- Email: [email protected]
Radiological assessment of post laryngectomy speech
ClinicalRadiology(1990) 41, 312-316
Radiological Assessment of Post Laryngectomy Speech J. McIVOR*, P. F. EVANS*, A. PERRY'~ and A. D. CHEESMAN~
Departments of* Radiology, t Speech Therapy and ~ Ear Nose and Throat Surgery, Charing Cross Hospital, London Laryngectomy for carcinoma of the larynx has a five year survival rate of approximately 80% but the operation has a high morbidity as most patients do not develop effective alaryngeal speech. The published literature states that poor speech is occasionally due to anatomical or physiological problems in the reconstructed pharynx but is usually due to psychological, sociological or educational difficulties. We devised a radiological technique using a combination of video-fluoroscopy and static or spot films to assess the anatomy and function of the reconstructed pharynx of postlaryngectomy patients, and investigated 24 good and 134 poor or failed speakers. We examined the pharynx during the acts of swallowing, attempted phonation and attempted phonation with air insuffiation. A single vibrating segment (P-E segment) in an otherwise dilated pharynx was present in all good speakers, but was absent in poor or failed speakers whom we classified as hypotonic (19), hypertonic (40), spastic (62) and strictured (13) according to the radiological appearances. This classification has been used to plan subsequent clinical management and 46 of the 55 patients treated to date (84%) have developed effective alaryngeal speech.
Laryngectomy for carcinoma of the larynx has a 5 year survival rate of approximately 80% (Berry, 1983) which is much higher than the survival rates after most operations for malignant disease. It is probably due to the fact that tumour extension beyond the larynx occurs late in most cases and laryngectomy usually removes the whole tumour~ There a-re no reliable figures-for the,number of laryngectomy operations carrie~ out in the UK each year, but it is probably in the region of 200, in which case the number of surviving patients is approximately 2000. Although the survival rate is high, the operation has a high morbidity as most patients do not develop effective speech and have a continuing and often demoralising problem in communication. The proportion who fail to develop speech was recorded as 88% by Harwood and Rawlinson (1983). Most papers quote a figure nearer to 75% (Schafer and Johns, 1982), but failure rates as low as 30% have also been reported (Smith et al., 1966). The main reason for this wide variation is that different criteria have been used by different authors to define 'good' speech. The causes of speech failure were reviewed by Perry (1983) and are usually classified as: anatomical; physiological; psychological; sociological and educational as suggested by Duguay (1966). Post-laryngectomy oesophageal speech is produced by regurgitating air from the oesophagus through a 'vibrating segment' (pharyngo-esophageal or P-E segment) in the reconstructed pharyngo-oesophagus. The volume of air which can be retained in the dilated oesophagus rarely exceeds 80 cc (Edels, 1983) which is much less than the Correspondenceto: Dr J. McIvor,X-rayDepartment,CharingCross Hospital, FulhamPalaceRoad, LondonW6 8RF.
normal vital capacity of four litres in men aged 60 years (Cotes, 1979) and this tends to produce short phrases of quiet, low pitched speech. Surgical procedures have been devised to divert air from the trachea into the reconstructed pharyngo-oeso. phagus and thus increase the flow of air through the P-E segment. In this way it was hoped to produce louder and more sustained speech. Staffieri (1980) described a surgical technique for creating a mucosal shunt between the trachea and the reconstructed pharynx so that digital occlusion of the tracheal opening (tracheostoma) during expiration caused air to pass into the pharynx. However, there were serious problems caused by aspiration of saliva and other fluids into the lungs. The next development was a valved prosthesis which could be passed through a surgically created fistula between the trachea and pharynx (Singer and Blom, 1980; Panje, 1981). The valve prevented aspiration into the lungs but it was soon realised that a correctly placed and functioning prosthesis did not invariably produce good speech. Attention then focused on the anatomy and physiology of the reconstructed pharyngo-oesophagus and the vibrating P-E segment. Blom et al. (1985) used the air insutttation test originally described by Seeman (1967) to assess the function of the P-E segment. A 14 gauge tube was passed trans-nasally into the pharynx and the tip placed beloW the P-E segment. Air was insufltated through the tube as the patient attempted to speak. The radiological appearances of the reconstructed pharynx after laryngectomy were described by Diedrich and Youngstrom (1966) and reviewed by Blom and Singer (1979). Cheesman et al. (1986) described the radiological and clinical procedures used to assess 50 post-laryngectomy patients who were being considered for the BornSinger procedure at Charing Cross Hospital. The aim of this paper is to describe the radiological technique in more detail, to record the classification of 134 poor or failed oesophageal speakers and indicate the results of treatment based on the radiological findings. PATIENTS AND METHODS We assessed 24 good and 134 poor speakers using a combination of fluoroscopy, video recording and static radiographs (spot films). Fluoroscopy allowed the movement of the reconstructed pharynx to be viewed while observing the patient, video recording provided an easily accessible permanent record of movement and the spot films gave a more anatomically detailed but static image of the pharynx and P-E segment. The examinations were carried out by a team consisting of a radiologist, speech therapist and otorhinolaryngologist, all of whom observed the fluoroscopy screen and the patient during the examination. We used a standard Siemans Sirescop 3 fluoroscoPY unit with the table in the vertical position and the foot rest removed. Most patients were examined whilst standing as
POST LARYNGECTOMY SPEECH
313
Table 1 - Rafliologieal features of the reconstructed pharynx in good and poor oesophageal speakers after laryngectomy
Hypotonic G o o d
HypertonicSpastic
Stricture
Barium swallow
Dilates normally
Dilates normally
Usually dilates normally
Usually narrowed posteriorly
Persistent narrowing
Attempted phonation
Dilates widely, no vibrating segment
Vibrating P-E segment, Air in oesophagus
Marked narrowing with 1 or 2 narrowed segments
Severe narrowing with 1 or 2 narrowed segments
Persistent narrowing. No air in oesophagus
Attempted phonation with air insufflation
Dilates widely, no vibrating segment
Vibrating P-E segment, Air in oesophagus
Marked narrowing as above, Oesophagus dilates
Severe narrowing as above. Oesophagus dilates widely
Persistent narrowing. Oesophagus dilates
No voice
Good voice
Variable voice
No voice
No voice
the shoulders tend to drop slightly in this position and the other examinations were carried out with the patients sitting. The distance between the table top and the image intensifier measured 45 cm which allowed most patients to be examined in the true lateral position but those with very broad shoulders were placed with the shoulders slightly oblique. The tracheostoma was marked with a radio-opaque marker on the skin. It can be difficult to obtain good lateral radiographs of the pharynx and upper oesophagus after laryngectomy, as the lower part of the neck is much thinner and more radiolucent than the upper part. However, two simple modifications made it possible to obtain satisfactory radiographs of the reconstructed pharynx in all our patients. Firstly the X-ray beam was filtered by a copper sheet (0.5 mm thick) fixed to the table top and secondly the field size reaching the image intensifier was reduced with a lead rubber sheet (0.5 mm lead equivalent) attached to the front of the explorator which masked out the non-attenuated part of the beam (Figs 1-5). In our experience, lateral fluoroscopy and films were adequate for diagnosis unless the patient had a surgical complication such as stricture or fistula, when additional antero-posterior films were helpful. The radiation dose was estimated at 5-10 cGy which is rather high for a diagnostic procedure but much less than the therapeutic dose of 4000-6000 cGy already received by most patients and the field was, of course, limited to the neck and thoracic inlet.
least two spot films were taken during maximum vocalisation.
Attempted Phonation With Air Insufliation A flexible rubber tube with a densely radio-opaque tip and two opposed side holes was passed through the nose and P-E segment. We used a specially modified 14 gauge bronchographic rubber catheter which could be readily seen on fluoroscopy (Fig. 4b). The tip of the tube was positioned below the level of the narrowest part of the P-E segment and air was passed through the tube at a rate of approximately 0.5 litres per minute. The pharynx was coated with a sip of barium and the patient was asked to mime counting from one to ten as before. The procedure was usually repeated with the
Swallowing The patient swallowed barium (Baritop) while the pharynx was screened in the lateral position. The movements were viewed on a TV monitor and were simultaneously recorded on videotape as was the whole examination. Spot films were taken on at least two occasions when the P-E segment was fully dilated by a bolus of barium.
Attempted Phonation The patient took a sip of barium to coat the mucosa of the pharynx, swallowed air to produce a reservoir in the Oesophagus and counted slowly from one to ten. The movements were recorded on videotape as before and at
(a)
(b)
Fig. 1 - (a) Lateral view of a good oesophageal speaker during phonation. The vibrating pharyngo-oesophageal (P-E) segment is marked with arrows. There is a slight narrowing below which is of no functional significance. The metallic marker is adjacent to the tracheostoma and the rectangular field size has been reduced anteriorly and inferiorly with a sheet of lead rubber attached to the explorator. (b) Antero-posterior view of another good oesophageal speaker during phonation with the vibrating P-E segment marked with arrows. The metallic marker is situated at the tracheostoma.
314
CLINICALRADIOLOGY
(a)
(b)
Fig. 2 - (a) Hypotonic patient during attempted phonation. The anterior and posterior walls of the reconstructed pharynx are widely separated and there is no vibrating P-E segment and no voice. A tracheostomy tube is present. (b) The same hypotonic patient during attempted phonation with digital pressure applied to the anterior surface of the neck. The anterior wall is displaced posteriorly producing a P-E segment (arrowed) and voice.
catheter tip at slightly different levels and spot films were taken during maximum vocalisation.
(a)
(b)
Fig. 4 (a) Spastic patient during attempted phonation. There are two severely narrowed segments which are narrower and longer than in the hypertonic patient shown in Fig. 3. The upper oesophagus is dilated (arrowed). A small pseudo-epiglottis (P) is present. There is a metallic marker at the tracheostoma. (b) The same spastic patient during attempted phonation with air insufflation. There is still severe narrowing of the reconstructed pharynx with no vibrating P-E segment and no voice. The oesophagus is more dilated than in Fig. 4a.
Classification
Poor speakers were divided into four categories: hypotonic, hypertonic, spastic (spasm), and stricture. The main features of these categories are summarised in Table 1.
(a)
Good OesophagealSpeakers The characteristic feature of this group is the presence of one vibrating P-E segment during phonation (Figs la, I b). The segment can measure up to 3 cm in length and the anterior and posterior walls are separated by a few millimetres during phonation. The pharynx dilates fully during swallowing. There are sometimes slight irregularities of the anterior and posterior walls but the most striking radiological feature is often a 'pseudo-epiglottis' which arises from the anterior wall of the P-E segment at the base of the tongue. It has
(b)
Fig. 3 - (a) Hypertonic patient during attempted phonation. There are two severely narrowed segments but no vibrating P-E segment and no voice. A pseudo-epiglottis (P) is present. (b) The same hypertonic patient during attempted phonation with air insufftation. The narrowed segments are unchanged and there is still no vibrating P-E segment and no voice. The pseudo-epiglottis is also unchanged. The tip of the tracheostoma marker is just visible inferiorly.
(a)
(b)
Fig. 5 - (a) Stricture (arrowed) on a lateral view of the reconstructed pharynx during a barium swallow. (b) Antero-posterior view of the same patient showing asymmetrical narrowing o f the reconstructed pharynx typical of stricture.
315
POST LARYNGECTOMY SPEECH
the appearance o f a thickened fixed epiglottis and is probably the site of surgical anastomosis between the tongue and anterior wall of the pharynx. This feature is common in good and poor oesophageal speakers and seems to be of no functional significance (Figs 3 and 4).
Hypotonic The characteristic feature of this group is that the anterior and posterior walls of the reconstructed pharynx remain widely separated during attempted phonation (Fig. 2a) and air insufltation. There is no evidence of a narrowed or vibrating segment unless pressure is applied to the anterior part of the neck, usually with the patient's finger or thumb when the anterior wall of the pharynx is displaced posteriorly so that it contacts the posterior wall (Fig. 2b). This produces a vibrating P-E segment and voicing occurs. The pharynx dilates widely on swallowing and although there are sometimes irregularities in the anterior and posterior walls, there is no hold up to the passage of barium.
Hypertonic The characteristic feature of this group is the presence of one or two severely narrowed segments during attempted speech which persist during air insufltation (Figs 3a, 3b). If two segments are present as is often the case, one is usually narrower than the other and seems to be more functionally significant. The outline of the pharynx during swallowing is variable but there is never any hold up to the passage of barium. The anterior wall may be irregular and a pseudoepiglottis is often present (Figs 3a, 3b). The posterior wall may demonstrate a filling defect on some swallows due to muscle spasm.
Spastic The features of the hypertonic group are all present but in a more extreme form and the more severely affected patients in the hypertonic group merge with the more mildly affected patients in the spastic group. The main feature which identifies the spastic patient is that the narrowed segment (or segments) which are present during attempted phonation (Fig. ,4a) become even narrower on air insufltation, occluding the pharynx and producing oesophageal dilatation (Fig. 4b). The air is then released in sudden explosive bursts and there is no speech. The dilated upper oesophagus can be clearly seen on fluoroscopy and the oesophageal contractions and explosive releases of air can be seen and heard by the examining team. Static or spot films taken during air insufflation show marked oesophageal dilatation and the narrowed segment (or segments) are narrowed and longer than in the hypertonic group. There is usually a filling defect arising from the posterior wall of the pharynx on swallowing and there is often some hold-up to the passage of barium. It can be difficult to distinguish between persistent muscle spasm and stricture, but if the patient can be persuaded to take multiple swallows of barium the spastic segment should dilate at least once and this can be recorded on videotape and viewed at leisure when the examination is complete.
Stricture The distinguishing feature of this group shows on the barium swallow, where there is constant narrowing of a short segment in the reconstructed pharynx with hold up to the passage of barium (Figs 5a, 5b). The narrowed segment has the same outline on swallowing and phonation and it never dilates. There is no vibrating P-E segment and no voice. It can be difficult or even impossible to pass a catheter through the stricture. Patients in this group always report dysphagia which is worse for solids than liquids.
RESULTS The results of 134 consecutive examinations carried out on poor oesophageal speakers between January 1980 and July 1988 is given in Table 2. DISCUSSION We were surprised when the radiological assessment demonstrated an anatomical or physiological reason for poor oesophageal speech in every case as this did not agree with the widely held view that poor speech often results from psychological, sociological and educational problems. In our series a vibrating P-E segment in an otherwise dilated pharynx was demonstrated in all good speakers and these features were never present in poor or failed speakers. The appearance of the pharynx was not affected by the presence of a naso-pharyngeal tube (Figs 3b and 4b). The main advantage o f the radiological assessment is that it places patients in treatment groups and identifies those who require further surgery to the pharynx if they are to develop effective alaryngeal speech. Patients with stricture always require surgery; usually a flap reconstruction of the pharynx to improve dysphagia as well as speech. Those with spasm are treated by surgical myotomy and it is our custom to place metallic liga clips at both ends of the myotomy site so that the effect of the operation can be readily assessed on subsequent barium examinations. Treatment of the hypertonic group varies: surgical myotomy is required for the most severely affected patients but some can be managed by a straightforward Blom-Singer procedure and the insertion of a prosthesis of the low resistance type. Patients classified as hypotonic are often helped by a tight collar which approximates the anterior and posterior walls of the reconstructed pharynx in the same way as digital pressure on the anterior part of the neck. Table 2 - Radiological classification of 134 initial examinations carried out on poor and failed oesophageal speakers between January 1980 and July 1988
Radiological classification Number Percentage (%) Itypotonic Hypertonic Spastic Stricture
19 40 62 13
14 30 46 10
134
100
316
CLINICAL RADIOLOGY
By the end o f 1988, 55 o f the 134 p o o r or failed speakers in the series had been m a n a g e d by prosthetic valve insertion combined with the m a n a g e m e n t outlined above and 46 (84%) had developed effective speech. This is m u c h higher than the success rates quoted in the introduction and compares f a v o u r a b l y with the success rate o f 69% reported by Maniglia et al. (1989) who assessed 62 patients one year after valve insertion. The reasons for failure in our nine cases were: alcoholism (3); failing eyesight (2) and there were four patients who disliked the deep tone p r o d u c e d by the reconstructed p h a r y n x and preferred to continue with an artificial vibrating larynx. The radiological appearances o f p o o r oesophageal speakers in this series has resulted in modifications to the technique o f laryngectomy, the m o s t significant being 'primary' tracheo-oesophageal puncture at the time o f laryngectomy with insertion o f a prosthetic valve soon afterwards. Patients treated in this way have a slightly higher p r o p o r t i o n o f g o o d speakers 6 m o n t h s after surgery (89%) than those treated by secondary puncture at a second operation (84%). The m a n a g e m e n t o f laryngectomy at Charing Cross Hospital is likely to change in the future as new data is obtained f r o m patients currently under treatment but the information already available indicates that g o o d postlaryngectomy speakers have a single vibrating P-E segment in an otherwise dilated p h a r y n x and that p o o r speakers can be placed in treatment groups according to the radiological appearances o f the reconstructed pharynx.
REFERENCES
Berry, RJ (1983). Radiotherapy and chemotherapy. In Laryngectomy: Diagnosis to Rehabilitation, ed. Edels Y. pp. 18-36. Croom-Helm, London.
Blom, ED & Singer, MI (1979). Surgical-prosthetic approaches for Post laryngectomy rehabilitation. In Laryngectomy Rehabilitation, Ed. Keith RL. pp. 251-277. College Hill Press, Houston, Texas. Blom, ED, Singer, MI & Hamaker, RC (1985). An improved oeso. phageal insufflation test. Archives of Otolaryngology, 111, 211212. Cheesman, AD, Knight, J, Mclvor, J & Perry, A. (1986). Tracheo. oesophageal 'puncture speech' an assessment technique for failed oesophageal speakers. Journal of Laryngology and Otology, 100, 191-199. Cotes, JE. (1979). Lung Function Assessment and Application in Medicine. p. 371. Blackwell Scientific Publications, Oxford. Diedrich, WM & Youngstrom, KA (1966). Alaryngeal Speech. CC Thomas, Springfield, Illinois. Duguay, M (1966). Pre-operative ideas of speech after laryngectomy Archives of Otolaryngology, 83, 237-240. Edels, Y (1983). Pseudo-voice: its theory and practice. Laryngectomy: Diagnosis to Rehabilitation, Ed. Edels, Y. pp. 107-141. CroomHelm, London. Harwood, AR & Rawlinson, E (1983). The quality of life of patients following treatment for laryngeal cancer. International Journal of Radiation Oncology Biology and Physics, 9, 335-338. Maniglia, AJ, Lundy, DS, Casiano, RC & Swim, SC (1989). Speech restoration and complications of primary versus secondary tracheoesophageal puncture following total laryngectomy. Laryngoscope, 99, 489-491. Panje, WR (1981). Prosthetic vocal rehabilitation following laryngectomy--the voice button. Annals of Otology, Rhinology and Laryngology, 90, 116-120. Perry, A (1983). Surgical approach to speech rehabilitation. In Laryngectomy Diagnosis to Rehabilitation, Ed. Edels, Y. pp. 271-289. Croom Helm, London. Schaefer, SD & Johns, DF (1982). Attaining functional oesophageal speech Archives of Otolaryngology, 108, 647 649. Seeman, M (1967). Rehabilitation of laryngectomised subjects. Acta Otolaryngology, 64, 235-241. Singer, M I & Blom, ED (1980). An endoscopic technique for restoration of voice after laryngectomy. Annals of Otology, Rhinology and Laryngology, 89, 529 533. Smith, J, Rise, EN & Gralnek, DE (1966). Speech recovery in the laryngectomised patient. Laryngoscope, 76, 1540-1546. Staffieri, M (1980). New surgical approaches for speech rehabilitation after total laryngectomy. In Surgical and Prosthetic Approaches to Speech Rehabilitation, Ed. Shedd, DP & Weinberg, B. pp. 77-119. G.K. Hall, Boston.
Radiological Assessment of Post Laryngectomy Speech J. McIVOR*, P. F. EVANS*, A. PERRY'~ and A. D. CHEESMAN~
Departments of* Radiology, t Speech Therapy and ~ Ear Nose and Throat Surgery, Charing Cross Hospital, London Laryngectomy for carcinoma of the larynx has a five year survival rate of approximately 80% but the operation has a high morbidity as most patients do not develop effective alaryngeal speech. The published literature states that poor speech is occasionally due to anatomical or physiological problems in the reconstructed pharynx but is usually due to psychological, sociological or educational difficulties. We devised a radiological technique using a combination of video-fluoroscopy and static or spot films to assess the anatomy and function of the reconstructed pharynx of postlaryngectomy patients, and investigated 24 good and 134 poor or failed speakers. We examined the pharynx during the acts of swallowing, attempted phonation and attempted phonation with air insuffiation. A single vibrating segment (P-E segment) in an otherwise dilated pharynx was present in all good speakers, but was absent in poor or failed speakers whom we classified as hypotonic (19), hypertonic (40), spastic (62) and strictured (13) according to the radiological appearances. This classification has been used to plan subsequent clinical management and 46 of the 55 patients treated to date (84%) have developed effective alaryngeal speech.
Laryngectomy for carcinoma of the larynx has a 5 year survival rate of approximately 80% (Berry, 1983) which is much higher than the survival rates after most operations for malignant disease. It is probably due to the fact that tumour extension beyond the larynx occurs late in most cases and laryngectomy usually removes the whole tumour~ There a-re no reliable figures-for the,number of laryngectomy operations carrie~ out in the UK each year, but it is probably in the region of 200, in which case the number of surviving patients is approximately 2000. Although the survival rate is high, the operation has a high morbidity as most patients do not develop effective speech and have a continuing and often demoralising problem in communication. The proportion who fail to develop speech was recorded as 88% by Harwood and Rawlinson (1983). Most papers quote a figure nearer to 75% (Schafer and Johns, 1982), but failure rates as low as 30% have also been reported (Smith et al., 1966). The main reason for this wide variation is that different criteria have been used by different authors to define 'good' speech. The causes of speech failure were reviewed by Perry (1983) and are usually classified as: anatomical; physiological; psychological; sociological and educational as suggested by Duguay (1966). Post-laryngectomy oesophageal speech is produced by regurgitating air from the oesophagus through a 'vibrating segment' (pharyngo-esophageal or P-E segment) in the reconstructed pharyngo-oesophagus. The volume of air which can be retained in the dilated oesophagus rarely exceeds 80 cc (Edels, 1983) which is much less than the Correspondenceto: Dr J. McIvor,X-rayDepartment,CharingCross Hospital, FulhamPalaceRoad, LondonW6 8RF.
normal vital capacity of four litres in men aged 60 years (Cotes, 1979) and this tends to produce short phrases of quiet, low pitched speech. Surgical procedures have been devised to divert air from the trachea into the reconstructed pharyngo-oeso. phagus and thus increase the flow of air through the P-E segment. In this way it was hoped to produce louder and more sustained speech. Staffieri (1980) described a surgical technique for creating a mucosal shunt between the trachea and the reconstructed pharynx so that digital occlusion of the tracheal opening (tracheostoma) during expiration caused air to pass into the pharynx. However, there were serious problems caused by aspiration of saliva and other fluids into the lungs. The next development was a valved prosthesis which could be passed through a surgically created fistula between the trachea and pharynx (Singer and Blom, 1980; Panje, 1981). The valve prevented aspiration into the lungs but it was soon realised that a correctly placed and functioning prosthesis did not invariably produce good speech. Attention then focused on the anatomy and physiology of the reconstructed pharyngo-oesophagus and the vibrating P-E segment. Blom et al. (1985) used the air insutttation test originally described by Seeman (1967) to assess the function of the P-E segment. A 14 gauge tube was passed trans-nasally into the pharynx and the tip placed beloW the P-E segment. Air was insufltated through the tube as the patient attempted to speak. The radiological appearances of the reconstructed pharynx after laryngectomy were described by Diedrich and Youngstrom (1966) and reviewed by Blom and Singer (1979). Cheesman et al. (1986) described the radiological and clinical procedures used to assess 50 post-laryngectomy patients who were being considered for the BornSinger procedure at Charing Cross Hospital. The aim of this paper is to describe the radiological technique in more detail, to record the classification of 134 poor or failed oesophageal speakers and indicate the results of treatment based on the radiological findings. PATIENTS AND METHODS We assessed 24 good and 134 poor speakers using a combination of fluoroscopy, video recording and static radiographs (spot films). Fluoroscopy allowed the movement of the reconstructed pharynx to be viewed while observing the patient, video recording provided an easily accessible permanent record of movement and the spot films gave a more anatomically detailed but static image of the pharynx and P-E segment. The examinations were carried out by a team consisting of a radiologist, speech therapist and otorhinolaryngologist, all of whom observed the fluoroscopy screen and the patient during the examination. We used a standard Siemans Sirescop 3 fluoroscoPY unit with the table in the vertical position and the foot rest removed. Most patients were examined whilst standing as
POST LARYNGECTOMY SPEECH
313
Table 1 - Rafliologieal features of the reconstructed pharynx in good and poor oesophageal speakers after laryngectomy
Hypotonic G o o d
HypertonicSpastic
Stricture
Barium swallow
Dilates normally
Dilates normally
Usually dilates normally
Usually narrowed posteriorly
Persistent narrowing
Attempted phonation
Dilates widely, no vibrating segment
Vibrating P-E segment, Air in oesophagus
Marked narrowing with 1 or 2 narrowed segments
Severe narrowing with 1 or 2 narrowed segments
Persistent narrowing. No air in oesophagus
Attempted phonation with air insufflation
Dilates widely, no vibrating segment
Vibrating P-E segment, Air in oesophagus
Marked narrowing as above, Oesophagus dilates
Severe narrowing as above. Oesophagus dilates widely
Persistent narrowing. Oesophagus dilates
No voice
Good voice
Variable voice
No voice
No voice
the shoulders tend to drop slightly in this position and the other examinations were carried out with the patients sitting. The distance between the table top and the image intensifier measured 45 cm which allowed most patients to be examined in the true lateral position but those with very broad shoulders were placed with the shoulders slightly oblique. The tracheostoma was marked with a radio-opaque marker on the skin. It can be difficult to obtain good lateral radiographs of the pharynx and upper oesophagus after laryngectomy, as the lower part of the neck is much thinner and more radiolucent than the upper part. However, two simple modifications made it possible to obtain satisfactory radiographs of the reconstructed pharynx in all our patients. Firstly the X-ray beam was filtered by a copper sheet (0.5 mm thick) fixed to the table top and secondly the field size reaching the image intensifier was reduced with a lead rubber sheet (0.5 mm lead equivalent) attached to the front of the explorator which masked out the non-attenuated part of the beam (Figs 1-5). In our experience, lateral fluoroscopy and films were adequate for diagnosis unless the patient had a surgical complication such as stricture or fistula, when additional antero-posterior films were helpful. The radiation dose was estimated at 5-10 cGy which is rather high for a diagnostic procedure but much less than the therapeutic dose of 4000-6000 cGy already received by most patients and the field was, of course, limited to the neck and thoracic inlet.
least two spot films were taken during maximum vocalisation.
Attempted Phonation With Air Insufliation A flexible rubber tube with a densely radio-opaque tip and two opposed side holes was passed through the nose and P-E segment. We used a specially modified 14 gauge bronchographic rubber catheter which could be readily seen on fluoroscopy (Fig. 4b). The tip of the tube was positioned below the level of the narrowest part of the P-E segment and air was passed through the tube at a rate of approximately 0.5 litres per minute. The pharynx was coated with a sip of barium and the patient was asked to mime counting from one to ten as before. The procedure was usually repeated with the
Swallowing The patient swallowed barium (Baritop) while the pharynx was screened in the lateral position. The movements were viewed on a TV monitor and were simultaneously recorded on videotape as was the whole examination. Spot films were taken on at least two occasions when the P-E segment was fully dilated by a bolus of barium.
Attempted Phonation The patient took a sip of barium to coat the mucosa of the pharynx, swallowed air to produce a reservoir in the Oesophagus and counted slowly from one to ten. The movements were recorded on videotape as before and at
(a)
(b)
Fig. 1 - (a) Lateral view of a good oesophageal speaker during phonation. The vibrating pharyngo-oesophageal (P-E) segment is marked with arrows. There is a slight narrowing below which is of no functional significance. The metallic marker is adjacent to the tracheostoma and the rectangular field size has been reduced anteriorly and inferiorly with a sheet of lead rubber attached to the explorator. (b) Antero-posterior view of another good oesophageal speaker during phonation with the vibrating P-E segment marked with arrows. The metallic marker is situated at the tracheostoma.
314
CLINICALRADIOLOGY
(a)
(b)
Fig. 2 - (a) Hypotonic patient during attempted phonation. The anterior and posterior walls of the reconstructed pharynx are widely separated and there is no vibrating P-E segment and no voice. A tracheostomy tube is present. (b) The same hypotonic patient during attempted phonation with digital pressure applied to the anterior surface of the neck. The anterior wall is displaced posteriorly producing a P-E segment (arrowed) and voice.
catheter tip at slightly different levels and spot films were taken during maximum vocalisation.
(a)
(b)
Fig. 4 (a) Spastic patient during attempted phonation. There are two severely narrowed segments which are narrower and longer than in the hypertonic patient shown in Fig. 3. The upper oesophagus is dilated (arrowed). A small pseudo-epiglottis (P) is present. There is a metallic marker at the tracheostoma. (b) The same spastic patient during attempted phonation with air insufflation. There is still severe narrowing of the reconstructed pharynx with no vibrating P-E segment and no voice. The oesophagus is more dilated than in Fig. 4a.
Classification
Poor speakers were divided into four categories: hypotonic, hypertonic, spastic (spasm), and stricture. The main features of these categories are summarised in Table 1.
(a)
Good OesophagealSpeakers The characteristic feature of this group is the presence of one vibrating P-E segment during phonation (Figs la, I b). The segment can measure up to 3 cm in length and the anterior and posterior walls are separated by a few millimetres during phonation. The pharynx dilates fully during swallowing. There are sometimes slight irregularities of the anterior and posterior walls but the most striking radiological feature is often a 'pseudo-epiglottis' which arises from the anterior wall of the P-E segment at the base of the tongue. It has
(b)
Fig. 3 - (a) Hypertonic patient during attempted phonation. There are two severely narrowed segments but no vibrating P-E segment and no voice. A pseudo-epiglottis (P) is present. (b) The same hypertonic patient during attempted phonation with air insufftation. The narrowed segments are unchanged and there is still no vibrating P-E segment and no voice. The pseudo-epiglottis is also unchanged. The tip of the tracheostoma marker is just visible inferiorly.
(a)
(b)
Fig. 5 - (a) Stricture (arrowed) on a lateral view of the reconstructed pharynx during a barium swallow. (b) Antero-posterior view of the same patient showing asymmetrical narrowing o f the reconstructed pharynx typical of stricture.
315
POST LARYNGECTOMY SPEECH
the appearance o f a thickened fixed epiglottis and is probably the site of surgical anastomosis between the tongue and anterior wall of the pharynx. This feature is common in good and poor oesophageal speakers and seems to be of no functional significance (Figs 3 and 4).
Hypotonic The characteristic feature of this group is that the anterior and posterior walls of the reconstructed pharynx remain widely separated during attempted phonation (Fig. 2a) and air insufltation. There is no evidence of a narrowed or vibrating segment unless pressure is applied to the anterior part of the neck, usually with the patient's finger or thumb when the anterior wall of the pharynx is displaced posteriorly so that it contacts the posterior wall (Fig. 2b). This produces a vibrating P-E segment and voicing occurs. The pharynx dilates widely on swallowing and although there are sometimes irregularities in the anterior and posterior walls, there is no hold up to the passage of barium.
Hypertonic The characteristic feature of this group is the presence of one or two severely narrowed segments during attempted speech which persist during air insufltation (Figs 3a, 3b). If two segments are present as is often the case, one is usually narrower than the other and seems to be more functionally significant. The outline of the pharynx during swallowing is variable but there is never any hold up to the passage of barium. The anterior wall may be irregular and a pseudoepiglottis is often present (Figs 3a, 3b). The posterior wall may demonstrate a filling defect on some swallows due to muscle spasm.
Spastic The features of the hypertonic group are all present but in a more extreme form and the more severely affected patients in the hypertonic group merge with the more mildly affected patients in the spastic group. The main feature which identifies the spastic patient is that the narrowed segment (or segments) which are present during attempted phonation (Fig. ,4a) become even narrower on air insufltation, occluding the pharynx and producing oesophageal dilatation (Fig. 4b). The air is then released in sudden explosive bursts and there is no speech. The dilated upper oesophagus can be clearly seen on fluoroscopy and the oesophageal contractions and explosive releases of air can be seen and heard by the examining team. Static or spot films taken during air insufflation show marked oesophageal dilatation and the narrowed segment (or segments) are narrowed and longer than in the hypertonic group. There is usually a filling defect arising from the posterior wall of the pharynx on swallowing and there is often some hold-up to the passage of barium. It can be difficult to distinguish between persistent muscle spasm and stricture, but if the patient can be persuaded to take multiple swallows of barium the spastic segment should dilate at least once and this can be recorded on videotape and viewed at leisure when the examination is complete.
Stricture The distinguishing feature of this group shows on the barium swallow, where there is constant narrowing of a short segment in the reconstructed pharynx with hold up to the passage of barium (Figs 5a, 5b). The narrowed segment has the same outline on swallowing and phonation and it never dilates. There is no vibrating P-E segment and no voice. It can be difficult or even impossible to pass a catheter through the stricture. Patients in this group always report dysphagia which is worse for solids than liquids.
RESULTS The results of 134 consecutive examinations carried out on poor oesophageal speakers between January 1980 and July 1988 is given in Table 2. DISCUSSION We were surprised when the radiological assessment demonstrated an anatomical or physiological reason for poor oesophageal speech in every case as this did not agree with the widely held view that poor speech often results from psychological, sociological and educational problems. In our series a vibrating P-E segment in an otherwise dilated pharynx was demonstrated in all good speakers and these features were never present in poor or failed speakers. The appearance of the pharynx was not affected by the presence of a naso-pharyngeal tube (Figs 3b and 4b). The main advantage o f the radiological assessment is that it places patients in treatment groups and identifies those who require further surgery to the pharynx if they are to develop effective alaryngeal speech. Patients with stricture always require surgery; usually a flap reconstruction of the pharynx to improve dysphagia as well as speech. Those with spasm are treated by surgical myotomy and it is our custom to place metallic liga clips at both ends of the myotomy site so that the effect of the operation can be readily assessed on subsequent barium examinations. Treatment of the hypertonic group varies: surgical myotomy is required for the most severely affected patients but some can be managed by a straightforward Blom-Singer procedure and the insertion of a prosthesis of the low resistance type. Patients classified as hypotonic are often helped by a tight collar which approximates the anterior and posterior walls of the reconstructed pharynx in the same way as digital pressure on the anterior part of the neck. Table 2 - Radiological classification of 134 initial examinations carried out on poor and failed oesophageal speakers between January 1980 and July 1988
Radiological classification Number Percentage (%) Itypotonic Hypertonic Spastic Stricture
19 40 62 13
14 30 46 10
134
100
316
CLINICAL RADIOLOGY
By the end o f 1988, 55 o f the 134 p o o r or failed speakers in the series had been m a n a g e d by prosthetic valve insertion combined with the m a n a g e m e n t outlined above and 46 (84%) had developed effective speech. This is m u c h higher than the success rates quoted in the introduction and compares f a v o u r a b l y with the success rate o f 69% reported by Maniglia et al. (1989) who assessed 62 patients one year after valve insertion. The reasons for failure in our nine cases were: alcoholism (3); failing eyesight (2) and there were four patients who disliked the deep tone p r o d u c e d by the reconstructed p h a r y n x and preferred to continue with an artificial vibrating larynx. The radiological appearances o f p o o r oesophageal speakers in this series has resulted in modifications to the technique o f laryngectomy, the m o s t significant being 'primary' tracheo-oesophageal puncture at the time o f laryngectomy with insertion o f a prosthetic valve soon afterwards. Patients treated in this way have a slightly higher p r o p o r t i o n o f g o o d speakers 6 m o n t h s after surgery (89%) than those treated by secondary puncture at a second operation (84%). The m a n a g e m e n t o f laryngectomy at Charing Cross Hospital is likely to change in the future as new data is obtained f r o m patients currently under treatment but the information already available indicates that g o o d postlaryngectomy speakers have a single vibrating P-E segment in an otherwise dilated p h a r y n x and that p o o r speakers can be placed in treatment groups according to the radiological appearances o f the reconstructed pharynx.
REFERENCES
Berry, RJ (1983). Radiotherapy and chemotherapy. In Laryngectomy: Diagnosis to Rehabilitation, ed. Edels Y. pp. 18-36. Croom-Helm, London.
Blom, ED & Singer, MI (1979). Surgical-prosthetic approaches for Post laryngectomy rehabilitation. In Laryngectomy Rehabilitation, Ed. Keith RL. pp. 251-277. College Hill Press, Houston, Texas. Blom, ED, Singer, MI & Hamaker, RC (1985). An improved oeso. phageal insufflation test. Archives of Otolaryngology, 111, 211212. Cheesman, AD, Knight, J, Mclvor, J & Perry, A. (1986). Tracheo. oesophageal 'puncture speech' an assessment technique for failed oesophageal speakers. Journal of Laryngology and Otology, 100, 191-199. Cotes, JE. (1979). Lung Function Assessment and Application in Medicine. p. 371. Blackwell Scientific Publications, Oxford. Diedrich, WM & Youngstrom, KA (1966). Alaryngeal Speech. CC Thomas, Springfield, Illinois. Duguay, M (1966). Pre-operative ideas of speech after laryngectomy Archives of Otolaryngology, 83, 237-240. Edels, Y (1983). Pseudo-voice: its theory and practice. Laryngectomy: Diagnosis to Rehabilitation, Ed. Edels, Y. pp. 107-141. CroomHelm, London. Harwood, AR & Rawlinson, E (1983). The quality of life of patients following treatment for laryngeal cancer. International Journal of Radiation Oncology Biology and Physics, 9, 335-338. Maniglia, AJ, Lundy, DS, Casiano, RC & Swim, SC (1989). Speech restoration and complications of primary versus secondary tracheoesophageal puncture following total laryngectomy. Laryngoscope, 99, 489-491. Panje, WR (1981). Prosthetic vocal rehabilitation following laryngectomy--the voice button. Annals of Otology, Rhinology and Laryngology, 90, 116-120. Perry, A (1983). Surgical approach to speech rehabilitation. In Laryngectomy Diagnosis to Rehabilitation, Ed. Edels, Y. pp. 271-289. Croom Helm, London. Schaefer, SD & Johns, DF (1982). Attaining functional oesophageal speech Archives of Otolaryngology, 108, 647 649. Seeman, M (1967). Rehabilitation of laryngectomised subjects. Acta Otolaryngology, 64, 235-241. Singer, M I & Blom, ED (1980). An endoscopic technique for restoration of voice after laryngectomy. Annals of Otology, Rhinology and Laryngology, 89, 529 533. Smith, J, Rise, EN & Gralnek, DE (1966). Speech recovery in the laryngectomised patient. Laryngoscope, 76, 1540-1546. Staffieri, M (1980). New surgical approaches for speech rehabilitation after total laryngectomy. In Surgical and Prosthetic Approaches to Speech Rehabilitation, Ed. Shedd, DP & Weinberg, B. pp. 77-119. G.K. Hall, Boston.