REVIEW URRENT C OPINION
Magnetic resonance for laryngeal cancer Roberto Maroldi, Marco Ravanelli, and Davide Farina
Purpose of review This review summarizes the most recent experiences on the integration of magnetic resonance in assessing the local extent of laryngeal cancer and detecting submucosal recurrences. Recent findings Advances in magnetic resonance have been characterized by the development of technical solutions that shorten the acquisition time, thereby reducing motion artifacts, and increase the spatial resolution. Phasedarray surface coils, directly applied to the neck, enable the use of parallel-imaging techniques, which greatly reduce the acquisition time, and amplify the signal intensity, being closer to the larynx. One of the most important drawbacks of this technique is the small field-of-view, restricting the imaged area to the larynx. Furthermore, diffusion-weighted imaging (DWI) has increased the set of magnetic resonance sequences. Differently from computed tomography (CT), which has only two variables (precontrast/ postcontrast), magnetic resonance is based on a multiparameter analysis (T2-weighting and T1-weighting, DWI, and postcontrast acquisition). This multiparameter approach amplifies the contrast resolution. It has, also, permitted to differentiate scar tissue (after laser resection) from submucosal recurrent disease. In addition, DWI sequences have the potential of a more precise discrimination of peritumoral edema from neoplastic tissue, which may lead to improve the assessment of paraglottic space invasion. Summary Magnetic resonance of the larynx is technically challenging. The use of surface coils and motion-reducing techniques is critical to achieve adequate image quality. The intrinsic high-contrast resolution is further increased by the integration of information from different sequences. When CT has not been conclusive, magnetic resonance is indicated in the pretreatment local assessment and in the suspicion of submucosal recurrence. Keywords follow-up, laryngeal cancer, magnetic resonance, staging
INTRODUCTION In the work-up of laryngeal cancer, multislice computed tomography (CT) is currently the mainstay in the assessment of submucosal tumor spread [1]. Altthough MRI has the potential of a better discrimination of submucosal tissue changes and cartilage abnormalities, this technique requires a longer acquisition time, thus challenging patient cooperation and also hampering its more diffuse utilization [2,3,4 ]. To exploit the potential of magnetic resonance, the technique of examination has to be shaped to optimize the contrast and spatial resolution while maintaining the whole examination time affordable. &
MRI studies published in the last decade points out how the spatial resolution largely varies in slice thickness and in-plane resolution (Table 1) [2,4 , 5–7]. In our opinion, a slice thickness greater than 3 mm may result in ‘partial volume effect’ artifacts. For example, three key structures of the central larynx – the main prominence of false and true vocal cords (respectively, mainly fat and muscle) and the ventricle (air) in between – are comprised within a vertical distance of about 7–9 mm. To adequately separate the content of the three structures (fat, muscle, and air), their signal should be sampled in, at least, three consecutive slices. &
Department of Radiology, University of Brescia, Brescia, Italy
TECHNIQUE OF EXAMINATION Spatial resolution is a key issue when small structures have to be carefully scrutinized – such as those within the larynx. An analysis of the most relevant
Correspondence to Roberto Maroldi, MD, Department of Radiology, University of Brescia, Piazza Spedali Civili 1, 25123 Brescia, Italy. Tel: +39 30 395900; e-mail: [email protected] Curr Opin Otolaryngol Head Neck Surg 2014, 22:131–139 DOI:10.1097/MOO.0000000000000036
1068-9508 ß 2014 Wolters Kluwer Health | Lippincott Williams & Wilkins
www.co-otolaryngology.com
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
Head and neck oncology
KEY POINTS Magnetic resonance of the larynx is technically challenging, and the use of surface coils and motionreducing techniques is a promising solution. Thanks to its intrinsic high-contrast resolution and the possibility to integrate information from different sequences, magnetic resonance is preferable to CT in those cases in which tissue discrimination is crucial (pretreatment staging when CT has not been conclusive, follow-up after surgical and nonsurgical organ preservation treatments). DWI is a novel promising tool owning the potential to increase the precision of morphological sequences in defining tumor borders (separation between tumor and peritumor inflammation) and to increase the specificity in submucosal recurrence detection.
Of course, thin slices can be obtained, but at the cost of a decreased signal-to-noise ratio. This solution implies, also, an increased acquisition time, as more slices have to be acquired. To overcome this limitation, a solution is to use the coils directly applied to the patient’s neck (surface coils), being the signal intensity proportional to the distance between the structure examined and the receiver coil. Because of the increased signal and the small field of view (15–18 cm), both slice thickness and in-plane-resolution can thus be optimized. A limitation of this solution is that the effective signal is confined to a radius of 5 8 cm (depth height). Thereby, this limitation restricts to the larynx the area that can be imaged [2,4 ]. Nevertheless, some magnetic resonance equipment permits to use, simultaneously, surface and neck coils. The use of such ‘concentric set’ of coils enables both local high-resolution imaging and more global-region imaging [8]. A critical technical challenge for MRI of the larynx is the artifacts related to movement, namely breathing, swallowing, and vessel pulsation. Patients undergoing magnetic resonance examination are &
often unable to restrain from swallowing or moving for at least 90 s, particularly if a bulky tumor narrows the air space or patients have already been treated. Hence, it is necessary to obtain key information in the shortest possible examination time. Although technical improvements are continuous, until now it has not been possible to acquire high-resolution images while the patient maintains the apnea. Although movements related to breathing are the main cause of artifacts, a regular, rhythmic and not too deep breathing may avoid degrading significantly the image quality. To overcome the artifacts due to carotid arteries pulsation, presaturation bands located caudally to the larynx can be used to null the signal of the fast speed blood protons entering the volume of study. The development of parallel imaging techniques, possible with phased-array coils, enables to ‘split’ the coil into independent parts and acquire simultaneously the signal from each one [8]. This solution permits a significant reduction of the acquisition time. Another strategy available for motion artifacts reduction is to use the motion correction techniques, such as periodically rotated overlapping parallel lines with enhanced reconstructions (commercially named BLADE or PROPELLER, depending on the manufacturer), these sequences are sometimes the only way to obtain diagnostic images in very uncooperative patients [4 ]. &
THE INFLUENCE OF PATTERNS OF TUMOR SPREAD ON MRI TECHNIQUE OF EXAMINATION A precise assessment of submucosal tumor extent plays a crucial role in the treatment planning of laryngeal cancers. Particularly, both transoral laser excision and open partial laryngectomies owe part of their recent growth and diffusion to more and more accurate radiological assessment of tumor extent toward the paraglottic and preepiglottic spaces and the detection of cartilage invasion. On CT and magnetic resonance, the particular geometry of the paraglottic space – both its
Table 1. Case series in the English-language literature (case reports excluded) using different coils and spatial resolution for the magnetic resonance study of the larynx Author
Magnetic field
Coil
Thickness
In-plane resolution
Verduijn et al. 2009 [2]
1.5 T, 3.0 T
Neck coil and surface coils
4 mm
0.4–0.7 0.4 mm
Ravanelli et al. 2013 [4 ]
1.5 T
Surface coils
3 mm
0.3 0.5 mm
Ljumanovic et al. 2007 [5]
1.0 T, 1.5 T
Surface coils
3–5 mm
0.8 1.0 mm
Banko et al. 2011 [6]
1.5 T
Neck coil
4 mm
0.9 0.9 mm
Becker et al. 2008 [7]
1.5 T
Neck coil and surface coils
3–4 mm
From 0.4 0.4 to 0.8 0.8 mm
&
132
www.co-otolaryngology.com
Volume 22 Number 2 April 2014
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
Magnetic resonance for laryngeal cancer Maroldi et al.
transverse and vertical extent – needs to be explored by axial and coronal planes. In fact, this space is not simply a fat tissue stripe deep to the vocal cords. It is a complex volume filled with adipose and loose connective tissue [9,10]. Laterally, the space is confined by the thyroid cartilage, medially by the quadrangular lamina and the ventricle, inferiorly by the conus elasticus, and posteriorly by the piriform sinus. It is important to note that the space has ill-defined boundaries with the preepiglottic space (a true fibrous septum is often absent) and with the prelaryngeal tissues lateral to the crico-thyroid membrane. It is, also, in continuity with extralaryngeal soft tissues laterally to conus elasticus, thus being a pathway of extralaryngeal spread through the crico-thyroid membrane for advanced glottic and subglottic tumours (Figs. 1 and 2). The paraglottic space is also posteriorly in continuity with the piriform sinus. Invasion of the posterior part of the paraglottic space, behind the vocal process of the arytenoid cartilage, is a key information in treatment planning, regardless of the T class, because it contraindicates transoral laser surgery. Fat tissue indentations may extend from the paraglottic space inside the thyroarytenoid muscle. In its vertical orientation, the paraglottic space acts as a ‘lift’ for laryngeal tumors, leading to transglottic spread. Finally, growing laterally within the paraglottic space, tumors may reach and invade the thyroid cartilage [5].
TUMOR AND PERITUMORAL INFLAMMATION At the paraglottic space, magnetic resonance discriminates better than CT the fat tissue, hyperintense on both T2 and T1-weighted sequences, from the signal of the intrinsic muscles, hypointense on T2, intermediate on T1, and from tumor tissue – more hyperintense than muscles on T2 but steadily hypointense than fat on T1 [11]. Frequently, a combination of a coronal T2 (especially the turbo spin echo T2 sequence) and an axial T1 (spin echo T1) can accurately describe the relationship between tumor and paraglottic space, without the need of contrast agent administration (Fig. 3). Zbaren et al. [12] have reported a slightly greater sensitivity of magnetic resonance versus CT (97 versus 93%) in detecting paraglottic space invasion [11,13]. However, both magnetic resonance and CT showed a very low specificity. At present, there are no studies reporting the performance of more advanced magnetic resonance techniques, such as diffusionweighted imaging (DWI). The overestimation of the ‘conventional’ magnetic resonance protocol is probably related to the presence of peritumoral
FIGURE 1. Surface-coil MR of the larynx. Paraglottic space extent demonstrated on a TSE T2-weighted image in the coronal plane. The vertical extent of the paraglottic space is inside the dotted line. The curve arrow indicates the potential extralaryngeal pathway through the inferior paraglottic space. Arrowheads point toward the conus elasticus. MR, magnetic resonance; TSE, turbo spin echo.
inflammation, which ‘amplifies/inflates’ the boundaries of abnormal tissues [6] (Fig. 4). The same limitation has been advocated to explain the high sensitivity, but low specificity, of magnetic resonance in assessing thyroid cartilage invasion, prior to the proposal of new criteria [7]. The basic idea behind the development of the criteria is that the peritumoral inflammation close to the bone marrow inside the ossified thyroid cartilage translates into edema (increased content of interstitial water) and increased vascularization. In the adult, inside the ossified cartilage there is yellow marrow. It almost exclusively contains fat cells and very few capillaries [14]. More water within fat marrow will account for a further increase of the expected T2 hyperintensity, but – most of all – for a hypointensity on T1, instead of a hyperintensity (Fig. 5). As the postcontrast enhancement of the bone marrow might be due either to hyperemia related to inflammation or to
1068-9508 ß 2014 Wolters Kluwer Health | Lippincott Williams & Wilkins
www.co-otolaryngology.com
133
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
Head and neck oncology
(a)
(b)
FIGURE 2. Right glottic-subglottic scc. (a) On coronal CT MPR postcontrast, the tumor invades the right side of the cricoid cartilage (arrowheads). (b) The TSE T2-weighted coronal plane clearly shows the extralaryngeal spread through the inferior paraglottic space (white arrows), widened. Invasion of cricoid cartilage is also shown (arrowheads). CT MPR, computed tomography multiplanar reconstructions; TSE, turbo spin echo.
vascularized tumor tissue, the new criteria suggest a comparison of the grade of enhancement. If greater than in the adjacent tumor, inflammation is the most probable explanation. Here comes the potential of the recent extracranial application of DWI. In the larynx, DWI can be applied to discriminate tumor from inflammation inside the paraglottic space [4 ,15 ] (Fig. 6). The advantage introduced by DWI sequences consists &
(a)
&
in obtaining information about the cellularity of tissues. That is, squamous cell carcinoma, which is packed with cells, can be separated from adjacent inflammatory edema, because – in edematous tissues – cells are greatly sparse, there is more interstitial water content, and water moves without restriction in the interstitium. It is mostly in this setting that multislice CT may be insufficient to define the boundaries between neoplastic tissue
(b)
FIGURE 3. Left transglottic scc. (a) Coronal CT MPR postcontrast. The epicenter of tumor (T) is in the supraglottis, the solid enhancing mass surrounds a sclerotic aryitenoid cartilage (a) and contacts the thyroid lamina. The inferior paraglottic space (asterisk) is enlarged and more hypodense than the tumor. (b) On the TSE T2-weighted coronal plane, the intermediate signal of tumor (T) invading the (upper) posterior paraglottic space is well demarcated from the more hyperintense (asterisk) inferior paraglottic space, indicating edema. Although the normal lateral crico-arytenoid muscle (white arrow) is very hypointense, the edematous changes make the left muscle less hypointense. CT MPR, computed tomography multiplanar reconstructions; TSE, turbo spin echo. 134
www.co-otolaryngology.com
Volume 22 Number 2 April 2014
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
Magnetic resonance for laryngeal cancer Maroldi et al.
(a)
(b)
(c)
(d)
FIGURE 4. Right glottic scc. (a) Coronal CT MPR postcontrast. Tumor (asterisk) extends into the ventricle and false vocal cord. The fat tissue of the paraglottic space (white arrows) is replaced by tissue density greater than fat but lower than tumor. (b) TSE T2-weighted coronal plane. Tumor (asterisk) spreads inside the vocal part of the thyroarytenoid muscle and follows its muscular bundles lateral to the ventricle to reach the false vocal cord. The paraglottic space (black arrows) adjacent to the tumor is hyperintense: fat or more water into fat? (c) On the gradient-echo T1-weighted sequence postcontrast, the tumor (asterisk) and the adjacent paraglottic space enhance very similarly. (d) The ADC map shows that while tumor (asterisk) has low signal (restriction to water molecules movement), the adjacent paraglottic space has high signal, indicating edema. Arrowheads indicate the external boundaries of the thyroid laminae. ADC, apparent diffusion coefficient; CT MPR, computed tomography multiplanar reconstructions; TSE, turbo spin echo.
and edema, especially when the tumor does not enhance significantly after contrast administration.
MAGNETIC RESONANCE AND CARTILAGE INVASION The invasion of laryngeal cartilages influences the choice of surgical approach and the outcome of radiation therapy, leading to a greater risk of local recurrence and necrosis [16]. In glottic tumors, the
(a)
invasion of the inner cortical rim of the thyroid cartilage or cartilage penetration (T3) contraindicates transoral laser excision [17]. Either T3 or extralaryngeal cartilage tumor extent (T4a) can be treated by conservative surgical procedures, such as supracricoid laryngectomies. Although CT is characterized by an excellent ability of demonstrating the ossified rim of hyaline laryngeal cartilages, a recent study by Beitler et al. [18] challenges the predictive value of CT in detecting
(b)
(c)
FIGURE 5. Right glottic scc. The glottic level is analyzed by three different sequences: TSE T2-weighted (a), SE T1-weighted precontrast (b) and postcontrast (c). The tumor (T) invades the anterior two thirds of the true vocal cord. Marked edema (e) of the posterior vocal muscle, posterior paraglottic space and lateral cricoarytenoid muscle (lcam) is characterized by hyperintense signal on TSE T2 (a), hypointensity on SE T1 (c), greater enhancement than tumor on postcontrast SE T1 (c). Arrowheads on (a) and (c) delineate the boundary between tumor and edema. Similar changes are observed in the right thyroid lamina (asterisk), indicating edema. SE, spin echo; TSE, turbo spin echo. 1068-9508 ß 2014 Wolters Kluwer Health | Lippincott Williams & Wilkins
www.co-otolaryngology.com
135
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
Head and neck oncology
(a)
(b)
(c)
(d)
(e)
(f)
FIGURE 6. Right glottic scc. The glottic level is demonstrated by different sequences: TSE T2-weighted (a), DWI b1000 (b), SE T1-weighted postcontrast (c) and precontrast (d), DWI ADC map (e), 3D gradient echo fat-sat T1-weighted postcontrast (f). The greater contrast between tumor and adjacent structures is provided by TSE T2 and DWI sequences. On TSE T2 (a), the tumor (T) appears more hyperintense than both vocal (tvc) and lateral cricoarytenoid (lcam) muscles. On DWI, the tumor shows high signal on b1000 (b) and a very low signal on ADC map (e). This sequence amplifies the difference between tumor signal and edematous changes in the posterior vocal cord muscle (double-heads arrows on (b) and (e)). Extralaryngeal spread through the thyroid cartilage at the anterior commissure (T4a) is better depicted on T1 sequence and DWI b1000 (arrows on (c) and (b)). Intra-cartilage edema (arrowheads on (a)–(f)) is also demonstrated: inside the anterior right thyroid lamina the fat marrow is more hypointense on precontrast T1 (b), slighly more hyperintense on TSE T2 (a). Although the anatomy is less defined on DWI b1000, the signal inside the lamina is different from the neoplastic tissue. No significant postcontrast enhancement is seen on SE (c) and 3D gradient echo fat-sat (f) T1 sequences. ADC, apparent diffusion coefficient; DWI, diffusion-weighted imaging; TSE, turbo spin echo.
extralaryngeal spread. In their series of 107 consecutive previously untreated laryngectomy specimens, state of the art CT identified only 59% of pathologically documented thyroid cartilage invasion (T3) and 49% of extralaryngeal spread lesions (Fig. 7). The positive predictive values for thyroid cartilage penetration and extralaryngeal spread were 74 and 81%. Where does magnetic resonance stand? The new magnetic resonance criteria proposed by Becker et al. in 2008 [7] raised the specificity for thyroid cartilage invasion (T3 or T4a) from 54 to 75% while maintaining both a high sensitivity (96%) and high negative predictive value (96%). Again, the integration of DWI into the magnetic resonance protocol has the potential to increase the specificity.
MAGNETIC RESONANCE IN THE FOLLOWUP OF THE TREATED LARYNGEAL CANCER The follow-up of laryngeal cancer treated by conservative surgical approaches or (chemo)radiation 136
www.co-otolaryngology.com
therapy can be more challenging than the primary lesion, both for clinical assessment and for imaging studies [19]. Submucosal recurrences can be missed by endoscopy if a bulging is not present. In addition, deep biopsies guided by endoscopy alone are hindered by many false negatives [20]. In the posttreatment setting, the challenge for imaging techniques consists of the discrimination of persistent/recurrent tumor from inflammatory changes secondary to surgery or (chemo)radiation therapy: edema; granulation tissue/scar; radiationrelated fibrosis or necrosis of soft tissues; and cartilages. The greater contrast resolution of magnetic resonance permits a more precise classification of the complex tissue pattern within a laryngeal anatomy that can be distorted because of treatment. However, compliance and cooperation of treated patients are often more challenging than in pretreatment work-up. The edema of laryngeal soft tissues is an expected finding after surgery and radiation therapy, the latter Volume 22 Number 2 April 2014
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
Magnetic resonance for laryngeal cancer Maroldi et al.
(a)
(b)
FIGURE 7. Recurrent right glottic scc. Thirteen months post trans oral excision. On postcontrast axial 3D gradient echo fat-sat T1-weighted sequence (a) both the superficial recurrent tumor and the deep invasion through the thyroid lamina (arrows) are clearly shown. (b) The corresponding postcontrast CT does not demonstrate the superficial recurrence nor the T4a extent, although the irregular inner surface of the nonossified right thyroid lamina (arrowheads) can raise suspicion of abnormality at this level.
usually leading to a more persisting edema (months). As in the pretreatment peritumoral edema, a high signal on T2 and variable enhancement are usually observed. Scar is, indeed, a characteristic consequence of surgical excision, and it is quite confidently separable from other tissues by magnetic resonance because of very low signal intensity shown on T2-weighted images. T2 signal of scar tissue is typically lower than that showed by laryngeal
&
muscles [4 ]. Information regarding contrast enhancement may be puzzling. In fact, even if the absence of enhancement may indicate scar tissue and exclude recurrence, a variable grade of contrast enhancement can be observed in scar tissue. DWI is effective in the separation of edema, recurrence and scar, supporting T2-weighted information [15 ] (Fig. 8). Furthermore, DWI has the potential to improve recurrence detection being able to &
(a)
(b)
(c)
(d)
FIGURE 8. Recurrent left glottic scc. Fourteen months after type Va trans oral cordectomy. The ventricle level is demonstrated by four different sequences: DWI b1000 (a), TSE T2-weighted (b), DWI ADC map (c), 3D gradient echo fat-sat T1-weighted postcontrast (d). The recurrence (longer arrow) is better demonstrated on both DWI images (b1000 and ADC map, (a) and (c)) while almost missed on 3D gradient echo fat-sat T1-weighted postcontrast (d). On the TSE T2-weighted (b) image the recurrent tumor shows intermediate signal, brighter than the very low signal of postsurgical scar (arrows). On b1000 (a), the mature scar tissue is confirmed by the very low signal (arrows). ADC, apparent diffusion coefficient; DWI, diffusion-weighted imaging; TSE, turbo spin echo. 1068-9508 ß 2014 Wolters Kluwer Health | Lippincott Williams & Wilkins
www.co-otolaryngology.com
137
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
Head and neck oncology
(a)
(b)
(c)
(d)
FIGURE 9. Right transglottic scc. Submucosal persistent mass 4 months after chemoradiation therapy. The axial images – TSE T2-weighted (a), DWI ADC map (b) and 3D gradient echo fat-sat T1-weighted postcontrast (c) – are obtained at the level of the ventricle. The persistent tumor (T) shows heterogeneous signal on (a) and (c). Restriction on DWI ADC map (b) indicates that the residual tumor is vital. Peritumoral rim enhancement helps to detect the boundaries of tumor, which is surrounded by postradiation therapy edema. The signal inside the right thyroid lamina is hyperintense on (a) (arrowheads) and intermediate on (c), indicating ossified cartilage bone marrow edema. In the coronal TSE T2-weighted (d), the transglottic extent of the submucosal persistent tumor is associated with mass effect on the paraglottic space with medial and downward displacement of the true vocal cord (arrows). ADC, apparent diffusion coefficient; DWI, diffusion-weighted imaging; TSE, turbo spin echo.
differentiate recurrence from granulation tissue – both usually display an intermediate signal on T2 sequences [4 ]. In a series of 31 glottic cancers treated by transoral laser surgery, the combination of T2-weighted sequences and DWI showed an optimal diagnostic performance in detecting submucosal recurrences [4 ]. When the postradiotherapy and postchemotherapy larynx has to be imaged, two additional conditions may make imaging interpretation more challenging: (chondro)radionecrosis and residual mass. The CT pattern indicating chondro-radionecrosis is well established (cartilage destruction surrounded by air bubbles) and allows correct diagnosis in most cases. However, (chondro)radionecrosis and recurrent tumor may coexist. Fluorodeoxyglucose positron emission tomography has a high negative predictive value but is burdened with a high number of false positives [21–23]. DWI has the potential to differentiate radionecrosis and tumor recurrence [24] although – occasionally – fungal infection, pus, and dense necrotic debris along the walls of ulcerated lesions may cause signal restriction leading to false positive interpretation. To the best of our knowledge, no comparative studies including DWI have been carried out. After chemotherapy, residual masses may persist. In this case, imaging could be required to give information about residual mass viability. There are no studies investigating this issue. It is well known &
that PET-CT is not reliable before 3–4 months after the end of treatment [25]. In our – quite limited – experience in patients previously treated by chemo (radio)therapy and subsequently operated, DWI appeared as a promising tool in assessing tumor viability (Fig. 9).
&
138
www.co-otolaryngology.com
CONCLUSION State of the art magnetic resonance has several advantages in respect to multislice CT and few limitations, among which the most relevant is the degradation of image quality related to motion artifacts in noncooperative patients. A significant advantage consists in the possibility of analyzing laryngeal tissues by several different sequences. Their combination permits a more precise definition of signals indicating tumor, edema, and scar. When the clinical challenge depends on the precise local mapping of primary or relapsing tumor, magnetic resonance can provide information that is presently beyond the capability of CT. Acknowledgements None. Conflicts of interest Conflict of interest statement: None. Volume 22 Number 2 April 2014
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
Magnetic resonance for laryngeal cancer Maroldi et al.
REFERENCES AND RECOMMENDED READING Papers of particular interest, published within the annual period of review, have been highlighted as: & of special interest && of outstanding interest 1. Gilbert K, Dalley RW, Maronian N, et al. Staging of laryngeal cancer using 64-channel multidetector row CT: comparison of standard neck CT with dedicated breath-maneuver laryngeal CT. AJNR Am J Neuroradiol 2009; 31:251–256. 2. Verduijn GM, Bartels LW, Raaijmakers CP, et al. Magnetic resonance imaging protocol optimization for delineation of gross tumor volume in hypopharyngeal and laryngeal tumors. Int J Radiat Oncol Biol Phys 2009; 74:630–636. 3. Bertrand M, Tollard E, Francois A, et al. CT scan, MR imaging and anatomopathologic correlation in the glottic carcinoma T1-T2. Rev Laryngol Otol Rhinol (Bord) 2010; 131:51–57. 4. Ravanelli M, Farina D, Rizzardi P, et al. MR with surface coils in the follow-up & after endoscopic laser resection for glottic squamous cell carcinoma: feasibility and diagnostic accuracy. Neuroradiology 2013; 55:225–232. Original approach combining surface coils and multiple sequences. High-spatial resolution turbo spin echo T2 and DWI sequences allow high sensitivity and specificity in the detection of submucosal recurrences after trans oral surgery. 5. Ljumanovic R, Langendijk JA, van Wattingen M, et al. MR imaging predictors of local control of glottic squamous cell carcinoma treated with radiation alone. Radiology 2007; 244:205–212. 6. Banko B, Dukic V, Milovanovic J, et al. Diagnostic significance of magnetic resonance imaging in preoperative evaluation of patients with laryngeal tumors. Eur Arch Otorhinolaryngol 2011; 268:1617–1623. 7. Becker M, Zbaren P, Casselman JW, et al. Neoplastic invasion of laryngeal cartilage: reassessment of criteria for diagnosis at MR imaging. Radiology 2008; 249:551–559. 8. Casselman JW. High resolution imaging of the skull base and larynx. In: Schoenberg SO, Dietrich O, Reiser MF, editors. Parallel imaging in clinical MR applications. Berlin Heidelberg New York: Springer; 2007. 9. Reidenbach MM. The paraglottic space and transglottic cancer: anatomical considerations. Clin Anat 1996; 9:244–251. 10. Reidenbach MM. Borders and topographic relationships of the paraglottic space. Eur Arch Otorhinolaryngol 1997; 254:193–195. 11. Becker M, Burkhardt K, Allal AS, et al. Pretherapeutic and posttherapeutic laryngeal imaging. Radiologe 2009; 49:43–58.
12. Zbaren P, Becker M, Lang H. Pretherapeutic staging of laryngeal carcinoma. Clinical findings, computed tomography, and magnetic resonance imaging compared with histopathology. Cancer 1996; 77:1263–1273. 13. Becker M. Diagnose und Stadieneinteilung von Larynxtumoren mittels CT und MRT. Radiologe 1998; 38:93–100. 14. Vogler JB 3rd, Murphy WA. Bone marrow imaging. Radiology 1988; 168:679–693. 15. Tshering Vogel DW, Zbaeren P, Geretschlaeger A, et al. Diffusion-weighted & MR imaging including bi-exponential fitting for the detection of recurrent or residual tumour after (chemo)radiotherapy for laryngeal and hypopharyngeal cancers. Eur Radiol 2013; 23:562–569. The article explores the role of DWI in the detection of residual/recurrent tumors after (chemo)radiation. The special imaging technique used is very promising, but still difficult to be transferred into everyday clinical practice. 16. Wagner MM, Cure JK, Caudell JJ, et al. Prognostic significance of thyroid or cricoid cartilage invasion in laryngeal or hypopharyngeal cancer treated with organ preserving strategies. Radiat Oncol 2012; 7:219. 17. Peretti G, Piazza C, Ansarin M, et al. Transoral CO2 laser microsurgery for TisT3 supraglottic squamous cell carcinomas. Eur Arch Otorhinolaryngol 2010; 267:1735–1742. 18. Beitler JJ, Muller S, Grist WJ, et al. Prognostic accuracy of computed tomography findings for patients with laryngeal cancer undergoing laryngectomy. J Clin Oncol 2010; 28:2318–2322. 19. Zbaren P, Weidner S, Thoeny HC. Laryngeal and hypopharyngeal carcinomas after (chemo)radiotherapy: a diagnostic dilemma. Curr Opin Otolaryngol Head Neck Surg 2008; 16:147–153. 20. McGuirt WF, Greven KM, Keyes JW, et al. Laryngeal radionecrosis versus recurrent cancer: a clinical approach. Ann Otol Rhinol Laryngol 1998; 107:293–296. 21. Chu MM, Kositwattanarerk A, Lee DJ, et al. FDG PET with contrast-enhanced CT: a critical imaging tool for laryngeal carcinoma. Radiographics 2010; 30:1353–1372. 22. Gilbert MR, Branstetter BFT, Kim S. Utility of positron-emission tomography/ computed tomography imaging in the management of the neck in recurrent laryngeal cancer. Laryngoscope 2012; 122:821–825. 23. de Bree R, van der Putten L, Brouwer J, et al. Detection of locoregional recurrent head and neck cancer after (chemo)radiotherapy using modern imaging. Oral Oncol 2009; 45:386–393. 24. Vandecaveye V, de Keyzer F, Vander Poorten V, et al. Evaluation of the larynx for tumour recurrence by diffusion-weighted MRI after radiotherapy: initial experience in four cases. Br J Radiol 2006; 79:681–687. 25. Ong SC, Schoder H, Lee NY, et al. Clinical utility of 18F-FDG PET/CT in assessing the neck after concurrent chemoradiotherapy for Locoregional advanced head and neck cancer. J Nucl Med 2008; 49:532–540.
1068-9508 ß 2014 Wolters Kluwer Health | Lippincott Williams & Wilkins
www.co-otolaryngology.com
139
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
Magnetic resonance for laryngeal cancer Roberto Maroldi, Marco Ravanelli, and Davide Farina
Purpose of review This review summarizes the most recent experiences on the integration of magnetic resonance in assessing the local extent of laryngeal cancer and detecting submucosal recurrences. Recent findings Advances in magnetic resonance have been characterized by the development of technical solutions that shorten the acquisition time, thereby reducing motion artifacts, and increase the spatial resolution. Phasedarray surface coils, directly applied to the neck, enable the use of parallel-imaging techniques, which greatly reduce the acquisition time, and amplify the signal intensity, being closer to the larynx. One of the most important drawbacks of this technique is the small field-of-view, restricting the imaged area to the larynx. Furthermore, diffusion-weighted imaging (DWI) has increased the set of magnetic resonance sequences. Differently from computed tomography (CT), which has only two variables (precontrast/ postcontrast), magnetic resonance is based on a multiparameter analysis (T2-weighting and T1-weighting, DWI, and postcontrast acquisition). This multiparameter approach amplifies the contrast resolution. It has, also, permitted to differentiate scar tissue (after laser resection) from submucosal recurrent disease. In addition, DWI sequences have the potential of a more precise discrimination of peritumoral edema from neoplastic tissue, which may lead to improve the assessment of paraglottic space invasion. Summary Magnetic resonance of the larynx is technically challenging. The use of surface coils and motion-reducing techniques is critical to achieve adequate image quality. The intrinsic high-contrast resolution is further increased by the integration of information from different sequences. When CT has not been conclusive, magnetic resonance is indicated in the pretreatment local assessment and in the suspicion of submucosal recurrence. Keywords follow-up, laryngeal cancer, magnetic resonance, staging
INTRODUCTION In the work-up of laryngeal cancer, multislice computed tomography (CT) is currently the mainstay in the assessment of submucosal tumor spread [1]. Altthough MRI has the potential of a better discrimination of submucosal tissue changes and cartilage abnormalities, this technique requires a longer acquisition time, thus challenging patient cooperation and also hampering its more diffuse utilization [2,3,4 ]. To exploit the potential of magnetic resonance, the technique of examination has to be shaped to optimize the contrast and spatial resolution while maintaining the whole examination time affordable. &
MRI studies published in the last decade points out how the spatial resolution largely varies in slice thickness and in-plane resolution (Table 1) [2,4 , 5–7]. In our opinion, a slice thickness greater than 3 mm may result in ‘partial volume effect’ artifacts. For example, three key structures of the central larynx – the main prominence of false and true vocal cords (respectively, mainly fat and muscle) and the ventricle (air) in between – are comprised within a vertical distance of about 7–9 mm. To adequately separate the content of the three structures (fat, muscle, and air), their signal should be sampled in, at least, three consecutive slices. &
Department of Radiology, University of Brescia, Brescia, Italy
TECHNIQUE OF EXAMINATION Spatial resolution is a key issue when small structures have to be carefully scrutinized – such as those within the larynx. An analysis of the most relevant
Correspondence to Roberto Maroldi, MD, Department of Radiology, University of Brescia, Piazza Spedali Civili 1, 25123 Brescia, Italy. Tel: +39 30 395900; e-mail: [email protected] Curr Opin Otolaryngol Head Neck Surg 2014, 22:131–139 DOI:10.1097/MOO.0000000000000036
1068-9508 ß 2014 Wolters Kluwer Health | Lippincott Williams & Wilkins
www.co-otolaryngology.com
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
Head and neck oncology
KEY POINTS Magnetic resonance of the larynx is technically challenging, and the use of surface coils and motionreducing techniques is a promising solution. Thanks to its intrinsic high-contrast resolution and the possibility to integrate information from different sequences, magnetic resonance is preferable to CT in those cases in which tissue discrimination is crucial (pretreatment staging when CT has not been conclusive, follow-up after surgical and nonsurgical organ preservation treatments). DWI is a novel promising tool owning the potential to increase the precision of morphological sequences in defining tumor borders (separation between tumor and peritumor inflammation) and to increase the specificity in submucosal recurrence detection.
Of course, thin slices can be obtained, but at the cost of a decreased signal-to-noise ratio. This solution implies, also, an increased acquisition time, as more slices have to be acquired. To overcome this limitation, a solution is to use the coils directly applied to the patient’s neck (surface coils), being the signal intensity proportional to the distance between the structure examined and the receiver coil. Because of the increased signal and the small field of view (15–18 cm), both slice thickness and in-plane-resolution can thus be optimized. A limitation of this solution is that the effective signal is confined to a radius of 5 8 cm (depth height). Thereby, this limitation restricts to the larynx the area that can be imaged [2,4 ]. Nevertheless, some magnetic resonance equipment permits to use, simultaneously, surface and neck coils. The use of such ‘concentric set’ of coils enables both local high-resolution imaging and more global-region imaging [8]. A critical technical challenge for MRI of the larynx is the artifacts related to movement, namely breathing, swallowing, and vessel pulsation. Patients undergoing magnetic resonance examination are &
often unable to restrain from swallowing or moving for at least 90 s, particularly if a bulky tumor narrows the air space or patients have already been treated. Hence, it is necessary to obtain key information in the shortest possible examination time. Although technical improvements are continuous, until now it has not been possible to acquire high-resolution images while the patient maintains the apnea. Although movements related to breathing are the main cause of artifacts, a regular, rhythmic and not too deep breathing may avoid degrading significantly the image quality. To overcome the artifacts due to carotid arteries pulsation, presaturation bands located caudally to the larynx can be used to null the signal of the fast speed blood protons entering the volume of study. The development of parallel imaging techniques, possible with phased-array coils, enables to ‘split’ the coil into independent parts and acquire simultaneously the signal from each one [8]. This solution permits a significant reduction of the acquisition time. Another strategy available for motion artifacts reduction is to use the motion correction techniques, such as periodically rotated overlapping parallel lines with enhanced reconstructions (commercially named BLADE or PROPELLER, depending on the manufacturer), these sequences are sometimes the only way to obtain diagnostic images in very uncooperative patients [4 ]. &
THE INFLUENCE OF PATTERNS OF TUMOR SPREAD ON MRI TECHNIQUE OF EXAMINATION A precise assessment of submucosal tumor extent plays a crucial role in the treatment planning of laryngeal cancers. Particularly, both transoral laser excision and open partial laryngectomies owe part of their recent growth and diffusion to more and more accurate radiological assessment of tumor extent toward the paraglottic and preepiglottic spaces and the detection of cartilage invasion. On CT and magnetic resonance, the particular geometry of the paraglottic space – both its
Table 1. Case series in the English-language literature (case reports excluded) using different coils and spatial resolution for the magnetic resonance study of the larynx Author
Magnetic field
Coil
Thickness
In-plane resolution
Verduijn et al. 2009 [2]
1.5 T, 3.0 T
Neck coil and surface coils
4 mm
0.4–0.7 0.4 mm
Ravanelli et al. 2013 [4 ]
1.5 T
Surface coils
3 mm
0.3 0.5 mm
Ljumanovic et al. 2007 [5]
1.0 T, 1.5 T
Surface coils
3–5 mm
0.8 1.0 mm
Banko et al. 2011 [6]
1.5 T
Neck coil
4 mm
0.9 0.9 mm
Becker et al. 2008 [7]
1.5 T
Neck coil and surface coils
3–4 mm
From 0.4 0.4 to 0.8 0.8 mm
&
132
www.co-otolaryngology.com
Volume 22 Number 2 April 2014
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
Magnetic resonance for laryngeal cancer Maroldi et al.
transverse and vertical extent – needs to be explored by axial and coronal planes. In fact, this space is not simply a fat tissue stripe deep to the vocal cords. It is a complex volume filled with adipose and loose connective tissue [9,10]. Laterally, the space is confined by the thyroid cartilage, medially by the quadrangular lamina and the ventricle, inferiorly by the conus elasticus, and posteriorly by the piriform sinus. It is important to note that the space has ill-defined boundaries with the preepiglottic space (a true fibrous septum is often absent) and with the prelaryngeal tissues lateral to the crico-thyroid membrane. It is, also, in continuity with extralaryngeal soft tissues laterally to conus elasticus, thus being a pathway of extralaryngeal spread through the crico-thyroid membrane for advanced glottic and subglottic tumours (Figs. 1 and 2). The paraglottic space is also posteriorly in continuity with the piriform sinus. Invasion of the posterior part of the paraglottic space, behind the vocal process of the arytenoid cartilage, is a key information in treatment planning, regardless of the T class, because it contraindicates transoral laser surgery. Fat tissue indentations may extend from the paraglottic space inside the thyroarytenoid muscle. In its vertical orientation, the paraglottic space acts as a ‘lift’ for laryngeal tumors, leading to transglottic spread. Finally, growing laterally within the paraglottic space, tumors may reach and invade the thyroid cartilage [5].
TUMOR AND PERITUMORAL INFLAMMATION At the paraglottic space, magnetic resonance discriminates better than CT the fat tissue, hyperintense on both T2 and T1-weighted sequences, from the signal of the intrinsic muscles, hypointense on T2, intermediate on T1, and from tumor tissue – more hyperintense than muscles on T2 but steadily hypointense than fat on T1 [11]. Frequently, a combination of a coronal T2 (especially the turbo spin echo T2 sequence) and an axial T1 (spin echo T1) can accurately describe the relationship between tumor and paraglottic space, without the need of contrast agent administration (Fig. 3). Zbaren et al. [12] have reported a slightly greater sensitivity of magnetic resonance versus CT (97 versus 93%) in detecting paraglottic space invasion [11,13]. However, both magnetic resonance and CT showed a very low specificity. At present, there are no studies reporting the performance of more advanced magnetic resonance techniques, such as diffusionweighted imaging (DWI). The overestimation of the ‘conventional’ magnetic resonance protocol is probably related to the presence of peritumoral
FIGURE 1. Surface-coil MR of the larynx. Paraglottic space extent demonstrated on a TSE T2-weighted image in the coronal plane. The vertical extent of the paraglottic space is inside the dotted line. The curve arrow indicates the potential extralaryngeal pathway through the inferior paraglottic space. Arrowheads point toward the conus elasticus. MR, magnetic resonance; TSE, turbo spin echo.
inflammation, which ‘amplifies/inflates’ the boundaries of abnormal tissues [6] (Fig. 4). The same limitation has been advocated to explain the high sensitivity, but low specificity, of magnetic resonance in assessing thyroid cartilage invasion, prior to the proposal of new criteria [7]. The basic idea behind the development of the criteria is that the peritumoral inflammation close to the bone marrow inside the ossified thyroid cartilage translates into edema (increased content of interstitial water) and increased vascularization. In the adult, inside the ossified cartilage there is yellow marrow. It almost exclusively contains fat cells and very few capillaries [14]. More water within fat marrow will account for a further increase of the expected T2 hyperintensity, but – most of all – for a hypointensity on T1, instead of a hyperintensity (Fig. 5). As the postcontrast enhancement of the bone marrow might be due either to hyperemia related to inflammation or to
1068-9508 ß 2014 Wolters Kluwer Health | Lippincott Williams & Wilkins
www.co-otolaryngology.com
133
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
Head and neck oncology
(a)
(b)
FIGURE 2. Right glottic-subglottic scc. (a) On coronal CT MPR postcontrast, the tumor invades the right side of the cricoid cartilage (arrowheads). (b) The TSE T2-weighted coronal plane clearly shows the extralaryngeal spread through the inferior paraglottic space (white arrows), widened. Invasion of cricoid cartilage is also shown (arrowheads). CT MPR, computed tomography multiplanar reconstructions; TSE, turbo spin echo.
vascularized tumor tissue, the new criteria suggest a comparison of the grade of enhancement. If greater than in the adjacent tumor, inflammation is the most probable explanation. Here comes the potential of the recent extracranial application of DWI. In the larynx, DWI can be applied to discriminate tumor from inflammation inside the paraglottic space [4 ,15 ] (Fig. 6). The advantage introduced by DWI sequences consists &
(a)
&
in obtaining information about the cellularity of tissues. That is, squamous cell carcinoma, which is packed with cells, can be separated from adjacent inflammatory edema, because – in edematous tissues – cells are greatly sparse, there is more interstitial water content, and water moves without restriction in the interstitium. It is mostly in this setting that multislice CT may be insufficient to define the boundaries between neoplastic tissue
(b)
FIGURE 3. Left transglottic scc. (a) Coronal CT MPR postcontrast. The epicenter of tumor (T) is in the supraglottis, the solid enhancing mass surrounds a sclerotic aryitenoid cartilage (a) and contacts the thyroid lamina. The inferior paraglottic space (asterisk) is enlarged and more hypodense than the tumor. (b) On the TSE T2-weighted coronal plane, the intermediate signal of tumor (T) invading the (upper) posterior paraglottic space is well demarcated from the more hyperintense (asterisk) inferior paraglottic space, indicating edema. Although the normal lateral crico-arytenoid muscle (white arrow) is very hypointense, the edematous changes make the left muscle less hypointense. CT MPR, computed tomography multiplanar reconstructions; TSE, turbo spin echo. 134
www.co-otolaryngology.com
Volume 22 Number 2 April 2014
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
Magnetic resonance for laryngeal cancer Maroldi et al.
(a)
(b)
(c)
(d)
FIGURE 4. Right glottic scc. (a) Coronal CT MPR postcontrast. Tumor (asterisk) extends into the ventricle and false vocal cord. The fat tissue of the paraglottic space (white arrows) is replaced by tissue density greater than fat but lower than tumor. (b) TSE T2-weighted coronal plane. Tumor (asterisk) spreads inside the vocal part of the thyroarytenoid muscle and follows its muscular bundles lateral to the ventricle to reach the false vocal cord. The paraglottic space (black arrows) adjacent to the tumor is hyperintense: fat or more water into fat? (c) On the gradient-echo T1-weighted sequence postcontrast, the tumor (asterisk) and the adjacent paraglottic space enhance very similarly. (d) The ADC map shows that while tumor (asterisk) has low signal (restriction to water molecules movement), the adjacent paraglottic space has high signal, indicating edema. Arrowheads indicate the external boundaries of the thyroid laminae. ADC, apparent diffusion coefficient; CT MPR, computed tomography multiplanar reconstructions; TSE, turbo spin echo.
and edema, especially when the tumor does not enhance significantly after contrast administration.
MAGNETIC RESONANCE AND CARTILAGE INVASION The invasion of laryngeal cartilages influences the choice of surgical approach and the outcome of radiation therapy, leading to a greater risk of local recurrence and necrosis [16]. In glottic tumors, the
(a)
invasion of the inner cortical rim of the thyroid cartilage or cartilage penetration (T3) contraindicates transoral laser excision [17]. Either T3 or extralaryngeal cartilage tumor extent (T4a) can be treated by conservative surgical procedures, such as supracricoid laryngectomies. Although CT is characterized by an excellent ability of demonstrating the ossified rim of hyaline laryngeal cartilages, a recent study by Beitler et al. [18] challenges the predictive value of CT in detecting
(b)
(c)
FIGURE 5. Right glottic scc. The glottic level is analyzed by three different sequences: TSE T2-weighted (a), SE T1-weighted precontrast (b) and postcontrast (c). The tumor (T) invades the anterior two thirds of the true vocal cord. Marked edema (e) of the posterior vocal muscle, posterior paraglottic space and lateral cricoarytenoid muscle (lcam) is characterized by hyperintense signal on TSE T2 (a), hypointensity on SE T1 (c), greater enhancement than tumor on postcontrast SE T1 (c). Arrowheads on (a) and (c) delineate the boundary between tumor and edema. Similar changes are observed in the right thyroid lamina (asterisk), indicating edema. SE, spin echo; TSE, turbo spin echo. 1068-9508 ß 2014 Wolters Kluwer Health | Lippincott Williams & Wilkins
www.co-otolaryngology.com
135
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
Head and neck oncology
(a)
(b)
(c)
(d)
(e)
(f)
FIGURE 6. Right glottic scc. The glottic level is demonstrated by different sequences: TSE T2-weighted (a), DWI b1000 (b), SE T1-weighted postcontrast (c) and precontrast (d), DWI ADC map (e), 3D gradient echo fat-sat T1-weighted postcontrast (f). The greater contrast between tumor and adjacent structures is provided by TSE T2 and DWI sequences. On TSE T2 (a), the tumor (T) appears more hyperintense than both vocal (tvc) and lateral cricoarytenoid (lcam) muscles. On DWI, the tumor shows high signal on b1000 (b) and a very low signal on ADC map (e). This sequence amplifies the difference between tumor signal and edematous changes in the posterior vocal cord muscle (double-heads arrows on (b) and (e)). Extralaryngeal spread through the thyroid cartilage at the anterior commissure (T4a) is better depicted on T1 sequence and DWI b1000 (arrows on (c) and (b)). Intra-cartilage edema (arrowheads on (a)–(f)) is also demonstrated: inside the anterior right thyroid lamina the fat marrow is more hypointense on precontrast T1 (b), slighly more hyperintense on TSE T2 (a). Although the anatomy is less defined on DWI b1000, the signal inside the lamina is different from the neoplastic tissue. No significant postcontrast enhancement is seen on SE (c) and 3D gradient echo fat-sat (f) T1 sequences. ADC, apparent diffusion coefficient; DWI, diffusion-weighted imaging; TSE, turbo spin echo.
extralaryngeal spread. In their series of 107 consecutive previously untreated laryngectomy specimens, state of the art CT identified only 59% of pathologically documented thyroid cartilage invasion (T3) and 49% of extralaryngeal spread lesions (Fig. 7). The positive predictive values for thyroid cartilage penetration and extralaryngeal spread were 74 and 81%. Where does magnetic resonance stand? The new magnetic resonance criteria proposed by Becker et al. in 2008 [7] raised the specificity for thyroid cartilage invasion (T3 or T4a) from 54 to 75% while maintaining both a high sensitivity (96%) and high negative predictive value (96%). Again, the integration of DWI into the magnetic resonance protocol has the potential to increase the specificity.
MAGNETIC RESONANCE IN THE FOLLOWUP OF THE TREATED LARYNGEAL CANCER The follow-up of laryngeal cancer treated by conservative surgical approaches or (chemo)radiation 136
www.co-otolaryngology.com
therapy can be more challenging than the primary lesion, both for clinical assessment and for imaging studies [19]. Submucosal recurrences can be missed by endoscopy if a bulging is not present. In addition, deep biopsies guided by endoscopy alone are hindered by many false negatives [20]. In the posttreatment setting, the challenge for imaging techniques consists of the discrimination of persistent/recurrent tumor from inflammatory changes secondary to surgery or (chemo)radiation therapy: edema; granulation tissue/scar; radiationrelated fibrosis or necrosis of soft tissues; and cartilages. The greater contrast resolution of magnetic resonance permits a more precise classification of the complex tissue pattern within a laryngeal anatomy that can be distorted because of treatment. However, compliance and cooperation of treated patients are often more challenging than in pretreatment work-up. The edema of laryngeal soft tissues is an expected finding after surgery and radiation therapy, the latter Volume 22 Number 2 April 2014
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
Magnetic resonance for laryngeal cancer Maroldi et al.
(a)
(b)
FIGURE 7. Recurrent right glottic scc. Thirteen months post trans oral excision. On postcontrast axial 3D gradient echo fat-sat T1-weighted sequence (a) both the superficial recurrent tumor and the deep invasion through the thyroid lamina (arrows) are clearly shown. (b) The corresponding postcontrast CT does not demonstrate the superficial recurrence nor the T4a extent, although the irregular inner surface of the nonossified right thyroid lamina (arrowheads) can raise suspicion of abnormality at this level.
usually leading to a more persisting edema (months). As in the pretreatment peritumoral edema, a high signal on T2 and variable enhancement are usually observed. Scar is, indeed, a characteristic consequence of surgical excision, and it is quite confidently separable from other tissues by magnetic resonance because of very low signal intensity shown on T2-weighted images. T2 signal of scar tissue is typically lower than that showed by laryngeal
&
muscles [4 ]. Information regarding contrast enhancement may be puzzling. In fact, even if the absence of enhancement may indicate scar tissue and exclude recurrence, a variable grade of contrast enhancement can be observed in scar tissue. DWI is effective in the separation of edema, recurrence and scar, supporting T2-weighted information [15 ] (Fig. 8). Furthermore, DWI has the potential to improve recurrence detection being able to &
(a)
(b)
(c)
(d)
FIGURE 8. Recurrent left glottic scc. Fourteen months after type Va trans oral cordectomy. The ventricle level is demonstrated by four different sequences: DWI b1000 (a), TSE T2-weighted (b), DWI ADC map (c), 3D gradient echo fat-sat T1-weighted postcontrast (d). The recurrence (longer arrow) is better demonstrated on both DWI images (b1000 and ADC map, (a) and (c)) while almost missed on 3D gradient echo fat-sat T1-weighted postcontrast (d). On the TSE T2-weighted (b) image the recurrent tumor shows intermediate signal, brighter than the very low signal of postsurgical scar (arrows). On b1000 (a), the mature scar tissue is confirmed by the very low signal (arrows). ADC, apparent diffusion coefficient; DWI, diffusion-weighted imaging; TSE, turbo spin echo. 1068-9508 ß 2014 Wolters Kluwer Health | Lippincott Williams & Wilkins
www.co-otolaryngology.com
137
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
Head and neck oncology
(a)
(b)
(c)
(d)
FIGURE 9. Right transglottic scc. Submucosal persistent mass 4 months after chemoradiation therapy. The axial images – TSE T2-weighted (a), DWI ADC map (b) and 3D gradient echo fat-sat T1-weighted postcontrast (c) – are obtained at the level of the ventricle. The persistent tumor (T) shows heterogeneous signal on (a) and (c). Restriction on DWI ADC map (b) indicates that the residual tumor is vital. Peritumoral rim enhancement helps to detect the boundaries of tumor, which is surrounded by postradiation therapy edema. The signal inside the right thyroid lamina is hyperintense on (a) (arrowheads) and intermediate on (c), indicating ossified cartilage bone marrow edema. In the coronal TSE T2-weighted (d), the transglottic extent of the submucosal persistent tumor is associated with mass effect on the paraglottic space with medial and downward displacement of the true vocal cord (arrows). ADC, apparent diffusion coefficient; DWI, diffusion-weighted imaging; TSE, turbo spin echo.
differentiate recurrence from granulation tissue – both usually display an intermediate signal on T2 sequences [4 ]. In a series of 31 glottic cancers treated by transoral laser surgery, the combination of T2-weighted sequences and DWI showed an optimal diagnostic performance in detecting submucosal recurrences [4 ]. When the postradiotherapy and postchemotherapy larynx has to be imaged, two additional conditions may make imaging interpretation more challenging: (chondro)radionecrosis and residual mass. The CT pattern indicating chondro-radionecrosis is well established (cartilage destruction surrounded by air bubbles) and allows correct diagnosis in most cases. However, (chondro)radionecrosis and recurrent tumor may coexist. Fluorodeoxyglucose positron emission tomography has a high negative predictive value but is burdened with a high number of false positives [21–23]. DWI has the potential to differentiate radionecrosis and tumor recurrence [24] although – occasionally – fungal infection, pus, and dense necrotic debris along the walls of ulcerated lesions may cause signal restriction leading to false positive interpretation. To the best of our knowledge, no comparative studies including DWI have been carried out. After chemotherapy, residual masses may persist. In this case, imaging could be required to give information about residual mass viability. There are no studies investigating this issue. It is well known &
that PET-CT is not reliable before 3–4 months after the end of treatment [25]. In our – quite limited – experience in patients previously treated by chemo (radio)therapy and subsequently operated, DWI appeared as a promising tool in assessing tumor viability (Fig. 9).
&
138
www.co-otolaryngology.com
CONCLUSION State of the art magnetic resonance has several advantages in respect to multislice CT and few limitations, among which the most relevant is the degradation of image quality related to motion artifacts in noncooperative patients. A significant advantage consists in the possibility of analyzing laryngeal tissues by several different sequences. Their combination permits a more precise definition of signals indicating tumor, edema, and scar. When the clinical challenge depends on the precise local mapping of primary or relapsing tumor, magnetic resonance can provide information that is presently beyond the capability of CT. Acknowledgements None. Conflicts of interest Conflict of interest statement: None. Volume 22 Number 2 April 2014
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
Magnetic resonance for laryngeal cancer Maroldi et al.
REFERENCES AND RECOMMENDED READING Papers of particular interest, published within the annual period of review, have been highlighted as: & of special interest && of outstanding interest 1. Gilbert K, Dalley RW, Maronian N, et al. Staging of laryngeal cancer using 64-channel multidetector row CT: comparison of standard neck CT with dedicated breath-maneuver laryngeal CT. AJNR Am J Neuroradiol 2009; 31:251–256. 2. Verduijn GM, Bartels LW, Raaijmakers CP, et al. Magnetic resonance imaging protocol optimization for delineation of gross tumor volume in hypopharyngeal and laryngeal tumors. Int J Radiat Oncol Biol Phys 2009; 74:630–636. 3. Bertrand M, Tollard E, Francois A, et al. CT scan, MR imaging and anatomopathologic correlation in the glottic carcinoma T1-T2. Rev Laryngol Otol Rhinol (Bord) 2010; 131:51–57. 4. Ravanelli M, Farina D, Rizzardi P, et al. MR with surface coils in the follow-up & after endoscopic laser resection for glottic squamous cell carcinoma: feasibility and diagnostic accuracy. Neuroradiology 2013; 55:225–232. Original approach combining surface coils and multiple sequences. High-spatial resolution turbo spin echo T2 and DWI sequences allow high sensitivity and specificity in the detection of submucosal recurrences after trans oral surgery. 5. Ljumanovic R, Langendijk JA, van Wattingen M, et al. MR imaging predictors of local control of glottic squamous cell carcinoma treated with radiation alone. Radiology 2007; 244:205–212. 6. Banko B, Dukic V, Milovanovic J, et al. Diagnostic significance of magnetic resonance imaging in preoperative evaluation of patients with laryngeal tumors. Eur Arch Otorhinolaryngol 2011; 268:1617–1623. 7. Becker M, Zbaren P, Casselman JW, et al. Neoplastic invasion of laryngeal cartilage: reassessment of criteria for diagnosis at MR imaging. Radiology 2008; 249:551–559. 8. Casselman JW. High resolution imaging of the skull base and larynx. In: Schoenberg SO, Dietrich O, Reiser MF, editors. Parallel imaging in clinical MR applications. Berlin Heidelberg New York: Springer; 2007. 9. Reidenbach MM. The paraglottic space and transglottic cancer: anatomical considerations. Clin Anat 1996; 9:244–251. 10. Reidenbach MM. Borders and topographic relationships of the paraglottic space. Eur Arch Otorhinolaryngol 1997; 254:193–195. 11. Becker M, Burkhardt K, Allal AS, et al. Pretherapeutic and posttherapeutic laryngeal imaging. Radiologe 2009; 49:43–58.
12. Zbaren P, Becker M, Lang H. Pretherapeutic staging of laryngeal carcinoma. Clinical findings, computed tomography, and magnetic resonance imaging compared with histopathology. Cancer 1996; 77:1263–1273. 13. Becker M. Diagnose und Stadieneinteilung von Larynxtumoren mittels CT und MRT. Radiologe 1998; 38:93–100. 14. Vogler JB 3rd, Murphy WA. Bone marrow imaging. Radiology 1988; 168:679–693. 15. Tshering Vogel DW, Zbaeren P, Geretschlaeger A, et al. Diffusion-weighted & MR imaging including bi-exponential fitting for the detection of recurrent or residual tumour after (chemo)radiotherapy for laryngeal and hypopharyngeal cancers. Eur Radiol 2013; 23:562–569. The article explores the role of DWI in the detection of residual/recurrent tumors after (chemo)radiation. The special imaging technique used is very promising, but still difficult to be transferred into everyday clinical practice. 16. Wagner MM, Cure JK, Caudell JJ, et al. Prognostic significance of thyroid or cricoid cartilage invasion in laryngeal or hypopharyngeal cancer treated with organ preserving strategies. Radiat Oncol 2012; 7:219. 17. Peretti G, Piazza C, Ansarin M, et al. Transoral CO2 laser microsurgery for TisT3 supraglottic squamous cell carcinomas. Eur Arch Otorhinolaryngol 2010; 267:1735–1742. 18. Beitler JJ, Muller S, Grist WJ, et al. Prognostic accuracy of computed tomography findings for patients with laryngeal cancer undergoing laryngectomy. J Clin Oncol 2010; 28:2318–2322. 19. Zbaren P, Weidner S, Thoeny HC. Laryngeal and hypopharyngeal carcinomas after (chemo)radiotherapy: a diagnostic dilemma. Curr Opin Otolaryngol Head Neck Surg 2008; 16:147–153. 20. McGuirt WF, Greven KM, Keyes JW, et al. Laryngeal radionecrosis versus recurrent cancer: a clinical approach. Ann Otol Rhinol Laryngol 1998; 107:293–296. 21. Chu MM, Kositwattanarerk A, Lee DJ, et al. FDG PET with contrast-enhanced CT: a critical imaging tool for laryngeal carcinoma. Radiographics 2010; 30:1353–1372. 22. Gilbert MR, Branstetter BFT, Kim S. Utility of positron-emission tomography/ computed tomography imaging in the management of the neck in recurrent laryngeal cancer. Laryngoscope 2012; 122:821–825. 23. de Bree R, van der Putten L, Brouwer J, et al. Detection of locoregional recurrent head and neck cancer after (chemo)radiotherapy using modern imaging. Oral Oncol 2009; 45:386–393. 24. Vandecaveye V, de Keyzer F, Vander Poorten V, et al. Evaluation of the larynx for tumour recurrence by diffusion-weighted MRI after radiotherapy: initial experience in four cases. Br J Radiol 2006; 79:681–687. 25. Ong SC, Schoder H, Lee NY, et al. Clinical utility of 18F-FDG PET/CT in assessing the neck after concurrent chemoradiotherapy for Locoregional advanced head and neck cancer. J Nucl Med 2008; 49:532–540.
1068-9508 ß 2014 Wolters Kluwer Health | Lippincott Williams & Wilkins
www.co-otolaryngology.com
139
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
