Ernst-Stecken A, et al. Neck dissectionfollowing radiochemotherapy of advanced head and neck cancerdforselected cases only? Radiother Oncol 2003;66:57e63.

8. Esteller E, Vega MC, Lopez M, et al. Salvage surgery after locoregionalfailure in head and neck carcinoma patients treated with chemo-radiotherapy. Eur Arch Otorhinolaryngol 2011;268:295e301.

9. Zundel MT, Michel MA, Schultz CJ, et al. Comparison of physical exam-ination and fluorodeoxyglucose positron emission tomography/computed tomography 4e6 months after radiotherapy to assess residualhead-and-neck cancer. Int J Radiat Oncol Biol Phys. 2011;81:e825e32.

10. McHam SA, Adelstein DJ, Rybicki LA, et al. Who merits a neck dissectionafter definitive chemoradiotherapy for N2eN3 squamous cell head andneck cancer? Head Neck 2003;25:791e8.

11. van den Brekel MW, Stel HV, Castelijns JA, et al. Cervical lymph nodemetastasis: assessment of radiologic criteria. Radiology 1990;177:379e84.

12. Porceddu SV, Pryor DI, Burmeister E, et al. Results of a prospective studyof positron emission tomography-directed management of residualnodal abnormalities in node-positive head and neck cancer afterdefinitive radiotherapy with or without systemic therapy. Head Neck2011;33:1675e82.

13. Mak D, Corry J, Lau E, et al. Role of FDG-PET/CT in staging and follow-upof head and neck squamous cell carcinoma. Q J Nucl Med Mol Imaging2011;55:487e99.

14. Paccagnella A, Ghi MG, Loreggian L, et al. Concomitant chemo-radiotherapy versus induction docetaxel, cisplatin and 5 fluorouracil(TPF) followed by concomitant chemoradiotherapy in locally advancedhead and neck cancer: a phase II randomized study. Ann Oncol2010;21:1515e22.

15. Gupta T, Master Z, Kannan S, et al. Diagnostic performance of post-treatment FDG PET or FDG PET/CT imaging in head and neck cancer:a systematic review and meta-analysis. Eur J Nucl Med Mol Imaging2011;38:2083e95.

16. Prestwich RJ, Kancherla K, Oksuz DC, et al. A single centre experiencewith sequential and concomitant chemoradiotherapy in locallyadvanced stage IV tonsillar cancer. Radiat Oncol 2010;5:121.

17. Prestwich RJ, Oksuz DC, Dyker K, et al. Feasibility and efficacy ofinduction docetaxel, cisplatin, and 5-fluorouracil chemotherapycombined with cisplatin concurrent chemoradiotherapy for non-metastatic stage iv head-and-neck squamous cell carcinomas. Int J RadiatOncol Biol Phys. 2011;81:e237e43.

18. Corry J, Peters L, Fisher R, et al. N2eN3 neck nodal control withoutplanned neck dissection for clinical/radiologic complete responders-results of Trans Tasman Radiation Oncology Group Study 98.02. HeadNeck 2008;30:737e42.

19. Therasse P, Arbuck SG, Eisenhauer EA, et al. New guidelines to evaluatethe response to treatment in solid tumors. European Organization forResearch and Treatment of Cancer, National Cancer Institute of theUnited States, National Cancer Institute of Canada. J Natl Cancer Inst2000;92:205e16.

20. Ceulemans G, Voordeckers M, Farrag A, et al. Can 18-FDG-PET duringradiotherapy replace post-therapy scanning for detection/demonstra-tion of tumor response in head-and-neck cancer? Int J Radiat Oncol BiolPhys. 2011;81:938e42.

21. Ong SC, Schoder H, Lee NY, et al. Clinical utility of 18F-FDG PET/CT inassessing the neck after concurrent chemoradiotherapy for locoregionaladvanced head and neck cancer. J Nucl Med 2008;49:532e40.

22. Ware RE, Matthews JP, Hicks RJ, et al. Usefulness of fluorine-18 fluo-rodeoxyglucose positron emission tomography in patients witha residual structural abnormality after definitive treatment for squa-mous cell carcinoma of the head and neck. Head Neck 2004;26:1008e17.

23. Branstetter 4th BF, Blodgett TM, Zimmer LA, et al. Head and neckmalignancy: is PET/CT more accurate than PET or CT alone? Radiology2005;235:580e6.

24. Syed R, Bomanji JB, Nagabhushan N, et al. Impact of combined (18)F-FDG PET/CT in head and neck tumours. Br J Cancer 2005;92:1046e50.

25. Gordin A, Daitzchman M, Doweck I, et al. Fluorodeoxyglucose-positronemission tomography/computed tomography imaging in patients withcarcinoma of the larynx: diagnostic accuracy and impact on clinicalmanagement. Laryngoscope 2006;116:273e8.

26. Schoder H, Yeung HW, Gonen M, et al. Head and neck cancer: clinicalusefulness andaccuracyof PET/CT image fusion.Radiology2004;231:65e72.

27. Isles MG, McConkey C, Mehanna HM. A systematic review and meta-analysis of the role of positron emission tomography in the follow up

R.J.D. Prestwich et al. / Clinical Radiology 67 (2012) 966e975 975

of head and neck squamous cell carcinoma following radiotherapy orchemoradiotherapy. Clin Otolaryngol 2008;33:210e22.

28. Kyzas PA, Evangelou E, Denaxa-Kyza D, et al. 18F-fluorodeoxyglucosepositron emission tomography to evaluate cervical node metastases inpatients with head and neck squamous cell carcinoma: a meta-analysis.J Natl Cancer Inst 2008;100:712e20.

29. Greven KM, Williams 3rd DW, McGuirt Sr WF, et al. Serial positronemission tomography scans following radiation therapy of patients withhead and neck cancer. Head Neck 2001;23:942e6.

30. Lowe VJ, Boyd JH, Dunphy FR, et al. Surveillance for recurrent head andneck cancer using positron emission tomography. J Clin Oncol2000;18:651e8.