- Email: [email protected]
Observation of tumour thickness and resection margin at surgical excision of primary oral squamous cell carcinoma—assessment by ultrasound
Int. J. Oral Maxillofac. Surg. 2006; 35: 324–331 doi:10.1016/j.ijom.2005.07.019, available online at http://www.sciencedirect.com
Clinical Paper Head and Neck Oncology
Observation of tumour thickness and resection margin at surgical excision of primary oral squamous cell carcinoma— assessment by ultrasound
A. K. Songra, S. Y. Ng, P. Farthing, I. L. Hutchison, P. F. Bradley Maxillofacial Unit, Royal London Hospital, Barts and The London NHS Trust, London E1 2BB, UK
A. K. Songra, S. Y. Ng, P. Farthing, I. L. Hutchison, P. F. Bradley: Observation of tumour thickness and resection margin at surgical excision of primary oral squamous cell carcinoma—assessment by ultrasound. Int. J. Oral Maxillofac. Surg. 2006; 35: 324–331. # 2005 International Association of Oral and Maxillofacial Surgeons. Published by Elsevier Ltd. All rights reserved. Abstract. Tumour thickness and the status of resection margins are of prognostic significance in the treatment of oral cancer. In a single blind prospective study, 14 patients with biopsy proven oral squamous cell carcinoma had intraoral ultrasound imaging done preoperatively to measure tumour thickness, and intraoperatively to measure the deep surgical margin half way during resection. The cut surface was demonstrated on ultrasound by placing a metal, ultrasoundreflective, retractor into the surgical cut. The ultrasound measurements were compared to the subsequent histological measurements. Using the threshold of 5 mm as indicator of margin clearance, there was agreement in 10 out of 14 cases between ultrasound and histology. Ultrasound detection of close surgical margins had a sensitivity of 83% and a specificity of 63%. For preoperative tumour thickness measurement, ultrasound imaging showed a high degree of correlation with histology (Pearson correlation coefficient = 0.95, P < 0.01). This original paper demonstrates that high resolution ultrasound imaging applied intraorally is a reliable tool in objectively assessing both the tumour thickness and the surgical margin clearance at the time of surgery.
Among the large number of prognostic factors demonstrated to be of importance in the treatment of oral squamous cell carcinoma are tumour thickness and the status of the resection margins. Achieving tumour clearance at the primary site 0901-5027/040324 + 08 $30.00/0
is a major problem but a very important aspect of cancer surgery. It has been shown by numerous studies that patients demonstrating invasive carcinoma at resection margins have a higher incidence of loco-regional recurrence and
Key words: tumour thickness; margin clearance; oral cancer; intraoral ultrasound imaging; intraoperative ultrasonography; intraoperative guidance. Accepted for publication 26 July 2005 Available online 13 December 2005
reduced survival rate12. Controversy exists regarding the value of postoperative radiotherapy following incomplete excision12,22. Complete removal of the primary tumour at the first attempt will obviate the need for adjuvant
# 2005 International Association of Oral and Maxillofacial Surgeons. Published by Elsevier Ltd. All rights reserved.
Resection margin by ultrasound radiotherapy in most cases and maximise prognosis. Visual inspection and palpation at the time of surgery in addition to a number of pre-operative modes of imaging all have their limitations in ensuring complete resection of a tumour mass. Frozen section control although widely used has its problems3,14,15. It cannot be used to study an entire margin and cannot demonstrate the amount of clearance. There are difficulties in deciding where to take the sample and in localising the biopsy site to the specimen. Processing samples is very time-consuming and may preclude its routine use. One of the current limitations of achieving margin clearance is the lack of an imaging technique to measure the thickness of the primary tumour. Clinical judgement has been shown to be unreliable, as evidenced by the number of resections that are reported with involved margins10,12. Imaging techniques such as CT and MRI do not have the resolution to demonstrate the thickness of primary oral cancer accurately17. Lack of pre-operative information regarding tumour thickness leads not only to possibly inadequate resection, but increases the risk of local recurrence and reduces the survival rate12,18. High resolution diagnostic ultrasound imaging is becoming well established in the field of head and neck oncology5,13,21. There are many studies describing the ultrasound features of metastatic cervical lymph nodes23, but not of the primary site. Transcutaneous extra-oral (through skin and muscles of submental region) ultrasound imaging of the tongue has been carried out in some early studies2. However, such an extra-oral approach can only measure approximate thickness and only when the tumour is large. Transcutaneous ultrasonography is considered inferior to intra-oral ultrasonography8. With improvements in imaging technology and availability of high frequency, high resolution, intra-oral ultrasound probes, it is now possible to make accurate measurement of the thickness of primary oral cancer. SHINTANI et al.16 showed that there is good correlation between intra-oral ultrasound thickness measurement of tumours and histological thickness of tumours. SHINTANI et al.17 showed that ultrasound is superior to CT and MRI for measurements of tumour thickness, especially those of less than 5 mm20. HELBIG et al.7 carried out a small study in five patients using ultrasound for intraoperative visualization and marking of tumour margins prior to resection. There is an increasing number of studies investigating the usefulness of ultrasonography for tumour thickness measure-
ments and a trend to use various imaging modalities for the guidance of tumour surgery. However, there is no study showing the value of intra-oral, intra-operative, ultrasound imaging in guiding resection of oral cancer. The purpose of our original study is to assess the depth of invasion or thickness of oral cancers with the use of intra-oral ultrasonography and to observe the deep surgical margin at the time of surgery, halfway during resection. This is an observational study to see whether ultrasound can accurately predict if the surgical margin is involved, close or clear of the tumour at the time of resection, and then correlated with subsequent histological analysis. As an observational study and to avoid subject bias (Hawthorn effect), the surgeons were asked to perform their resection as per normal and were not informed of the ultrasound results. This methodology has not previously been reported.
Method
Ethical approval from East London and City Health Authority was obtained for the use of ultrasound imaging intra-operatively in oral cancer patients. This was a single blind, prospective study performed between 1997 and 2002. Pre-operative tumour thickness measurements by ultrasound imaging were performed in 26 patients with biopsy proven oral SCC and who were previously untreated. In 14 of these patients (Table 1), we then carried out intra-operative ultrasound imaging during resection. Case selection depended on the availability of an ultrasound specialist (SN). The
325
ultrasound machine used was HDI 5000 (Advanced Technologies Ltd., Seattle). The ultrasound probe used was the broadband, linear 5–10 MHz Small Parts probe, with a footprint of 26 mm. This was designed for intra-operative use and was small enough to be used intra-orally. It was contra-angled which allowed good access to most parts of the oral cavity. The intra-operative ultrasound scanning was performed in order to: 1. measure the tumour thickness, 2. measure the deep margin clearance. A trained radiologist (SN) and a trained surgeon (AS) carried out all the scanning. Wherever possible, the ultrasound measurements were carried out using a noncontact technique, i.e., without the ultrasound probe touching the tumour surface. Any gaps between the probe and the tumour surface were filled with normal saline, as ultrasound waves do not travel through air. Access to posteriorly located tumours on the tongue was achieved by retraction aided by a suture. For tumours of the tongue (n = 11), tumour thickness and deep margin clearance measurements were always made in the axial plane and also, where accessible, the coronal plane. The deep margin assessment was carried out half way through surgical resection. In order to demonstrate the cut margin on ultrasound, an echogenic surface, such as a metal retractor, was placed in the surgical cut (Figs 1–3). Only light pressure was applied when holding the ultrasound probe next to the tumour surface, so as to avoid compressing the tissues. In some cases, when a gap appeared between the cut margin and the metal retractor; it would be
Table 1. Ultrasound and histological measurements of tumour thickness and of surgical margin clearance (n = 14) Tumour thickness Patient code 1 2 3 4 5 6 7 8 9 10 11 12 13 14
Tumour site Tongue Tongue Alveolar mucosa Tongue Lip Tongue Tongue Tongue Tongue Tongue Tongue Floor of mouth Tongue Tongue
Surgical margin
TNM stage
U/s
Hist.
U/s
Hist.
T1 T1 T1 T2 T1 T4 T1 T2 T2 T2 T1 T1 T2 T1
15.0 2.3 2.0 8.0 4.0 25.0 4.4 10.0 8.0 13.4 7.3 0.5 12.3 4.4
14.0 1.5 1.4 11.8 3.0 21.0 2.2 11.0 9.6 10.0 7.0 4.0 9.0 5.0
4.0 1.5 1.0 1.9 3.0 2.0 3.2 6.6 15.6 3.7 7.0 6.0 7.8 7.2
7.0 1.7 0.9 3.2 4.5 4.0 8.0 7.5 13.0 7.5 5.0 5.0 3.0 15.0
N0 M0 N0 M0 N0 M0 N2a M0 N0 M0 N3 M0 N0 M0 N0 M0 N2a M0 N0 M0 N0 M0 N0 M0 N1 M0 N0 M0
Patient 1 is featured in Fig. 2a–c. Patient 9 is featured in Fig. 1a–c. All measurements are in millimetres.
326
Songra et al.
Fig. 1. (a) Partially resected tongue tumour (B) with metal instrument (C) in the cut to provide a surface to reflect ultrasound. Ultrasound probe (A) is placed on the surface of tongue (labels match diagram in Fig. 3). Patient 9 in Table 1. (b) Ultrasound image of part (a). Note 2 pairs of electronic cursors to measure the distance from tongue surface to the deep margin of the tumour, and from the latter to the surgical cut. Field width: 26 mm. (c) Line diagram of part (b). S: tongue surface. T: tumour. M: metal instrument (bright white line). D: deep surgical margin. The surgical margin is clearly shown to be separate from the hypoechoic tumour. The gap between D and M is filled with water during ultrasound imaging. Note 2 pairs of cursors to show where measurements were made.
filled with water thus excluding air (Fig. 1b and c). The probe was carefully angulated to obtain an image which clearly showed the tumour deep margin, surgical resection
margin and the metal retractor (Fig. 3). The best images were acquired when the latter three planes were parallel to each other and the ultrasound beam was at 908 to all of
them. Patience and skill were always required. Once a good image was obtained, pairs of electronic cursors were placed on the (pre-calibrated) ultrasound screen and
Resection margin by ultrasound
327
Fig. 2. (a) Partially resected tongue tumour with ultrasound-reflective instrument in the cut. Patient 1 in Table 1. (b) Ultrasound image of part (a). In this case, the surgical cut is quite close to the deep margin of the tumour, rendering the latter difficult to delineate (see Fig. 4 for macroscopic specimen). Field width = 26 mm. (c) Line diagram of part (b). T: tumour. M: metal instrument. Note 2 pairs of cursors to show where measurements were made.
the following measurements made along a line at 908 to the tumour surface: 1. from tumour surface to deepest point on deep tumour margin; 2. from deepest point on deep tumour margin to surgical margin. As this was an observational study, the ultrasound measurements were not communicated to the surgeon performing the
operation. In this single blind study, the operating surgeon performed surgery as per usual, without being told of the ultrasound measurements of marginal clearance. The histopathologist was not given prior knowledge of any of the ultrasound measurements. Following resection, the fresh specimen was scanned by ultrasound to assess the entire tumour and its relationship to all the margins. The surface of the tumour was
marked with indelible ink opposite the site of the greatest depth, at which ultrasound measurements had been made. After the specimen was fixed, it was re-scanned by ultrasound. When possible, the specimen was cut by the histopathologist in the presence of the surgeon or the radiologist in order to guide the histologist to specific areas of interest (Fig. 4). This did not contravene the single blind nature of this study as only specimen orientation information
328
Songra et al.
Fig. 3. Diagram of partially resected tumour (B) being assessed by ultrasound for depth clearance, with a reflective instrument (C) in the surgical cut. A: ultrasound probe (contra-angle).
was discussed, not tumour depth and clearance information. After histological preparation, the histopathologist measured the tumour thickness and the surgical margin clearance at the point marked with
indelible ink without knowledge of the ultrasound measurements. Routine histological examination was then performed, including checking for clearance at all margins using the histological criteria recom-
mended by the Royal College of Pathologists for histological grading of surgical margin. For a tumour to be classified as completely excised, it must have a histological tumour-free margin of greater than 5 mm. A tumour-free margin of less than 5 mm is considered to be close, and that less than 1 mm is considered to be involved. The ultrasound findings were compared to the histological findings and various statistical analyses performed. Scatter plots of data with regression lines were obtained. Two correlation tests were performed on the distance measurements data in order to determine the overall correlation between them. The Kappa test was done to calculate the sensitivity, specificity and predictive value of using ultrasound imaging to determine whether the surgical margin was close or clear, where a close margin was defined as less than 5 mm, and a clear margin was greater or equal to 5 mm.
Fig. 4. (a–c) Macroscopic specimen (a) and diagram (b) of excised tongue squamous cell carcinoma (same case as in Fig. 2). In part (b), note proximity of deep tumour margin T to the surgical margin Q. S: surface of tongue and of tumour. Part (c) is the ultrasound image of the tumour before excision. Note how well the tumour outline matches that in the macroscopic specimen.
Resection margin by ultrasound
Fig. 5. Scatter plot of tumour thickness measurements by ultrasound and by histology. Note the plotted points nearly all lie on a straight line, showing a strong positive relationship (P < 0.01).
Results
Table 1 shows the measurements of tumour thickness and of deep margin clearance by ultrasound scanning and by histology. Figure 5 is a scatter plot of tumour thickness measurements by ultrasound and by histology. This graph shows the reliability of ultrasound for evaluation of tumour thickness against histological measurement. The plotted points nearly all lie on a straight line, thus showing a very high degree of correlation (significant at the 0.01 level). Calculation of the Pearson correlation coefficient (r = 0.948, P < 0.01) and the Spearman rank correlation coefficient (r = 0.913, P < 0.01) confirm this. The intraclass correlation coefficient is about 0.95 which implies really good reliability. For deep margin clearance, ultrasound has good correlation with histology, as shown by the Pearson correlation coefficient of 0.648 (P < 0.01). The Kappa statistic value of 0.44 shows ultrasound to have moderate predictive value. Table 2. Crosstabulation of histology and ultrasound detection of close and clear surgical margins Histology Ultrasound – Total
Close Clear –
Close
Clear
Total
5 1 6
3 5 8
8 6 14
Ultrasound imaging showed sensitivity of 83%, specificity of 63%, positive predictive value of 63% and negative predictive value of 83%. Detection of close and clear surgical margins by histology and by ultrasound imaging. Close margin < 5 mm, clear margin 5 mm.
Applying the threshold of 5 mm to indicate whether the deep surgical margin was clear of tumour, there was agreement in 10 out of 14 cases between ultrasound and histology (Table 2). Using ultrasound to detect close surgical margins resulted in one false negative and three false positives, giving a sensitivity of 83%, specificity of 63%, positive predictive value of 63% and negative predictive value of 83%. Discussion
Most deaths from oral squamous cell cancer (SCC) occur in association with failure to achieve local or regional disease control11. A smaller percentage of deaths occur as a result of distant metastatic disease. Although many adverse prognostic factors are recognised relating either to the clinical or histopathological features of a given tumour, failure to achieve surgical clearance can result in local or regional recurrence despite radical post-op radiotherapy. Achieving surgical clearance is the single prognostic factor controlled by the surgeon. It is well established that histological clearance of 5 mm or more is required in oral cancer resection in order to reduce local recurrence and improve survival10. In order to achieve 5 mm of histological clearance and to allow for specimen shrinkage and surgical error/underestimate9, a margin of 10 mm is advocated at the time of surgery. Despite this, the literature shows that approximately 20– 50% of resections are reported histologically as not clear that is close or involved margins19. This is due to underestimation of the true size of the tumour. YUEN
329
et al.24 have shown that even a 10 mm margin is not safe. They advocate a minimum of 15 mm and a maximum of 20 mm. A significant proportion of resections are reported inadequately cleared usually at the deep margins as this is clinically difficult to assess. Operatively surgeons will use two of their sensory faculties, notably sight and touch, to assess tumour clearance. However, both of these assessment techniques have significant disadvantages. The tumour’s vertical depth of invasion, like an iceberg, is much more difficult to assess both pre-operatively with CT or MRI and clinically at surgery17. It is for this reason that inadequate resections occur. In order to achieve deep clearance the thickness of tumour needs to be known or visualised at the time of resection. At present, CT and MRI do not have the resolution to demonstrate the thickness of primary T1, T2 oral tumours with certainty6,17. Tumour thickness is an important prognostic factor for neck metastasis1,4. Furthermore, YUEN et al.25 reported that tumour thickness is the only important factor that had significant predictive value for subclinical nodal metastases, local recurrence and survival. The difficulties of assessing oral tumours with extra-oral ultrasound measurements in some early studies have now been overcome by using high resolution intra-oral probes. More recent studies with intra-oral ultrasound imaging has been shown to be useful for evaluating tumour thickness in tongue carcinoma7,16,17. The advantages of diagnostic ultrasound are that it is non-invasive, does not use ionising radiation, is quick to perform and is repeatable. The machine is portable and can be taken to the operating theatre, thus real time imaging can be carried out at the time of surgery. This study sought ways to assess tumour clearance at the deep margin more objectively by using ultrasound scanning. Our study demonstrated that depth clearance can be assessed by ultrasound at the time of surgery. However, we could only use ultrasound for tumours on the anterior two thirds of tongue, floor of mouth and other regions accessible by the intraoral ultrasound probe. Posterior third of tongue could not be accessed by our ultrasound probe. The ultrasound measurements showed a high degree of reliability with histological measurements for tumour thickness. For deep margin clearance, ultrasound had a moderate predictive value. The discrepancies in deep margin clearance measurements can be accounted for by the following reasons:
330
Songra et al.
1. The histology and the ultrasound measurements might not have been made at exactly the same point. 2. There was some shrinkage of specimen after resection and after fixation in formalin. 3. Histological tissue processing took samples of tissue at the tumour margin, rather than the whole margin. Thus it was possible that a tumour with a deeply infiltrating margin at only one point could be missed on histology. 4. In order to hold the partly resected specimen in its original position when the ultrasound measurement was carried out mid-resection, slight compression at the surface was applied with the ultrasound probe. 5. There was some additional tissue compression when a metal retractor had to be placed in the surgical cut in order to provide an ultrasound reflector.
The study by HELBIG et al.7 assessed the accuracy of intra-operative ultrasound for the visualization of tumour size. They only studied five cases and statistical analysis was not possible on such small sample. They used a suture, placed under ultrasound guidance, to mark the deep tumour margin prior to any resection. Whilst this provided a definitive physical reference for histological registration of depth, the method had potential drawbacks such as tissue distortion and possible interference with the cells at the margin and ‘‘seeding’’ of tumour cells. HELBIG et al.7 did not assess surgical margin clearance of tumour. During the course of our current study we became aware of the relative ease with which ultrasound imaging could be applied to the surface, mucosal margins. We have carried out a separate study of this, which will be the subject of a future publication. To our knowledge we are the first to report the use of intraoperative, intra-oral ultrasound imaging to measure surgical margin clearance of tumour and to do so at the time of resection. With the patient under general anaesthesia, positioning of the ultrasound probe is easier, particularly if the tumour is posteriorly placed. There is also less patient movement artefact, leading to more accurate measurement, especially for the tongue. In addition to making measurements, the real time nature of ultrasound imaging lends itself to providing guidance at the time of surgery. This will have major impact on surgical practice. If this proves successful the need for frozen sections may diminish.
We chose to image the tumour partially resected but still attached because this would facilitate interpretation of the ultrasound image and give fixed points of reference. We deliberately did not inform the surgeon and the pathologist of the ultrasound results in order to avoid bias (Hawthorn effect). With the pathologist we had to show them where to cut (or not to cut) the specimen in order to avoid accidental damage to the site of interest. Having obtained encouraging results in this study, we plan to perform further studies using ultrasound imaging as an intraoperative guidance tool, where we will intentionally influence the surgeon’s actions. Now that a reliable and accurate tool in the form of high frequency, high resolution diagnostic ultrasound is available, it is possible to objectively assess the surgical margin and to achieve higher rates of clearance. The establishment of accurate tumour thickness measurements will enable more precise surgery to be performed in the future. This will allow less mutilating surgery, reduce morbidity and decrease risks of recurrence. Conclusion
This original paper shows that high resolution ultrasound imaging, applied intraorally to tumours accessible by the ultrasound probe, is a good reliable tool in objectively assessing: 1. tumour thickness; 2. surgical margin clearance at the time of operation. Acknowledgements. We are very grateful to ATL Ltd. for the loan of an HDI 5000 ultrasound machine and to the charity ‘‘The Facial Surgery Research Foundation-Saving Faces’’ and the Special Trustees of Barts and the London NHS Trust for eventual purchase of this ultrasound machine. We also thank our Consultant colleagues Mr. John Carter and Mr. Peter Hardee in Oral & Maxillofacial Surgery for permitting access to their patients.
References 1. Asakage T, Yokose T, Mukai K, Tsugane S, Tsubono Y, Asai M, Ebihara S. Tumour thickness predicts cervical metastasis in patients with stage I/II carcinoma of the tongue. Cancer 1998: 82: 1443–1448. 2. Bruneton JN, Roux P, Caramella E, Manzino JJ, Vallicioni J, Demard F. Tongue and tonsil cancer: staging with US. Radiology 1986: 158: 743–746.
3. Byers RM, Bland KI, Borlase B, Luna M. The prognostic and therapeutic value of frozen section determinations in the surgical treatment of squamous carcinoma of the head and neck. Am J Surg 1978: 136: 525–528. 4. Fukano H, Matsuura H, Hasegawa Y, Nakamura S. Depth of invasion as a predictive factor for cervical lymph node metastasis in tongue carcinoma. Head Neck 1997: 19: 205–210. 5. Hardee PSG, Carter JLB, Piper KM, Ng SY. Metachronous bilateral primary adenocarcinoma of the submandibular glands. Oral Surg Oral Med Oral Path Oral Rad Endo 2001: 91: 455–461. 6. Heissler E, Steinkamp HJ, Heim T, Zwicker C, Felix R, Bier J. Value of magnetic resonance imaging in staging carcinomas of the oral cavity and oropharynx. Int J Oral Maxillofac Surg 1994: 23: 22–27. 7. Helbig M, Flechtenmacher C, Hansmann J, Dietz A, Tasman A-J. Intraoperative B-mode endosonography of tongue carcinoma. Head Neck 2001: 23: 233–237. 8. Iro H, Nitsche N. Intra-oral sonography in neoplasms of the mouth and base of the tongue. HNO 1989: 37: 329–332. 9. Johnson RE, Sigman JD, Funk GF, Robinson RA, Hoffman HT. Quantification of surgical margin shrinkage in the oral cavity. Head Neck 1997: 19: 281– 286. 10. Kirita T, Okabe S, Izumo T, Sugimura M. Risk factors for the postoperative local recurrence of tongue carcinoma. J Oral Maxillofac Surg 1994: 52: 149–154. 11. Looser KG, Shah JP, Strong EW. The significance of positive margins in surgically resected epidermoid carcinomas. Head Neck Surg 1978: 1: 107–111. 12. Loree TR, Strong EW. Significance of positive margins in oral cavity squamous carcinoma. Am J Surg 1990: 160: 410– 414. 13. Ng SY, Songra AK, Ali N, Carter JLB. Ultrasound features of osteosarcoma of the mandible—a first report. Oral Surg Oral Med Oral Pathol Oral Radiol Endod 2001: 92: 582–586. 14. Ord RA, Aisner S. Accuracy of frozen sections in assessing margins in oral cancer resection. J Oral Maxfac Surg 1997: 55: 669–671. 15. Scholl P, Byers RM, Batsakis JG, Wolf P, Santini H. Microscopic cut through of cancer in the surgical treatment of squamous carcinoma of the tongue Prognostic and therapeutic implications. Am J Surg 1986: 152: 354–360. 16. Shintani S, Nakayama B, Matsuura H, Hasegawa Y. Intraoral ultrasonography is useful to evaluate tumor thickness in tongue carcinoma. Am J Surg 1997: 173: 345–347. 17. Shintani S, Yoshihama Y, Ueyama Y, Terakado N, Kamei S, Fijimoto Y,
Resection margin by ultrasound Hasegawa Y, Matsuura H, Matsumura T. The usefulness of intraoral ultrasonography in the evaluation of oral cancer. Int J Oral Maxillofac Surg 2001: 30: 139–143. 18. Spiro RH, Huvos AG, Wong GY, Spiro JD, Gnecco CA, Strong EW. Predictive value of tumour thickness in squamous carcinomas confined to the tongue and floor of the mouth. Am J Surg 1986: 152: 345–350. 19. Sutton DN, Brown JS, Rogers SN, Vaughan ED, Woolgar JA. The prognostic implications of the surgical margin in oral squamous cell carcinoma. Int J Oral Maxfac Surg 2003: 32: 30–34. 20. Tetsumura A, Yoshino N, Amagasa T, Nagumo K, Okada N, Sasaki T. High resolution magnetic resonance imaging of
21.
22.
23. 24.
squamous cell carcinoma of the tongue: an in vitro study. Dentomaxillofacial Radiol 2001: 30: 14–21. Van den Brekel MWM, Castelijins JA, Snow GB. Imaging of cervical lymphadenopathy. Neuroimaging Clin N Am 1996: 6: 417–434. Vikram B, Strong EW, Shah JP, Spiro R. Failure at the primary site following multi-modality treatment in advanced head and neck cancer. Head Neck Surg 1984: 6: 720–723. Ying M, Ahuja A. Sonography of neck lymph nodes Part I: normal lymph nodes. Clin Rad 2003: 58: 351–358. Yuen PW, Lam KY, Chan AC, Wei WI, Lam LK. Clinicopathological analysis of local spread of carcinoma of the tongue. Am J Surg 1998: 175: 242–244.
331
25. Yuen PW, Lam KY, Lam LK, Ho CM, Wong A, Chow TL, Yuen WF, Wei WI. Prognostic factors of clinically stage I and II oral tongue carcinoma—a comparative study of stage, thickness, shape, growth pattern, invasive front malignancy grading, Martinez-Gimeno score, and pathologic features. Head Neck 2002: 24: 513– 520. Address: Suk Y. Ng Department of Dental Radiology Guy’s, King’s and St Thomas’ Dental Institute King’s College London Caldecot Road London SE5 9RW, UK. E-mail: [email protected]
Clinical Paper Head and Neck Oncology
Observation of tumour thickness and resection margin at surgical excision of primary oral squamous cell carcinoma— assessment by ultrasound
A. K. Songra, S. Y. Ng, P. Farthing, I. L. Hutchison, P. F. Bradley Maxillofacial Unit, Royal London Hospital, Barts and The London NHS Trust, London E1 2BB, UK
A. K. Songra, S. Y. Ng, P. Farthing, I. L. Hutchison, P. F. Bradley: Observation of tumour thickness and resection margin at surgical excision of primary oral squamous cell carcinoma—assessment by ultrasound. Int. J. Oral Maxillofac. Surg. 2006; 35: 324–331. # 2005 International Association of Oral and Maxillofacial Surgeons. Published by Elsevier Ltd. All rights reserved. Abstract. Tumour thickness and the status of resection margins are of prognostic significance in the treatment of oral cancer. In a single blind prospective study, 14 patients with biopsy proven oral squamous cell carcinoma had intraoral ultrasound imaging done preoperatively to measure tumour thickness, and intraoperatively to measure the deep surgical margin half way during resection. The cut surface was demonstrated on ultrasound by placing a metal, ultrasoundreflective, retractor into the surgical cut. The ultrasound measurements were compared to the subsequent histological measurements. Using the threshold of 5 mm as indicator of margin clearance, there was agreement in 10 out of 14 cases between ultrasound and histology. Ultrasound detection of close surgical margins had a sensitivity of 83% and a specificity of 63%. For preoperative tumour thickness measurement, ultrasound imaging showed a high degree of correlation with histology (Pearson correlation coefficient = 0.95, P < 0.01). This original paper demonstrates that high resolution ultrasound imaging applied intraorally is a reliable tool in objectively assessing both the tumour thickness and the surgical margin clearance at the time of surgery.
Among the large number of prognostic factors demonstrated to be of importance in the treatment of oral squamous cell carcinoma are tumour thickness and the status of the resection margins. Achieving tumour clearance at the primary site 0901-5027/040324 + 08 $30.00/0
is a major problem but a very important aspect of cancer surgery. It has been shown by numerous studies that patients demonstrating invasive carcinoma at resection margins have a higher incidence of loco-regional recurrence and
Key words: tumour thickness; margin clearance; oral cancer; intraoral ultrasound imaging; intraoperative ultrasonography; intraoperative guidance. Accepted for publication 26 July 2005 Available online 13 December 2005
reduced survival rate12. Controversy exists regarding the value of postoperative radiotherapy following incomplete excision12,22. Complete removal of the primary tumour at the first attempt will obviate the need for adjuvant
# 2005 International Association of Oral and Maxillofacial Surgeons. Published by Elsevier Ltd. All rights reserved.
Resection margin by ultrasound radiotherapy in most cases and maximise prognosis. Visual inspection and palpation at the time of surgery in addition to a number of pre-operative modes of imaging all have their limitations in ensuring complete resection of a tumour mass. Frozen section control although widely used has its problems3,14,15. It cannot be used to study an entire margin and cannot demonstrate the amount of clearance. There are difficulties in deciding where to take the sample and in localising the biopsy site to the specimen. Processing samples is very time-consuming and may preclude its routine use. One of the current limitations of achieving margin clearance is the lack of an imaging technique to measure the thickness of the primary tumour. Clinical judgement has been shown to be unreliable, as evidenced by the number of resections that are reported with involved margins10,12. Imaging techniques such as CT and MRI do not have the resolution to demonstrate the thickness of primary oral cancer accurately17. Lack of pre-operative information regarding tumour thickness leads not only to possibly inadequate resection, but increases the risk of local recurrence and reduces the survival rate12,18. High resolution diagnostic ultrasound imaging is becoming well established in the field of head and neck oncology5,13,21. There are many studies describing the ultrasound features of metastatic cervical lymph nodes23, but not of the primary site. Transcutaneous extra-oral (through skin and muscles of submental region) ultrasound imaging of the tongue has been carried out in some early studies2. However, such an extra-oral approach can only measure approximate thickness and only when the tumour is large. Transcutaneous ultrasonography is considered inferior to intra-oral ultrasonography8. With improvements in imaging technology and availability of high frequency, high resolution, intra-oral ultrasound probes, it is now possible to make accurate measurement of the thickness of primary oral cancer. SHINTANI et al.16 showed that there is good correlation between intra-oral ultrasound thickness measurement of tumours and histological thickness of tumours. SHINTANI et al.17 showed that ultrasound is superior to CT and MRI for measurements of tumour thickness, especially those of less than 5 mm20. HELBIG et al.7 carried out a small study in five patients using ultrasound for intraoperative visualization and marking of tumour margins prior to resection. There is an increasing number of studies investigating the usefulness of ultrasonography for tumour thickness measure-
ments and a trend to use various imaging modalities for the guidance of tumour surgery. However, there is no study showing the value of intra-oral, intra-operative, ultrasound imaging in guiding resection of oral cancer. The purpose of our original study is to assess the depth of invasion or thickness of oral cancers with the use of intra-oral ultrasonography and to observe the deep surgical margin at the time of surgery, halfway during resection. This is an observational study to see whether ultrasound can accurately predict if the surgical margin is involved, close or clear of the tumour at the time of resection, and then correlated with subsequent histological analysis. As an observational study and to avoid subject bias (Hawthorn effect), the surgeons were asked to perform their resection as per normal and were not informed of the ultrasound results. This methodology has not previously been reported.
Method
Ethical approval from East London and City Health Authority was obtained for the use of ultrasound imaging intra-operatively in oral cancer patients. This was a single blind, prospective study performed between 1997 and 2002. Pre-operative tumour thickness measurements by ultrasound imaging were performed in 26 patients with biopsy proven oral SCC and who were previously untreated. In 14 of these patients (Table 1), we then carried out intra-operative ultrasound imaging during resection. Case selection depended on the availability of an ultrasound specialist (SN). The
325
ultrasound machine used was HDI 5000 (Advanced Technologies Ltd., Seattle). The ultrasound probe used was the broadband, linear 5–10 MHz Small Parts probe, with a footprint of 26 mm. This was designed for intra-operative use and was small enough to be used intra-orally. It was contra-angled which allowed good access to most parts of the oral cavity. The intra-operative ultrasound scanning was performed in order to: 1. measure the tumour thickness, 2. measure the deep margin clearance. A trained radiologist (SN) and a trained surgeon (AS) carried out all the scanning. Wherever possible, the ultrasound measurements were carried out using a noncontact technique, i.e., without the ultrasound probe touching the tumour surface. Any gaps between the probe and the tumour surface were filled with normal saline, as ultrasound waves do not travel through air. Access to posteriorly located tumours on the tongue was achieved by retraction aided by a suture. For tumours of the tongue (n = 11), tumour thickness and deep margin clearance measurements were always made in the axial plane and also, where accessible, the coronal plane. The deep margin assessment was carried out half way through surgical resection. In order to demonstrate the cut margin on ultrasound, an echogenic surface, such as a metal retractor, was placed in the surgical cut (Figs 1–3). Only light pressure was applied when holding the ultrasound probe next to the tumour surface, so as to avoid compressing the tissues. In some cases, when a gap appeared between the cut margin and the metal retractor; it would be
Table 1. Ultrasound and histological measurements of tumour thickness and of surgical margin clearance (n = 14) Tumour thickness Patient code 1 2 3 4 5 6 7 8 9 10 11 12 13 14
Tumour site Tongue Tongue Alveolar mucosa Tongue Lip Tongue Tongue Tongue Tongue Tongue Tongue Floor of mouth Tongue Tongue
Surgical margin
TNM stage
U/s
Hist.
U/s
Hist.
T1 T1 T1 T2 T1 T4 T1 T2 T2 T2 T1 T1 T2 T1
15.0 2.3 2.0 8.0 4.0 25.0 4.4 10.0 8.0 13.4 7.3 0.5 12.3 4.4
14.0 1.5 1.4 11.8 3.0 21.0 2.2 11.0 9.6 10.0 7.0 4.0 9.0 5.0
4.0 1.5 1.0 1.9 3.0 2.0 3.2 6.6 15.6 3.7 7.0 6.0 7.8 7.2
7.0 1.7 0.9 3.2 4.5 4.0 8.0 7.5 13.0 7.5 5.0 5.0 3.0 15.0
N0 M0 N0 M0 N0 M0 N2a M0 N0 M0 N3 M0 N0 M0 N0 M0 N2a M0 N0 M0 N0 M0 N0 M0 N1 M0 N0 M0
Patient 1 is featured in Fig. 2a–c. Patient 9 is featured in Fig. 1a–c. All measurements are in millimetres.
326
Songra et al.
Fig. 1. (a) Partially resected tongue tumour (B) with metal instrument (C) in the cut to provide a surface to reflect ultrasound. Ultrasound probe (A) is placed on the surface of tongue (labels match diagram in Fig. 3). Patient 9 in Table 1. (b) Ultrasound image of part (a). Note 2 pairs of electronic cursors to measure the distance from tongue surface to the deep margin of the tumour, and from the latter to the surgical cut. Field width: 26 mm. (c) Line diagram of part (b). S: tongue surface. T: tumour. M: metal instrument (bright white line). D: deep surgical margin. The surgical margin is clearly shown to be separate from the hypoechoic tumour. The gap between D and M is filled with water during ultrasound imaging. Note 2 pairs of cursors to show where measurements were made.
filled with water thus excluding air (Fig. 1b and c). The probe was carefully angulated to obtain an image which clearly showed the tumour deep margin, surgical resection
margin and the metal retractor (Fig. 3). The best images were acquired when the latter three planes were parallel to each other and the ultrasound beam was at 908 to all of
them. Patience and skill were always required. Once a good image was obtained, pairs of electronic cursors were placed on the (pre-calibrated) ultrasound screen and
Resection margin by ultrasound
327
Fig. 2. (a) Partially resected tongue tumour with ultrasound-reflective instrument in the cut. Patient 1 in Table 1. (b) Ultrasound image of part (a). In this case, the surgical cut is quite close to the deep margin of the tumour, rendering the latter difficult to delineate (see Fig. 4 for macroscopic specimen). Field width = 26 mm. (c) Line diagram of part (b). T: tumour. M: metal instrument. Note 2 pairs of cursors to show where measurements were made.
the following measurements made along a line at 908 to the tumour surface: 1. from tumour surface to deepest point on deep tumour margin; 2. from deepest point on deep tumour margin to surgical margin. As this was an observational study, the ultrasound measurements were not communicated to the surgeon performing the
operation. In this single blind study, the operating surgeon performed surgery as per usual, without being told of the ultrasound measurements of marginal clearance. The histopathologist was not given prior knowledge of any of the ultrasound measurements. Following resection, the fresh specimen was scanned by ultrasound to assess the entire tumour and its relationship to all the margins. The surface of the tumour was
marked with indelible ink opposite the site of the greatest depth, at which ultrasound measurements had been made. After the specimen was fixed, it was re-scanned by ultrasound. When possible, the specimen was cut by the histopathologist in the presence of the surgeon or the radiologist in order to guide the histologist to specific areas of interest (Fig. 4). This did not contravene the single blind nature of this study as only specimen orientation information
328
Songra et al.
Fig. 3. Diagram of partially resected tumour (B) being assessed by ultrasound for depth clearance, with a reflective instrument (C) in the surgical cut. A: ultrasound probe (contra-angle).
was discussed, not tumour depth and clearance information. After histological preparation, the histopathologist measured the tumour thickness and the surgical margin clearance at the point marked with
indelible ink without knowledge of the ultrasound measurements. Routine histological examination was then performed, including checking for clearance at all margins using the histological criteria recom-
mended by the Royal College of Pathologists for histological grading of surgical margin. For a tumour to be classified as completely excised, it must have a histological tumour-free margin of greater than 5 mm. A tumour-free margin of less than 5 mm is considered to be close, and that less than 1 mm is considered to be involved. The ultrasound findings were compared to the histological findings and various statistical analyses performed. Scatter plots of data with regression lines were obtained. Two correlation tests were performed on the distance measurements data in order to determine the overall correlation between them. The Kappa test was done to calculate the sensitivity, specificity and predictive value of using ultrasound imaging to determine whether the surgical margin was close or clear, where a close margin was defined as less than 5 mm, and a clear margin was greater or equal to 5 mm.
Fig. 4. (a–c) Macroscopic specimen (a) and diagram (b) of excised tongue squamous cell carcinoma (same case as in Fig. 2). In part (b), note proximity of deep tumour margin T to the surgical margin Q. S: surface of tongue and of tumour. Part (c) is the ultrasound image of the tumour before excision. Note how well the tumour outline matches that in the macroscopic specimen.
Resection margin by ultrasound
Fig. 5. Scatter plot of tumour thickness measurements by ultrasound and by histology. Note the plotted points nearly all lie on a straight line, showing a strong positive relationship (P < 0.01).
Results
Table 1 shows the measurements of tumour thickness and of deep margin clearance by ultrasound scanning and by histology. Figure 5 is a scatter plot of tumour thickness measurements by ultrasound and by histology. This graph shows the reliability of ultrasound for evaluation of tumour thickness against histological measurement. The plotted points nearly all lie on a straight line, thus showing a very high degree of correlation (significant at the 0.01 level). Calculation of the Pearson correlation coefficient (r = 0.948, P < 0.01) and the Spearman rank correlation coefficient (r = 0.913, P < 0.01) confirm this. The intraclass correlation coefficient is about 0.95 which implies really good reliability. For deep margin clearance, ultrasound has good correlation with histology, as shown by the Pearson correlation coefficient of 0.648 (P < 0.01). The Kappa statistic value of 0.44 shows ultrasound to have moderate predictive value. Table 2. Crosstabulation of histology and ultrasound detection of close and clear surgical margins Histology Ultrasound – Total
Close Clear –
Close
Clear
Total
5 1 6
3 5 8
8 6 14
Ultrasound imaging showed sensitivity of 83%, specificity of 63%, positive predictive value of 63% and negative predictive value of 83%. Detection of close and clear surgical margins by histology and by ultrasound imaging. Close margin < 5 mm, clear margin 5 mm.
Applying the threshold of 5 mm to indicate whether the deep surgical margin was clear of tumour, there was agreement in 10 out of 14 cases between ultrasound and histology (Table 2). Using ultrasound to detect close surgical margins resulted in one false negative and three false positives, giving a sensitivity of 83%, specificity of 63%, positive predictive value of 63% and negative predictive value of 83%. Discussion
Most deaths from oral squamous cell cancer (SCC) occur in association with failure to achieve local or regional disease control11. A smaller percentage of deaths occur as a result of distant metastatic disease. Although many adverse prognostic factors are recognised relating either to the clinical or histopathological features of a given tumour, failure to achieve surgical clearance can result in local or regional recurrence despite radical post-op radiotherapy. Achieving surgical clearance is the single prognostic factor controlled by the surgeon. It is well established that histological clearance of 5 mm or more is required in oral cancer resection in order to reduce local recurrence and improve survival10. In order to achieve 5 mm of histological clearance and to allow for specimen shrinkage and surgical error/underestimate9, a margin of 10 mm is advocated at the time of surgery. Despite this, the literature shows that approximately 20– 50% of resections are reported histologically as not clear that is close or involved margins19. This is due to underestimation of the true size of the tumour. YUEN
329
et al.24 have shown that even a 10 mm margin is not safe. They advocate a minimum of 15 mm and a maximum of 20 mm. A significant proportion of resections are reported inadequately cleared usually at the deep margins as this is clinically difficult to assess. Operatively surgeons will use two of their sensory faculties, notably sight and touch, to assess tumour clearance. However, both of these assessment techniques have significant disadvantages. The tumour’s vertical depth of invasion, like an iceberg, is much more difficult to assess both pre-operatively with CT or MRI and clinically at surgery17. It is for this reason that inadequate resections occur. In order to achieve deep clearance the thickness of tumour needs to be known or visualised at the time of resection. At present, CT and MRI do not have the resolution to demonstrate the thickness of primary T1, T2 oral tumours with certainty6,17. Tumour thickness is an important prognostic factor for neck metastasis1,4. Furthermore, YUEN et al.25 reported that tumour thickness is the only important factor that had significant predictive value for subclinical nodal metastases, local recurrence and survival. The difficulties of assessing oral tumours with extra-oral ultrasound measurements in some early studies have now been overcome by using high resolution intra-oral probes. More recent studies with intra-oral ultrasound imaging has been shown to be useful for evaluating tumour thickness in tongue carcinoma7,16,17. The advantages of diagnostic ultrasound are that it is non-invasive, does not use ionising radiation, is quick to perform and is repeatable. The machine is portable and can be taken to the operating theatre, thus real time imaging can be carried out at the time of surgery. This study sought ways to assess tumour clearance at the deep margin more objectively by using ultrasound scanning. Our study demonstrated that depth clearance can be assessed by ultrasound at the time of surgery. However, we could only use ultrasound for tumours on the anterior two thirds of tongue, floor of mouth and other regions accessible by the intraoral ultrasound probe. Posterior third of tongue could not be accessed by our ultrasound probe. The ultrasound measurements showed a high degree of reliability with histological measurements for tumour thickness. For deep margin clearance, ultrasound had a moderate predictive value. The discrepancies in deep margin clearance measurements can be accounted for by the following reasons:
330
Songra et al.
1. The histology and the ultrasound measurements might not have been made at exactly the same point. 2. There was some shrinkage of specimen after resection and after fixation in formalin. 3. Histological tissue processing took samples of tissue at the tumour margin, rather than the whole margin. Thus it was possible that a tumour with a deeply infiltrating margin at only one point could be missed on histology. 4. In order to hold the partly resected specimen in its original position when the ultrasound measurement was carried out mid-resection, slight compression at the surface was applied with the ultrasound probe. 5. There was some additional tissue compression when a metal retractor had to be placed in the surgical cut in order to provide an ultrasound reflector.
The study by HELBIG et al.7 assessed the accuracy of intra-operative ultrasound for the visualization of tumour size. They only studied five cases and statistical analysis was not possible on such small sample. They used a suture, placed under ultrasound guidance, to mark the deep tumour margin prior to any resection. Whilst this provided a definitive physical reference for histological registration of depth, the method had potential drawbacks such as tissue distortion and possible interference with the cells at the margin and ‘‘seeding’’ of tumour cells. HELBIG et al.7 did not assess surgical margin clearance of tumour. During the course of our current study we became aware of the relative ease with which ultrasound imaging could be applied to the surface, mucosal margins. We have carried out a separate study of this, which will be the subject of a future publication. To our knowledge we are the first to report the use of intraoperative, intra-oral ultrasound imaging to measure surgical margin clearance of tumour and to do so at the time of resection. With the patient under general anaesthesia, positioning of the ultrasound probe is easier, particularly if the tumour is posteriorly placed. There is also less patient movement artefact, leading to more accurate measurement, especially for the tongue. In addition to making measurements, the real time nature of ultrasound imaging lends itself to providing guidance at the time of surgery. This will have major impact on surgical practice. If this proves successful the need for frozen sections may diminish.
We chose to image the tumour partially resected but still attached because this would facilitate interpretation of the ultrasound image and give fixed points of reference. We deliberately did not inform the surgeon and the pathologist of the ultrasound results in order to avoid bias (Hawthorn effect). With the pathologist we had to show them where to cut (or not to cut) the specimen in order to avoid accidental damage to the site of interest. Having obtained encouraging results in this study, we plan to perform further studies using ultrasound imaging as an intraoperative guidance tool, where we will intentionally influence the surgeon’s actions. Now that a reliable and accurate tool in the form of high frequency, high resolution diagnostic ultrasound is available, it is possible to objectively assess the surgical margin and to achieve higher rates of clearance. The establishment of accurate tumour thickness measurements will enable more precise surgery to be performed in the future. This will allow less mutilating surgery, reduce morbidity and decrease risks of recurrence. Conclusion
This original paper shows that high resolution ultrasound imaging, applied intraorally to tumours accessible by the ultrasound probe, is a good reliable tool in objectively assessing: 1. tumour thickness; 2. surgical margin clearance at the time of operation. Acknowledgements. We are very grateful to ATL Ltd. for the loan of an HDI 5000 ultrasound machine and to the charity ‘‘The Facial Surgery Research Foundation-Saving Faces’’ and the Special Trustees of Barts and the London NHS Trust for eventual purchase of this ultrasound machine. We also thank our Consultant colleagues Mr. John Carter and Mr. Peter Hardee in Oral & Maxillofacial Surgery for permitting access to their patients.
References 1. Asakage T, Yokose T, Mukai K, Tsugane S, Tsubono Y, Asai M, Ebihara S. Tumour thickness predicts cervical metastasis in patients with stage I/II carcinoma of the tongue. Cancer 1998: 82: 1443–1448. 2. Bruneton JN, Roux P, Caramella E, Manzino JJ, Vallicioni J, Demard F. Tongue and tonsil cancer: staging with US. Radiology 1986: 158: 743–746.
3. Byers RM, Bland KI, Borlase B, Luna M. The prognostic and therapeutic value of frozen section determinations in the surgical treatment of squamous carcinoma of the head and neck. Am J Surg 1978: 136: 525–528. 4. Fukano H, Matsuura H, Hasegawa Y, Nakamura S. Depth of invasion as a predictive factor for cervical lymph node metastasis in tongue carcinoma. Head Neck 1997: 19: 205–210. 5. Hardee PSG, Carter JLB, Piper KM, Ng SY. Metachronous bilateral primary adenocarcinoma of the submandibular glands. Oral Surg Oral Med Oral Path Oral Rad Endo 2001: 91: 455–461. 6. Heissler E, Steinkamp HJ, Heim T, Zwicker C, Felix R, Bier J. Value of magnetic resonance imaging in staging carcinomas of the oral cavity and oropharynx. Int J Oral Maxillofac Surg 1994: 23: 22–27. 7. Helbig M, Flechtenmacher C, Hansmann J, Dietz A, Tasman A-J. Intraoperative B-mode endosonography of tongue carcinoma. Head Neck 2001: 23: 233–237. 8. Iro H, Nitsche N. Intra-oral sonography in neoplasms of the mouth and base of the tongue. HNO 1989: 37: 329–332. 9. Johnson RE, Sigman JD, Funk GF, Robinson RA, Hoffman HT. Quantification of surgical margin shrinkage in the oral cavity. Head Neck 1997: 19: 281– 286. 10. Kirita T, Okabe S, Izumo T, Sugimura M. Risk factors for the postoperative local recurrence of tongue carcinoma. J Oral Maxillofac Surg 1994: 52: 149–154. 11. Looser KG, Shah JP, Strong EW. The significance of positive margins in surgically resected epidermoid carcinomas. Head Neck Surg 1978: 1: 107–111. 12. Loree TR, Strong EW. Significance of positive margins in oral cavity squamous carcinoma. Am J Surg 1990: 160: 410– 414. 13. Ng SY, Songra AK, Ali N, Carter JLB. Ultrasound features of osteosarcoma of the mandible—a first report. Oral Surg Oral Med Oral Pathol Oral Radiol Endod 2001: 92: 582–586. 14. Ord RA, Aisner S. Accuracy of frozen sections in assessing margins in oral cancer resection. J Oral Maxfac Surg 1997: 55: 669–671. 15. Scholl P, Byers RM, Batsakis JG, Wolf P, Santini H. Microscopic cut through of cancer in the surgical treatment of squamous carcinoma of the tongue Prognostic and therapeutic implications. Am J Surg 1986: 152: 354–360. 16. Shintani S, Nakayama B, Matsuura H, Hasegawa Y. Intraoral ultrasonography is useful to evaluate tumor thickness in tongue carcinoma. Am J Surg 1997: 173: 345–347. 17. Shintani S, Yoshihama Y, Ueyama Y, Terakado N, Kamei S, Fijimoto Y,
Resection margin by ultrasound Hasegawa Y, Matsuura H, Matsumura T. The usefulness of intraoral ultrasonography in the evaluation of oral cancer. Int J Oral Maxillofac Surg 2001: 30: 139–143. 18. Spiro RH, Huvos AG, Wong GY, Spiro JD, Gnecco CA, Strong EW. Predictive value of tumour thickness in squamous carcinomas confined to the tongue and floor of the mouth. Am J Surg 1986: 152: 345–350. 19. Sutton DN, Brown JS, Rogers SN, Vaughan ED, Woolgar JA. The prognostic implications of the surgical margin in oral squamous cell carcinoma. Int J Oral Maxfac Surg 2003: 32: 30–34. 20. Tetsumura A, Yoshino N, Amagasa T, Nagumo K, Okada N, Sasaki T. High resolution magnetic resonance imaging of
21.
22.
23. 24.
squamous cell carcinoma of the tongue: an in vitro study. Dentomaxillofacial Radiol 2001: 30: 14–21. Van den Brekel MWM, Castelijins JA, Snow GB. Imaging of cervical lymphadenopathy. Neuroimaging Clin N Am 1996: 6: 417–434. Vikram B, Strong EW, Shah JP, Spiro R. Failure at the primary site following multi-modality treatment in advanced head and neck cancer. Head Neck Surg 1984: 6: 720–723. Ying M, Ahuja A. Sonography of neck lymph nodes Part I: normal lymph nodes. Clin Rad 2003: 58: 351–358. Yuen PW, Lam KY, Chan AC, Wei WI, Lam LK. Clinicopathological analysis of local spread of carcinoma of the tongue. Am J Surg 1998: 175: 242–244.
331
25. Yuen PW, Lam KY, Lam LK, Ho CM, Wong A, Chow TL, Yuen WF, Wei WI. Prognostic factors of clinically stage I and II oral tongue carcinoma—a comparative study of stage, thickness, shape, growth pattern, invasive front malignancy grading, Martinez-Gimeno score, and pathologic features. Head Neck 2002: 24: 513– 520. Address: Suk Y. Ng Department of Dental Radiology Guy’s, King’s and St Thomas’ Dental Institute King’s College London Caldecot Road London SE5 9RW, UK. E-mail: [email protected]