Impact of resection depth on voice outcomes following transoral laser microsurgery for early glottic lesions

Introduction

Laryngeal squamous cell carcinoma (SCC) accounts for approximately 0.7% of all malignancies and is the most common non-cutaneous malignant tumour of the head and neck region.^1,2 Strongly associated with tobacco and alcohol use, it shows a clear male predominance and most frequently affects the vocal folds, accounting for about 75% of cases.³ As glottic tumours typically compromise voice quality early in the disease course, they are often detected at an early stage.⁴

A range of treatment options is used for early glottic SCC, from open partial laryngectomy to more conservative approaches such as transoral cold steel resection, radiofrequency ablation, transoral laser excision, and radiation therapy.⁵ Oncological outcomes across these modalities are favourable and largely comparable, yet consensus regarding optimal voice preservation remains unclear.^5,6,7 Among the different treatment modalities, transoral laser microsurgery (TLM) has emerged as a preferred strategy for early-stage glottic carcinoma, when anatomically feasible, due to its favourable profile in terms of cost-effectiveness, shorter hospitalisation, lower complication rates and effective organ preservation.⁸ TLM, as classified by the European Laryngological Society (ELS), encompasses a range of cordectomy types tailored to the depth and extent of tumour invasion. Resections range from superficial epithelial removal to extended excisions involving the anterior commissure and both vocal folds.⁹

However, while most studies focus primarily on oncological control following TLM, relatively few have systematically addressed functional outcomes, particularly voice quality, which remains a crucial determinant of postoperative quality of life, especially for professionally voice-dependent individuals. Among published studies on functional outcomes, most compare voice outcomes following more conservative (type I–II or I–III) versus more radical (type III–VI or IV–VI) cordectomies.^10,11,12,13

Our study aimed to assess and compare voice quality using validated objective and subjective measures in patients undergoing the most conservative laser cordectomy types classified as type I, versus complete removal of the vocal ligament and/ or part of the vocal muscle (type II/III cordectomy).

Patients and methods

Results

A total of 65 patients were enrolled in the study. Follow-up voice analysis was not obtained in 36 patients, who did not meet further inclusion criteria for the following reasons: 17 had low-grade dysplasia, 10 were treated with radiotherapy, 3 required type IV or higher cordectomy, one underwent repeated laser cordectomies due to residual disease, and 5 declined further participation. Of the 29 patients who completed the study, type I and type II/III cordectomy were performed in 13 and 16 cases, respectively. In Group 1, we found that 1 patient underwent a full-length-vocal-fold mucosal resection, while all others underwent laser removal of the lesion and surrounding mucosa. Patient enrolment, exclusions, and final study-group allocation is shown in Figure 1.

FIGURE 1.

Both groups were comparable in demographic and clinical characteristics, as presented in Table 1.

Subjective voice assessment results

Patient-reported outcomes are summarised in Table 2. According to the VAS scale, Group 2 patients reported the poorest subjective voice quality at 1 month postoperatively (mean 5.7), whereas ratings in the Group 1 were lower and did not show a significant decline (mean 3.9). By 12 months, both groups had improved substantially, with scores converging to 3.0 in the Group 1 and 3.6 in the Group 2, with no significant difference between the groups (Figure 2).

TABLE 2.

Patient assessment of voice quality and satisfaction; by study group and time of measurement

Parameter / Patient group	Before treatment	1 month after treatment	3 months after treatment	12 months after treatment	P-value 1
Voice quality (VAS)
Cordectomy Type I	4.8 (0.5)	3.9 (0.5)	3.5 (0.5)	3.0 (0.4)	0.040
Cordectomy Type II/III	5.6 (0.3)	5.7 (0.6)	4.3 (0.6)	3.6 (0.5)	0.031
P-value 2	0.342	0.032	0.279	0.367	P-value3 = 0.635
Voice handicap index score (VHI)
Cordectomy Type I	26.9 (7.9)	19.5 (4.9)	15.7 (5.2)	9.2 (3.6)	0.041
Cordectomy Type II/III	31.6 (6.0)	41.0 (6.8)	24.4 (6.3)	26.6 (7.5)	0.016
P-value 2	0.404	0.039	0.210	0.195	P-value3 = 0.263

1 Comparison between measurement within patient group;

2 Comparison between patient groups within measurement;

3 Interaction between main factors (patient group, measurement); VAS = visual analog scale; VHI = voice handicap index

Data are mean values (standard errors)

FIGURE 2.

Similar trends were observed in the VHI-30 questionnaire (Figure 3). The greatest handicap was perceived at 1 month in the Group 2 (mean score 41), while scores in the Group 1 were lower at the same time point (mean 19.5). At 12 months, VHI-30 values decreased markedly in both groups, reaching 9.2 in the Group 1 and 26.6 in the Group 2, both below the clinical threshold of 30 for moderate handicap. Although overall improvement over time was statistically significant in both groups, changes over time did not differ significantly between them.

FIGURE 3.

Perceptual voice analysis by two independent phoniatricians using the GRBAS scale showed improvement in the overall dysphonia grade from G2 preoperatively to G1 at 12 months postoperatively for both surgical types. Parameters R and B also improved from 2 to 1 in both groups. Parameters A and S remained consistently low (score = 1) throughout the follow-up period in both groups. Inter-rater agreement was satisfactory, with Cohen’s kappa values ranging from 0.30 to 0.97 across all GRBAS components.

Signs of MTD were stable in the Group 1 but increased over time in the Group 2, without reaching statistical significance. Scarring was predominantly mild in both groups, although moderate to severe scarring was more frequent after type II/III cordectomy.

Discussion

TLM is an established treatment modality for early glottic carcinoma, offering excellent oncological outcomes. However, significant variability persists in postoperative voice function, underscoring the need to clarify the factors influencing voice recovery.^5,6,7,8 Assessing voice outcome is a complex, multidimensional process. Studies vary considerably in their methodologies and choice of outcome measures, which range from perceptual and acoustic analyses to patient-reported outcome measures. These are often used independently rather than as part of a comprehensive evaluation. Furthermore, differences in follow-up duration limit comparisons between studies and may obscure the true course of voice recovery.^14,15

Heterogeneity in tumour- and patient-related characteristics also contributes to variability across studies. Tumour extent, depth of resection, involvement of the vocal ligament or muscle, and patient factors such as age, smoking status, baseline voice quality, and professional voice demands all significantly influence postoperative functional outcomes. As a result, findings from mixed cohorts may limit the ability to determine voice outcomes specific to the type of transoral laser cordectomy.^10,11,16

The earliest studies stratified patients by tumour TNM classification.^17,18,19 This approach offered limited insight into functional outcomes, as detailed information on the extent of resection and the volume of excised tissue was often lacking. Consequently, these early reports could not establish a reliable association between the amount of vocal fold tissue removed and postoperative voice quality.

Later studies suggested that postoperative voice parameters vary according to the type of cordectomy, highlighting the importance of documenting the surgical extent of tissue removal.^20,21 Nevertheless, even widely adopted cordectomy classifications may not fully capture the functional complexity of vocal fold surgery, as procedures grouped within the same category can differ substantially in tissue involvement, biomechanical alteration, and their impact on vibratory function. Several studies have reported improvements in voice quality following type I–III laser cordectomies and less favourable voice quality after more extensive resections (types IV–VI).^10,11,12 However, direct comparative results between less extensive cordectomy types (I–III) remain unclear.^22,23

In this context, the present study focused on a direct comparison between type I cordectomy and more extensive type II/III resections. In clinical practice, we observe that in most cases, in procedures that we define as laser cordectomy type I, at least a small segment of preserved mucosa remains on the vocal fold. With earlier lesion detection, removal of the entire vocal fold mucosa is rarely necessary, so the question arises at this point whether it is necessary to consider a more detailed division of type I cordectomy into a subtype with at least partially preserved mucosa and a subtype where the mucosa is removed along the entire length of the vocal fold. Theoretically, the vibratory properties of a vocal fold that is decorticated along its entire length are different than those with partially preserved mucosa.

An exploratory descriptive comparison within Group 1 demonstrated that the single patient undergoing classical full-length type I cordectomy showed substantially greater impairment of acoustic parameters related to vibratory stability, including jitter, shimmer, NHR, and VTI, compared with patients undergoing more limited mucosal resections with partial preservation of the vocal fold mucosa. Although interpretation is inherently limited by the single-patient nature of this comparison, these findings may further support the importance of maximal mucosal preservation for maintenance of postoperative vibratory function. Interestingly, in this individual patient, the marked objective acoustic deviations were not accompanied by similarly pronounced deterioration in subjective voice outcomes, highlighting the potential discrepancy between objective voice analysis and patient-perceived vocal function.

The favourable recovery of the mucosal wave observed in patients from Group 1 may, at least in part, be attributable to preservation of the maculae flavae. Sato et al. described these dense elliptical structures, located at the anterior and posterior ends of the vocal folds, as specialized sites involved in the production and maintenance of the extracellular matrix responsible for the unique elastic and vibratory properties of the vocal fold mucosa. Furthermore, the maculae flavae have been proposed to function as a stem cell microenvironment containing cells involved in vocal fold tissue regeneration. During more extensive cordectomies involving the anterior commissure, vocal process, or deeper layers of the vocal fold, these structures may be partially or completely disrupted, potentially impairing regenerative capacity and promoting fibrotic healing. In contrast, preservation of mucosal regions containing the maculae flavae during limited resections may facilitate more favourable restoration of vibratory function through improved mucosal healing and reduced scar formation. This hypothesis is consistent with the findings of the present study, in which recovery of the mucosal wave was observed predominantly following the most tissue-preserving surgical approaches.^24,25

By specifically contrasting procedures limited to the epithelium and superficial lamina propria with those involving the vocal ligament and/ or underlying muscle fibres, we aimed to clarify whether increasing surgical depth and mucosal area results in clinically meaningful differences in postoperative voice outcomes. Our study addresses a relevant gap in the existing literature, as functional comparisons between less extensive cordectomy types remain limited despite their widespread application in the management of early glottic carcinoma.

A key strength of this study is the standardised use of a multidimensional, ELS-based protocol for voice quality assessment across all participants.²⁶ This comprehensive evaluation included a validated patient-reported outcome measure (VHI-30), perceptual voice assessment using the GRBAS scale, aerodynamic measurements, and detailed acoustic analysis.

While several studies have limited postoperative follow-up to six months, our findings highlight the importance of longer-term monitoring in patients undergoing cordectomies types I/III, as certain voice parameters continue to improve beyond 12 months. Staníková et al. reported persistent elevation of GRBAS parameters (G, R, and B) at six months following cordectomies.¹⁰ In contrast, in our cohort, G, R, and B scores continued to improve up to 12 months after type II/III cordectomy. Similarly, MPT returned to preoperative values at one year. These observations are consistent with reports from other authors who noted delayed functional improvement occurring after 12 months or longer.^27,28

Age may further influence the course of voice recovery. Lechien et al. demonstrated a slower improvement in voice quality among patients older than 60 years.¹¹ The mean age of our cohort was over 60 years, which may explain why statistically significant improvements in several voice parameters were observed only after one year of follow-up.

An important limitation of the present study is that all postoperative voice outcomes were compared with preoperative measurements obtained in the presence of pre-existing vocal fold pathology. Consequently, postoperative improvement should be interpreted as recovery relative to a pathological baseline rather than definitive restoration of normal vocal function. Although comparison with matched healthy controls or normative acoustic data may provide an external reference, voice parameters are highly individual and influenced by age, sex, habitual voice use, smoking status, reflux-related symptoms, and recording conditions. Therefore, follow-up within the same patients over time allows assessment of postoperative recovery trajectories, but does not permit definitive conclusions regarding complete normalization of voice function.

Published normative acoustic data indicate lower jitter, shimmer, and NHR values in healthy individuals than those observed in the present cohort, even at late follow-up. This suggests that postoperative recovery should primarily be interpreted as improvement relative to a pathological baseline rather than complete restoration of normal voice function. Nevertheless, patients undergoing more limited resections generally remained closer to established normative ranges throughout follow-up.^29,30,31,32

An important finding of the present study is that several objective and subjective voice parameters demonstrated partial convergence between the groups at 12 months postoperatively. Although patients undergoing more limited mucosal resections generally exhibited faster functional recovery and more favourable early postoperative voice outcomes, differences between the groups became less pronounced over time for several clinically relevant parameters, including VAS, VHI-30, MPT, jitter, NHR, and the mucosal wave. These findings suggest that, within the conservative resection range evaluated in the present study, resection depth may primarily influence the trajectory and speed of postoperative recovery rather than definitive medium-term voice outcome.

The role of aerodynamic and acoustic measurements after laser cordectomy I–III remains controversial.¹² Despite less favourable early postoperative stroboscopic and acoustic outcomes after type II/III cordectomy compared to type I, patient-reported outcomes – including overall satisfaction assessed by VAS and VHI-30 – were comparable between groups at 12 months and showed significant improvement over time.

Similar to our findings, previous studies have reported a discrepancy between objective acoustic findings and patient-perceived voice quality following laser cordectomy. Long-term follow-up studies demonstrated that patients may report high subjective satisfaction with voice quality despite persistent mild-to-moderate objective dysphonia. The authors proposed successful oncological treatment, preservation of laryngeal function, and adaptation to residual dysphonia as possible explanations for this discrepancy.^33,34

Several studies have also emphasized that objective acoustic measures and patient-perceived voice quality represent distinct and only partially overlapping constructs. While objective voice analysis primarily evaluates vibratory and acoustic characteristics of phonation, subjective assessment additionally reflects patient adaptation, communicative sufficiency, treatment expectations, and overall quality of life. These considerations may partially explain the divergence between persistent objective acoustic abnormalities and favourable patient-reported outcomes observed in the present study.^35,36,37

In addition, most of our patients were retired men without activities requiring high-level vocal performance. Accordingly, preservation of sufficient vocal loudness for functional communication appears to be a relevant outcome in this population.

Some authors recommend at least six months of postoperative voice therapy after scar maturation, with medialisation procedures reserved for selected cases when indicated. Nevertheless, only a minority of patients ultimately undergo additional interventions to improve voice quality.^38,39 In our cohort, patients who underwent type II/III cordectomy showed more pronounced scarring and signs of muscle tension dysphonia during the first postoperative year. Increased muscular tension appears to be a compensatory mechanism facilitating effective glottic closure following more extensive tissue resection. This explanation is supported by a decrease in SPI values, a gradual reduction in the s/z ratio, and a concomitant increase in MPT at 12 months postoperatively. In accordance with good clinical practice, all eligible patients were offered voice therapy during follow-up.⁴⁰ However, none chose to undergo voice therapy, suggesting satisfactory perceived voice outcomes. Their decision was accordingly documented and was not influenced by unavailability of the service. Consequently, despite persistent differences in several objective voice-quality measures, patient-reported voice outcomes remained comparable between the two study groups.

Type I cordectomy is traditionally defined as a subepithelial resection limited to the epithelium and superficial lamina propria, sparing both the vocal ligament and the vocalis muscle, and is commonly employed for vocal fold lesions suspicious for premalignant or early malignant transformation involving only mucosa of the vocal fold. With the increasingly early detection of glottic lesions, conventional type I cordectomy, which involves removal of the entire length of the vocal fold mucosa, is being performed less and less frequently. In early lesions it is possible to preserve at least some segment of unaffected mucosa. As a result, the group of patients undergoing type I cordectomy is becoming increasingly heterogeneous, as it is more often possible to preserve at least part of the mucosa, which certainly has a significant impact on the vibratory properties of the vocal fold. More extensive subepithelial resection may theoretically predispose to scarring of the underlying vocal ligament, potentially impairing mucosal wave propagation and adversely affecting voice outcomes. However, in the present study, procedures classified as type I cordectomy most often involved some mucosal preservation, which may partly explain the generally very favourable objective and subjective functional results. In this context, it may be reasonable to consider that type I cordectomy encompasses a spectrum of surgical extents with potentially different functional implications. Further differentiation between different extents of subepithelial resection may therefore be relevant when evaluating postoperative voice outcomes, even if they are mostly favourable. Conversely, when a type II cordectomy or more extensive resection is required, thorough preoperative counselling and multidisciplinary evaluation are warranted.⁴¹

This is consistent with previous studies showing that removal of small, well-circumscribed vocal fold lesions is associated with full regeneration of the mucosal wave, largely independent of the surgical technique.⁴² These findings may be particularly relevant in the management of professional voice users with precancerous or early malignant vocal fold lesions, whose livelihood directly depends on voice quality. Accordingly, when preservation of some mucosal segment is possible in terms of oncological safety, it permits preservation of the vibratory properties of the vocal fold. Consequently, postoperative voice outcomes are more favourable, therefore it should be considered an appropriate therapeutic option.

In the present study, differences between the two groups were less pronounced in patient-reported voice outcomes than in objective voice quality parameters at the end of the observation period. This finding highlights the importance of psychological and perceptual dimensions of voice assessment and supports the need for an individualised, patient-centred approach. Future research should incorporate standardised psychological instruments to better assess patients’ expectations, emotional status, and their impact on subjective voice-related outcomes.

Limited sample size is an important limitation of this study. Hence, it may be underpowered for some comparisons. However, cohort sizes in existing studies evaluating voice outcomes after laser cordectomy are also highly variable, ranging from a few patients to more than one hundred. Notably, statistically significant differences in voice parameters have been observed even in small cohorts (n < 15 per group), with findings comparable to those reported in larger studies.^10,42,43

Taken together, our findings suggest that in patients with early glottic carcinoma, the extent of laser cordectomy primarily influences the course of postoperative voice recovery rather than the ultimate level of patient-perceived voice quality. While more extensive resections are associated with delayed functional improvement and less favourable objective voice parameters, long-term subjective voice outcomes remain satisfactory for most patients. These findings support the importance of long-term follow-up, individualised patient counselling, and a multidimensional approach to voice assessment when selecting the optimal treatment strategy.

Conclusions

Objective voice measures demonstrated faster postoperative recovery in Group 1, whereas subjective voice outcomes improved substantially in both groups by 12 months.

Although the present findings provide important insight into medium-term postoperative voice recovery following conservative TLM, longer follow-up and larger prospective cohorts are needed to determine whether these functional differences persist over time.

In modern clinical practice, premalignant and early malignant vocal fold lesions are often diagnosed and treated at an early stage. As a result, segments of the vibratory apparatus can be preserved more often with laser cordectomies. Consequently, the population of patients undergoing type I cordectomy is becoming increasingly heterogeneous.

With the oncologically safe option of preserving mucosal segments, there is a greater likelihood of maintaining a higher level of quality of voice function parameters. This approach reflects a shift toward function-preserving strategies in the context of earlier detection of glottic lesions. Hence, there is a need for a more detailed definition of the area of removed vocal fold mucosa in the case of type I cordectomy and, as part of this, a revision of the established ELS criteria, but that requires a multi-centre validation or a much larger, homogeneous cohort.