Japanese Journal of Clinical Oncology 34:171-175 (2004)
© 2004 Foundation for Promotion of Cancer Research
Magnetic Resonance for T-staging of Nasopharyngeal Carcinoma—The Most Informative Pair of Sequences
1 Department of Radiology and 2 Department of Clinical Oncology, Pamela Youde Nethersole Eastern Hospital, and 3 Department of Radiology, Hong Kong Baptist Hospital, Hong Kong
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
Objective: To evaluate the most informative pair of sequences in magnetic resonance (MR) for T-staging of nasopharyngeal carcinoma (NPC).
Methods: The MR images of 134 patients with newly diagnosed NPC, from 1996 to 2002, were retrospectively reviewed. All the patients were scanned using 1.5 Tesla MR systems. The images of the nasopharynx were reviewed by two qualified radiologists to determine the positive findings and the T-stage by UICC (6th edition) System, using each sequence separately. The T-stage derived from a single MR sequence was then compared with the T-stage based on the five selected sequences to assess the number and percentage of patients who were being understaged. Therefore, the overall percentage accuracy of each single sequence could be determined. A pair of sequences providing information to achieve almost 100% diagnostic accuracy was then derived.
Results: The overall percentage accuracy of five individual sequences of the nasopharynx is as follows: contrast-enhanced (CE) fat suppression (FS) axial T1 (94.8%), CE FS coronal T1 (88.1%), FS axial T2 (85.8%), non-contrast enhanced (NE) axial T1 (78.4%) and non-contrast enhanced (NE) coronal T1 (77.6%). CE FS axial T1 has the best accuracy. All the structures that are missed in CE FS axial T1, which lead to apparent understaging, are appreciated in NE axial T1-weighted images.
Conclusion: Individual sequences supplement each other in the NPC staging. CE FS axial T1 is the most informative individual sequence. Combination of CE FS axial T1 and NE axial T1 of the nasopharynx provides sufficient information to achieve almost 100% diagnostic accuracy in T-staging; therefore, both should be included in the MR-staging protocol.
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
Complex human anatomy, a variety of soft tissues in the head and neck regions and propensity for skull base infiltration in nasopharyngeal carcinoma (NPC) make accurate radiological interpretation of NPC a diagnostic challenge. Apart from the histological diagnosis, accurate radiological staging of the disease is crucial in pretreatment planning and prognosis. The advent of magnetic resonance imaging (MRI), with good tissue contrast and multi-planar capability (1–3), allows greater sensitivity and specificity than computed tomography (CT). A high incidence of NPC has been observed in Hong Kong. Our institution performs MRI staging for all newly diagnosed NPC patients. The aim of this study is to evaluate the most informative individual sequence in magnetic resonance (MR) for T-staging NPC in routine clinical practice, and thereby determine the overall percentage accuracy of each individual sequence. Thus, a pair of sequences that provide information to achieve almost 100% diagnostic accuracy was derived.
|
METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
We retrospectively reviewed 134 patients with newly diagnosed NPC (histologically proven), who were referred by the Department of Clinical Oncology from December 1996 to November 2002. There were 84 males and 50 females, and their age ranged from 17–83 years (mean, 49.5 years). In all patients, the following five sequences of the nasopharynx were obtained: non-contrast enhanced (NE) axial T1, fat suppression (FS) axial T2, NE coronal T1, contrast enhanced (CE) FS axial T1 and CE FS coronal T1-weighted images. Additionally, CE FS axial T1-weighted imaging from the neck down to the supraclavicular fossae was performed on 47 patients. Adjunctive FS sagittal T2 and CE FS sagittal T1-weighted imaging of the clivus was performed on 16 patients. Sagittal T1 localizer images were included in all the patients. All images were obtained from 1.5 Tesla MRI systems (Signa, GE Medical Systems, Milwaukee, Wisconsin, USA or Symphony, Siemens Medical Systems, Erlangen, Germany). A head coil was used in all the sequences. In addition to the head coil, a neck coil was also used in NE coronal T1 and CE FS coronal T1 of the nasopharynx and CE FS axial T1-images of the neck.
The following parameters were used for imaging of the nasopharynx: NE axial T1-images were obtained with SE 450/15, 90, 2 excitations, a 22-cm field of view (FOV), a 256 x 256 frequency matrix, a 3-mm-thick section and a 0.9-mm gap. FS axial T2-images were obtained with TSE 4500/88, 180, 3 excitations, a 22-cm FOV, a 256 x 256 frequency matrix, a 3-mm-thick section and a 0.9-mm gap. NE coronal T1-images were obtained with SE 450/15, 90, 2 excitations, a 21-cm FOV, a 256 x 256 frequency matrix, a 4-mm-thick section and a 1.0-mm gap. CE FS axial T1-images were obtained with SE of 553/17, 90, 1 excitation, a 22-cm FOV, a 256 x 256 frequency matrix, a 3-mm-thick section and a 0.9-mm gap. CE FS coronal T1-images were obtained with SE of 461/17, 90, 1 excitation, a 21-cm FOV, a 256 x 256 frequency matrix, a 4-mm-thick section and a 1.0-mm gap.
The following parameters were used for imaging of the clivus: FS sagittal T2-images were obtained with TSE 4000/105, 180, 3 excitations, a 20-cm FOV, a 256 x 256 frequency matrix, a 3-mm-thick section and a 0-mm gap. CE FS sagittal T1-images were obtained with SE 507/17, 90, 2 excitations, a 22-cm FOV, a 256 x 256 frequency matrix, a 3-mm-thick section and a 0-mm gap. Sagittal T1 localizer images were obtained with SE 450/15, 90, 2 excitations, a 30-cm FOV, a 256 x 256 frequency matrix, a 3-mm-thick section and a 0-mm gap.
The following parameters were used for imaging of the neck down to the supraclavicular region: CE FS axial T1-images were obtained with SE 461/17, 90, 1 excitation, a 22-cm FOV, a 256 x 256 frequency matrix, a 5-mm-thick section and a 1-mm gap.
A 0.1 mmol/Kg of body weight bolus injection of Omniscan (Nycomed Imaging AS, Oslo, Norway) was administered for the post-gadolinium series.
A database was designed with reference to the 2002 American Joint Committee on Cancer (AJCC) (6th edition) criteria for NPC (4) (Table 1). The MRI findings were classified according to the AJCC/UICC criteria for NPC staging, and only those findings concurred by two qualified radiologists (KY, WK) were regarded as positive. The positive findings and the T-stage derived from each sequence were recorded. The T-stages determined by a single sequence were then compared with the T-stages based on the five sequences, which was regarded as a gold standard, to assess the number and percentage of patients being understaged. Therefore, the overall percentage accuracy of individual sequences was determined by percentage of the accepted stage for all T-stages.
|
The sagittal images were not included in this study as they were obtained only for the clivus; therefore, the entire skull base was not included. Nevertheless, the same pair of radiologists gave their consensus on whether the 16 adjunctive sagittal images provided additional information leading to changes in staging.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
The T-stage based on the findings of all five sequences is regarded as a gold standard. Thirty-five patients were in T1, 32 in T2, 36 in T3 and 31 in T4-stage. The overall percentage accuracy of various individual sequences of the nasopharynx is as follows: 94.8% (127/134) for CE FS axial T1, 88.1% (118/134) for CE FS coronal T1, 85.8% (115/134) for FS axial T2, 78.4% (105/134) for NE axial T1 and 77.6% (104/134) for NE coronal T1. The percentage accuracy of the individual sequences in predicting T-stage is summarized in Table 2. CE FS axial T1 has the best accuracy, i.e., same stage as the gold standard. When seven patients (5.2%) who were understaged on the basis of CE FS axial T1-weighted image alone were reviewed, all the missed structures in four of seven patients (all pterygoid plates) were appreciated in NE axial T1-weighted images (Fig. 1a and b). Three of seven patients (two medial pterygoid muscles and one longus capitus muscle) were apparently understaged in CE FS axial T1-weighted imaging because these muscles were misinterpreted as being involved only on viewing a single NE axial T1-weighted sequence (Fig. 2a and b). Therefore, the inaccuracies of CE FS axial T1-weighted imaging were associated with only four patients (3.0%). Although NE axial T1-weighted imaging alone may lead to false positive results, as in these three patients, these findings, which lead to apparent overstaging, are compensated by CE FS axial T1.
|
|
|
If the overall MR stage derived from all the five sequences is subdivided into T1, T2, T3 and T4, the T2 patients are more frequently understaged than the T3 and T4 patients (T1 is not included as T1 disease cannot be understaged).
The stages derived from the five sequences of 16 patients were not affected when the adjunctive FS sagittal T2 and CE FS sagittal T1-weighted images of the clivus were reviewed.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
NPC is a common head and neck neoplasm in Southern China. In Hong Kong, NPC ranked fourth for males and ninth for females in terms of incidence in 1999, and it ranked sixth for males and eleventh for females in terms of cancer deaths (5). In 1999, 1118 cases of newly diagnosed NPC, with a male to female ratio of 2.5 to 1 (5), were recorded. From our observation, NPC has a great propensity for deep infiltration, especially at the base of the skull. Although histological diagnosis is crucial, accurate staging is equally important for treatment planning and long-term survival. Cross-sectional imaging plays a critical role in tumor staging. Good tissue contrast and multi-planar capability of MR enabled us to perform MR staging in all newly diagnosed NPC patients (1–3,6–8). The anatomy of the nasopharynx is complicated and accurate interpretation is a challenge. The objective of this study was to evaluate the most informative individual MR sequence, and thereby derive a pair of sequences that provide sufficient information to achieve almost 100% accuracy in T-staging of NPC. To the best of our knowledge, studies regarding the most informative sequences in MR imaging of NPC patients have not yet been described.
On reviewing the literature, studies comparing different sequences in MR evaluation of head and neck tumors have been described: fast spin echo T2-weighted images with FS and CE FS T1 conventional spin echo sequences (9); CE FS T1-weighted images, conventional spin echo T2-weighted and fast spin echo FS T2-weighted images (10). Vogl et al. (6) reported MR imaging in evaluation of nasopharyngeal tumors by comparing NE and CE axial and coronal T1, axial and coronal proton density T2-weighted images without FS. Other researchers have compared T2-weighted turbo spin echo images with FS and CE T1-weighted spin echo images for NPC without FS in CE series (11) and fast short time inversion recovery images (FSTIR) and CE FS T1-weighted spin echo images (12).
Due to the abundance of fat and complex anatomy of the head and neck, FS MRI is considered to improve the detection and delineation of head and neck lesions (13–16). Although FS is associated with some disadvantages, such as uneven FS in a large field of view and in areas with sharp changes in anatomy, an increase in susceptibility artifact around the air-containing structures and low signal-to-noise ratio, they are outweighed by the advantage of increased contrast (14). Our observation of the nasopharyngeal region reveals that no substantial adverse effect is noted in the FS series. With regard to the studies involving FS (9,10,12), Dubin et al. (9) described that FS T2-weighted images offer better contrast between tumors and adjacent muscle, fat and mucosa as compared with CE FS T1-weighted images. Nevertheless, CE FS T1-weighted image may offer complementary information on the precise characterization of complex tumors such as tumor necrosis (9) and on the potential determination of tumor extent such as perineural spread (9,17) and meningeal infiltration (9) with axial and coronal planes supplementing each other. King et al. (18,19) performed only FS axial T1, without FS in CE coronal and sagittal series. On the other hand, Ross et al. (10) described that CE FS T1-weighted images provided the highest contrast-to-noise ratio for head and neck tumors and abnormal lymph nodes, and suggested that CE FS T1-weighted images and FS T2-weighted images should be included as part of the MR imaging protocol for head and neck tumors. Yamamoto et al. (12) observed that both FSTIR and CE FS T1-weighted sequences showed the same performance for detection of NPC; however, it was useful to combine these two imaging sequences as a pre-treatment evaluation.
In the present study, we compared all five sequences of the nasopharynx, three axial and two coronal planes for evaluation of NPC. We found that the number and percentage of patients who were being understaged in an individual MR sequence, in ascending order of frequency, are CE FS axial T1, CE FS coronal T1, FS axial T2, NE axial T1 and NE coronal T1-weighted images. If the overall MR stages are subdivided into T1, T2, T3 and T4, the number of T2 patients who are being understaged is greater than the stage T3 and T4 patients. Each individual sequence of the nasopharynx shows 100% accuracy in T1-stage patients. Single CE FS axial T1 sequence has 100% accuracy in T4-stage patients. Single coronal sequence is less accurate in NPC staging when compared with single axial sequence, especially in T2-stage. This is because T2-stage structures are not well delineated; therefore, interpretation in the coronal plane alone is more difficult than in the axial plane. The result of the present study demonstrates a significant impact on the staging of T2 patients, because a greater difference is observed in treatment planning and prognosis of these patients if they are being understaged.
Additionally, CE FS axial T1 is the most accurate single sequence (127 patients, 94.8%, with the same stage as the gold standard). All the structures missed by CE FS axial T1 in four of seven patients were pterygoid plates, which were appreciated in NE axial T1. This suggests that pterygoid plate is the most easily misinterpreted skull base structure in single CE FS axial T1 sequence, and loss of marrow fat in the pterygoid plate in T1-weighted images without contrast is more sensitive in detecting tumor involvement of the pterygoid plate. On the other hand, two medial pterygoid muscles and one longus capitus muscle are misinterpreted as tumor involvement on viewing single NE axial T1-weighted sequence alone. This is because the nasopharynx and the adjacent muscles show similar signal intensity in NE axial T1-images. Therefore, the inaccuracies of CE FS axial T1-weighted images should occur in only four patients (3.0%) and CE FS axial T1 should have an accuracy of 97.0%. Although NE axial T1-weighted images alone may lead to false positive results, as in the case of these three patients, these findings that lead to apparent overstaging are compensated by CE FS axial T1.
A suitable protocol in MR staging of NPC in routine clinical practice should provide accurate information for staging and be effective and informative. After evaluation of the most informative individual sequence in MR staging of NPC, it was concluded that CE FS axial T1 and NE axial T1-images should be performed because a combination of these sequences will approach almost 100% diagnostic accuracy.
The order of assessing individual single sequence may affect the outcome. Moreover, misinterpretation in any individual sequence may lead to overstaging. Therefore, these are the drawbacks of our study. In conclusion, different sequences supplement each other in NPC staging. CE FS axial T1 and NE axial T1 provide sufficient information to achieve almost 100% diagnostic accuracy in T-staging. Though FS axial T2 was less informative in T-staging in the present study, FS axial T2 or FSTIR axial were useful for T-staging (10,12). Contrary to King et al. (19), we suggest that supraclavicular fossae should be included in the T-staging protocol because supraclavicular nodes are important prognostic factors (20). However, as the air–tissue contrast in supraclavicular fossae of thin patients may lead to misinterpretation, further investigation needs to be conducted to determine the most suitable sequence for neck nodes. The present study provides useful suggestions for selecting the most effective and informative sequences.
|
FOOTNOTES |
|---|
+ For reprints and all correspondence: John Chan, Department of Radiology, Pamela Youde Nethersole Eastern Hospital, 3 Lok Man Road, Chai Wan, Hong Kong. E-mail: chankamwai{at}sinaman.com
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
1 Ng SH, Chang TC, Ko SF, Yen PS, Wan YL, Tang LM, et al. Nasopharyngeal carcinoma: MRI and CT assessment. Neuroradiology 1997;39:741–6.[CrossRef][Web of Science][Medline]
2 Olmi P, Fallai C, Colagrande S, Giannardi G. Staging and follow-up of nasopharyngeal carcinoma: magnetic resonance imaging versus computerized tomography. Int J Radiat Oncol Biol Phys 1995;32(3):795–800.[CrossRef][Web of Science][Medline]
3 Poon PY, Tsang VH, Munk PL. Tumor extent and T stage of nasopharyngeal carcinoma: a comparison of magnetic resonance imaging and computed tomographic findings. Can Assoc Radiol J 2000;51(5):287–95.[Web of Science][Medline]
4 Greene FL, Page DL, et al. 2002 AJCC Cancer Staging Handbook, 6th edn. Springer-Verlag Publication, New York, 50.
5 Cancer incidence and mortality in Hong Kong in 1998–1999, Hospital Authority 2000, 38.
6 Vogl T, Dresel S, Bilaniuk LT, Grevers G, Kang K, Lissner J. Tumors of the nasopharynx and adjacent areas: MR imaging with Gd-DTPA. AJR Am J Roentgenol 1990;154:585–92.
7 Chong VFH, Fan YF. Skull base erosion in nasophayngeal carcinoma: detection by CT and MRI. Clin Radiol 1996;51:625–31.[CrossRef][Web of Science][Medline]
8 Chong VFH, Fan YF, Khoo JBK. Nasopharyngeal carcinoma with intracranial spread: CT and MR characteristics. J Comput Assist Tomogr 1996;20(4):563–9.[CrossRef][Web of Science][Medline]
9 Dubin MD, Teresi LM, Bradley WG Jr., Jordan JE, Pema PJ, Goergen SK, et al. Conspicuity of tumors of the head and neck on fat-suppressed MR images: T2-weighted fast-spin-echo versus contrast-enhanced T1-weighted conventional spin-echo sequences. AJR Am J Roentgenol 1995;164:1213–21.
10 Ross MR, Schomer DF, Chappell P, Enzmann DR. MR imaging of head and neck tumors: comparison of T1 weighted contrast-enhanced fat suppressed images with conventional T2 weighted and fast spin-echo T2 weighted images. AJR 1994;163:173–8.
11 King AD, Lam WW, Leung SF, Chan YL, Metreweli C. Comparison of T2 weighted fat suppressed turbo spin echo and contrast enhanced T1 weighted spin echo MRI in nasopharyngeal carcinoma. Brit J Radiol 1997;70:1208–14.[Abstract]
12 Yamamoto S, Takano H, Motoori K, Ueda T, Ikeda M, Kimura S, et al. Detection of nasopharyngeal carcinoma: fast short time inversion recovery images compared with fat suppression, contrast enhanced T1 weighted spin echo images. Brit J Radiol 2001;74:805–10.
13 Tien RD, Hesselink JR, Chu PK, Szumowski J. Improved detection and delineation of head and neck lesions with fat suppression spin-echo MR imaging. AJNR 1991;12:19–24.[Abstract]
14 Tien RD, Robbins KT. Correlation of clinical, surgical, pathologic and MR fat suppression results for head and neck cancer. Head Neck 1992;14:278–84.[Web of Science][Medline]
15 Barakos JA, Dillon WP, Chew WM. Orbit, skull base and pharynx: contrast-enhanced fat suppression MR imaging. Radiology 1991;179:191–8.
16 Tien RD. Fat suppression MR imaging in neuroradiology: techniques and clinical application. AJR 1991;158:369–79.
17 Su CY, Lui CC. Perineural invasion of the trigeminal nerve in patients with nasopharyngeal carcinoma. Imaging and clinical correlations. Cancer 1996;78:2063–9.[CrossRef][Web of Science][Medline]
18 King AD, Lam WW, Leung SF, Chan YL, Teo P, Metreweli C. MRI of local disease in nasopharyngeal carcinoma: tumor extent vs tumor stage. Brit J Radiol 1999;72:734–41.[Abstract]
19 King AD, Ahuja AT, Leung SF, Lam WWM, Teo P, Chan YL, et al. Neck node metastasis from nasopharyngeal carcinoma: MR imaging of patterns of disease. Head Neck 2000;22:275–81.[CrossRef][Web of Science][Medline]
20 Som PM, Curtin HD, Mancuso AA. An imaging-based classification for the cervical nodes designed as an adjunct to recent clinically based nodal classifications. Arch Otolaryngol Head Neck Surg 1999;125:388–96.
Received November 2, 2003; accepted January 27, 2004
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  M. Comoretto, L. Balestreri, E. Borsatti, M. Cimitan, G. Franchin, and M. Lise Detection and Restaging of Residual and/or Recurrent Nasopharyngeal Carcinoma after Chemotherapy and Radiation Therapy: Comparison of MR Imaging and FDG PET/CT Radiology, October 1, 2008; 249(1): 203 - 211. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. M. Lell, C. Gmelin, C. Panknin, K. T. Eckel, M. Schmid, W. A. Bautz, and H. Greess Thin-Slice MDCT of the Neck: Impact on Cancer Staging Am. J. Roentgenol., March 1, 2008; 190(3): 785 - 789. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A Siddiqui and S E J Connor Imaging of the pharynx and larynx Imaging, March 1, 2007; 19(1): 83 - 103. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||