Using CT to Localize Side and Level of Vocal Cord Paralysis

(1)

Using CT to Localize Side and Level of Vocal Cord Paralysis

OBJECTIVE. The purpose of our study was to assess the relative accuracy of imaging findings related to peripheral recurrent nerve paralysis on axial CT studies of the neck. Also assessed were imaging findings of a central vagal neuropathy.

MATERIALS AND METHODS. We retrospectively identified 40 patients who had clinically diagnosed vocal cord paralysis and had undergone CT.Eight imaging signs of vocal cord paralysis were assessed, and an imaging distinction between a central or peripheral vagal neuropathy was made by evaluating asymmetric dilatation of the oropharynx with thinning of the constrictor muscles. In two patients, we studied the use of reformatted coronal images from a multidetector CT scanner.

RESULTS. For unilateral vocal cord paralysis, the most sensitive imaging findings were ipsilateral pyriform sinus dilatation, medial positioning and thickening of the ipsilateral aryepiglottic fold, and ipsilateral laryngeal ventricle dilatation. In two patients, coronal reformatted images aided the diagnosis by better showing flattening of the subglottic arch. Imag- ing findings allowed localization of a central vagal neuropathy in four patients.

CONCLUSION. Three reliable imaging findings associated with vocal cord paralysis were identified on routine axial CT studies: ipsilateral pyriform sinus dilatation, medial positioning and thickening of the ipsilateral aryepiglottic fold, and ipsilateral laryngeal ventricle dilatation. Coronal reformatted images of the larynx may be helpful, but they are not necessary in 95% of patients. Ipsilateral pharyngeal constrictor muscle atrophy is a helpful imaging finding to localize a more central vagal neuropathy. Our findings can aid radiologists in iden- tifying peripheral and central vagal neuropathy in patients who present for CT of the neck who have a normal voice and are without a history suggestive of a vagal problem.

he causes of vocal cord paralysis are varied, and nearly half of the cases are reported as being either toxic or idiopathic [1]. That is, in at least half of all patients with vocal cord paralysis, sectional imaging studies may fail to identify a lesion along the course of either the vagus nerve or the recurrent laryngeal nerves. In these patients, our data indicate that the imaging differentiation of a central vagal neuropathy from a purely recurrent laryngeal nerve (peripheral) neuropathy may be possible by observing the effects of a pharyngeal plexus neuropathy on the ipsilateral pharyngeal constrictor muscles. Once the distinction is made, further workup may be specifi- cally focused either just below the skull base or at the posterior fossa.

The imaging identification of a lesion either along the course of the recurrent laryngeal nerves or the main vagus nerve may allow the

initiation of specific treatment. However, once the nerve is compromised in these patients, function rarely returns. If no causative lesion is identified on imaging, either an idiopathic or toxic cause is diagnosed. If a toxic cause is identified, treatment may be directed to a causative disease; however, complete return of nerve function is rare. Thus, the primary reason to image these patients is that failure to identify a causative lesion on imaging allows the clinician to follow a course of conservative management with more confidence.

At least 10 findings associated with a recurrent laryngeal nerve paralysis have been previ- ously noted on coronal contrast-enhanced laryngographic studies [2]. No statistical analysis of the cases that formed these findings in Landman’s work is available. The same findings have not been systematically studied on routine axial imaging studies of the neck to as- Shy-Chyi Chin¹

Simon Edelstein² Cheng-Yu Chen¹ Peter M. Som²

Received May 16, 2002; accepted after revision August 29, 2002.

1Department of Radiology, Tri-Service General Hospital, National Defense Medical Center, Taipei, Taiwan.

2Department of Radiology, Box 1234, The Mount Sinai Hospital and School of Medicine, One Gustave Levy Pl., New York, NY 10029. Address correspondence to P. M. Som.

AJR 2003;180:1165–1170 0361–803X/03/1804–1165

T

(2)

sess which ones are the most reliable in pre- dicting the presence of vocal cord paralysis.

Complicating the use of these findings is the observation that axial CT scans often are not properly aligned in the plane of the true vocal cords; as a result, partial visualization of por- tions of the ventricles, true vocal cords, and the subglottic larynx may be encountered on specific images. Although such misalignment of the scans may make application of the signs of a vocal cord paralysis difficult, we asked which, if any, of these findings could nonethe- less be used with confidence.

The purpose of our study was to assess for the first time on routine axial CT scans of the neck the relative accuracy of eight of the 10 findings noted by Landman [2] to be associated with a recurrent laryngeal nerve paralysis.

In addition, the finding of a dilated oropharynx with thinning of the pharyngeal constrictor muscle was assessed as a sign of a central vagal neuropathy.

Materials and Methods

We retrospectively reviewed the records of 40 consecutive patients with clinically identified vocal cord paralysis who had undergone CT of the neck in the previous 18 months. CT in these patients was performed either as contrast-enhanced helical studies with 2.5-mm contiguous scans obtained with a pitch of 0.7, or as 3-mm contiguous helical scans of the neck obtained on single-detec- tor scanners. The CT scanners were HiSpeed and LightSpeed Ultra units (General Electric Medical Systems, Milwaukee, WI).

Eight imaging findings associated with a vocal cord paralysis were evaluated, and each case was reviewed by two neuroradiologists and one head and neck radiologist to evaluate whether each sign was present, the side of vocal cord paralysis as assessed on the images, and whether there was imaging evidence of a central vagal neuropathy or only a peripheral neuropathy. Identification of a mass that might account for the paralysis was also noted. Any differ- ences in assessment were resolved by consensus among the three radiologists. For the radiologic assessment, all radiologists were unaware of the clini-

cal findings; only after the imaging diagnosis was made was the imaging assessment compared with the clinical evaluation.

Results

Of the 40 total patients, four patients (10%) clinically had bilateral vocal cord paralysis, 15 patients (37.5%) had a right vocal cord paralysis, and 21 patients (52.5%) had a left vocal cord paralysis. The clinical side of paralysis was correctly diagnosed on imaging in all 36 patients with unilateral disease. In four of these patients (10%), findings were mixed on the axial images as to the side affected. In all four of these patients, the correct clinical side was diagnosed on the basis of the preponderance of the imaging findings as to side. In two of these equivocal examinations, coronal reformatted images of the larynx revealed flattening of the subglottic arch on the affected side, aiding the diagnosis. Overall, of the four patients with clinically bilateral paralysis, one was diag-

A B

C

Fig. 1.—45-year-old woman with hoarseness and right recurrent laryngeal nerve paral- ysis who underwent axial contrast-enhanced CT.

A–C, Scans show dilatation of right vallecula (arrow, A) and dilatation of right pyriform si- nus (A), dilatation of right pyriform sinus and thickening and medial positioning of right aryepiglottic fold (B), and dilatation of right laryngeal ventricle (white arrow, C) and ante- rior positioning of right arytenoid cartilage (black arrow, C).

(3)

nosed on imaging as having a right paralysis;

one, as having a left paralysis; and two, as probably having bilateral paralysis.

Four (10%) of the 40 patients had pharyngeal plexus neuropathy and 36 patients (90%) had recurrent laryngeal nerve (peripheral) neuropathy. A lesion was identified along the course of the recurrent laryngeal nerve in five patients (12.5%): one aneurysm (arotic arch), two thyroid masses (right recurrent nerve), one pyriform sinus carcinoma (left recurrent nerve), and one apical lung tu- mor (left recurrent nerve). A mass was also seen in relation to the vagus nerve in one tu- mor of the skull base (schwannoma of the right vagus). Thus, in 34 (85%) of our patients, no lesions were identified along the course of the vagus nerve or the recurrent la-

ryngeal nerve. On the basis of clinical assessment and imaging, the neuropathy in these 34 patients was diagnosed as being idiopathic. None of these patients had a complete recovery of nerve function.

With regard to the specific imaging findings of a vocal cord paralysis in the 40 patients, ipsilateral medial positioning and thickening of the aryepiglottic fold were seen in 31 patients (77.5%); ipsilateral pyriform sinus dilatation, in 31 patients (77.5%); ipsilateral laryngeal ventricle dilatation, in 31 patients (77.5%); anteromedial positioning of the ipsilateral arytenoid cartilage, in 20 patients (50%); fullness of the ipsilateral true vocal cord, in 18 patients (45%); ipsilateral subglottic fullness, in nine patients (22.5%);

and ipsilateral vallecular dilatation, in seven

patients (17.5%) (Figs. 1–4). Dilatation of the ipsilateral pharyngeal wall was seen in four patients (10%). Flattening of the subglottic arch was identified in the two patients with unilateral paralysis for whom we had coronal reformatted multidetector images (Fig. 3).

Of these imaging findings, the finding was on the contralateral side in two (6.4%) of 31 patients with pyriform sinus dilatation, in one (5%) of 20 patients with arytenoid dis- placement, and in one (12.5%) of eight patients with subglottic fullness. In the four patients with bilateral vocal cord paralysis, laryngeal ventricle dilatation was absent on one side in two patients, pyriform sinus dilatation was absent on one side in one patient, and vallecular dilatation was absent on one side in one patient.

A B

C

Fig. 2.—54-year-old man with hoarseness and right vocal cord paralysis who underwent axial contrast-enhanced CT.

A–D, Scans show dilatation of right vallecula (A), dilatation of right pyriform sinus and thickening and medial positioning of right aryepiglottic fold (B), fullness of right true vocal cord and slight anterior positioning of right arytenoid cartilage (C), and right subglottic fullness (D).

D

(4)

For the 31 patients with unilateral vocal cord paralysis, three findings (ipsilateral pyriform sinus dilatation, medial positioning and thickening of the ipsilateral aryepiglottic fold, and ipsilateral laryngeal ventricle dilatation) were seen in 77.5% of patients. When two of these three findings were in agreement, all patients were correctly diagnosed as to the side affected.

Fullness of the affected vocal cord was seen in only 45% of the patients in our study. However, this concordance of findings did not hold up for the four patients with bilateral vocal cord paralysis; in two of these patients, a unilateral paralysis was diagnosed on imaging.

Discussion

The recurrent laryngeal nerve may become paralyzed as a result of disease any-

where along its course from the brainstem to the caudal margins of each recurrent nerve.

Because of its longer course and its exten- sion into the mediastinum, the left side is more often affected than the right side (52.5%

of the unilateral peripheral neuropathies in this series). Peripheral vocal cord paralysis is more common than a central cause, with only 10% of cases being central in one large series [3] (10%

in our series). Overall, vocal cord paralysis has been reported, among other causes, to be the result of a peripheral neuritis associated with al- coholism, viruses, acute bacterial infections, and drug toxicities [1, 3]. Neuropathies associated with multiple sclerosis, polio, myasthenia gravis, Parkinson’s disease, amyotrophic lateral sclerosis, cerebrovascular diseases, and compli- cations of acromegaly have also been implicated [1, 3].

Lesions affecting the brainstem, skull base, and carotid sheath have been associated with vagal neuropathy, whereas specific lesions affecting the recurrent laryngeal nerves include thyroid and cervical esophageal tumors. Only rarely has benign thyroid disease been cited.

Overall, approximately 4% of patients with a unilateral vocal cord paralysis have thyroid disease, but only 0.7% of patients with benign thyroid disease have a recurrent laryngeal nerve paralysis [1, 3]. On the left side, aortic aneurysm, cardiomegaly, and upper lobe tumors have been implicated as potential causes, whereas on the right side, supraclavicular tumors and aneurysm of the subclavian artery may be the cause.

Recurrent laryngeal paralysis may also result from trauma or prior surgery [1, 3].

Clinically, localizing the affected side of the larynx is relatively straightforward: the true

A B

C

Fig. 3.—33-year-old woman with weak voice and left vocal cord paralysis.

A–C, Axial contrast-enhanced CT scans show dilatation of right pyriform sinus (large arrow, A), thickening and medial positioning of left aryepiglottic fold (small arrow, A), dilatation of left laryngeal ventricle (arrow, B), and fullness of right true vocal cord (arrow, C).

D, Coronal reformatted CT scan shows flattening of left subglottic arch (small arrow) and dilatation of left laryngeal ventricle (large arrow).

D

(5)

vocal cord can be seen either to be immobile or to have sluggish mobility. However, imaging of some patients with a unilateral recurrent nerve paralysis may have few, if any, sectional imaging findings; 35% of these patients may be asymptomatic and may have a normal voice [3]. In our series, all of the patients had hoarseness. If the history provided to the radiologist at the time of the imaging study includes the suspected cause and the side affected by the vocal cord paralysis, specific focus can be made during the imaging study to evaluate the localized area of neurologic damage or, in the proper clinical setting, the presence of recurrent disease.

Unfortunately, a specific history is often not available at the time of imaging assessment, and only a history of hoarseness or vocal cord paralysis is provided. In addition, because some patients with a vocal cord paralysis may have a normal voice, the imaging identification of such a paralysis may be an incidental and new finding for the clinician. The usefulness of this study is that radiologists now have imaging findings to help them reliably suggest the presence of a vocal cord paralysis on routine CT of the neck in patients who present without a history suggestive of a paralyzed vocal cord.

Such a finding can then alert the clinician to further evaluate the patient.

A causative lesion was not seen on the imaging study in 85% of our patients. The differ- ential diagnosis and potential treatment differ for patients with a central vagal neuropathy and those with a specific recurrent laryngeal nerve problem. For this reason, an imaging distinction at the time of the study could help

the radiologist further focus the study. One dis- tinguishing feature of a central vagal problem (seen in four patients in our series) is the paralysis or paresis of the ispilateral pharyngeal constrictor muscles that occurs when the pharyngeal plexus is affected.

The pharyngeal plexus is formed from branches of the cranial nerves IX, X, XI and rami from the sympathetic trunk. The vagal branches arise from the nodose ganglion situ- ated just below the skull base and enter the pharyngeal muscles along the upper border of the middle constrictor (just caudal relative to the level of the hard palate), sending branches to the superior and inferior constrictors. The inferior constrictor is also supplied from branches of the superior and recurrent laryngeal nerves [4]. A lesion affecting the pharyngeal plexus will cause paresis or paralysis of the ipsilateral constrictor muscles, which eventually become thinner as they atrophy. This finding is especially noted when the thickness of these muscles is compared with the normal thickness of the contralateral pharyngeal constrictor muscles.

The nasopharynx is supported, in part, by the pharyngobasilar fascia, a thick fascia that arises from the superior edge of this muscle and attaches to the base of the skull. It has been suggested that the purpose of this fascia is to maintain the configuration of the nasopharynx during breathing, thereby maximiz- ing pressure equalization in the ear through the eustacian tube [5]. Whether this premise is true or not, this strong fascia tends to maintain the nasopharyngeal configuration during breathing so that asymmetry is usually not seen in cases of unilateral pharyngeal plexopathy.

However, no such supporting fascia is present at the levels of the middle and inferior pharyngeal constrictor muscles, and a unilateral pharyngeal plexopathy leads not only to muscle wasting, but also to dilatation of the ipsilateral pharynx. This outcome is associated with loss of constrictor function, and it is also reflective of the fact that the primary cause of pharyngeal dilatation is increased intrapharyngeal pressure. Thus, the findings on imaging of an outward bowing of the oropharyngeal and hy- popharyngeal contour in conjunction with thinning of the constrictor muscle are evidence of ipsilateral pharyngeal plexus damage (Fig.

4). As such, these findings localize the abnor- mality to a level either just below the skull base or more cranially in the brainstem.

With regard to the larynx and the imaging identification of vocal cord paralysis, at least 10 findings have been associated with such paralysis as described by Landman [2]. These include incomplete abduction of both the true and false cords in quiet breathing, dilated ipsilateral laryngeal ventricle, a flattened ipsilateral subglottic arch, the interarytenoid notch displaced to the normal side during phonation, the paralyzed arytenoid cartilage anteriorly positioned and abutting or crossing the midline, thinner-than- normal edge of the paralyzed true cord, dilated ipsilateral pyriform sinus (associated with medial folding of the aryepiglottic fold), dilated ipsilateral vallecula, the paralyzed cord lower than the normal cord during inspiration but higher than the normal cord during phonation, and flattened ipsilateral lateral wall of the vestibule [2].

These observations were derived from frontal contrast laryngography and cine studies.

A B

Fig. 4.—64-year-old man with hoarseness, left vocal cord paralysis, and left-sided palatal weakness who underwent axial contrast-enhanced CT.

A and B, Scans show dilated left oropharynx with thinning and atrophy of pharyngeal wall (arrow A) and dilatation of left pyriform sinus and thickening and medial posi- tioning of left aryepiglottic fold (B).

(6)

Routine neck CT is performed in the axial plane. Of the 10 findings associated with vocal cord paralysis [2], tilting of the interarytrenoid notch, a thin edge of the paralyzed cord, and the changing craniocaudal position of the paralyzed cord in phonation and inspiration were considered unlikely to be consistently identified on CT. On axial CT, the findings thought to be most consistently seen in a patient with vocal cord paralysis include thickening and medial positioning of the ipsilateral aryepiglottic fold, dilatation of the ipsilateral pyriform sinus, dilatation of the ipsilateral laryngeal ventricle, anterior and medial positioning of the ipsilateral arytenoid cartilage, fullness of the ipsilateral true vocal cord, ipsilateral subglottic fullness, dilatation of the ipsilasteral val- leculla, and flattening of the ipsilateral subglottic arch. Of these eight findings, only three (medial positioning and thickening of the ipsilateral aryepiglottic fold, ipsilateral pyriform sinus dilatation, and ipsilateral laryngeal ventricle) were seen in more than 75% of the patients in our study. Two other findings (anteromedial positioning of the ipsilateral arytenoid cartilage and fullness of the ipsilateral true vocal cord) were seen in more than 45% of the patients.

With the advent of multidetector CT, high- resolution reformatted coronal images are now be routinely available. This capability has allowed more definitive analysis of the larynx for determining the presence of ipsilateral vocal cord paralysis. In particular, flattening of

the subglottic arch is better and more consistently seen on reformatted coronal images than on axial images. However, such images were not necessary for diagnosis in 38 patients (95%); thus, we cannot suggest that reformatted images are routinely necessary.

Because the typical paralyzed vocal cord fails to completely abduct during quiet breathing, one might expect that the most common imaging finding would be asymmetry in the appearance of the true vocal cords. However, a fullness of the paralyzed vocal cord was seen in only 45% of our patients. The explanation of the relatively low frequency of this imaging finding may be the failure in many cases of technicians to precisely align the scanning plane with the true vocal cords. It appears that the routine nature of the neck CT study and the pressure placed on technicians to achieve rapid patient throughput may affect such precise plane alignment.

Although some of the imaging findings discussed in this study may be known to radiologists, a paucity of documentation of all these imaging signs—much less an assessment of their significance—is found in the literature.

The purpose of our study was to determine what CT findings on axial studies of the neck best allowed the radiologist to diagnose the presence of a vocal cord paralysis. Our study suggests that the indirect findings of ipsilateral pyriform sinus dilatation, medial positioning and thickening of the ipsilateral aryepiglottic

fold, and ipsilateral laryngeal ventricle dilatation are more reliable imaging criteria than the appearance of the true vocal cord itself for as- sessing unilateral vocal cord paralysis. In difficult cases, the use of coronal reformatted images of the larynx may further refine diagnosis by revealing ipsilateral flattening of the subglottic arch, but these reformatted images were used in only two patients. We cannot form a reliable conclusion regarding their use in all patients as part of the routine examina- tion; such a statement requires further studies.

Because a specific history is not always available at the time of imaging interpretation, the presence of the findings we have discussed may help the radiologist avoid missing the diagnosis of vocal cord paralysis, central vagal neuropathy, or both.

References

1. Thornell WC. Vocal cord paralysis. In: Paparella MM, Shumrick DA, eds. Otolaryngology, vol. 3. Head and neck. Philadelphia: Saunders, 1973:649–657 2. Landman GHM. Laryngography: cine laryngogra-

phy. Baltimore: Williams & Wilkins, 1970:62–75 3. Levine HL, Tucker HM. Surgical management of

the paralyzed larynx. In: Baily BJ, Biller HF, eds.

Surgery of the larynx. Philadelphia: Saunders, 1985:117–134

4. Williams PL, Bannister LH, Berry MM. Gray’s anatomy, 38th ed. Edinburgh: Churchill Living- stone, 1999:1251–1252, 1729–1733

5. Last RJ. Anatomy regional and applied, 6th ed.

Edinburgh: Churchill Livingstone, 1978:414–418