|Year : 2015 | Volume
| Issue : 4 | Page : 746-751
Improving the accuracy of target volume delineation by combined use of computed tomography, magnetic resonance imaging and positron emission tomography in head and neck carcinomas
Deepika Chauhan1, Sheh Rawat1, Manoj K Sharma1, Parveen Ahlawat1, Manoj Pal1, Gourav Gupta1, Abhinav Dewan1, Manoj Gupta1, Shelly Sharma2, Chandragouda Dodagoudar3, Anjali Pahuja1, Swarupa Mitra1, Surender K Sharma1
1 Department of Radiation Oncology, Rajiv Gandhi Cancer Institute and Research Centre, New Delhi, India
2 Department of Radiodiagnosis, Rajiv Gandhi Cancer Institute and Research Centre, New Delhi, India
3 Department of Medical Oncology, Rajiv Gandhi Cancer Institute and Research Centre, New Delhi, India
|Date of Web Publication||15-Feb-2016|
1 B/6, Rang Rasayan Apartment, Sector 13, Rohini, New Delhi - 110 085
Source of Support: None, Conflict of Interest: None
Objective: Conformal radiation therapy mandates accurate delineation of target volumes, which requires incorporation of modern imaging modalities like magnetic resonance imaging (MRI) and positron emission tomography (PET) in addition to conventionally used computed tomography (CT). This can resolve discrepancies in target delineation in head and neck carcinomas resulting in better local control. We hereby report the comparison of Gross Tumor Volumes (GTVs) (primary) drawn using PET, CT and MRI and their concordance indices.
Methods and Material: Twenty five patients with head and neck cancer were taken into this study. MRI, PET and CT planning scans were done as per standard guidelines. Three sets of primary GTVs namely GTV- PET, GTV-CT and GTV-MRI were contoured on fused images. All the three volumes and concordances among the volumes were analyzed.
Result: The mean GTV-CT, GTV-PET and GTV-MRI volumes were 29.65 cc ± 31.27, 32.05 cc ± 33.75 and 24.85 cc ± 25.28 respectively. There was a significant difference in the GTV-MRI & GTV-CT volumes (P = 0.023) and GTV-PET & GTV-MRI volumes (P = 0.049). However, there was no significant difference in the GTV-PET & GTV-CT volume (P = 0.468). The mean CI (PET-MRI), CI (CT-MRI) and CI (PET-CT) was 0.42, 0.46 and 0.47 respectively, which depicts a moderate concordance.
Conclusion: PET and MRI are useful imaging tools in head and neck malignancies and should be used in conjunction with CT scan for improved target volume delineation.
Keywords: Concordance index, fusion, head and neck carcinomas
|How to cite this article:|
Chauhan D, Rawat S, Sharma MK, Ahlawat P, Pal M, Gupta G, Dewan A, Gupta M, Sharma S, Dodagoudar C, Pahuja A, Mitra S, Sharma SK. Improving the accuracy of target volume delineation by combined use of computed tomography, magnetic resonance imaging and positron emission tomography in head and neck carcinomas. J Can Res Ther 2015;11:746-51
|How to cite this URL:|
Chauhan D, Rawat S, Sharma MK, Ahlawat P, Pal M, Gupta G, Dewan A, Gupta M, Sharma S, Dodagoudar C, Pahuja A, Mitra S, Sharma SK. Improving the accuracy of target volume delineation by combined use of computed tomography, magnetic resonance imaging and positron emission tomography in head and neck carcinomas. J Can Res Ther [serial online] 2015 [cited 2019 Sep 15];11:746-51. Available from: http://www.cancerjournal.net/text.asp?2015/11/4/746/163679
| > Introduction|| |
Head and neck cancers (HNCs) are the sixth most common cancers worldwide causing 300,000 deaths annually.  It accounts for 3% of all cancers in the developed countries like USA and 30% of all cancers in developing countries such as India. Radiation therapy along with concurrent chemotherapy is the standard treatment for most of the locally advanced head and neck squamous cell carcinoma. Computed tomography (CT) is the conventionally used imaging modality to delineate the tumor. Increasing role of positron emission tomography (PET) and magnetic resonance imaging (MRI) in delineation of accurate extent of disease is in vogue in this era of highly precise radiotherapy, which leads to better local control and quality of life due to dose intensification to gross tumor volume (GTV) and lesser dose to the surrounding normal tissues. , MRI has advantages over CT, having superior soft tissue contrast and fewer dental artifacts, but it cannot distinguish between the tumor and peritumoral edema. ,, Molecular imaging can pinpoint the tumor territory as well as segregate the tumor edema. ,,, Nowadays, improved computer algorithms can provide a better fusion of PET, CT, and MRI. We conducted a study to compare the primary GTV volumes drawn individually using PET, MRI, and CT scan, and also calculated and compared the concordance between these imaging modalities in head and neck squamous cell carcinoma.
| > Materials and methods|| |
Twenty-one patients with head and neck squamous cell carcinomas were recruited and treated with concurrent chemoradiotherapy between May 2011 and March 2012. The study was approved by the Institutional Review Board. The patient characteristics are shown in [Table 1].
Methodology for patient setup
Patients were taken in the CT simulator room on an immobilization board with both arms by the side. Orfit cast was made in the treatment position and planning scan with intravenous contrast was done with 5 mm cut sections from the vertex of the skull to the fourth thoracic vertebra. Customized head support cushions and laser position system were used to achieve maximum reproducibility in position.
Standard procedure for positron emission tomography/computed tomography
PET-CT scans were performed using full ring dedicated PET scanner (Siemens Biograph scanner, Lutetium Oxyorthosilicate crystal based 40 slice scanner) in 3D mode. Non-contrast CT scan (120 kV, 80 mA) was performed for attenuation correction and anatomical localization. Standard whole body PET-CT scan was acquired from the base of the skull to mid-thigh with hands in upright position. The acquisition time was 120-180 s/bed position.
After completion of the diagnostic scan, the table couch was changed to flat table couch. The patients were immobilized with orfit cast and head rest and after proper alignment with the help of lasers, planning scan was done with 3 mm cut sections from the vertex of the skull to the fourth thoracic vertebra. The regional scan was taken with the same set up criteria as that of planning CT scan.
Standard procedures for magnetic resonance imaging scan
MRI scans were performed using a standard procedure. MRI Scans were acquired on (1.5 T Siemens Magnetom Avanto scanner 26647). Patients were aligned straight with the orfit cast and neck matrix coil. The positional replication was done. Gadolinium contrast (1.5-2 ml/kg body weight) was given intravenously during the scan. The extent of scan and number of slices were kept same as that of planning CT scan. The t2-weighted sequence in MRI was used for fusion though the relevant information obtained from pre- and post-contrast T1-weighted sequences was also incorporated.
All data sets were sent by way of Digital Imaging and Communications in Medicine to SomaVision v10 (Varian Medical Systems, Palo Alto, CA) workstation and various target volumes and organs at risk (OARs) were contoured.
Fusion and delineation of target volumes
The contrast-enhanced CT simulation scans were fused with MRI and PET-CT scans. The fusion between the various datasets (CT/MRI and CT/PET) was done in ECLIPSE version 10 (Varian Medical Systems, Palo Alto, CA) using automatic image registration algorithm. Once we had the automatic registered sequences, a manual fine adjustment was also done to ensure the accuracy of the fusion. We had excluded those cases from the study wherein the fusion was suboptimal despite these measures. After fusion, GTV-CT was contoured on contrast-enhanced CT scan and information of endoscopy and physical examination was also taken into account. In addition, we have also incorporated the information from the standard contouring guidelines. , Similarly, GTV-MRI and GTV-PET were contoured on the fused images [Figure 1]a-c]. The contours drawn on a particular modality were not referred to while contouring on a different modality. This was done to reduce the bias in contouring.
|Figure 1: (a) Oropharyngeal lesion seen on computed tomography. (b) Same lesion appreciated on magnetic resonance imaging but with reduced area. (c) Positron emission tomography image of the same patient but demonstrating larger area of uptake|
Click here to view
We have taken 40% of the SUVmax as a reference for contouring GTV on PET. , The air or bone, if encompassed in the 40% SUVmax, was edited out from the contour. The volumes once drawn were approved by a senior radiation oncologist and then were validated by a qualified radiologist and nuclear medicine physician. The fusion studies harboring a mismatch of more than 3 mm were excluded from the study.
Using Boolean operator, which is an inbuilt application in the ECLIPSE version 10 contouring system, common and composite volumes were generated individually for CT/MRI, PET/MRI, and PET/CT. In order to know the similarity and harmony between GTVs drawn on different imaging modalities, concordance index (CI) was calculated. The ratio of the common and composite volume is designated as the CI between the chosen modalities [Figure 2]a]. If A and B denote two different imaging modalities, the CI  between them will be
|Figure 2: (a) Concordance index calculated by dividing common volume by composite volume. (b) Imaging modalities depicting poor concordance. (c) Moderate concordance. (d) Good concordance|
Click here to view
We, therefore, had three sets of volumes and three sets of CIs for comparison. All the volumes were measured in cc by an inbuilt system application.
Continuous variables such as GTV volumes and CIs were presented as mean ± standard deviation (SD). A Pearson product-moment correlation was run to determine the relationship between the three sets of GTV and CI. Paired t-test was run for volumetric comparison between the GTV volumes and CIs. Statistical analysis was performed using the SPSS, version 19.0 for Windows software (SPSS, Chicago, IL, USA). For all statistical tests, P values smaller than 0.05 were considered as significant.
| > Results|| |
The mean GTV-CT, GTV-PET, and GTV-MRI volumes were 29.65 cc ± 31.27 (range 1.58-115.33 cc), 32.05 cc ± 33.75 (range 1.48-108.20 cc), and 24.85 cc ± 25.28 (range 0.90-92.18 cc), respectively [Table 2]. A Pearson product-moment correlation was run to determine the relationship between (GTV-CT and GTV-MRI), (GTV-PET and GTV-MRI), and (GTV-CT and GTV-PET). The data showed no violation of normality, linearity or homoscedasticity. There was a very strong, positive correlation between the above mentioned sets of volumes which was also statistically significant (r = 0.973, P = 0.000; r = 0.897, P = 0.000; and r = 0.898, P = 0.000, respectively) [Figure 3]a-c]. This indicates the consistency in the delineation of target volumes on all the imaging modalities used.
|Figure 3: (a) A simple scatter plot showing relationship between GTV-CT and GTV-PET volumes. (b) GTV-CT and GTV-MRI volumes. (c) GTV-PET and GTV-MRI volumes|
Click here to view
Paired t-tests were run for volumetric comparison between the GTV volumes. There was a significant difference in the GTV-CT (M = 29.65 cc, SD ± 31.27) and GTV-MRI (M = 24.85 cc, SD ± 25.28); t(20) = −2.471, P = 0.023, and in the GTV-PET (M = 32.05 cc, SD ± 33.75) and GTV-MRI (M = 24.85 cc, SD ± 25.28); t(20) = 2.096, P = 0.049, thus a mean decrease in GTV volume by 4.79 cc on MRI and a mean increase in GTV volume by 7.20 cc on PET. However, there was no statistically significant difference in the GTV-PET (M = 32.05 cc, SD = ±33.75) and GTV-CT (M = 29.65 cc, SD = ±31.27); t(20) = 2.096, P = 0.468. The results suggest that if GTV is delineated on three different modalities, the volumes differed significantly, when compared on CT and MRI, and MRI and PET; however volumes did not differ significantly on CT and PET.
The mean CI (PET-CT), (CT-MRI), and (PET-MRI), were 0.47 SD ± 0.23, 0.42 SD ± 0.17, and 0.46 SD ± 0.15, respectively. Paired t-tests were used to compare all the three CIs. There was no significant difference between CI (CT-MRI) versus (PET-CT), CI (PET-CT) versus (PET-MRI), and between CI (PET-MRI) versus (CT-MRI), P = 0.964, 0.265, and 0.193, respectively [Table 3]. The mean Dice similarity coefficient (PET-CT), (CT-MRI) and (PET-MRI), were 0.61 SD ± 0.23, 0.63 SD ± 0.23, and 0.57 SD ± 0.16, respectively. Paired t-tests were used to compare all the three DSCs. There was no significant difference between DSC (CT-MRI) versus (PET-CT), DSC (PET-CT) versus (PET-MRI), and between CI (PET-MRI) versus (CT-MRI), P = 0.608, 0.438, and 0.083, respectively. This signifies the very importance of each individual modality used and emphasizes their combined use for precise delineation.
|Table 3: Comparison of concordance indices between paired imaging modalities |
Click here to view
| > Discussion|| |
It has been observed that precision in the delineation of the target volume is of utmost importance as IMRT is the standard technique for radiation delivery for HNC in the present era.
The emerging popularity of IMRT in HNCs in recent years is due to its ability to achieve much better dose conformity than conventional radiotherapy techniques. , There is significant dose reduction to the surrounding critical structures while simultaneously escalating the dose to the GTV.  It also has the potential to prescribe different doses to multiple target volumes with the very accurate delivery of radiation dose due to onboard imaging capability. This accuracy in dose delivery and sparing of surrounding normal tissue has resulted in improved loco-regional control probability with reduced morbidity. However, to avail these advantages of IMRT, a precise and correct delineation of targets and OARs is mandatory. The interobserver variability in head and neck IMRT is well known. Various authors have reported magnitudes of both inter and intraobserver variability in multiple studies. In order to overcome variability in target volume measurement, several researchers have developed semi-automated or automated systems for tumor boundary outlining and target volume measurement. Even these techniques have been criticized for their unproven validity. In recent years, computer-based automated or semi-automated tumor segmentation has been developed with an aim to minimize intraobserver and interobserver variability.
There have been many studies emphasizing the incorporation of various imaging modalities for target delineation in HNCs. 
CT scan presently is the basic imaging modality for conformal radiation treatment planning purposes as it provides intrinsic information on the electron density of various tissues which in turn is used by various dose calculation algorithms. In addition to it, CT scans are widely available and lack geometric distortion of the images. MRI is now frequently being incorporated in image segmentation as it provides superior soft tissue contrast and has fewer dental inlay artifacts as compared with CT.  The typical disadvantages of MRI include its inferior spatial accuracy and longer scan time which can introduce patient motion artifacts. In a study of advanced HNC patients, it was observed that CT displayed 30% larger volumes than MRI.  Our study depicts similar results showing 20% larger CT volumes compared to MRI. The explanation for such results could be due to the inclusion of the areas which are doubtful on CT scan in the target volume and lesser inter and intraobserver variations in MRI, which could lead to the exclusion of such areas. In a study by Geets et al.  MRI and CT were compared for the delineation of pharyngeal-laryngeal tumors and concluded that MRI did not show a clinical advantage over CT in terms of volume determination and interobserver variability. Therefore, a need arises of a molecular imaging modality such as PET, which can further provide useful information on target delineation.
Molecular imaging modalities use various physiological alterations in a cancer cell as a target. The cancer cell has high glycolytic rate as compared to the normal cell. Being a glucose marker and its availability, 18-fluoro-deoxy-glucose (18-FDG) is the most commonly used PET tracer to identify and localize malignant cells. FDG PET-CT also reduces the interobserver variability in GTV delineation and identifies the necrotic parts of the GTV potentially requiring an additional radiation dose. However, there are few objections to the routine use of this modality as there can be false positivity due to inflammation, limited spatial resolution, and lack of a standard method for segmentation. 
The method of tumor delineation also has an impact on the GTV volume. Schinagl et al.  in their study proposed five different GTVs, based on visual delineation (GTVvis), threshold based (GTV40% and GTV50%), SUV based (GTVsuv) and signal background ratio based (GTVsbr). All the methods showed inconsistencies. SBR being the most and applying an isocontour of SUV of 2.5 was least preferable for GTV delineation. In our study, we have used 40% of the SUVmax as the segmentation method. A fixed threshold that is, 40% SUVmax is suitable for phantom based experiments using symmetrical volumes with homogenous activity and a sharp demarcation from the background activity, but in our set of patients with asymmetrical volumes and heterogeneous distribution of activity within the tumor, its reliability is questionable. In addition, this method is not dependent on background activity which is one more limitation of using this as a segmentation tool.
The volume of GTV can increase or decrease based upon the imaging modality (PET/CT) used for delineation. Daisne et al.  showed a reduction of the GTV volume using 18F-FDG PET. The surgical specimen of the tumor corresponded most closely with the PET-based volume and rest all other modalities (CT, MRI) overestimated the extent of the tumor. In another study by Riegel et al.,  the GTV-PET volumes on an average were larger than the corresponding CT based volumes. In our study also, the GTV-PET was larger than the GTV-CT in more than 60% of cases. This may have been secondary to presence of metabolically active areas which might appear as edema/inflammation on other imaging modalities, or PET avid areas such as the tonsils, base of tongue, muscles of mastication, thyroid gland, and salivary glands.
To further analyze the agreement between the volumes drawn with the help of three imaging modalities, CIs were calculated as mentioned previously. Concordance between the two imaging modalities signifies the equivalence in the localization of a particular target. The basic criteria to be fulfilled before calculating concordance are to maintain the positional replication and to ensure a good fusion of PET, MRI, and CT scan images. To achieve the same, consistent position and immobilization techniques were used while doing individual scans and also the time interval for doing these scans was not more than 2 days. In a similar study,  there was a longer time period between the scans.
The concordance can be poor, moderate or good depending upon the ratio of the overlap between the common and composite volume [Figure 2]b-d]. However, not much has been discussed in the literature regarding the same. We have devised a formulary to grade concordance on a scale of 0 to 1, wherein 0 denotes no agreement between the target delineation by two individual modalities used, and 1 signifies the perfect overlap. Based on this scale, we have three sets of values ranging from 0 to 0.33, 0.34 to 0.67, and 0.68 to 1, denoting poor, moderate and good CI, respectively. A good concordance signify the perfect alignment and symmetry between the two volumes as depicted by the two different imaging modalities used, hence it can be inferred that the information provided by individual modality is adequate, however, a poor concordance merits the integral use of both the imaging modalities. In a study by Thiagarajan et al.,  moderate concordance (0.54, 0.55, and 0.62) was seen for primary GTV between the three imaging modalities. Our study similarly outlines a moderate CI between the imaging modalities used for target delineation, reinforcing their combined use.
The shortcomings of the present study include the lack of histopathological correlation, limited recruitment of patients and a possible introduction of error during fusion of images.
Our study emphasizes the importance of integration of all the modalities (PET, CT, and MRI) to reduce the chances of uncertainties in target volume delineation in HNCs and thereby to effectively eliminate the risk of local recurrence.
| > Conclusion|| |
It is difficult to ascertain the impact of a particular imaging modality in the delineation of target volumes. PET and MRI have shown promising results in identifying the tumor in situations where the use of CT as the sole modality might have introduced errors. Our study revealed that CT, PET, and MRI are complementary to each other, at the same time revealing specific information attributed to their sole use. The use of a single imaging modality for target volume delineation is inadequate with limitations pertaining to the modality used. Combined use of imaging modalities indeed enhances the delineation, and one imaging modality compensates for the lacunae associated with the other, resulting in superior delineation. To substantiate, we recommend the usage of one additional imaging modality that is, either MRI or PET to complement and validate the CT scan, however one should not forget the value of good physical examination in delineation of target volume.
So, we recommend the use of information provided by each modality in the delineation of the target volume. However, further studies are needed to substantiate the data provided by us and also to know its impact on local control. One should also keep in mind the cost-effectiveness and logistics associated with the combined usage of these imaging modalities.
Financial support and sponsorship
Conflicts of interest
There are no conflicts of interest.
| > References|| |
Boyle P, Levin B. World Cancer Report 2008. Lyon: International Agency for Research on Cancer; 2008.
Hunt MA, Zelefsky MJ, Wolden S, Chui CS, LoSasso T, Rosenzweig K, et al.
Treatment planning and delivery of intensity-modulated radiation therapy for primary nasopharynx cancer. Int J Radiat Oncol Biol Phys 2001;49:623-32.
Kam MK, Chau RM, Suen J, Choi PH, Teo PM. Intensity-modulated radiotherapy in nasopharyngeal carcinoma: Dosimetric advantage over conventional plans and feasibility of dose escalation. Int J Radiat Oncol Biol Phys 2003;56:145-57.
Castelijns JA, Gerritsen GJ, Kaiser MC, Valk J, van Zanten TE, Golding RG, et al.
Invasion of laryngeal cartilage by cancer: Comparison of CT and MR imaging. Radiology 1988;167:199-206.
Becker M, Zbären P, Casselman JW, Kohler R, Dulguerov P, Becker CD. Neoplastic invasion of laryngeal cartilage: Reassessment of criteria for diagnosis at MR imaging. Radiology 2008;249:551-9.
Castelijns JA, van den Brekel MW, Hermans R. Imaging of the larynx. Semin Roentgenol 2000;35:31-41.
Wang D, Schultz CJ, Jursinic PA, Bialkowski M, Zhu XR, Brown WD, et al.
Initial experience of FDG-PET/CT guided IMRT of head-and-neck carcinoma. Int J Radiat Oncol Biol Phys 2006;65:143-51.
Van Baardwijk A, Baumert BG, Bosmans G, van Kroonenburgh M, Stroobants S, Gregoire V, et al.
The current status of FDG-PET in tumour volume definition in radiotherapy treatment planning. Cancer Treat Rev 2006;32:245-60.
Guido A, Fuccio L, Rombi B, Castellucci P, Cecconi A, Bunkheila F, et al.
Combined 18F-FDG-PET/CT imaging in radiotherapy target delineation for head-and-neck cancer. Int J Radiat Oncol Biol Phys 2009;73:759-63.
Schwartz DL, Ford E, Rajendran J, Yueh B, Coltrera MD, Virgin J, et al.
FDG-PET/CT imaging for preradiotherapy staging of head-and-neck squamous cell carcinoma. Int J Radiat Oncol Biol Phys 2005;61:129-36.
Nuyts S. Defining the target for radiotherapy of head and neck cancer. Cancer Imaging 2007;7:S50-5.
Rasch C, Steenbakkers R, van Herk M. Target definition in prostate, head, and neck. Semin Radiat Oncol 2005;15:136-45.
Bradley J, Thorstad WL, Mutic S, Miller TR, Dehdashti F, Siegel BA, et al.
Impact of FDG-PET on radiation therapy volume delineation in non-small-cell lung cancer. Int J Radiat Oncol Biol Phys 2004;59:78-86.
Ciernik IF, Huser M, Burger C, Davis JB, Szekely G. Automated functional image-guided radiation treatment planning for rectal cancer. Int J Radiat Oncol Biol Phys 2005;62:893-900.
Van Baardwijk A, Bosmans G, Boersma L, Buijsen J, Wanders S, Hochstenbag M, et al.
PET-CT-based auto-contouring in non-small-cell lung cancer correlates with pathology and reduces interobserver variability in the delineation of the primary tumor and involved nodal volumes. Int J Radiat Oncol Biol Phys 2007;68:771-8.
Eisbruch A, Dawson LA, Kim HM, Bradford CR, Terrell JE, Chepeha DB, et al.
Conformal and intensity modulated irradiation of head and neck cancer: The potential for improved target irradiation, salivary gland function, and quality of life. Acta Otorhinolaryngol Belg 1999;53:271-5.
Wu Q, Manning M, Schmidt-Ullrich R, Mohan R. The potential for sparing of parotids and escalation of biologically effective dose with intensity-modulated radiation treatments of head and neck cancers: A treatment design study. Int J Radiat Oncol Biol Phys 2000;46:195-205.
Fogliata A, Bolsi A, Cozzi L, Bernier J. Comparative dosimetric evaluation of the simultaneous integrated boost with photon intensity modulation in head and neck cancer patients. Radiother Oncol 2003;69:267-75.
Nishioka T, Shiga T, Shirato H, Tsukamoto E, Tsuchiya K, Kato T, et al.
Image fusion between 18FDG-PET and MRI/CT for radiotherapy planning of oropharyngeal and nasopharyngeal carcinomas. Int J Radiat Oncol Biol Phys 2002;53:1051-7.
Rasch C, Keus R, Pameijer FA, Koops W, de Ru V, Muller S, et al.
The potential impact of CT-MRI matching on tumor volume delineation in advanced head and neck cancer. Int J Radiat Oncol Biol Phys 1997;39:841-8.
Geets X, Daisne JF, Arcangeli S, Coche E, De Poel M, Duprez T, et al.
Inter-observer variability in the delineation of pharyngo-laryngeal tumor, parotid glands and cervical spinal cord: Comparison between CT-scan and MRI. Radiother Oncol 2005;77:25-31.
Paulino AC, Koshy M, Howell R, Schuster D, Davis LW. Comparison of CT- and FDG-PET-defined gross tumor volume in intensity-modulated radiotherapy for head-and-neck cancer. Int J Radiat Oncol Biol Phys 2005;61:1385-92.
Schinagl DA, Vogel WV, Hoffmann AL, van Dalen JA, Oyen WJ, Kaanders JH. Comparison of five segmentation tools for 18F-fluoro-deoxy-glucose-positron emission tomography-based target volume definition in head and neck cancer. Int J Radiat Oncol Biol Phys 2007;69:1282-9.
Daisne JF, Duprez T, Weynand B, Lonneux M, Hamoir M, Reychler H, et al.
Tumor volume in pharyngolaryngeal squamous cell carcinoma: Comparison at CT, MR imaging, and FDG PET and validation with surgical specimen. Radiology 2004;233:93-100.
Riegel AC, Berson AM, Destian S, Ng T, Tena LB, Mitnick RJ, et al.
Variability of gross tumor volume delineation in head-and-neck cancer using CT and PET/CT fusion. Int J Radiat Oncol Biol Phys 2006;65:726-32.
Thiagarajan A, Caria N, Schöder H, Iyer NG, Wolden S, Wong RJ, et al.
Target volume delineation in oropharyngeal cancer: Impact of PET, MRI, and physical examination. Int J Radiat Oncol Biol Phys 2012;83:220-7.
[Figure 1], [Figure 2], [Figure 3]
[Table 1], [Table 2], [Table 3]