Radiotherapy and Oncology 106 (2013) 283–287

Contents lists available at SciVerse ScienceDirect

Radiotherapy and Oncology

journal homepage: www.thegreenjournal .com

Tumor progression before treatment

Oesophageal tumour progression between the diagnostic 18F-FDG-PET and the18F-FDG-PET for radiotherapy treatment planning

Christina T. Muijs ⇑, Jan Pruim, Jannet C. Beukema, Maaike J. Berveling, John Th. Plukker,Johannes A. LangendijkUniversity of Groningen, University Medical Center Groningen, Department of Radiation Oncology, Groningen, The Netherlands

a r t i c l e i n f o a b s t r a c t

Article history:Received 3 August 2012Received in revised form 25 October 2012Accepted 27 October 2012Available online 27 November 2012

Keywords:Oesophageal cancerPETTumour progression

0167-8140/$ - see front matter � 2012 Elsevier Irelanhttp://dx.doi.org/10.1016/j.radonc.2012.10.015

⇑ Corresponding author. Address: Department of RaMedical Center Groningen, P.O. Box 30001, 9300 RB G

E-mail address: c.t.muijs@rt.umcg.nl (C.T. Muijs).

Background and purpose: To test whether the interval between diagnostic and therapeutic FDG-PET-scan-ning is associated with early tumour progression.Material and methods: All patients (n = 45) underwent two PET scans, one for staging (‘baseline PET’)using an HR+ positron camera or PET/CT-scanner and one for radiotherapy planning (‘therapeutic PET’)using a PET/CT-scanner.Material and methods: All images were reviewed in random order by an experienced nuclear physician. Ifthere were any discrepancies, the images were also compared directly. SUVmax, tumour length, lymphnode metastases and distant metastases were assessed.Results: The median time between the PET scans was 22 days (range: 8–49). The SUVmax increased (>10%)(19 patients, 42%) or decreased (11 patients, 24%). Fourteen patients (31%) showed tumour length pro-gression (>1 cm). TNM progression was found in 12 patients (27%), with newly detected mediastinalnodes (N) in eight patients (18%) and newly detected distant metastases (M) in six patients (13%). No sig-nificant prognostic factors were found. However, a trend was noted towards TNM progression for thetype of PET-camera (p = 0.05, 95% CI 0.01–0.66) and for the interval between the PET scans (p = 0.09,95% CI �0.9 to 12.5).Conclusion: This study suggests rapid oesophageal tumour progression. Therefore, the interval betweenrelevant imaging and start of the radiotherapy should be minimized. Furthermore, ‘state of the art’ PETscanners should be used.

� 2012 Elsevier Ireland Ltd. All rights reserved. Radiotherapy and Oncology 106 (2013) 283–287

In the last two decades, major efforts have been made toimprove survival by exploring new and emerging treatmentstrategies in oesophageal carcinoma [1,2]. Neo-adjuvant chemora-diotherapy (CRT) followed by surgery has been shown to signifi-cantly improve the survival rates as compared to surgery alone[3]. However, despite the introduction of these multimodalitytreatment strategies into routine clinical practice, treatment fail-ure, including both loco-regional recurrences and distant metasta-ses are frequently observed, even after complete local tumourresponse.

In oesophageal cancer, TNM-stage, standardized uptake value(SUV) and tumour length are important prognostic factors for over-all survival. More specifically, small sized tumours with low SUVvalues without lymph node metastases are associated with betteroverall survival. Although tumour length is not included in the cur-rent TNM staging system, the association between tumour length

d Ltd. All rights reserved.

diation Oncology, Universityroningen, The Netherlands.

and outcome in terms of loco-regional tumour control and overallsurvival has been well recognized [4,5].

Tumour progression can be identified by an increased SUV onFDG-PET. A high SUV is commonly associated with tumour aggres-siveness as expressed by an increased tumour length and nodalinvolvement [6]. As tumour progression growth continues duringboth the diagnostic and the preparation phase for irradiation, itis worthwhile to examine the clinical impact of tumour progres-sion in this time interval. The routine introduction of an increasednumber of sophisticated diagnostic procedures to improve thestaging and accuracy of target definition and delineation for radio-therapy has certainly led to an improvement of quality, but often atthe expense of increased time required to perform all these proce-dures. Moureau-Zabotto et al. [7] found new distant metastases onFDG-PET/CT-scans made for radiotherapy treatment planning in 6%of the patients which consequently lead to a change of treatmentstrategy from curative to palliative intent.

Therefore, in the current study, we tested the hypothesis thatthe time interval between the diagnostic and therapeutic FDG-PET-scan performed for radiotherapy treatment planning is associ-ated with tumour progression in terms of increased SUV-values

284 Early oesophageal tumour progression

and tumour lengths, newly diagnosed pathologic lymph nodes anddistant metastases,

Methods and materials

Patients

The study population was composed of patients who met thefollowing eligibility criteria: histologically confirmed oesophagealcancer (adeno- or squamous cell carcinoma); stage T2-T4a/N0 orT1-T4a/N1-N3, M0; selected for curative (neo-adjuvant) CRT, andno previous treatment or active infection.

All patients were staged according to the 7th edition of theTNM-system of the Union International Contre le Cancer (UICC)[8], based on the following procedures: physical examination,endoscopic ultrasonography (EUS), cervical/thoracic/abdominalCT, and whole body FDG-PET. Fine needle aspiration biopsy (FNAB)or other additional investigations were carried out only when indi-cated. After the collection of all diagnostic information, the pa-tients were discussed in a multidisciplinary tumour board.

All patients were included in a prospective trial to assess theimpact of a planning-PET/CT compared to the standard planning-CT on the delineation of target volumes for curative (CH-)RT ofoesophageal cancer. For this purpose, all patients underwent a sec-ond, i.e. a research, FDG-PET(/CT)-scan for future radiotherapytreatment planning. This study was approved by the local ethicscommittee and all patients provided written informed consent.

Table 1Patient and tumour characteristics.

Characteristics n = 45 (%)

GenderMale 33 (73)Female 12 (27)

Age (years)Median (range) 63 (41–85)

HistologyAC 34 (76)SC 11 (24)

Tumour localizationHigh 1 (2)Mid 3 (7)Distal/GEJ 41 (91)

Clinical stageT2N0M0 4 (9)T2N1M0 4 (9)T3N0M0 4 (9)T3N1M0 12 (27)T3N2M0 11 (24)T3N3M0 4 (9)T4aN0M0 2 (4)T4aN1M0 2 (4)T4aN2M0 2 (4)

Imaging

Staging FDG-PET(/CT)-scans were performed after the initialdiagnosis of the primary oesophageal tumour. These FDG-PET-scans will be referred to as ‘baseline PET’. For radiotherapy treat-ment planning, a second PET/CT scan was performed in treatmentposition over a limited field of view within 2 weeks after referral tothe department of Radiation Oncology, and will be referred to as‘therapeutic PET’.

The baseline PET scans were performed using an ECAT HR + PETcamera (Siemens/CTI, Knoxville, TN, USA) or an integrated PET/CTscanner (Biograph mCT 4-64 PET/CT, Siemens, Knoxville, TN,USA). The therapeutic PET images were all acquired using the inte-grated PET/CT scanner. All PET-scans were reconstructed accordingto a validated and standardized protocol to minimize the variabil-ity of SUV measures due to the use of different PET-cameras [9,10].

Serum glucose levels were measured after the patients had fas-tened for at least 4 h. Depending on body weight (5 MBq/kg forHR+ and 3 MBq/kg for integrated PET/CT), a median dose of378 MBq FDG (range 170–618) was administered intravenously.Emission scans were obtained 60 min after the injection of FDGand were performed for 5 min per bed position on the HR+ and2–3 min per bed position on the integrated PET/CT. The baselinePET included the skull to mid femur. For the therapeutic PET theneck, thorax and the upper abdomen, including the liver, wereincluded.

All images were reviewed in random order and without knowl-edge of the patient’s details by an experienced nuclear physician(JP). For quantitative analysis of the FDG uptake, the maximumstandardized uptake value (SUVmax) based on body weight and cor-rected for serum glucose was used [10,11]. Serum glucose levelswere measured before each PET examination, using a calibratedglucose metre.

For further analyses, the following features were used: visualtumour length, tumour length using 70% of the SUVmax (SUV70%),and visual interpretation of lymph nodes involvement and/or thepresence of distant metastases.

If there were any discrepancies after the independent scoringbetween the two FDG-PET scans of one individual patient, theimages were also compared directly in order to verify the originof the discrepancies.

Follow up

Histo[cyto]pathological confirmation of the tumour progressionas detected by the therapeutic FDG-PET[/CT] was not performed,since both PET-scans were reviewed and compared during or afterthe neo-adjuvant chemoradiation. Therefore, we compared theoverall survival for the patients with and without TNM-progres-sion. Routine follow up was performed every 3 months in the firstyear and 6 months in the second year, followed by a yearly control.In the first 2 years, a CT-scan of the thorax and abdomen was partof this follow up for non-metastasized patients.

Statistical analysis

Univariate analysis was carried out to assess the association be-tween the interval between prognostic determinants and TNMprogression, tumour length progression or an increased SUV. Anindependent t-test was performed for dichotomised variables,while a chi-square test was used for continuous variables.

Overall survival time was calculated from the last day of exter-nal radiotherapy, according to the Kaplan–Meier method. Survivalcurves were compared using the log rank test. To perform thesecalculations, SPSS version 18.0 was used.

Results

Between March 2009 and June 2011, a total of 45 patients wereincluded, with a median age of 64 years (range: 41–85 years). Mosttumours were adenocarcinoma (76%) and located at the distaloesophagus (91%). Patients and tumour characteristics are listedin Table 1. All patients underwent a second FDG-PET/CT scan,which was performed after a median interval of 22 days (range:8–49 days) after baseline PET.

The therapeutic FDG-PET/CT-scan revealed 18 new pathologiclymph nodes, which resulted in a changed N-stage in eight patients

C.T. Muijs et al. / Radiotherapy and Oncology 106 (2013) 283–287 285

(18%). Histology, SUVmax, tumour length, sex, type of PET camera orinterval between the 2 PET-scans were not significant prognosticfactors for the finding of new pathologic lymph nodes.

Two lymph nodes, that were suspicious at baseline PET, were nolonger visible at the therapeutic FDG-PET/CT, suggesting a non-malignant origin. In one of these patients, the N-status changedfrom N1 to N0. In this patient, the FDG-uptake in a lymph nodeat the minor curvature with an initial SUV of 2.9 was no longer ob-served at the therapeutic FDG-PET/CT.

All patients were initially staged as M0 on their baseline FDG-PET(/CT). However, new distant metastases became manifest insix patients (13%), making them ineligible for curative treatment.

Overall, a total of 12 patients (27%) showed tumour progressionin terms of a worse TNM-stage (Fig. 1). There characteristics arelisted in Table 2. We found no significant prognostic factors forTNM progression. However, a trend was noted towards TNM pro-gression for the type of PET-camera used (p = 0.05, 95% CI 0.01–0.66) and for the interval between the two PET scans (p = 0.09,95% CI �0.9 to 2.5).

At visual interpretation 14 patients (31%) showed progressionin tumour length with >1 cm. A tumour length reduction wasfound in two patients (4%). For the other 29 patients (65%), the tu-mour length at the baseline FDG-PET was similar to the tumourlength at the therapeutic FDG–PET/CT. Univariate analyses showedno significant prognostic factors for tumour length progression.

The corrected SUVmax of the primary tumour increased by morethan 10% in 19 patients (42%), while a decrease of the correctedSUVmax by more than 10% was seen 11 patients (24%). The meanand median corrected SUVmax of the primary tumour of the base-line PET and the second PET were comparable.

The SUVmax and the length of the primary tumour at the base-line PET images were significant prognostic factors for an increaseof the SUV at the second PET. Patients with an increased SUV had

Fig. 1. Lymph node progression was found within an interval of 31 days between the dPET/CT scanner.

smaller tumours (38 vs 50 mm) with a lower SUV (11 vs 16) atthe baseline PET.

The newly diagnosed pathologic lymph nodes at the secondPET-scan had a median SUV of 4.5 (range 2.6–9.9). Nine patientsshowed an increased SUV (mean increase 2.9) of previously diag-nosed lymph nodes. Six of them also had an increased SUVmax ofthe primary tumour. However, two patients showed a decreaseof the SUVmax of the primary tumour.

Three patients showed a decreased SUV of pathologic lymphnodes diagnosed at the baseline PET. In all these patients we foundmore than 10% increase of the SUV of the primary tumour and/oran increased number of pathologic lymph nodes. In one patientthe SUVmax remained the same.

At the time of analysis, 15 patients (33.3%) had died within 2–22 months after the start of treatment. The other 30 patients werestill alive (66.7%) at a median follow up time of 14 months. The 1-,2- and 3-year overall survival (OS) rates were 84%, 72%, and 59%,respectively. The mean survival was 18 months (95% CI; 15–21 months).

TNM progression between the two PET-scans and an increasednumber of pathologic lymph nodes were significant prognostic fac-tors for OS (Table 1). For the TNM-progression group the mean sur-vival was 12 months (95% CI; 6–17 months) as compared to20 months for the group without TNM-progression (95% CI; 17–23 months).

The causes of death were not significantly different between thegroup with or without TNM progression (p = 0.35). In the groupwith TNM-progression two patients (17%) died post-operatively.The other five patients (42%) died of progressive disease. For pa-tients without TNM-progression the causes of death were in hospi-tal in two patients (6%), intercurrent disease in one patient (3%)and progressive disease in five patients (15%). Another patient(3%) died because of pulmonary complications and decompensatio

iagnostic and therapeutic PET scans, which were both performed on the integrated

Table 2Characteristics of patients with TNM progression.

No. Age Sex Histology Diagnostic Time interval(days)

DSUVmax

correctedD Number oflymph nodes

Therapeutic

SUVmax

correctedTumourlength (cm)

cN cM Type ofcamera

cN cM Location M

19 64 Male PCC 9.4 c 2 0 IntegratedPET

29 1.3 -1 1 1 Pleural left

30 68 Male AC 11.1 7.0 1 0 HR+ 33 1.7 0 1 1 Thoracicvertebra

46 65 Female PCC 13.6 6.2 1 0 HR+ 49 0.7 0 1 1 Lung(2x)49 63 Male AC 11.3 3.2 0 0 Integrated

PET18 1.4 1 1 0 –

50 70 Male AC 16.0 4.8 2 0 HR+ 29 6.8 0 2 1 Diffusehepatogenic

56 75 Male AC 12.9 10 2 0 IntegratedPET

44 8.3 1 2 1 Lung LBK

68 80 Male AC 42.1 5.4 1 0 IntegratedPET

31 �8.7 4 3 0 –

74 62 Male AC 28.6 6.4 0 0 HR+ 14 �3.0 2 1 1 Lung right72 66 Male AC 13.7 4.0 0 0 HR+ 32 1.7 1 1 0 –51 63 Female AC 46.2 5.7 1 0 HR+ 42 �17.9 1 1 0 –41 64 Male PCC 15.5 4.6 0 0 HR+ 15 �2.3 0 0 0 –80 70 Male AC 14.1 5.5 0 0 HR+ 20 �4.1 1 1 0

286 Early oesophageal tumour progression

cordis within 90 days after surgical resection, which might havebeen treatment related as well.

Discussion

In the present study, tumour progression between baselineFDG-PET and a research/therapeutic FDG-PET/CT, in terms of in-creased tumour length and/or more advanced TNM-stage, wasfound in a substantial portion of the patients (31% and 27%, respec-tively), despite the relatively short interval between the two PETscans (median 22 days).

Studies on repeated PET-CT imaging prior to the neo-adjuvanttreatment are scarce. However, several studies reported on post-neoadjuvant distant metastases, with incidences varying from 8%to 17% [12–16]. Interval distant metastases may be explained bytumour progression during and/or after neo-adjuvant treatment.In the two largest studies by Bruzzi et al. and Blom et al., intervaldistant metastases were described in 8% of the patients. All of thesepatients were diagnosed with distal adenocarcinomas with posi-tive nodal stages, suggesting lymph node involvement to be a riskfactor for the development of interval metastases. Progression ofthe metastatic tumour burden prior to the treatment might alsoexplain the post-neoadjuvantly detected metastases, since the ex-pected systemic effect of the chemotherapy regimen used for theneoadjuvant chemoradiation seems limited [3].

In the present study, newly detected pathologic lymph nodes ordistant metastases were found in pre-treatment in 27% of the pa-tients, after comparison of the 2 PET-scans. However, progressionover time seemed not the only factor for the detection of pre-treat-ment tumour progression. There was also a trend noted for TNMprogression and the type of PET camera used. Despite the use ofa validated and standardized protocol [10] to minimize the vari-ability for both cameras, the spatial resolution of the integratedPET/CT scanner is higher at the visual interpretation settings. Thiscould explain some of the differences in the detectability of smalltumour lesions. Furthermore, additional information of the CT-images will also increase the visualization of both regional lymphnodes and distant metastases [17,18]. However, in the currentstudy, the CT images were available for both the diagnostic andthe therapeutic scans, in comparison with PET images in the caseof possible TNM progression. These findings underline the impor-

tance of the use of ‘state of the art’ techniques and PET scannersto prevent the missing of relevant tumour lesions.

The current study revealed newly detected distant metastasesin six patients (13%), which developed within 49 days. Conse-quently, the treatment intent changed from curative to palliativeintent since metastatic disease implies that curative treatmentwas no longer possible [19]. A curatively intended treatment, con-sisting of neo-adjuvant chemoradiation followed by a surgicalresection, is very intense with the subsequent risk of (post-)opera-tive morbidity and/or mortality, and should therefore be preventedin patients that cannot be cured. Therefore, for the detection ofinterval metastases, it remains important to reevaluate the M-stage before surgical resection.

Newly diagnosed pathologic lymph nodes were found in eightpatients in the current study. Accurate identification of regionalpathologic lymph nodes, and thus adequate inclusion in the grosstumour volume (GTV) in order to ensure optimal dose coverage,is essential for the successful radiotherapy. A treatment plan basedon the diagnostic PET information would have missed these newlyinvolved lymph nodes as detected on the therapeutic PET, whichmight result in a geographic miss [19–22] and might lead to recur-rence or residual tumour [23]. Given the rapid progression found inthe study, it is important to minimize the time between the rele-vant imaging and start of the radiotherapy.

FDG-PET has some limitations regarding its use for the detectionof regional lymph node metastases. Nonspecific inflammation in themediastinum may result in a false-positive uptake of FDG on the PETimages. This seemed the case in two patients in which suspectedlymph nodes at the baseline PET were no longer visible at the secondPET, suggesting temporarily inflammatory involvement. Potentially,this flaw can be avoided by including late time imaging in the proto-col, as it has been shown that in time the uptake in inflammatory le-sions tends to decline, whereas in most tumours the uptake is stillrising after one hour after the injection [24,25]. Furthermore, thespatial resolution of FDG-PET is limited. Luketisch et al. [26] statedthat small regional lymph node metastases with mean greatestdimensions ranging from 2 to 10 mm could not be detected byFDG-PET, while Kato et al. [27] found a threshold size of 6–8 mmfor lymph node metastases. The detection of metastatic diseaseFDG-PET faces the same limitations. Regional or metastatic diseasecould already be present for quite some time before it becomes vis-ible. However, the results of the current study suggest rather fastvisualization and thus the progression of oesophageal cancer.

C.T. Muijs et al. / Radiotherapy and Oncology 106 (2013) 283–287 287

This was confirmed by the survival rates of the group with TNMprogression which were significantly worse as compared to thegroup without TNM progression. In the current study, a total offour patients died in hospital after surgery, and one patient diedbecause of pulmonary complications and decompensatio cordiswithin 90 days after surgical resection, which had its effect onthe survival rates. Although the causes of death between thegroups with or without TNM progression were not significantlydifferent, the percentage of patients who died of post-operativecomplications was slightly higher in the group with TNM progres-sion. It seems reasonable that metastatic tumour progressionmight influence the risk of post-operative complications.

The results of the current study suggest rapid oesophageal tu-mour progression, in terms of an increased tumour length and/oran advanced TNM-stage. Therefore, the interval between relevantimaging and start of the radiotherapy should be minimized, includ-ing re-evaluation if the start of the treatment is delayed. Further-more, to prevent the miss of tumour lesions, ‘state of the art’ PETscanners should be used with high spatial resolution.

References

[1] Geh JL, Bond SJ, Bentzen SM, Glynne-Jones R. Systematic overview ofpreoperative (neoadjuvant) chemoradiotherapy trials in oesophageal cancer:evidence of a radiation and chemotherapy dose response. Radiother Oncol2006;78:236–44.

[2] Smith GL, Smith BD, Buchholz TA, et al. Patterns of care and locoregionaltreatment outcomes in older esophageal cancer patients: the SEER-medicarecohort. Int J Radiat Oncol Biol Phys 2009.

[3] Gebski V, Burmeister B, Smithers BM, Foo K, Zalcberg J, Simes J. Survivalbenefits from neoadjuvant chemoradiotherapy or chemotherapy inoesophageal carcinoma: a meta-analysis. Lancet Oncol 2007;8:226–34.

[4] Eloubeidi MA, Desmond R, Arguedas MR, Reed CE, Wilcox CM. Prognosticfactors for te survival of patients with esophageal carcinoma in the U.S.: theimportance of tumor length and lymph node status. Cancer 2002;95:1434–43.

[5] Griffiths EA, Brummell Z, Gorthi G, Pritchard SA, Welch IM. Tumor length as aprognostic factor in esophageal malignancy: univariate and multivariatesurvival analyses. J Surg Oncol 2006;93:258–67.

[6] Suzuki A, Xiao L, Hayashi Y, et al. Prognostic significance of baseline positronemission tomography and importance of clinical complete response in patientswith esophageal or gastroesophageal junction cancer treated with definitivechemoradiotherapy. Cancer 2011.

[7] Moureau-Zabotto L, Touboul E, Lerouge D, et al. Impact of CT and 18F-deoxyglucose positron emission tomography image fusion for conformalradiotherapy in esophageal carcinoma. Int J Radiat Oncol Biol Phys2005;63:340–5.

[8] Edge SB, Byrd DR, Compton CC, et al. AJCC cancer staging manual. 7th ed. NewYork: Springer; 2009.

[9] Boellaard R, O’Doherty MJ, Weber WA, et al. FDG PET and PET/CT: EANMprocedure guidelines for tumour PET imaging: version 1.0. Eur J Nucl Med MolImaging 2010;37:181–200.

[10] Boellaard R, Oyen WJ, Hoekstra CJ, et al. The Netherlands protocol forstandardisation and quantification of FDG whole body PET studies in multi-centre trials. Eur J Nucl Med Mol Imaging 2008;35:2320–33.

[11] Boellaard R. Standards for PET image acquisition and quantitative dataanalysis. J Nucl Med 2009;50(Suppl. 1):11S–20S [Epub 2009 Apr 20].

[12] Bruzzi JF, Swisher SG, Truong MT, et al. Detection of interval distantmetastases: clinical utility of integrated CT-PET imaging in patients withesophageal carcinoma after neoadjuvant therapy. Cancer 2007;109:125–34.

[13] Cerfolio RJ, Bryant AS, Ohja B, Bartolucci AA, Eloubeidi MA. The accuracy ofendoscopic ultrasonography with fine-needle aspiration, integrated positronemission tomography with computed tomography, and computedtomography in restaging patients with esophageal cancer after neoadjuvantchemoradiotherapy. J Thorac Cardiovasc Surg 2005;129:1232–41.

[14] Flamen P, Van Cutsem E, Lerut A, et al. Positron emission tomography forassessment of the response to induction radiochemotherapy in locallyadvanced oesophageal cancer. Ann Oncol 2002;13:361–8.

[15] Blom RL, Schreurs WM, Belgers HJ, Oostenbrug LE, Vliegen RF, Sosef MN. Thevalue of post-neoadjuvant therapy PET-CT in the detection of intervalmetastases in esophageal carcinoma. Eur J Surg Oncol 2011;37:774–8.

[16] Weber WA, Ott K, Becker K, et al. Prediction of response to preoperativechemotherapy in adenocarcinomas of the esophagogastric junction bymetabolic imaging. J Clin Oncol 2001;19:3058–65.

[17] Yuan S, Yu Y, Chao KS, et al. Additional value of PET/CT over PET in assessmentof locoregional lymph nodes in thoracic esophageal squamous cell cancer. JNucl Med 2006;47:1255–9.

[18] Schreurs LM, Pultrum BB, Koopmans KP, et al. Better assessment of nodalmetastases by PET/CT fusion compared to side-by-side PET/CT in oesophagealcancer. Anticancer Res 2008;28:1867–73.

[19] Graham AJ, Shrive FM, Ghali WA, et al. Defining the optimal treatment oflocally advanced esophageal cancer: a systematic review and decisionanalysis. Ann Thorac Surg 2007;83:1257–64.

[20] Muijs CT, Beukema JC, Pruim J, Mul VE, Groen H, Plukker JT, et al. A systematicreview on the role of FDG-PET/CT in tumour delineation and radiotherapyplanning in patients with esophageal cancer. Radiother Oncol2010;97:165–71.

[21] Muijs CT, Schreurs LM, Busz DM, et al. Consequences of additional use of PETinformation for target volume delineation and radiotherapy dose distributionfor esophageal cancer. Radiother Oncol 2009;93:447–53.

[22] Vrieze O, Haustermans K, De Wever W, et al. Is there a role for FGD-PET inradiotherapy planning in esophageal carcinoma? Radiother Oncol2004;73:269–75.

[23] Cai WJ, Xin PL. Pattern of relapse in surgical treated patients with thoracicesophageal squamous cell carcinoma and its possible impact on targetdelineation for postoperative radiotherapy. Radiother Oncol 2010;96:104–7.

[24] Basu S, Kung J, Houseni M, Zhuang H, Tidmarsh GF, Alavi A. Temporal profile offluorodeoxyglucose uptake in malignant lesions and normal organs overextended time periods in patients with lung carcinoma: implications for itsutilization in assessing malignant lesions. Q J Nucl Med Mol Imaging2009;53:9–19.

[25] Hu Y, Hu C, Zhang H, Ping Y, Chen LQ. How does the number of resected lymphnodes influence TNM staging and prognosis for esophageal carcinoma? AnnSurg Oncol 2010;17:784–90.

[26] Luketich JD, Schauer PR, Meltzer CC, et al. Role of positron emissiontomography in staging esophageal cancer. Ann Thorac Surg 1997;64:765–9.

[27] Kato H, Kuwano H, Nakajima M, et al. Comparison between positron emissiontomography and computed tomography in the use of the assessment ofesophageal carcinoma. Cancer 2002;94:921–8.