Laryngeal carcinoma is one of the most common head and neck cancers. It was estimated that 151,000 new cases (approximately 2.0 men and 0.3 women in every 100,000 population) of laryngeal carcinoma were diagnosed worldwide in 2008 [1]. Glottic carcinoma represents the majority of laryngeal cancer cases. It is mainly confined to the anterior portion of the vocal cord [2]. Early symptom of glottic carcinoma is hoarseness that can be diagnosed at early-stage of the disease. Laryngeal cancer has a fairly good prognosis among malignancies of the head and neck if prompt and appropriate treatment is provided during the early stage [3-5]. Early detection and treatment can lead to a cure rate of 90%–95% with maximal voice preservation [4].

Several studies using objective and subjective analyses have focused on voice quality of patients with glottic carcinomas after laser surgery (LS) and radiotherapy (RT). In previous meta-analyses investigating voice quality, posttreatment favored external radiation rather than transoral laser surgical excision [6]. One study has shown no clinical difference in voice quality between transoral LS and external radiation [7]. Huang et al. [8] have compared patients treated with LS and RT and found no clear difference in jitter or shimmer. However, in meta-analysis of maximum phonation time (MPT), fundamental frequency differences were observed between the two groups [8]. Patients who had undergone RT showed increased MPT and decreased frequency (F0) compared to those with LS. LS generally results in loss of tissue while RT leads to vocal fold fibrosis and muscle fatigue [9]. The objective of the present meta-analysis was to evaluate voice quality of LS and RT based on the grade, roughness, breathiness, asthenia, and strain (GRBAS) scale and acoustic analysis parameters including frequency–perturbation (jitter), measures of amplitude–perturbation (shimmer), and noise to harmonic ratio (NHR). Results of this study can allow patients diagnosed with glottic carcinomas to select an effective treatment in terms of voice quality, provided both LS and RT are equally effective in treating the cancer.

T1 and T2 early-stage glottic cancer is associated with lower morbidity and higher cure rate than advanced laryngeal cancer. Appropriate treatment can minimize its impact on voice. Thus, appropriate treatment plan needs to be established based on patient’s sex, age, and mental state. If possible, a treatment method that maintains good phonation and respiratory function is desirable. Treatment methods for early glottic cancer include laser cordectomy (LC) and RT that can preserve voice [3,4]. As early glottic cancer is a disease affecting voice, a study analyzing voice after treatment with these two methods could facilitate patient decision-making. RT is one of the primary treatment modalities in patients with early glottic cancer with favorable outcomes. It can result in voice preservation [5].

Hong et al. [9] have investigated delayed tissue effects of human vocal fold following radiation therapy and found fibrotic changes in human superficial lamina propria and thyroarytenoid muscle impairing vocal fold closure, pliability, and dynamic changes in tension and length. They reported that, compared to surgery, RT improved voice quality as laryngeal edema subsided after surgery and restored the original function and appearance of the larynx [9]. However, many complications are associated with RT as mentioned earlier.

LC is an endoscopic laryngeal surgery in which the lesion is excised using a laser after appropriate laryngeal exposure under general anesthesia. The range of cordectomy is determined according to the European Laryngological Society classification. In general, the lesion is excised with a 2-mm border and electro-cautery is used in the area that is beyond reach of the laser. LC is associated with a minimal risk of laryngeal edema after the surgery. It has low morbidity rate and less side effects. It can be performed repeatedly. Therefore, LC represents a better surgical option for patients with early glottic cancer because it has fewer complications. In addition, it is more likely to preserve voice in the long term [13]. LC does not exclude other treatment methods. It is also easy to detect cancer recurrence [6]. However, as LC leaves 1–3 mm residual lesion during resection to preserve the voice, it may not sufficiently remove the tumor. Thus, tumor recurrence is possible [14]. Several attempts have been made to minimize damage to voice quality by subdividing the resection range [15].

A number of studies have been conducted to determine voice quality of patients with early glottic cancer after RT and LC. Most of these studies focused on voice acoustic analysis, psychoacoustic analysis using GRBAS scale with analysis of voice handicap index, and comparison of quality of life after surgery [16]. However, in previous studies, no significant differences in F0, jitter, or shimmer were found, although MPT was slightly longer in the RT group than that in the LC group [7,8]. The difference in MPT and shimmer between the two groups suggests that patients who have received RT might have better voice than those who have received LC. A study comparing voice handicap index and the quality of life has shown that the group that underwent LC has statistically better results [15,16]. In addition, the LC group showed similar MPT and acoustic parameters to the normal group and voice after LC was similar to that after RT. Thus, LC can be recommended as a cost-effective intervention with a reasonable voice outcome [7]. The RT group showed more severe insufficiency of glottal closure due to laryngeal stroboscopy than the LC group, suggesting ventricular closure during phonation and hyperadduction of the arytenoid cartilage. According to a psychoacoustic analysis [10-12,17], patients who received RT had higher frequency of normal or mild coarse voice whereas those who received LC exhibited higher frequency of moderate or severe coarse voice, indicating that RT was better than LC [18-24]. However, Tamura et al. [20] have revealed that both groups show similar voice preservation results, suggesting that LC does not affect the voice considerably. Hong et al. [9] reported that even though evaluation of stop consonants was not always performed clinically and clinical difference might be subtle, significant differences were observed in certain measures at group level. Changes in voice onset time and vowel can result in decreased speech intelligibility in the RT group. Thus, LS might be superior to radiation therapy in terms of speech production.

In this study, we conducted a meta-analysis of two modalities for early-stage glottic cancer. Our results revealed better acoustic analysis parameters and perceptual outcome in patients treated with RT compared with those treated with LS. We compared parameters of acoustic analysis (jitter, shimmer, and NHR) and parameters of perceptual analysis (GRBAS) in early glottic cancer through a meta-analysis. The meta-analysis included the following studies: van Gogh et al. [17], Luo et al. [10], Czecior et al. [11], Ahn et al. [12], Nunez Batalla et al. [18], Wedman et al. [21], and McGuirt et al. [24]. We found no significant differences between LS and RT groups in acoustic analysis parameters. Kono et al. [13] and Krengli et al. [19] have shown significant differences in jitter, shimmer, and NHR between the two modalities, with more favorable values shown in the RT group. However, Rydell et al. [23] reported that RT was more effective than LS in jitter in acoustic analysis.

Our meta-analysis of acoustic parameters suggested significant differences between patients treated with LS and RT according to jitter, shimmer, and NHR. When comparing acoustic analysis between the two modalities, jitter, shimmer, NHR had more favorable values in the RT group than those in the LS group, similar to results Kono et al. [13] and Sjogren et al. [14]. With respect to GRBAS of perceptual analysis, Czecior et al. [11] and Sjogren et al. [14] have shown no significant differences between the two treatments. However, Ma et al. [16] found that patients treated with RT exhibited greater roughness and strain than patients treated with LS, although no differences in breathiness or asthenia were noted. Aaltonen et al. [15] found that RT results showed statistically significant differences in B and A than transoral LS, although G, R, and S were similar between the two treatment groups. Rydell et al. [23] found that RT treatment was better than LS in terms of B. By contrast, Rosier et al. [22] found no difference in B between the two options.

In summary, our study found significant differences in G and A of GRBAS scale between LS and RT. Voice quality in patients treated with RT tended to be better than that in patients who were treated with LS. Results of this meta-analysis suggest significant differences in voice quality following RT and LS for the treatment of early glottic cancer. However, this study has limitations related to a relatively small sample size for meta-analysis. In addition, most of the included studies were retrospective in design. Moreover, our study lacked data related to LS types and radiation doses for RT.

Our meta-analysis showed that voice quality in perceptual and acoustic analysis was significantly better in patients who were treated with RT compared with patients who underwent LS. RT may enhance voice quality in early-stage glottic cancer. Further well-designed, multicenter, and randomized controlled studies are needed. Widening the search to languages beyond English and Korean might return a significantly higher number of quality articles.