The Laryngoscope C 2014 The American Laryngological, V
Rhinological and Otological Society, Inc.
Primary Soft Tissue Reconstruction After Transoral Laryngeal Tumor Resection Peak Woo, MD, FACS Objectives/Hypothesis: Primary reconstruction by endoscopic buccal mucosal grafting can be considered in cancer patients with large defects and in patients with bilateral disease. We present our case series of primary endoscopic buccal grafting after cancer surgery. Study Design: Seventeen cases in 16 patients treated with buccal primary reconstruction (PR) are compared to 26 subjects without reconstruction (SO) and to 18 radiation patients (RO). The Voice Handicap Index, voice grade, and acoustic analysis were used to compare voice outcomes between groups. Methods: Case–control study. Results: There were more European Laryngological Association type II, III, and type Va resections in the PR group. Acoustic analysis, Voice Handicap Index short form, and perceptual ratings by three expert listeners showed no differences in voice outcomes between groups. There were voice-related differences between surgical patients, with the European Laryngological Association type I patients having better voices than European Laryngological Association type II, III, and Va groups combined. Voice outcomes showed mild to moderate dysphonia in all three groups. The presence of the buccal graft did not hinder tumor surveillance. Conclusions: Primary reconstruction appears to be safe in patients with large soft tissue defects. There were no detectable functional differences between the PR, SO, and RO groups. Key Words: Laryngeal cancer, voice outcome, buccal graft, laryngeal reconstruction, endoscopic surgery. Level of Evidence: 3. Laryngoscope, 00:000–000, 2014
INTRODUCTION Treatment for early stage glottic carcinoma (T1a, T1b, or T2) may include endoscopic surgery, radiation, or open surgery. With cures equivocal between treatment types,1 treatment often comes down to patient preference and regional variations in treatment patterns. Voice outcome after different treatments remains controversial. Voice after treatment often shows residual dysphonia.2 Late reconstruction after endoscopic surgery includes type I medialization laryngoplasty, injection laryngoplasty, and buccal graft reconstruction.3 These approaches are usually performed late. With cancers that involve the anterior commissure and larger glottis cancers, the problem of web and poor voice becomes especially acute. Resection of bilateral vocal folds and muscle volume can result in poor vocal function.4 Although larger defects can be reconstructed immediately during open surgical reconstruction after
From the Department of Otolaryngology, Head and Neck Surgery, Icahn School of Medicine, New York, New York, U.S.A. Editor’s Note: This Manuscript was accepted for publication October 12, 2014. Presented at the American Laryngological Association Annual Meeting, Las Vegas, Nevada, U.S.A., May 14–15, 2014. The author has no funding, financial relationships, or conflicts of interest to disclose. Send correspondence to Dr. Peak Woo, 300 Central Park West, 1-H, New York, NY 10024. E-mail: [email protected] DOI: 10.1002/lary.25019
Laryngoscope 00: Month 2014
partial laryngectomy,5 immediate reconstruction after endoscopic surgery is not popular. Buccal graft reconstruction is a technique that has been used for treatment of laryngeal stenosis.6 More recent reports have documented its ability to manage anterior glottic webs using an endoscopic technique.7–9 The technical aspects of this procedure have been published.10 Such an approach may be viable as primary reconstruction after cancer ablation. By primary reconstruction, we hope to improve voice function and reduce late reconstruction. This article reports on 17 primary reconstructions in 16 patients using the buccal graft. We compared the voice outcomes of this group to similar groups of patients treated by transoral surgery alone or by radiation. The article compares the voice outcome by type of resection, need for repeat surgery, and oncological outcome. We postulate that immediate reconstruction is safe in selected patients with laryngeal neoplasms.
MATERIALS AND METHODS From 1998 through 2013, 16 patients underwent resection of laryngeal neoplasm followed by buccal graft reconstruction. A total of 17 buccal grafts were placed. The indications for primary reconstruction were based on the resection defect and the site of tumor involvement. The indications for resection were T1a, T1b, or T2 cancer that required resection using European Laryngological Association (ELR) classification type II or greater, bilateral resection of carcinoma, and T1b cancer that required ELR type Va resection.11 We did not perform reconstruction in patients
Woo: Tissue Reconstruction
1
TABLE I. Summary of 16 Patients Treated by Primary Reconstruction. Case
Sex
Age, yr
ELR Type
FU, mo
Stage
Additional RX
Other Treatment
1
Male
85
Va
15
T1b
No
2 3
Female Male
40 55
Va II
41 84
Tcis bil T1a
Yes Yes
4
Male
48
II
43
T1a
No
5 6
Male Male
79 81
Va Va
36 144
T1 bil Tcis bil
No Yes
7
Male
63
III
12
T1a
No
8 9
Female Male
83 80
III II
72 84
T1a T1a
Yes No
10
Male
54
Va
84
Benign
No
11 12
Male Male
78 74
I Va
108 188
T1a T1 bil
Yes Yes
Biopsy Excision
13
Female
89
Va
154
T1b
Yes
Biopsy
14 15
Male Male
79 84
II Va
96 153
T1a T1 bil
No Yes
Excision
16
Male
44
II
69
Benign
No
Comment
Negotiator Cut web Biopsy
Speech pathologist Musician Chief executive officer
Biopsy
Psychiatrist
Excision Executive
Classical singer Priest
ELR 5 European Laryngological Association; FU 5 follow-up.
who were at high risk for general anesthesia, patients with poor margin control, or patients with inability for follow-up. The case series is listed in Table I. The technical aspects of the two-staged procedure have been published.10 For both unilateral and bilateral disease, only one side is treated with buccal graft reconstruction. Figure 1 shows the prepared graft and the Silastic splint with Prolene suture in place ready for placement into the donor site. Figure 2 shows the result in a patient with a right vocal fold buccal graft in place at 3 years after surgery.
Data Collection and Analysis Data were collected on the clinical status and pathology and whether additional procedures were done after the initial resection and reconstruction. For each patient available for follow-up (n 5 12 cases), we recorded their voice and collected the Voice Handicap Index short form (VHI-10). This group was categorized as the surgical group with primary reconstruction. This is the primary reconstruction (PR) group. For comparison, we did the same examination and voice measures for a group of patients with early glottic carcinoma treated with endoscopic surgery resection only (SO, n 5 26). A third group of patients with early glottic carcinoma treated with radiation therapy alone (RO, n 5 18) was also studied. Complete data are available for voice analysis for 56 subjects (PR 5 12, SO 5 26, RO 5 18). The group was also divided into the surgical treatment arm versus the radiation arm for further analysis. Thus, the two groupings for analysis for this series were: 1) surgery with primary grafting versus surgery alone versus radiation therapy and 2) surgery group (surgery with primary grafting and surgery alone) versus radiation therapy. Tabulation from each patient was made regarding demographics, follow-up, tumor staging, the type of resection, and whether additional surgery was necessary. Data were entered into a spreadsheet, and statistics were compiled (SPSS version 20; IBM, Armonk, NY).
Laryngoscope 00: Month 2014
2
Analysis of voice was done using three different measures. These measures were: patients’ perception of voice (VHI-10), perceptual evaluation of voice by expert listeners using the Grade, Roughness, Breathiness, Asthenia, Strain (GRBAS) scale, and computer analysis of voice using Cepstral peak prominence analysis. Each patient was asked to fill out the VHI-10 at the time of the last follow-up. The patient’s voice was recorded in a standard manner using Computerized Speech Laboratory (CSL) hardware (Analysis of Dysphonia in Speech and Voice [ADSV], model 5109; KayPENTAX, Montvale, NJ). The speech token “How hard did he hit him?” from the CAPE V sentence was recorded in 10 seconds. The patient repeated this token four to five times at the most comfortable pitch and loudness. Analysis of the token was done by the CSL firmware and ADSV software (KayPENTAX).
Fig. 1. Composite graft mounted on a Silastic splint ready for insertion. The buccal graft is secured on the reinforced Silastic splint. The 2-0 Prolene suture is placed through both the graft and the Silastic splint. The Prolene suture will be used to hold the graft and Silastic splint combination against the recipient site. [Color figure can be viewed in the online issue, which is available at www.laryngoscope.com.]
Woo: Tissue Reconstruction
Fig. 2. Stroboscopy video print showing a patient with buccal graft in place on the patient’s right fold at 3 years after surgery.
The software uses Cepstral peak analysis of the running speech and calculates the Cepstral/Spectral Index of Dysphonia (CSID) score.12 The CSID score from each subject was entered into the database as a measure of objective parameter of voice. Perception of voice was evaluated by three expert listeners. They were two PhD speech–language pathologists and a laryngologist. The PhD raters were specialists in voice disorders each with >20 years of experience. The laryngologist was a laryngology fellowship-trained physician. All were blinded as to the treatment arm and outcome data. Each listener was asked to listen to the voice token “How hard did he hit him?” recorded previously for acoustic analysis. The expert listeners were asked to rate the voice on the GRBAS scale. From the GRBAS scale, the total value for the overall grade from the three listeners was tabulated as the severity grade for each patient. Thus, if a patient had a grade score from the three GRBAS raters of G2, G2, and G1, the patient would have a summed dysphonia grade score of 5. The voice measures were analyzed against the treatment code of surgery and reconstruction versus surgery alone versus radiation, and against surgery versus radiation. Each of the voice parameters of overall grade, CSID score, and VHI-10 was analyzed against the code of primary reconstruction and surgery versus surgery alone versus radiation (PR vs. SO vs. RO). We also performed the same analysis with the surgery group versus the radiation group ([PR 1 SO] vs. RO). Subset analysis was carried out in surgery patients by analyzing the voice results against the ELR types of resection. Voice results were also analyzed for the T1a patients against the others (T1b, T1bil, and T2). This was done to see whether differences in voice could be predicted based on staging and resection type.
RESULTS There were 61 subjects in the study group. There were 17 buccal graft reconstructions in 16 patients. Twenty-six patients had surgery only. Eighteen patients had radiation therapy. Six patients in the radiation group (6 of 18) had an initial endoscopic surgery attempt but were referred for radiation therapy due to margin control or due to a second primary tumor. No patients in the PR group had radiation. There were no differences between cancer staging and the type of treatment (v2 analysis). Laryngoscope 00: Month 2014
The mean follow-up for the PR group was 86 months (standard deviation [SD], 652 months). The mean follow-up for the SO group was 52 months (SD, 631 months), and the mean follow-up for the RO group was 58 months (SD, 635 months). Table I lists the 16 PR patients and their case characteristics. One patient had a second primary tumor in the contralateral vocal fold 14 years after the initial buccal graft. He was treated by a second resection with buccal graft in 2012. He was free of disease as of March 2014. He was the only patient who had undergone a second buccal graft on the contralateral side. The indications for reconstruction were: 1) patients with greater than ELR type I resections (n 5 6) or ELR type Va (n 5 5) resections, and 2) patients with bilateral diseases that involved the anterior commissure (n 5 4). Eight of the 16 patients were voice professionals. The graft was placed to bulk up the resected vocal fold to prevent glottic incompetence. A second indication was to prevent anterior glottic web formation in patients with bilateral lesions or after T1b resections. We were unable to have standard voice recording in four of the subjects. Two of the cancer patients died. One died of prostate cancer and one from cardiac disease. The other 12 patients were still under the care of the author at the time of writing. No patients in this group had gone on to radiation therapy or open surgery. Eight of the 16 patients were voice professionals who needed voice as part of their work. All continued to perform their work. This included one classical singer and one professional musician. The type of resection done in the PR group showed a higher rate of ELR type II, type III, and type Va resections. There was only one patient with ELR type I resection. There were no graft failures. Endoscopy after surgery showed the graft to be an excellent, pliable soft tissue substitute for the vocal fold. A video photograph during phonation in of one of our patients during phonation is shown in Figure 2. The graft served to achieve glottic closure during phonation. The graft appeared to retain much of the pliability of cheek mucosa. Video stroboscopy examination showed good oscillatory property. The amplitude was slightly reduced compared to a normal vocal fold. The buccal graft could oscillate in synchrony with the opposite vocal fold, providing airflow modulation. In some cases, it even had a mucosal wave that could propagate along the surface of the buccal graft. In patients with bilateral resections where there was no pliable vocal fold mucosa, the buccal graft could serve as the primary oscillator. None of the 61 patients had died of laryngeal cancer or had gone on to laryngectomy. Secondary lesions noted on follow-up that required biopsy and excision were high. In the buccal graft group, 8 of 16 (50%) had additional biopsy or excision. Three of the 14 patients had neoplasms or severe dysplasia that went on to excision. All were able to be picked up on endoscopy. In the SO group, 16 of 28 needed additional surgery. This included three patients who underwent late reconstruction for glottic incompetence by type I thyroplasty. In the RO group, six of the 18 patients went on to additional Woo: Tissue Reconstruction
3
TABLE II. Summary Statistics of Age, Follow-up, Sum Total of GRBAS Score, Sum of Grade, CSID Score, and VHI-10 Score. Descriptive Statistics
N
Minimum
Maximum
Mean
SD
Age, yr
61
36
91
71.95
13.666
FU, yr Sum all GRBAS
61 56
9 8
188 31
62.89 18.46
41.049 6.257
Sum grade
56
3
9
5.18
1.722
CSID score VHI-10
56 56
3 0
88 32
35.59 12.16
17.255 8.166
CSID 5 Cepstral/Spectral Index of Dysphonia; FU 5 follow-up; GRBAS 5 Grade, Roughness, Breathiness, Asthenia, Strain; VHI-10 5 Voice Handicap Index short form.
treatment. There were no differences between the treatment groups in the age, sex, cancer staging, or professional voice user status. Although there were more patients in the reconstruction group with higher ELR resection type, significance was not reached.
Voice Results Table II summarizes the voice data for all three groups. There was a wide range of voice outcomes in all voice measures. The overall dysphonia grade was between mild and moderate. VHI-10 showed a mean value of 12, indicating mild disability. The CSID score was elevated at 35. This was consistent with mild to moderate aperiodicity, normal being up to 20. The median value of grade of dysphonia was 5, indicating mild to moderate dysphonia. In comparing the voice results between the PR, SO, and RO groups, there were no significant differences in any of the measures. Analysis showed no differences between groups in all parameters of patient perception, CSID score, and in grading scored by the expert listeners (analysis of variance [ANOVA]). There were no differences in the voice outcome between the surgery group and the radiation group. This was true for all three parameters of VHI-10, overall grade, and CSID score. There was a difference in voice results when comparing grade of dysphonia versus cancer stage and the type of resection. Patients with T1a cancer had better dysphonia grades than the combined T1b and T2 group (P
Rhinological and Otological Society, Inc.
Primary Soft Tissue Reconstruction After Transoral Laryngeal Tumor Resection Peak Woo, MD, FACS Objectives/Hypothesis: Primary reconstruction by endoscopic buccal mucosal grafting can be considered in cancer patients with large defects and in patients with bilateral disease. We present our case series of primary endoscopic buccal grafting after cancer surgery. Study Design: Seventeen cases in 16 patients treated with buccal primary reconstruction (PR) are compared to 26 subjects without reconstruction (SO) and to 18 radiation patients (RO). The Voice Handicap Index, voice grade, and acoustic analysis were used to compare voice outcomes between groups. Methods: Case–control study. Results: There were more European Laryngological Association type II, III, and type Va resections in the PR group. Acoustic analysis, Voice Handicap Index short form, and perceptual ratings by three expert listeners showed no differences in voice outcomes between groups. There were voice-related differences between surgical patients, with the European Laryngological Association type I patients having better voices than European Laryngological Association type II, III, and Va groups combined. Voice outcomes showed mild to moderate dysphonia in all three groups. The presence of the buccal graft did not hinder tumor surveillance. Conclusions: Primary reconstruction appears to be safe in patients with large soft tissue defects. There were no detectable functional differences between the PR, SO, and RO groups. Key Words: Laryngeal cancer, voice outcome, buccal graft, laryngeal reconstruction, endoscopic surgery. Level of Evidence: 3. Laryngoscope, 00:000–000, 2014
INTRODUCTION Treatment for early stage glottic carcinoma (T1a, T1b, or T2) may include endoscopic surgery, radiation, or open surgery. With cures equivocal between treatment types,1 treatment often comes down to patient preference and regional variations in treatment patterns. Voice outcome after different treatments remains controversial. Voice after treatment often shows residual dysphonia.2 Late reconstruction after endoscopic surgery includes type I medialization laryngoplasty, injection laryngoplasty, and buccal graft reconstruction.3 These approaches are usually performed late. With cancers that involve the anterior commissure and larger glottis cancers, the problem of web and poor voice becomes especially acute. Resection of bilateral vocal folds and muscle volume can result in poor vocal function.4 Although larger defects can be reconstructed immediately during open surgical reconstruction after
From the Department of Otolaryngology, Head and Neck Surgery, Icahn School of Medicine, New York, New York, U.S.A. Editor’s Note: This Manuscript was accepted for publication October 12, 2014. Presented at the American Laryngological Association Annual Meeting, Las Vegas, Nevada, U.S.A., May 14–15, 2014. The author has no funding, financial relationships, or conflicts of interest to disclose. Send correspondence to Dr. Peak Woo, 300 Central Park West, 1-H, New York, NY 10024. E-mail: [email protected] DOI: 10.1002/lary.25019
Laryngoscope 00: Month 2014
partial laryngectomy,5 immediate reconstruction after endoscopic surgery is not popular. Buccal graft reconstruction is a technique that has been used for treatment of laryngeal stenosis.6 More recent reports have documented its ability to manage anterior glottic webs using an endoscopic technique.7–9 The technical aspects of this procedure have been published.10 Such an approach may be viable as primary reconstruction after cancer ablation. By primary reconstruction, we hope to improve voice function and reduce late reconstruction. This article reports on 17 primary reconstructions in 16 patients using the buccal graft. We compared the voice outcomes of this group to similar groups of patients treated by transoral surgery alone or by radiation. The article compares the voice outcome by type of resection, need for repeat surgery, and oncological outcome. We postulate that immediate reconstruction is safe in selected patients with laryngeal neoplasms.
MATERIALS AND METHODS From 1998 through 2013, 16 patients underwent resection of laryngeal neoplasm followed by buccal graft reconstruction. A total of 17 buccal grafts were placed. The indications for primary reconstruction were based on the resection defect and the site of tumor involvement. The indications for resection were T1a, T1b, or T2 cancer that required resection using European Laryngological Association (ELR) classification type II or greater, bilateral resection of carcinoma, and T1b cancer that required ELR type Va resection.11 We did not perform reconstruction in patients
Woo: Tissue Reconstruction
1
TABLE I. Summary of 16 Patients Treated by Primary Reconstruction. Case
Sex
Age, yr
ELR Type
FU, mo
Stage
Additional RX
Other Treatment
1
Male
85
Va
15
T1b
No
2 3
Female Male
40 55
Va II
41 84
Tcis bil T1a
Yes Yes
4
Male
48
II
43
T1a
No
5 6
Male Male
79 81
Va Va
36 144
T1 bil Tcis bil
No Yes
7
Male
63
III
12
T1a
No
8 9
Female Male
83 80
III II
72 84
T1a T1a
Yes No
10
Male
54
Va
84
Benign
No
11 12
Male Male
78 74
I Va
108 188
T1a T1 bil
Yes Yes
Biopsy Excision
13
Female
89
Va
154
T1b
Yes
Biopsy
14 15
Male Male
79 84
II Va
96 153
T1a T1 bil
No Yes
Excision
16
Male
44
II
69
Benign
No
Comment
Negotiator Cut web Biopsy
Speech pathologist Musician Chief executive officer
Biopsy
Psychiatrist
Excision Executive
Classical singer Priest
ELR 5 European Laryngological Association; FU 5 follow-up.
who were at high risk for general anesthesia, patients with poor margin control, or patients with inability for follow-up. The case series is listed in Table I. The technical aspects of the two-staged procedure have been published.10 For both unilateral and bilateral disease, only one side is treated with buccal graft reconstruction. Figure 1 shows the prepared graft and the Silastic splint with Prolene suture in place ready for placement into the donor site. Figure 2 shows the result in a patient with a right vocal fold buccal graft in place at 3 years after surgery.
Data Collection and Analysis Data were collected on the clinical status and pathology and whether additional procedures were done after the initial resection and reconstruction. For each patient available for follow-up (n 5 12 cases), we recorded their voice and collected the Voice Handicap Index short form (VHI-10). This group was categorized as the surgical group with primary reconstruction. This is the primary reconstruction (PR) group. For comparison, we did the same examination and voice measures for a group of patients with early glottic carcinoma treated with endoscopic surgery resection only (SO, n 5 26). A third group of patients with early glottic carcinoma treated with radiation therapy alone (RO, n 5 18) was also studied. Complete data are available for voice analysis for 56 subjects (PR 5 12, SO 5 26, RO 5 18). The group was also divided into the surgical treatment arm versus the radiation arm for further analysis. Thus, the two groupings for analysis for this series were: 1) surgery with primary grafting versus surgery alone versus radiation therapy and 2) surgery group (surgery with primary grafting and surgery alone) versus radiation therapy. Tabulation from each patient was made regarding demographics, follow-up, tumor staging, the type of resection, and whether additional surgery was necessary. Data were entered into a spreadsheet, and statistics were compiled (SPSS version 20; IBM, Armonk, NY).
Laryngoscope 00: Month 2014
2
Analysis of voice was done using three different measures. These measures were: patients’ perception of voice (VHI-10), perceptual evaluation of voice by expert listeners using the Grade, Roughness, Breathiness, Asthenia, Strain (GRBAS) scale, and computer analysis of voice using Cepstral peak prominence analysis. Each patient was asked to fill out the VHI-10 at the time of the last follow-up. The patient’s voice was recorded in a standard manner using Computerized Speech Laboratory (CSL) hardware (Analysis of Dysphonia in Speech and Voice [ADSV], model 5109; KayPENTAX, Montvale, NJ). The speech token “How hard did he hit him?” from the CAPE V sentence was recorded in 10 seconds. The patient repeated this token four to five times at the most comfortable pitch and loudness. Analysis of the token was done by the CSL firmware and ADSV software (KayPENTAX).
Fig. 1. Composite graft mounted on a Silastic splint ready for insertion. The buccal graft is secured on the reinforced Silastic splint. The 2-0 Prolene suture is placed through both the graft and the Silastic splint. The Prolene suture will be used to hold the graft and Silastic splint combination against the recipient site. [Color figure can be viewed in the online issue, which is available at www.laryngoscope.com.]
Woo: Tissue Reconstruction
Fig. 2. Stroboscopy video print showing a patient with buccal graft in place on the patient’s right fold at 3 years after surgery.
The software uses Cepstral peak analysis of the running speech and calculates the Cepstral/Spectral Index of Dysphonia (CSID) score.12 The CSID score from each subject was entered into the database as a measure of objective parameter of voice. Perception of voice was evaluated by three expert listeners. They were two PhD speech–language pathologists and a laryngologist. The PhD raters were specialists in voice disorders each with >20 years of experience. The laryngologist was a laryngology fellowship-trained physician. All were blinded as to the treatment arm and outcome data. Each listener was asked to listen to the voice token “How hard did he hit him?” recorded previously for acoustic analysis. The expert listeners were asked to rate the voice on the GRBAS scale. From the GRBAS scale, the total value for the overall grade from the three listeners was tabulated as the severity grade for each patient. Thus, if a patient had a grade score from the three GRBAS raters of G2, G2, and G1, the patient would have a summed dysphonia grade score of 5. The voice measures were analyzed against the treatment code of surgery and reconstruction versus surgery alone versus radiation, and against surgery versus radiation. Each of the voice parameters of overall grade, CSID score, and VHI-10 was analyzed against the code of primary reconstruction and surgery versus surgery alone versus radiation (PR vs. SO vs. RO). We also performed the same analysis with the surgery group versus the radiation group ([PR 1 SO] vs. RO). Subset analysis was carried out in surgery patients by analyzing the voice results against the ELR types of resection. Voice results were also analyzed for the T1a patients against the others (T1b, T1bil, and T2). This was done to see whether differences in voice could be predicted based on staging and resection type.
RESULTS There were 61 subjects in the study group. There were 17 buccal graft reconstructions in 16 patients. Twenty-six patients had surgery only. Eighteen patients had radiation therapy. Six patients in the radiation group (6 of 18) had an initial endoscopic surgery attempt but were referred for radiation therapy due to margin control or due to a second primary tumor. No patients in the PR group had radiation. There were no differences between cancer staging and the type of treatment (v2 analysis). Laryngoscope 00: Month 2014
The mean follow-up for the PR group was 86 months (standard deviation [SD], 652 months). The mean follow-up for the SO group was 52 months (SD, 631 months), and the mean follow-up for the RO group was 58 months (SD, 635 months). Table I lists the 16 PR patients and their case characteristics. One patient had a second primary tumor in the contralateral vocal fold 14 years after the initial buccal graft. He was treated by a second resection with buccal graft in 2012. He was free of disease as of March 2014. He was the only patient who had undergone a second buccal graft on the contralateral side. The indications for reconstruction were: 1) patients with greater than ELR type I resections (n 5 6) or ELR type Va (n 5 5) resections, and 2) patients with bilateral diseases that involved the anterior commissure (n 5 4). Eight of the 16 patients were voice professionals. The graft was placed to bulk up the resected vocal fold to prevent glottic incompetence. A second indication was to prevent anterior glottic web formation in patients with bilateral lesions or after T1b resections. We were unable to have standard voice recording in four of the subjects. Two of the cancer patients died. One died of prostate cancer and one from cardiac disease. The other 12 patients were still under the care of the author at the time of writing. No patients in this group had gone on to radiation therapy or open surgery. Eight of the 16 patients were voice professionals who needed voice as part of their work. All continued to perform their work. This included one classical singer and one professional musician. The type of resection done in the PR group showed a higher rate of ELR type II, type III, and type Va resections. There was only one patient with ELR type I resection. There were no graft failures. Endoscopy after surgery showed the graft to be an excellent, pliable soft tissue substitute for the vocal fold. A video photograph during phonation in of one of our patients during phonation is shown in Figure 2. The graft served to achieve glottic closure during phonation. The graft appeared to retain much of the pliability of cheek mucosa. Video stroboscopy examination showed good oscillatory property. The amplitude was slightly reduced compared to a normal vocal fold. The buccal graft could oscillate in synchrony with the opposite vocal fold, providing airflow modulation. In some cases, it even had a mucosal wave that could propagate along the surface of the buccal graft. In patients with bilateral resections where there was no pliable vocal fold mucosa, the buccal graft could serve as the primary oscillator. None of the 61 patients had died of laryngeal cancer or had gone on to laryngectomy. Secondary lesions noted on follow-up that required biopsy and excision were high. In the buccal graft group, 8 of 16 (50%) had additional biopsy or excision. Three of the 14 patients had neoplasms or severe dysplasia that went on to excision. All were able to be picked up on endoscopy. In the SO group, 16 of 28 needed additional surgery. This included three patients who underwent late reconstruction for glottic incompetence by type I thyroplasty. In the RO group, six of the 18 patients went on to additional Woo: Tissue Reconstruction
3
TABLE II. Summary Statistics of Age, Follow-up, Sum Total of GRBAS Score, Sum of Grade, CSID Score, and VHI-10 Score. Descriptive Statistics
N
Minimum
Maximum
Mean
SD
Age, yr
61
36
91
71.95
13.666
FU, yr Sum all GRBAS
61 56
9 8
188 31
62.89 18.46
41.049 6.257
Sum grade
56
3
9
5.18
1.722
CSID score VHI-10
56 56
3 0
88 32
35.59 12.16
17.255 8.166
CSID 5 Cepstral/Spectral Index of Dysphonia; FU 5 follow-up; GRBAS 5 Grade, Roughness, Breathiness, Asthenia, Strain; VHI-10 5 Voice Handicap Index short form.
treatment. There were no differences between the treatment groups in the age, sex, cancer staging, or professional voice user status. Although there were more patients in the reconstruction group with higher ELR resection type, significance was not reached.
Voice Results Table II summarizes the voice data for all three groups. There was a wide range of voice outcomes in all voice measures. The overall dysphonia grade was between mild and moderate. VHI-10 showed a mean value of 12, indicating mild disability. The CSID score was elevated at 35. This was consistent with mild to moderate aperiodicity, normal being up to 20. The median value of grade of dysphonia was 5, indicating mild to moderate dysphonia. In comparing the voice results between the PR, SO, and RO groups, there were no significant differences in any of the measures. Analysis showed no differences between groups in all parameters of patient perception, CSID score, and in grading scored by the expert listeners (analysis of variance [ANOVA]). There were no differences in the voice outcome between the surgery group and the radiation group. This was true for all three parameters of VHI-10, overall grade, and CSID score. There was a difference in voice results when comparing grade of dysphonia versus cancer stage and the type of resection. Patients with T1a cancer had better dysphonia grades than the combined T1b and T2 group (P
