CT in Carcinoma of the Larynx and Pyriform Sinus:

Transcription

1 577 Gordon Gamsu1 W. Richard Webb Joel B. Shallit Albert A. Moss Received April 24, 1 980; accepted after revision October 7, All authors: Department of Radiology, University of California Medical Center, San Francisco, CA Address reprint requests to G. Gamsu (room M-396). AJR 1 36: , March X/81 / $00.00 American Roentgen Ray Society CT in Carcinoma of the Larynx and Pyriform Sinus: Value of Phonation Scans The structural and functional information obtained from CT performed during quiet breathing and phonation of the letter E was investigated in 25 patients with carcinoma of the larynx on pyniform sinus. Significant additional information was obtained from the phonation scans in all patients. In 1 6 patients, vocal cord dysfunction was found on the phonation scans. In 14 patients, phonation CT demonstrated an abnormal anyepiglottic fold better than CT during quiet breathing. A pynifonm sinus was distorted or displaced in 1 1 patients and CT during phonation was more accurate than laryngoscopy or CT during quiet breathing in detecting abnormalities deep to this region. CT scans were the most accurate method of detecting tumor extension Into the subglottic space, and into the preepiglottic space; each extension was seen in seven patients. Thyroid cartilage destruction was detected only by CT in six patients. CT scans during phonation should be an integral part of laryngeal CT and in conjunction with laryngoscopy could possibly replace laryngography for the evaluation of patients with laryngeal carcinoma. Computed tomography is rapidly gaining acceptance in the evaluation of laryngeal disorders [1-6]. To preserve function, limited surgical techniques have been developed for the removal of laryngeal cancer [7-9]. The determination of tumor extent has thus become critical in patient selection. CT can directly evaluate penetration of tumor into the laryngeal soft tissues and cartilages; whereas laryngoscopy and contrast laryngography can only infer deep-seated abnormalities from changes in surface contour of the laryngeal cavity [5, 6]. Prior studies of laryngeal CT have reported on scans obtained during quiet breathing or breath holding. In the relaxed state during quiet breathing, the vocal cords are abducted, and vocal cord function cannot be evaluated. The subglottic area is not clearly delineated from the glottis. The pyriform sinuses are collapsed, and aryepiglottic folds are not well seen. For these reasons, optimum evaluation of the larynx by CT is not obtained in many patients. To extend the structural and functional information obtained from CT of the larynx, 25 patients with laryngeal or pyriform sinus carcinomas were investigated using CT scans obtained both during quiet breathing and during phonation of the letter E. Subjects and Methods Twenty-five unselected patients with biopsy proven squamous cell carcinoma of the larynx or pyriform sinus were studied. All the patients had direct laryngoscopy for clinical staging of their tumor using the Joint Committee for Cancer Staging Manual [1 0]. CT scans of the larynx were obtained on the GE Research Scanner using 1 20 kvp, ma, 2 msec pulse widths, 5 mm collimation, and 2.4 sec scan times. The patient was placed supine with his head (orbitomeatal line) extended 250_300 from the vertical and immobilized with a Picker Head Holder with Vacuum Base. Two separate series of scans were obtained, one during quiet breathing and the other during phonation of the letter E. Both series of scans started at the inferior border of the cricoid cartilage and progressed toward the head. Twelve to seventeen 5 mm contiguous scans were sufficient to encompass the larynx and supralaryngeal area.

2 578 GAMSU ET AL. AJR:136, March 1981 A B C Fig. 1 -A, CT scan in normal subject during quiet breathing at level of true vocal cords. Arytenoid cartilages (arrows) cleanly visible in close proximity to inner margin of thyroid cartilage. Right anytenoid slightly larger than left. B, During phonation of letter E, arytenoid cartilages adduct and rotate For the quiet breathing series, the patient was instructed to breathe quietly through his mouth without stopping at end-inspiration on end-expiration. With the 2.4 sec scan time and 1 sec interscan pause, the series took sec to complete. For the phonation scans, the patient was instructed to phonate the letter E. The time that the patient could phonate was recorded and the series of scans was performed in increments to conform to the patient s ability to phonate. Usually, three or four increments were required to obtain the scans needed. All the scans were magnified two to three times and photographed at a window level of H and a window width of H. Coronal and sagittal transformatted images through the cavity of the larynx were routinely generated. The CT scans were evaluated by two radiologists without any clinical information beyond knowing that the patients had laryngeal carcinoma. The quiet breathing series of scans were evaluated first for the presence and extent of tumor, cartilage destruction, and cervical lymph node enlargement. Only after the quiet breathing series of scans was completed were the phonation scans reviewed in conjunction with the quiet breathing scans for additional information. The coronal and sagittal transformatted images were viewed next, to determine whether additional information could be obtained as to the extent of the tumor. Only after the CT scans had been interpreted were the patients records reviewed and operative and pathologic data correlated. Results On the basis of clinical and endoscopic findings in the 25 patients, seven had gbottic tumors, five had supragbottic tumors, six had transgbottic tumors involving both the gbottic and supragbottic regions, and four had circumgbottic tumors crossing the midline to involve both sides of the larynx. Three patients had lesions that arose in a pyriform sinus. In all patients, the side and level of the tumor determined at endoscopy correlated exactly with the findings from the CT scans. In most cases, CT showed more extensive disease I inwards. Distended pyniform sinuses (P) are seen lateral to arytenoid cartilages. C. At lower window level and wider window width than B. Anytenoid cartilages no longer visible. Vocal cords now seen spanning larynx from front to back. than seen at baryngoscopy. The endoscopic staging of the tumors using the Joint Committee TNM classification, showed four T1 lesions, seven T2 lesions, 1 0 T3 lesions, and three T4 lesions; one lesion was not classified. In nine patients, the CT scan changed the tumor classification: in eight of the nine to a more extensive stage of disease. In six of these patients, clinically unsuspected cartilage destruction was found on CT which made the lesion a stage T4. This finding was confirmed in all six cases at laryngectomy. In seven patients, extension of tumor into the preepigbottic space was visible on the CT scans and not clinically appreciated. In one patient, a neck mass thought to be an ertlarged lymph node was shown on CT scans to be direct extension from a supragbothc tumor, causing a stage T2 lesion to be converted to a T4 lesion. Enlarged lymph nodes were visible in the neck on the CT scans in six patients. In three of the six, the nodes were also palpable. In two of the other three cases, the clinically nonpalpable nodes seen on CT were found to be enlarged and involved with tumor at surgery. In one case, a clinically enlarged submental lymph node was not within the field of the CT scans. The areas in which the CT scans during phonation were of assistance were the vocal cords and arytenoid cartilages, the aryepigbottic folds, the pyriform sinus, the preepigbottic space, and the subgbottic region. These will be dealt with separately. Vocal Cords and Arytenoid Cartilages In all patients, the normal arytenoid cartilages were visible on the CT scans both during quiet breathing and during phonation. The exact level of the true vocal cords can only be determined on quiet breathing when the vocal process

3 AJR:136, March 1981 CT OF LARYNX CARCINOMA DURING PHONATION 579 Fig. 2.-A, Scan during quiet breathing at level of true vocal cords in night. B, During phonation. Left anytenoid moves normally. Right arytenoid patient with clinical fixation of right true cord. Cordal mass (arrows) displaces remains fixed in abducted position. vocal process of right arytenoid cartilage medially. Enlarged lymph node on Fig. 3.-A, Normal scan through supraglottic larynx at level of thyroid notch. Top of right arytenoid is seen. Laryngeal vestibule visible, but pyniform sinuses are collapsed. B, Phonation in same subject. Distension of pyriform sinuses (P) with improved visualization of anyepiglottic struts (S). of the arytenoid cartilage is visible. In our series, the vocal process of the arytenoid was seen in about half of the cases. During quiet breathing, the vocal cords are relaxed and the normal arytenoid cartilages are situated laterally above the signet part of the cricoid cartilage (fig. 1A) [1 1 ]. The arytenoid is separated from the inner surface of the thyroid cartilage by no more than a few millimeters and the two sides are symmetrical. A small condensation of soft tissue extends anteriorly from the vocal process of the arytenoid, constituting the relaxed vocal cord. An obvious difference in size of apparently normal arytenoid cartilages was observed in seven patients. During phonation the arytenoid cartilages adduct and rotate toward the midline (figs. 1 B and 1 C). Symmetry of movement is a key finding. In all of the 25 patients in this study, additional information regarding arytenoid and vocal cord function was evident on the CT scans performed during phonation. In 1 6, abnormal mobility was identified; and in nine, normal mobility was demonstrated. Fifteen patients had unilateral or bilateral cord fixation or decreased mobility at laryngoscopy. In 14 of these 1 5, the CT scans during phonation demonstrated no movement or a marked decreased movement of the relevant arytenoid cartilage (fig. 2). In one patient, the arytenoid clearly moved on the CT scans during phonation and the clinical diagnosis of fixation did not correlate with the CT findings. In two additional patients, the CT clearly showed decreased or absent movement of the arytenoid cartilage on phonation. Endoscopic examination either did

4 580 GAMSU ET AL. AJR:136, March 1981 Fig. 4.-A, Tumor involves right supraglottic larynx. Thickening and increased density on night. B, During phonation. Right pyriform sinus now visible. It Is smaller and displaced posteriorly. Right aryepiglottic strut (5) is thickened. Neither feature was visible on quiet breathing. not visualize the vocal cord or the cords were only poorly seen because of a bulky supraglottic tumor mass. The CT findings were thus not confirmed. In seven patients with cord fixation, the arytenoid cartilage was in the lateral, abducted position during quiet breathing and demonstrated decreased or absent medial movement. In nine patients, the CT scans showed the respective arytenoid cartilage in the paramedian position during quiet breathing and absence of movement during phonation. Cord fixation was associated with extensive tumor; subgbottic extension was seen in four patients and cartilage destruction in five. Aryepiglottic Folds During quiet breathing, the supraglottic larynx is relaxed and pyriform sinuses are partially collapsed. The postero- Fig. 5.-A, Quiet breathing in patient with right pyniform sinus carcinoma. Laryngeal vestibule is distorted, right pyriform sinus is not visible, and thyroid cartilage is distorted and destroyed (arrows). B, Phonation scan. Tumor fills right pyriform sinus. Appendix of night laryngeal ventricle (arrow) displaced medially by tumor. Left appendix is normal (open arrow). lateral wall of the larynx between the pyriform sinus and laryngeal vestibule is not well delineated (fig. 3A). This part of the laryngeal wall has been referred to as the aryepiglottic fold; on CT scans, however, they appear more as pillars or struts. The free margin of the aryepigbottic folds are thinner and are visualized at a higher level on CT scans. With phonation, the pyriform sinuses are distended and the aryepiglottic folds are well seen (fig. 3B). The free margin of the aryepigbottic folds are equally well seen with quiet breathing and phonation. Additional useful information regarding the aryepigbottic folds was found on the phonation CT scans in almost all patients. In eight patients, an aryepiglottic fold was considered to be involved with tumor at direct laryngoscopy. In all of these patients, the CT scan during phonation showed a thickened strut on the relevant side or sides (fig. 4). The phonation scans greatly improved assessment of the ary-

5 AJR:136, March 1981 CT OF LARYNX CARCINOMA DURING PHONATION 581 epigbottic folds. In 1 4 patients, the phonation scans demonstrated to better effect that an aryepigbottic fold was abnormal. In the other 1 1 cases, the phonation scans demonstrated that the struts were free of disease. Pyriform Sinuses The pyriform sinuses are collapsed during quiet breathing and distended with phonation. In this series of patients, three tumors arose within a pyriform sinus. In all three, the CT scans during phonation were better able to delineate the extent of tumor involvement (fig. 5). In an additional eight cases, a pyriform sinus was displaced or distorted by tumor as demonstrated by the phonation CT scans. In four of these cases, laryngoscopy was also abnormal, and in an additional one, probably abnormal. In the other three cases, the abnormality deforming the pyriform sinus was not appreciated at laryngoscopy but was involved by tumor in the two patients undergoing laryngectomy. In another five patients, tumor was in proximity to a pyriform sinus. The scans during quiet breathing were inadequate to exclude involvement by tumor, but during phonation the collapsed sinus distended normally. Preepiglottic Space In the normal larynx, the preepigbottic space starts 5-10 mm above the anterior commissure and increases in antenor/posterior depth in a superior direction. On CT scans the preepigbottic space is filled with uniform, low density, fibrofatty material that extends posteriorly into the aryepigbottic folds and superiorly as far as the valbeculae (fig. 6). In seven patients in this series, extension of tumor into the preepigbottic space was visible on the CT scans. In these seven patients, the normal low density tissue was replaced by higher tumor density (fig. 7A). In none of these patients was this finding visible at direct laryngoscopy, but in all four having a total laryngectomy, invasion of the preepiglottic space was confirmed. In an additional eight patients in whom the tumor could have invaded the preepiglottic area, the presence of low density material precluded tumor invasion. The sagittab transformatted scan through the midline confirmed preepigbottic invasion by tumor in all seven patients in which this finding was evident on the transverse axial scans (fig. 7B). A normal appearance of the preepigbottic space was also confirmed by the sagittal scans in the relevant cases. Subglottic Space During quiet breathing the true and false cords are abducted. Unless the vocal process of the arytenoid cartilage is visualized, the exact level of the true cord is difficult to ascertain. With phonation, the arytenoid cartilages and the true vocal cords adduct and are clearly visible. The level of the true cords can be precisely ascertained. Any thickening of the tissue within the cricoid cartilage extending inferiorby for more than 5 mm below the bevel of the true cord was considered to represent subgbottic extension of tumor (fig. 8). Seven patients had subgbottic extension by this criterion. Direct laryngoscopy showed the subgbottic tumor in only two of these cases. However, in one of the two and in another three of the seven cases in the series, subglottic extension of tumor was evident at laryngectomy. The other three cases were treated with radiation and confirmation is not available. Cartilage Invasion CT scans during phonation did not specifically aid in demonstrating involvement of the thyroid cartilage by tumor. It is included here to emphasize the importance of cartilage destruction in the management of patients with laryngeal carcinoma. Considerable variation exists in the normal degree of calcification and shape of the thyroid cartilage. Of the 25 patients, six demonstrated destruction of the thyroid cartilage. All six underwent laryngectomy and thyroid cartilage invasion was confirmed. Disruption of the thyroid cartilage may be gross with fragments displaced away from the normal contour of the cartilage (fig. 9), or it may be subtle and difficult to delineate. All six patients with destruction of the thyroid cartilage had CT evidence of vocal cord dysfunction with decreased mobility of the arytenoid cartilage. In four, the arytenoid cartilage was in the paramedian position on quiet breathing and in two, it was in the abducted lateral position on phonation. Five of the six were extensive transgbottic or circumgbottic lesions. Ofthe six patients with destruction of the thyroid cartilage, two also demonstrated buckling or deformity of the thyroid cartilage (fig. 1 0). However, there were three additional patients with buckling or tilting ofthe thyroid cartilage whose surgical or clinical follow-up indicated that the cartilage was not invaded by tumor. Discussion Computed tomography of the larynx provides unique information not available by other radiobogic techniques or by laryngoscopy [5, 6]. CT examination of the larynx in the relaxed state during quiet mouth breathing is essential to delineate the extent of a tumor mass, but provides incompbete information. Scans during phonation enable improved visualization of the mobile and pliable parts of the larynx. Important functional information is obtained. The precise level of the true vocal cords is also determined. Additional positive or negative information was obtained in all patients in this study. This study is the first time CT has been performed during a voluntary physiologic maneuver, besides suspended respiration. The true vocal cords constitute the laryngeal glottis. Mobility ofthe cords is of critical importance in the management of patients with laryngeal carcinoma. Fixation of the vocal cord has been recognized as a poor prognostic feature and an indication for surgical intervention [1 2, 1 3]. The assessment of true vocal cord fixation is subjectively made at endoscopy. Fixation may be the result of replacement of the vocalis muscle by tumor, subgbottic extension of tumor with

6 582 GAMSU ET AL. AJR:136, March 1981 Fig. 6.-A. Normal preepiglottic space contains low-density fibrofatty upward, cricoid cartilage (C), thyroid cartilage (1 ), hyoid bone (H), and at top tissue (arrows). Low-density material extends into anterior part of aryepiglot- left, mandible (M). Low-density preepiglottic space is evident (arrows). tic struts. B, Sagittal reformatted image through midline shows from below Fig. 7.-A, Scan in patient with supraglottic tumor on right. Laryngeal on right replaced by tumor-density tissue (arrows). B, Sagittal transformatted vestibule displaced and pyniform sinus pushed backward. Preepiglottic space image. Replacement of low-density preepiglottic space by tumor. cricoid attachment, invasion of the thyroid cartilage with fixation, or invasion of the cricoarytenoid joint. These causes of cord fixation cannot be separated preoperatively and the selection of patients for laryngectomy or hemilaryngectomy is at present unresolved [1 3]. In our study, six patients with a fixed cord had cartilage destruction and four showed subglottic extension of tumor. Either of these findings mdicates a need for total laryngectomy rather than hemilaryngectomy. Thus, CT with phonation can accurately diagnose a fixed vocal cord and assist in determining the type of surgery needed [14]. The normal aryepigbottic folds are relaxed during quiet breathing and often can not be distinguished by CT from the lateral wall of the collapsed pyriform sinuses. The folds extend upward from the false cords and arytenoid cartilages and are an important route of infiltration for tumor. Carcinomas arising from the laryngeal ventricle, false vocal cords, and region of the arytenoid cartilages tend to extend into the aryepigbottic folds [1 5]. During phonation, the aryepigbottic folds become thinner and are more easily evaluated by CT. In our study phonation scans were essential in visualizing or confirming an abnormality of an aryepiglottic fold. The detection of tumor deep to the pyriform sinuses is relatively difficult by direct laryngoscopy. The sinuses, lying lateral to the cavity of the larynx are not anatomically part of the larynx, but management of carcinoma arising from or extending into the pyriform sinus is similar to primary lesions of the larynx. Laryngeal tumors involving the medial wall of the pyriform sinus are considered stage T3 and surgery is often the treatment of choice [1 0]. On CT scans obtained during phonation, the distended pyriform sinus usually extends down to the level of the arytenoid cartilages. The lateral wall is closely applied to the inner margin of the thyroid cartilage with essentially no tissue plane visible between them. The medial wall is formed by the outer margin of the aryepiglottic strut. The distended pyriform sinus is therefore well seen on CT scans. CT is probably more

7 AJR:136, March 1981 CT OF LARYNX CARCINOMA DURING PHONATION 583 Fig. 8.-A, Scan during quiet breathing at level of true cords. Thickening of right true cord by tumor which extends to anterior commissure (arrow). B, Scan 1.5 cm lower. Tumor within cnicoid cartilage (arrows); indicates subglottic extension. Fig. 9.-Patient with massive tumor involvement of larynx. Thyroid cartilage exploded with fragments within large mass (arrows). Fig Buckling left side of thyroid cartilage (arrow) in patient with left true cord lesion. At laryngectomy, thyroid cartilage was not destroyed. sensitive in detecting a lesion deep to the pyriform sinus than is laryngoscopy. If the CT scan during phonation is abnormal, laryngoscopy should be carefully directed to the area of abnormality. The preepigbottic space is a silent area of the larynx. Tumor extension into this space is rarely appreciated by imaging methods other than CT. Lateral radiographs of the neck obtained during quiet breathing and during a Valsalva maneuver have been recommended. Lack of approximation of the hyoid bone to thyroid cartilage during a Valsalva maneuver suggests infiltration of tumor into the preepiglottic space [1 5]. In our experience, this radiobogic finding is insensitive. CT is the method of choice for delineating the preepiglottic space. Extension of tumor below the glottis is a common route of spread. The conus elasticus, a thick membrane between the true cord and the inner surface of the cricoid cartilage is a barrier to tumor growth [2]. Gbottic tumors tend to invade downward on the inner surface of the conus. Below the true cords the airway mucosa is very close to the cricoid cartilage. Any tissue seen on CT encroaching on the airway below the true cords should be considered abnormal. The difficulty with CT during quiet breathing is in defining the level of the true cords and the depth of their undersurface. The same problem is found in laryngoscopy, and laryngography is often used to define subglottic extension of tumor [1 6, 1 7]. CT scans obtained during phonation facilitate identification of the adducted true cords. The subglottic space starts 4-8 mm below the free margin of the true cords. On CT, any visible tissue within the cricoid cartilage more than 5 mm below the level determined as that of the true cords should be considered indicative of subgbottic extension of tumor. CT during phonation will probably prove more accurate than laryngoscopy in determining subglottic extension of laryngeal carcinoma. The detection of invasion of the thyroid cartilage by tumor

8 584 GAMSU ET AL. AJR:136, March 1981 is important in the management of patients with laryngeal carcinoma [1, 2, 1 8]. Cartilage involvement changes the clinical classification of the tumor to T4, reflecting the advanced nature of the disease. The response to radiation therapy is poor and most otolaryngobogists consider total laryngectomy as the only method of achieving control of the tumor. Neither palpation of the neck nor endoscopy is capable of detecting cartilage invasion. Plain radiographs and laryngograms are relatively insensitive in detecting destruction of the thyroid cartilage. CT is the only method capable of making this observation with sensitivity. In four patients in this study, the CT finding of cartilage destruction was a major factor in the decision to perform a laryngectomy, and not attempt radiation therapy. A bucent line within the inner margin of the thyroid cartilage has been described on CT [1, 1 8]. Obliteration of this plane has been suggested as an indicator of tumor invasion into the thyroid cartilage. In none of our patients was this line observed. We considered that it is most probably an artifact of CT images obtained with long scan times and relatively poorer resolution. CT of the larynx provides unique information on deep penetration of laryngeal tumors and is more accurate than laryngography in defining tumor extent [5]. The problem areas in which laryngography can assist direct laryngoscopy are in the subgbottic region, in evaluating the inferior aspect of bulky tumors, and in confirming vocal cord fixation [19-22]. CT using phonation as well as quiet breathing can evaluate all of these areas, and in our opinion may replace contrast laryngography. ACKNOWLEDGMENTS We thank Lauranne Cox for assistance with the CT examinations, and W. H. Berninger and A. W. Redington of the General Electric Corporation Research and Development Laboratory, Schenectady, NY for technical advice. REFERENCES 1. Mancuso AA, Hanafee WN, Juillard JF, Winter J, Calcaterra TC. The role of computed tomography in the management of cancer of the larynx. Radiology 1977;1 24 : Mancuso AA, Calcaterra TC, Hanafee WN. Computed tomography of the larynx. Radiol Clin North Am 1978; 1 51 : Archer CR, Friedman WH, Yeager VL, Katsantonis GP. Evaluation of laryngeal cancer by computed tomography. J Comput Assist Tomogr 1 978;2 : Mancuso AA, Hanafee WN. Computed tomography of the injured larynx. Radiology I 979; 133: Mancuso AA, Hanafee WN. A comparative evaluation of computed tomography and laryngography. Radiology 1 979;1 33: Ward PH, Hanafee W, Mancuso A, Shallit J, Berci G. Evaluation of computerized tomography, cinelaryngoscopy, and laryngography in determining the extent of laryngeal disease. Ann Otol Rhinol Laryngol 1979;88 : Ogura JH, Dedo HH. Gbottic reconstruction following subtotal gbottic-supraglottic laryngectomy. Laryngoscope 1 965;75: Ogura JH, Biller HF. Conservation surgery in cancer of the head and neck. Otolaryngol Clin North Am 1969;2: Ogura JH, Heeneman H. Conservation surgery of the larynx and hypopharynx-selection of patients and results. Can J Otolaryngol 1973;2: Larynx. In: Manual for Staging of Cancer Chicago: American Joint Committee, 1 978: Pernkopf E. Atlas of topographical and applied human anatomy, vol 1. Philadelphia: Saunders, Kirchner JA, Som ML. Clinical significance of fixed vocal cord. Laryngoscope 1971;81 : Lesinski SG, Bauer WC, Ogura JH. Hemilaryngectomy for T3 (fixed cord) epidermoid carcinoma of larynx. Laryngoscope 1976;86: Mancuso AA, Tamahawa Y, Hanafee WN. CT of the fixed cord. AJR 1980;1 35: Jing BS. Roentgen examination of laryngeal cancer: a critical evaluation. J Otolaryngol 1 975;4 : Bruce PD. The laryngeal subglottis. J Laryngol Otol 1 975;89: Lehmann OH, Fletcher GH. Contribution of the laryngogram to the management of malignant laryngeal tumors. Radiology 1 964;83 : Archer CR, Yeager VL. Evaluation of laryngeal cartilages by computed tomography. J Comput Assist Tomogr 1 979:3 : Holtz S, Powers WE, McGavran MH, Ogura J. Contrast examination of the larynx and pharynx. Glottic, infraglottic and transglottic tumors. AJR 1 963;89 : Brindle MJ, Stell PM. Radiological assessment of laryngeal carcinoma. Clin Radiol 1 968; 1 9 : Valvassori G, Goldstein JC. Radiographic evaluation of the larynx. Otolaryngol Clin North Am 1970;3: Hemmingsson A. Roentgenologic examination of the larynx: a clinical comparison. Acta Radiol [Diagn] (Stockh) 1972;1 2: