We Can Predict Postpalatoplasty Velopharyngeal Insufficiency in Cleft Palate Patients

Transcription

1 The Laryngoscope VC 2013 The American Laryngological, Rhinological and Otological Society, Inc. We Can Predict Postpalatoplasty Velopharyngeal Insufficiency in Cleft Palate Patients Jacques E. Leclerc, MD; Audrey Godbout, MD; Isabelle Arteau-Gauthier, MD; Sophie Lacour, MOA; Kati Abel, MOA; Elisa-Maude McConnell, MSc Objectives/Hypothesis: To find an anatomical measurement of the cleft palate (or a calculated parameter) that predicts the occurrence of velopharyngeal insufficiency (VPI) after palatal cleft repair. Study Design: Retrospective cohort study. Methods: Charts were reviewed from cleft palate patients who underwent palatoplasty by the Von Langenbeck technique for isolated cleft palate or Bardach two-flap palatoplasty for cleft lip-palate. Seven anatomical cleft parameters were prospectively measured during the palatoplasty procedure. Three blinded speech language pathologists retrospectively scored the clinically assessed VPI at 4 years of age. The recommendation of pharyngoplasty was also used as an indicator of VPI. Results: From 1993 to 2008, 67 patients were enrolled in the study. The best predicting parameter was the ratio a/(30 2 b1), in which a is defined as the posterior gap between the soft palate and the posterior pharyngeal wall and b1 is the width of the cleft at the hard palate level. An a/(30 2 b1) ratio >0.7 to 0.8 is associated with a higher risk of developing VPI (relative risk , sensitivity 5 72% 81%, P <.03). Conclusions: The width of the cleft at the hard palate level and the posterior gap between the soft palate and the posterior pharyngeal wall were found to be the most significant parameters in predicting VPI. The best correlation was obtained with the ratio a/(30 2 b1). Key Words: Cleft palate, hypernasality, velopharyngeal insufficiency, velopharyngeal dysfunction, palatoplasty, pharyngoplasty. Level of Evidence: 4 Laryngoscope, 124: , 2014 INTRODUCTION Velopharyngeal insufficiency (VPI) can be defined as an inadequate physiological barrier between the nasopharynx and oropharynx during speech. Nasal air escapes during the production of various phonemes, and affects speech intelligibility and therefore the patient s quality of life. VPI sometimes persists in cleft palate (CP) patients after palatoplasty. Factors such as the length and function of the soft palate, the depth and width of the nasopharynx, and the motion of the posterior and lateral pharyngeal walls 1 determine the quality of speech. The best evaluation of this speech problem remains the perceptual speech assessment performed by a trained speech language pathologist (SLP). 2 In some cases, speech therapy alone is insufficient to correct the From the Quebec University Hospital Center (CHUQ) (J.E.L.), Quebec City, Quebec, Canada, Department of Otolaryngology Head and Neck Surgery, Laval University (A.G., I.A.-G.), Quebec City, Quebec, Canada, Department of Speech Language Pathology, Quebec University Hospital Center (S.L., K.A., E.-M.M.), Quebec City, Quebec, Canada Editor s Note: This Manuscript was accepted for publication April 22, The authors have no other funding, financial relationships, or conflicts of interest to disclose. Financial support was obtained from Quebec University Hospital Center Foundation. Send correspondence to Jacques E. Leclerc, 2705 Boul. Laurier, Quebec, Qc. G1V 4G2 Canada. jeleclerc@ccapcable.com DOI: /lary problem. A secondary surgical procedure such as pharyngoplasty is conducted after reaching a consensus between the multidisciplinary group, the patient, and his family. However, drawbacks such as hyponasality or sleep apnea by overcorrection must always be taken into consideration. Objective The objective of this study was to determine which children are at risk to present VPI after palatoplasty. Our specific objective was to find a statistically significant anatomical measurement or calculated parameter of the palate or its cleft that can predict the occurrence of VPI. This might help to reduce the prevalence of VPI by the use of other surgical techniques, and could have relevance for determining the duration of speech therapy follow-up, the need for secondary surgery, and its inherent risks. This may also help to spare children from the psychological and social stigmata related to this problem. MATERIALS AND METHODS In a tertiary care academic university-based medical center, we retrospectively reviewed the charts of patients who underwent palatoplasty performed by the senior author (J.E.L.). From 1993 to 2008, all patients born with a CP (6cleft lip) were prospectively enrolled in a database including data on the pregnancy, birth weight, and types of associated malformations. At 561

2 Fig. 1. Cleft lip/palate measurements diagram (J.E.L.). [Color figure can be viewed in the online issue, which is available at wileyonlinelibrary.com.] the time of the palatoplasty procedure under general anesthesia, after positioning of the Dingman retractor, several anatomical parameters of the soft/hard palate and the cleft were measured (Fig. 1, Table I). The measurements in millimeters were obtained with an ophthalmologic caliper and a ruler and were included in the database. The inclusion criteria for our study were: All the patients who had undergone palatoplasty by the senior author (J.E.L.) between 9 and 13 months of age, either by the Von Langenbeck technique for CP patients or by Bardach two-flap palatoplasty for cleft lip-palate (CLP) patients. In both groups, the levator palatini was freed from the posterior edge of the hard palate and retropositioned to restore the levator sling and allow tension-free closure in the midline. Availability of a complete set of lip/palate measurements at the time of the repair procedure. A complete management and follow-up by our speech pathology team (child older than 4 years). Patients with obvious cognitive and developmental disorders or syndromes or other craniofacial malformations including the Pierre Robin sequence were excluded from the study. In addition to the previously described measurements, >20 other parameters using two or more measurements were created and tested in our quest to find the best possible predictor. Evaluation of Velopharyngeal Function The SLPs were blinded to the results of the anatomical measurements. Clinical charts were retrospectively reviewed for information regarding patients speech as related to velopharyngeal function by a team of three SLPs specializing in the evaluation of CP patients. The speech of each patient was assessed as close as possible to the age of 4 years. Some patients had remaining small hard palate fistulas but no soft palate fistulas. The quality of speech and resonance is acknowledged as being the main outcome measure for evaluating postpalatoplasty patients. 3 Perceptual evaluation remains the gold standard for evaluating speech, as well as the most commonly used method. 4 6 Variable a b b1 c d e f TABLE I. Description of the Palatal and Cleft Measurements. Description Distance between the posterior end of the soft palate and the posterior pharyngeal wall in the plane of the hard palate Largest width of the cleft at the soft palate level Width of the cleft at the posterior end of the hard palate Length of the soft palate, from its posterior end to the posterior end of the hard palate Largest width of the dental arch measured at the top of the crests Length of the hard palate, from its posterior end to the top of the anterior dental arch with a rigid ruler; the depth of the hard palate is not evaluated Total length of the cleft, from the posterior end of the soft palate to the anterior end of the cleft 562

3 TABLE II. Parameters Used by the Speech Language Pathologists. Perceptual Speech Evaluation Hypernasality (sentences) 0 5 within normal limits 1 5 mild, indicates hypernasality resonance perceived on vowels 2 5 moderate, hypernasality resonance perceived on vowels, approximants/semivowels, and liquids, and presence of weakened consonants 3 5 severe, indicates all of the above and the replacement of voiced consonants by their nasal equivalents * Audible nasal emission (sentences) 0 5 normal 1 5 present * Audible nasal turbulence (sentences) 0 5 normal 1 5 present * Speech Acceptability 0 5 normal 1 5 normal to mild 2 5 moderate (speech deviates from normal to a moderate degree) * 3 5 severe (speech deviates from normal to a severe degree) * Pharyngoplasty Recommendation Failed adequate speech therapy and anatomical limitation *Positive for velopharyngeal insufficiency. Speech Parameters and Rating Scales The data collected underwent a mapping process, involving the conversion of our local team s evaluation protocol to the predetermined set of chosen parameters based on Henningsson et al. 4 and Sell et al. 5 For the purpose of this study, three different criteria or sets of criteria were used to classify each case as to the presence of VPI: Perceptual speech evaluation (PSE) was considered positive if at least one of these criteria was fulfilled: severity score of 3 on the hypernasality scale or presence of audible nasal emission (score of 1 on the scale) or presence of nasal turbulence (score of 1 on the scale; Table II). Speech acceptability was assessed independently. Acceptability is defined as the degree to which speech calls attention to itself apart from the content of the spoken language, and it is closely related to patients quality of life. Scores of 2 and 3 were used as an indication of VPI (Table II). Pharyngoplasty recommendation (PR) was used as the third indicator of significant persisting VPI. A surgical treatment was offered to the patient when 1) speech therapy did not improve VPI after complete SLP follow-up and 2) an obvious anatomical defect could explain the problem (Table II). No pharyngoplasty was recommended or performed in the absence of VPI. Statistical Analysis In regard to each outcome, the discriminating capacity (dc) was evaluated for each variable by the c-index corresponding to the area under the receiver operating characteristic (ROC) curve: 0 dc 1 using the SAS LOGISTIC procedure (version 9.2.3; SAS Institute, Cary, NC). Only variables having dc 0.6 were retained for further analysis. Using the OUTROC option of the LOGISTIC procedure, each selected variable was dichotomized at a corresponding level that maximized the sensitivity and the specificity of the measurement. A relative risk ratio was calculated for each variable. All P values were determined using the v 2 test or the Fisher exact test. RESULTS The charts of 272 CLP patients who underwent palatoplasty between 1993 and 2008 were reviewed. Only 67 patients met all the inclusion criteria (Fig. 2). The clinical features of our series are shown in Table III. The CP measurement data are shown in Figure 3. The independent variables used in our analysis are depicted in Table IV. Our study has shown a respective incidence of Fig. 2. Flowchart for patient recruitment. 563

4 Characteristic TABLE III. Cleft Lip/Palate Series. Value TABLE IV. Incidence (%) of Velopharyngeal Insufficiency as Determined by Perceptual Speech Evaluation, Speech Acceptability, and Pharyngoplasty Recommendation. Total number 67 Gender, male %/female % 58.2/41.7 Weight at palatoplasty, kg 9.1 (61.4) Age palatoplasty, mo 10.6 (60.9) Age at evaluation by SPL, mo 49.6 (616.0) Age at PR, mo 73.9 (638.4) Isolated cleft palate 38 (57%) Unilateral cleft lip and palate 23 (34%) Bilateral cleft lip and palate 6 (9%) PR 5 pharyngoplasty recommendation; SPL 5 speech language pathologist. 20.9%, 30.8%, and 50% for PR, speech acceptability, and PSE. Added to our anatomical measurements, >20 combinations of the anatomical measurements were created as other parameters. For each dependent variable (PR, PSE, acceptability), these combinations of anatomical measurements were tested. The 10 that best discriminated according to the dc value of the ROC curve are presented in decreasing order in Tables V to VI. Tables V and VI, respectively, present the data of only 60 and 65 patients of 67. In their retrospective chart evaluation, the SLPs felt that they could not adequately classify the missing others according to the available information. We did not get b1 values for all the patients in each table. The missing values correspond to the number of patients with clefts limited to the soft palate. We elected to assign no value for b1 instead of zero. The f parameter was our marker for the anteroposterior extent of the cleft. There was an obvious clerical mistake for one c value that we did not use for any of our calculations. For all three dependent variables, the data in these tables show that the ratio a/(30 2 b1), in which a is defined as the posterior gap between the soft palate and the posterior pharyngeal wall and b1 is the width of the cleft at the hard palate level, had a good discriminating value and a globally better sensitivity that reached statistical significance. The cutting ratio values for PSE, Fig. 3. Cleft palate anatomical measurements in millimeters. Velopharyngeal Insufficiency Evaluation % Perceptual speech evaluation 50 Hypernasality * Audible nasal emission score * Audible nasal turbulence score * Speech acceptability 30.8 Pharyngoplasty recommendation 20.9 *Respective percentages for the criteria used for perceptual speech evaluation in the study. The sum of the percentages of the individual PSE criteria does not match the 50% total because some patients scored as positive >1 criterion. acceptability, and PR were respectively 0.73, 0.83, and Table VIII shows the calculated relative risk of developing VPI for these values. The parameter a 1 b1 was a close contender (Tables V). b1 and a presented the highest discriminating values among the single anatomical measurements, but did not get good rankings in all three dependent variables (Table VIII). The measurement b1 performed well in two of the three VPI evaluations and the depth of the nasopharynx (a) in one (PSE). They were also found within many of the top 10 parameters (Tables V). There was no association between age, sex and PSE, and PR and acceptability. DISCUSSION VPI Rates: Why Three Evaluation Modalities? We selected to use three VPI evaluation modalities. Within them, we defined VPI differently to be able to pick up the subtle as well as the more severe cases and therefore cover the widest possible range of VPI. Our ideal anatomical parameter or combination had to perform well in all three modalities. PR was the indicator of an obvious problem caused by a significant anatomical deficit that needed to be corrected. The procedure was found to be required in 20.9% of subjects in our series at approximately 4 years old. For the second modality, speech acceptability, we elected to use the scores 2 (moderate) and 3 (severe), which included a larger group of children. We found that 30.5% of the 4-year-olds presented a significant difficulty. Of this group, 12% had a moderate degree and 18.5% a severe degree of difficulty. With the third modality, perceptual evaluation, we opted to include all patients with even subclinical audible nasal emission or turbulence. We found that 50% of the patients were completely normal. The residual 50% included a wide range of VPI, from clinically normal to severe. These figures may seem elevated. However, 12 of 67 patients (18%) scored positive in only one of the two 564

5 TABLE V. Top 10 Discriminating Parameters (dc Scores) for Perceptual Speech Evaluation. PSE 5 50% (60 Patients) Variables Dc Cutting Values No. * % of VPI RR P Sensitivity, % Specificity, % Rank a 1 d < a/(50 2 d) < a 3 d < a/(c 1 e) < a < a/c < a/(30 2 b1) < a 1 b < a/(30 2 b) < a 3 b < *Number of cases above or below the cutting value. Percentage of VPI in the patients above or below the cutting value. v2 test. Parameter within top 10 in all three VPI evaluation modalities (P <.05). dc 5 discriminant capacity with associated ranking; PSE 5 perceptual speech evaluation; RR 5 relative risk; VPI 5 velopharyngeal insufficiency. parameters (constant or inconstant nasal turbulence or emission). They were classified as minimal VPI, but experienced no functional or social impact on everyday life. Of the remaining 32%, as previously mentioned, 2 = 3 of the cases were offered a pharyngoplasty (14 of 67 patients). We found a close correlation between the 20.9% pharyngoplasty rate and the score 3 hypernasality rate of 18.2% (Table IV; P <.0001). We are in the process of reviewing the nonoperated cases for a possible secondary palatal lengthening procedure such as the Furlow operation. The 32% rate of VPI is likely mostly related to the inherent limitations of the palatoplasty procedure. Looking retrospectively at the charts of this group, we found other factors that may contribute to the persisting VPI. Some cases had minor residual fistulas for which a surgical procedure was considered but was not already done at age 4 years. Some patients had other deleterious health problems not identified by our exclusion criteria, such as significant deafness and dyspraxia. These conditions may have interfered with the re-education of the child. Reported rates of persistent VPI following primary palatoplasty show a wide range in the literature, from 13% to 35% A number of factors such as the size of the cleft, age, and type of primary palatal repair performed as well as speech assessment procedures may be responsible for these differences in VPI rates. No data have clearly been identified to predict the occurrence of VPI. Conflicting data about the association between age, surgery, and classification of the cleft with VPI was noticeable in our literature review. Top Performing Parameters The goal of the study was to find an anatomical measurement of the CP or a calculated parameter that predicts the occurrence of VPI. We wanted to use linear measurements that are readily available to the surgeon. We were looking for a single ideal parameter to be applicable to the various forms of cleft malformations for this combination of two surgical techniques. We selected in our database a group of CP cases associated or not associated with a cleft lip, all treated surgically with an identical technique. We fully acknowledge that velopharyngeal closure is a dynamic process and other neuromuscular factors are significant in the development of VPI even in an ideal subject. We eliminated all cases with syndromes as well as other confounding factors and hoped this would reduce the effect of any neurologic component in the selected cases. The ratio a/(30 2 b1) was a top performer in all three tests and was considered the best parameter. The 565

6 TABLE VI. Top 10 Discriminating Parameters (dc Scores) for Acceptability. Acceptability, VPI % (65 patients) Variables dc Cutting Values No. * % of VPI RR P Sensitivity, % Specificity, % Rank a 3 b < a 1 b < a/c < a 1 d < a/(50 2 d) < a/(30 2 b1) < b < a 3 d < c < b1/d < *Number of cases above or below the cutting value. Percentage of VPI in the patients above or below the cutting value. v 2 test. Parameter within top 10 in all three VPI evaluation modalities (P <.05). dc 5 discriminant capacity with associated ranking; RR 5 relative risk; VPI 5 velopharyngeal insufficiency. number 30 corresponds to the largest distance in millimeters between the inner maxillary crests in our cohort. The subtraction 30 2 b1 represents the total width of the insertion of the muscular band on each side of the cleft. This finding indicates that wider clefts (with narrower muscular bands) with a large gap between the ridge of the soft palate and the posterior pharyngeal wall are prone to VPI. A ratio higher than 0.7 to 0.8 is associated with an increased risk of VPI as measured by all three modalities: PSE, acceptability, and PR. The anatomical measurement b and the combinations a 1 b, a 3 b, and a/(30 2 b) did not perform quite as well as b1, a 1 b1, a 3 b1, and a/(30 2 b1). It is impossible to draw definitive comparative conclusions, because the b group included all the cleft patients and the b1 group only the soft palate clefts. The combinations a 1 d, a 3 d, and a/(50 2 d) were designed to determine whether a larger nasopharyngeal distance (a) and a larger width of the dental arch (d) were associated to an increased risk of VPI. The number 50 (representing millimeters) was selected so that the subtraction 50 2 d would always produce a positive number. For example, an a/(50 2 d) ratio value 1.12 was a very good performer in PSE. As a group, the combinations including a and d generally outperformed the isolated anatomical a and d measurements. In the PSE evaluations, they showed high sensitivity and specificity. 566 The dental arch width d is partly determined by the width of the cleft b1. We think this is why both variables were good performers when associated with a in our study. In the top 10 lists of predictors (Tables V), the width of the hard palate (b1), and the depth of the nasopharynx (a) were very prominent in the top performing parameters. As shown in Table VIII, b1 > 9 mm and a > 16 mm were the top anatomical parameters. Lam and coworkers 13 studied isolated CP cases repaired with either a Furlow double-opposing Z-plasty or a combined Furlow palatoplasty and a V to Y pushback procedure. They also found that palatal width > 10 mm is associated with a higher risk of VPI. Further studies are required to determine whether our top parameters will be similar for other types of CP repair such as the Furlow procedure. It is likely they will remain the best predictors, but the cutting values of the ratios and measurements may be different. Moreover, with a larger cohort, we may be able to separate the various types of clefts (6cleft lip, unilateral vs. bilateral cleft lip) and get a specific ratio that would reach statistical significance. The length of the soft palate (c), the length of the hard palate (e), and the length of the cleft (f) did not show any association with VPI. Palatal length has been studied in an attempt to predict the need for

7 TABLE VII. Top 10 Discriminating Parameters (dc Scores) for PR. PR % (67 Patients) Variables dc Cutting Values No. % of VPI * RR P Sensitivity, % Specificity, % Rank a 3 b < b1/b < a 1 b < a/(30 2 b1) < b < a/c < a 3 d < a < b1/d < b/d < Number of cases above or below the cutting value. *Percentage of VPI in the patients above or below the cutting value. Fisher exact test. Parameter within top 10 in all three VPI evaluation modalities (P <.05). v 2 square test. dc 5 discriminant capacity with associated ranking; PR 5 pharyngoplasty recommendation; RR 5 relative risk; VPI 5 velopharyngeal insufficiency. pharyngoplasty. Randall et al. 14 measured the distance between the distal tips of both uvulae and the posterior pharyngeal wall. They concluded that for a cleft patient, if the tips of the uvulae did not reach the adenoids, a primary pharyngoplasty must be considered. Marrinan et al. 7 and Paliobeli et al. 15 suggested a working hypothesis that the vomeric muscular complex obtained by objective measurements of palatal musculature could be used to predict VPI. Kummer et al. 16 have looked at the possible relationship between the characteristics of speech and velopharyngeal gap size. They found that some information about the gap size can be predicted to a certain extent if the patient has nasal rustle. Lam and coworkers 17 in 2006 developed the concept of a gap area index. Their findings were that nasoendoscopic evaluation correlated better with VPI than multiview video fluoroscopy. TABLE VIII. Ratio a/(30 2 b1), b1, and a: Summary of Sensibility, p Value and RR of Velopharyngeal Insufficiency. Variables VPI Evaluation Sensitivity (P) RR of VPI a/(30 2 b1) 0.73 Perceptual Speech Evaluation 76% (0.014) 2.2 a/(30 2 b1) 0.83 Acceptability 72% (0.003) 3.3 a/(30 2 b1) 0.79 Pharyngoplasty Recommendation 81% (0.028) 5.1 b1 Perceptual Speech Evaluation NS b1 9 mm Acceptability 67% (0.01) 2.7 b1 9 mm Pharyngoplasty Recommendation 73% (0.036) 3.8 a 16 mm Perceptual Speech Evaluation 70% (0.01) 2.0 a Acceptability NS a Pharyngoplasty Recommendation NS NS 5 not significant; RR 5 relative risk; VPI 5 velopharyngeal insufficiency. 567

8 Study Limitations The SLPs had to classify the patients retrospectively according to the new rating systems: 1) Universal Parameters for Reporting Speech Outcomes in Individuals With Cleft Palate 4 and 2) the Great Ormond Street Speech Assessment. 5 They had to rely on the available information in the charts. Speech recordings or instrumental assessments were not available for all the patients and could not be used. Although the francophone speech protocol is not standardized, the same speech material was used across time to assess all the subjects in the study. In an attempt to facilitate crosscenter comparisons, the data collected in the present study underwent a mapping process involving the conversion of our team s evaluation protocol to a determined set of four parameters. The analysis carried out by three different professionals induces a bias into the results. Furthermore, inter-rater and intrarater reliability was not obtained, although consensus judgment was achieved on the first 15 charts reviewed. We cannot determine whether our results can be extended to other languages. The speech of each patient was assessed at the mean age of 4 years. In our experience, at this age, a comprehensive evaluation of speech and language is generally possible. Furthermore, children with documented satisfactory velopharyngeal sufficiency at this age are unlikely to develop subsequent VPI. 7 No audio/video recording or instrumental measures were obtained, although a listening tube was systematically used. 18 Instrumental measures such as nasopharyngoscopy would have been an important addition to our study, but were not available for all subjects. Among the factors that affect the surgical results, the exact surgical technique is important, and these cases were treated with freeing and retropositioning of the velar musculature. The majority of the cases were operated on at the beginning of the surgeon s career. The experience of the surgeon is a significant factor, and improvement of the results over time has not been studied. The problem of fistulas was not reviewed in the data. With the techniques that were used, the fistulas occurred only in the bony part of the repair (hard palate or maxillary arch). This is a possible bias in this study. No secondary procedure was added to lengthen the soft palate. Meaning of the Findings and Future Research The results of the study are preliminary and suggest that a surgeon undertaking a CP repair with the Von Langenbeck technique for CP patients or the Bardach two-flap palatoplasty for CLP patients can assess the risk of developing VPI by measuring the posterior gap behind the soft palate (a) and the hard palate cleft width (b1). If the ratio a/(30 2 b1) is higher than 0.7 to 0.8, the patient is more likely to develop VPI. b1 and a values >9 mm and >16 mm, respectively, may also be used. In the future, we will add measurements of the thickness of the soft palate muscles and the width of the posterior pharyngeal wall. We may be able to develop 568 algorithms including the palatal area, the cleft, and the nasopharyngeal surface, although they may not be as easy to use on site by the surgeon. The prediction of a suboptimal result may eventually warrant a change in the surgical technique. Similar studies must be conducted with other surgical techniques such as the Furlow procedure. CONCLUSION The conclusions of this exploratory retrospective study must be limited to CP patients undergoing palatoplasty by Von Langenbeck technique for isolated CP and Bardach two-flap palatoplasty for CLP, between the ages of 9 and 13 months. The speech evaluations were conducted in a French-speaking cleft population without other malformations. Our results suggest that: Based on all three VPI testing modalities, the best predicting parameter was the ratio a/(30 2 b1), in which a is defined as the distance between the posterior end of the soft palate and the posterior pharyngeal wall and b1 is defined as the width of the cleft at the hard palate level. An a/(30 2 b1) ratio >0.7 to 0.8 is associated with a higher risk of developing VPI (relative risk , sensitivity 5 72% 81%, P <.03). Based on two VPI testing modalities, the best anatomical parameter was a width of the cleft at the hard palate level >9 mm (relative risk , sensitivity 5 67% 73%, P <.04). A posterior gap between the soft palate and the posterior pharyngeal wall >16 mm (relative risk 5 2.0, sensitivity 5 70%) reached statistical significance (P 5.01) with only one VPI evaluation modality (PSE). Acknowledgments Members of the Committee of Orofacial Malformations of Quebec University Hospital Center would like to acknowledge the work of Dr. Georges Demers, who has been an inspiration for all of us throughout his career. BIBLIOGRAPHY 1. McWilliams BJ, Musgrave RH. Diagnosis of speech problems in patients with cleft palate. Br J Disord Commun 1971;6: Rudnick EF, Sie KC. Velopharyngeal insufficiency: current concepts in diagnosis and management. Curr Opin Otolaryngol Head Neck Surg 2008;16: Nyberg J, Westberg LR, Neovius E, Larson O, Henningsson G. Speech results after one-stage palatoplasty with or without muscle reconstruction for isolated cleft palate. Cleft Palate Craniofacial J 2009;47: Henningsson G, Kuehn DP, Sell D, Sweeney T, Trost-Cardamone JE, Whitehill TL. Universal parameters for reporting speech outcomes in individuals with cleft palate. Cleft Palate Craniofac J 2008;45: Sell D, Harding A, Grunwell P. GOS.SP.ASS. 98: an assessment for speech disorders associated with cleft palate and/or velopharyngeal dysfunction (revised). Int J Lang Commun Disord 1999;34: Lohmander A, Olsson M. Methodology for perceptual assessment of speech in patients with cleft palate: a critical review of the literature. Cleft Palate Craniofac J 2004;41: Marrinan EM, LaBrie RA, Mulliken JB. Velopharyngeal function in nonsyndromic cleft palate: relevance of surgical technique, age at repair, and cleft type. Cleft Palate Craniofac J 1998;35: Nyberg J, Westberg LR, Neovius E, Larson O, Henningsson G. Speech results after one-stage palatoplasty with or without muscle reconstruction for isolated cleft palate. Cleft Palate Craniofac J 2010;47: Phua YS, de Chalain T. Incidence of oronasal fistulae and velopharyngeal insufficiency after cleft palate repair: an audit of 211 children born between 1990 and Cleft Palate Craniofac J 2008;45: Van Lierde KM, Monstrey S, Bonte K, Van Cauwenberge P, Vinck B. The long-term speech outcome in Flemish young adults after two different types of palatoplasty. Int J Pediatr Otorhinolaryngol 2004;68: Inman DS, Thomas P, Hodgkinson PD, Reid CA. Oro-nasal fistula development and velopharyngeal insufficiency following primary cleft palate

9 surgery an audit of 148 children born between 1985 and Br J Plast Surg 2005;58: de Buys Roessingh AS, Cherpillod J, Trichet-Zbinden C, Hohlfeld J. Speech outcome after cranial-based pharyngeal flap in children born with total cleft, cleft palate, or primary velopharyngeal insufficiency. J Oral Maxillofac Surg 2006;64: Lam DJ, Chiu LL, Sie KC, Perkins JA. Impact of cleft width in clefts of secondary palate on the risk of velopharyngeal insufficiency. Arch Facial Plast Surg, 2012 Apr 16 Epub. 14. Randall P, LaRossa D, McWilliams BJ, Cohen M, Solot C, Jawad AF. Palatal length in cleft palate as a predictor of speech outcome. Plast Reconstr Surg 2000;106: ; discussion Paliobeli V, Psifidis A, Anagnostopoulos D. Hearing and speech assessment of cleft palate patients after palatal closure. Long-term results. Int J Pediatr Otorhinolaryngol 2005;69: Kummer, AW, Briggs M, Lee L. The relationship between the characteristics of speech and velopharyngeal gap size. Cleft Palate Craniofac J 2003;40: Lam DJ, Starr JR, Perkins JA, et al. A comparison of nasendoscopy and multiview videofluoroscopy in assessing velopharyngeal insufficiency. Otolaryngol Head Neck Surg 2006;134: Kummer AW. Perceptual assessment of resonance and velopharyngeal function. Semin Speech Lang 2011;32: