Skip to main content
The Journal of the Indian Prosthodontic Society logoLink to The Journal of the Indian Prosthodontic Society
. 2025 Jul 16;25(3):266–275. doi: 10.4103/jips.jips_64_25

A pilot study comparing conventional and digital impression techniques for speech analysis using Hindi vowels in maxillectomy patients rehabilitated with an obturator

Gunjan Chouksey 1,✉, Gurkaran Preet Singh 1, Vikas Gupta 2, Prasanta Sahoo 3, Rupali Choure 4, Anima Goyal 5
PMCID: PMC12370098  PMID: 40669000

Abstract

Aim:

Maxillectomy patients frequently have speech impairments resulting from the loss of the oral-nasal partition. Prosthodontic rehabilitation with an obturator helps restore speech intelligibility, with its success largely dependent on accurate impression recording of maxillary defects. This investigation evaluated the effectiveness of conventional versus digital impression techniques in the context of speech analysis, specifically using Hindi vowels, in maxillectomy patients rehabilitated with obturators.

Study Setting and Design:

This research, designed as a quasi-experimental study, was undertaken at a tertiary care hospital.

Material and Methods:

The study included 20 patients needing obturators, assigned to two groups: one received prostheses fabricated with conventional impressions, and the other with digital techniques. Speech parameters, including fundamental frequency, formant frequencies (F1, F2, F3), intensity, jitter, shimmer, and maximum phonation duration (MPD), were analyzed using Praat software before and three months after rehabilitation. Hindi vowels aa /a:/, ii /i:/, and uu /u:/ were sustained at a controlled intensity. Swallowing efficiency was assessed via the water swallow test.

Statistical Analysis Used:

The Wilcoxon Rank Sum test or exact test was used to compare the data, with a p-value < 0.05 considered significant.

Results:

The most common maxillectomy defects were Brown class 2b. After three months, significant improvement in speech parameters and swallowing efficiency was seen in both groups of participants. However, no statistically significant differences (P > 0.05) were found between the conventional and digital impression groups.

Conclusions:

This novel study compared conventional and digital impressions for speech analysis using Hindi vowels in maxillectomy patients after rehabilitation with an obturator. Voice recording and acoustic analysis using Hindi vowels provide valuable insights into speech rehabilitation outcomes in maxillectomy patients. Both conventional and digital impression techniques effectively fabricate obturators, improving speech characteristics and intelligibility. Both methods can be used for maxillectomy patients, allowing flexibility in clinical practice.

Keywords: Digital impressions, Hindi vowels, maxillectomy defects, obturator, speech analysis

INTRODUCTION

Patients with maxillectomy defects due to reasons such as cancer, trauma, and fungal infections often complain of impaired speech characteristics and intelligibility.[1,2] Speech is a unique and complex process that involves the coordinated interaction of several systems and structures. These include respiration, phonation, resonation, neurologic integration, audition, and articulation. Speech also requires the synchronized movements of both hard-tissue structures, such as the maxilla and mandible, and soft-tissue structures, including the lips, tongue, soft palate, and uvula.[3] In patients with maxillectomy, hypernasality is observed due to air leakage caused by the loss of the oral-nasal partition.[4,5,6,7] Articulation disorders are seen due to changes in the maxillary structure.[4]

Effective separation between oral and nasal cavities in patients with maxillectomy is achieved with the help of prosthodontic rehabilitation. Rehabilitation of these patients can be achieved through either prosthetic or surgical methods.[8] Obturators protect the underlying tissues and provide a surface for contact with the tongue for effective speech and swallowing, improving hypernasality.[9,10]

One way to investigate effective nasality correction after prosthodontic rehabilitation in patients with maxillectomy defects is to analyze the vowels instead of consonants. Vowels play a pivotal role in speech acoustics, serving as practical and representative markers of overall speech patterns.[1] Clinical assessments of nasality frequently rely on vowel sounds, as their acoustic properties are sensitive to even minor alterations in resonance and oral cavity volume. These variations are typically influenced by the degree of tongue constriction and its spatial relationship to the maxillary structures.[11]

Many previous studies have used vowels for speech analysis in maxillectomy patients.[1,9,12,13,14,15,16,17,18] Studies have highlighted a strong correlation between vowel quality and speech intelligibility.[5,14,17] Tobey and Lincks in 1989 examined vowel format frequencies using /i/ and /u/ in 5 patients with maxillectomy defects before and after prosthodontic rehabilitation and found differences amongst anterior versus posterior resections in speech adaptation.[9] Yoshida et al. in 2000 evaluated voice recordings using the /i/ vowel produced independently in 15 patients to evaluate hypernasality.[12] According to them, among male subjects, the fundamental frequency of the vowel /i/ within the 1/3-octave spectrum ranged from 160 to 200 Hz, while in females, it ranged from 200 to 250 Hz. A one-third-octave analysis was conducted by Lee et al. to measure the presence of hypernasality in adult speech was assessed for the vowel /i/, isolated from two Cantonese monosyllabic words patients with various etiologies, including dysarthria, maxillectomy, and cleft palate.[13] The study demonstrated that intrajudge reliability is high when utilizing one-third-octave band analysis, and applies to assessing hypernasality in the speech of adults with different aetiologies. Sumita et al. in 2002 calculated the formant frequencies after analyzing 12 male patients with maxillectomy defects and 12 healthy male participants using the vowels /a/, /e/, /i/, /o/, and /u/.[14] According to their findings, all five vowels exhibited a marked decrease in F2 values in patients with maxillectomy defects, while the /i/ vowel showed a significant increase in F1. Kwon et al. conducted a study in 2011 that used three Korean corner vowels, ⁄a⁄, ⁄i⁄, and ⁄ u⁄, to compare the speech function of patients with low-height bulb obturators to that of the same patients with high-height bulb obturators.[15] The F1 and F2 values for the high bulb obturators were compared with those for obturators with low bulb height. They found that variations in obturator bulb height did not result in significant changes in formant frequencies. However, hypernasality correction was commonly seen in patients with high bulb obturators. In 2018, Sumita et al. evaluated speech intelligibility in maxillectomy patients by analyzing vowels and calculating the formant frequency measurements for the vowels /a/, /e/, /i/, /o/, and /u/ using linear predictive coding.[5] They reported that assessing the F2 range of the five vowels can provide valuable clinical insights in predicting speech intelligibility after maxillectomy. Wu et al. conducted a study evaluating speech intelligibility and acoustic features in patients with palatal defects by analyzing Chinese vowel formants, acoustic vowel graphs (F1-F2 plots), and levels of nasalization. They compared outcomes among three groups: Patients using obturators, those without, and healthy controls.[16] Their findings showed a reduction in nasalization for the vowels /i/ and /u/ with obturator use, along with a gradual normalization of speech function. In addition, the F2 formant frequencies for six vowels in the obturator group were notably higher than in the nonobturator group and closely resembled those of normal subjects, suggesting that obturators help improve control over F2 frequency modulation. Elbashti et al. conducted a digital acoustic analysis of five Japanese vowels /a/, /e/, /i/, /o/, and /u/ with a speech analysis system and calculated the first formant (F1) and second formant (F2) frequencies and reported that a significantly higher F2 range was seen with Class II maxillectomy patients.[17] In 2022, Wang et al. in patients with a prosthesis after maxillectomy performed a digital acoustic analysis of the first three formant frequencies and recorded five Japanese vowels /a/, /e/, /i/, /o/, and /u/ in patients with a prosthesis after maxillectomy.[18] They concluded that wearing an obturator prosthesis altered both the first two formants and the third formant. In a 2023 study, Ali et al. and colleagues analyzed the formants of five Japanese vowels in patients with maxillectomy and reported strong positive correlations between the acoustic measurements obtained using Praat and those from the computerised speech lab (CSL), regardless of the Praat settings used.[1]

Hindi is the most commonly used language of India. It is written in the Devanagari script and is the most spoken language in India. It is also the 4th most spoken language in the world after Mandarin, Spanish, and English. There are 33 consonants and 13 vowels used for speaking and writing in Hindi. In Hindi, vowels can function as independent sounds; however, the consonants require vowels for pronunciation. The acoustics of Hindi oral and nasal vowels have been sparsely documented.[19]

Recording accurate impressions of the maxillectomy defects becomes important for an obturator to function well. The two ways of recording impressions are conventional and digital.[20,21,22,23,24,25,26] There is still a debate about digital impressions being as good as conventional impressions for recording maxillectomy defects.

No prior studies have compared conventional and digital impression techniques in terms of speech outcomes using Hindi vowels in maxillectomy patients rehabilitated with obturators. To address this gap, the present exploratory study aimed to evaluate and compare the impact of these two impression methods on speech, using acoustic analysis of Hindi vowels. This pilot study is novel in its focus on Hindi vowels for acoustic speech analysis, addressing the underrepresentation of non-English-speaking populations in prosthetic speech research, and in evaluating functional speech outcomes following obturator fabrication using both conventional and digital impression techniques. It is among the first to generate context-specific data from an Indian tertiary care setting, offering culturally relevant insights and laying the groundwork for future multicenter studies.

MATERIALS AND METHODS

This study included 20 maxillectomy patients with COVID-19-associated mucormycosis who reported to our institute for prosthodontic management following maxillectomy. Conducted between 2023 and 2024, all participants were native Hindi speakers between 30 and 58 years of age with normal hearing and no vocal cord abnormalities. Exclusion criteria included nonCOVID-19-related maxillary resections, bilateral total maxillectomy, hearing issues, vocal cord abnormalities, and inability to speak or understand Hindi. Ethical approval was obtained, and informed consent was received. Figure 1 shows the study flow chart.

Figure 1.

Figure 1

Study flow chart

A single investigator (G.C.) collected sociodemographic data and clinical details, including the number of remaining maxillary teeth and Brown classification of defects.[27] Speech and swallowing assessment was done at baseline before rehabilitation with an obturator. Patients were alternately assigned to two groups: In Group 1, conventional impressions of the maxillary arch, including the palate and defect, were obtained using a stainless-steel tray designed for dentulous patients and an irreversible hydrocolloid material (Imprint; DPI). These impressions were poured in Class III dental stone (Kaldent) to produce working models. In contrast, Group 2 employed a digital workflow: Intraoral scans were captured using the Medit i600 scanner, guided by Medit Link v3.3.3 software (MEDIT Corp., Seoul, South Korea) and adhering to the manufacturer’s recommended protocol for precision. Scanning commenced at the occlusal surfaces of the posterior teeth, progressed anteriorly in a zigzag pattern, and then covered the buccal and palatal surfaces similarly. Subsequently, the scanner was directed to capture the mucosal areas, including the palate and maxillectomy defect. The digital scan data were exported as STL files and processed with Medit Design v2.1.4 and Medit Model Builder v1.3.3 to create the virtual three-dimensional (3D) models. These were converted into physical resin models through rapid prototyping using a Dentiq 3D printer and CURO Element Model resin (Ackuretta, Taiwan). For all participants, mandibular arch impressions were made using the conventional method and poured in dental stone.

Definitive obturators were fabricated using the traditional techniques. Clasps were made from 21- gauge stainless steel wire (KC Smith and Co., UK) and adapted onto the maxillary casts. A record base of self-cure acrylic resin (DPI; Mumbai, India) with an occlusal rim was prepared, and jaw relations were recorded. Teeth were set in harmony with the patient’s existing occlusion, and wax contours were refined for anatomical fidelity. A trial insertion was conducted to assess esthetics and phonetics.

Final obturators [Figure 2], featuring hollow bulbs, were processed using heat-cured acrylic resin (DPI; Mumbai, India) for all patients in both groups. The fabrication protocol for the digital impression group mirrored that of the conventional group. The resin models produced via 3D printing were robust enough to withstand the dewaxing and heat polymerization process. Once finished and polished, the prostheses were delivered, ensuring a secure fit with no nasal leakage. Patients wore the obturators for a 3-month period prior to evaluation of speech and swallowing performance.

Figure 2.

Figure 2

Obturator fabricated using maxillary digital impression (a) and maxillary conventional impression (b)

Two audiologists and speech-language pathologists (P.S. and A.G.) did voice recording and acoustic analysis using a computer software program, Praat version 6.4.13, before and after 3 months of rehabilitation with obturators [Figure 3].[28] Praat is a widely used software for phonetic and acoustic analysis of speech, equipped with tools for waveform visualization, spectrogram generation (both wide and narrow-band), pitch and intensity tracking, speech synthesis, and audio editing. It enables the detailed examination of acoustic features such as pitch, intensity, and formant frequencies, allowing for in-depth phonetic assessments.

Figure 3.

Figure 3

Analysis of speech using Hindi vowels (aa/aː/, ii/iː/, uu/uː/) in maxillectomy patients using Praat version 6.4.13. (a) Before rehabilitation with an obturator. (b) After rehabilitation with an obturator

In this study, participants were seated comfortably in a sound-treated environment, facing a sound level meter. Voice recordings were captured using an omnidirectional MAX cm 903 electret condenser microphone. To maintain standardization, both the microphone and sound level meter were positioned at a 45° angle from the participant’s mouth, one on each side, at a consistent distance of 6 inches. A measuring tool was used for each subject to ensure uniform positioning during all the recordings.

Participants were asked to sustain the Hindi vowels aa /aː/, ii/iː/, and uu/uː/ for approximately 3 s at a consistent volume of around 75 dB (±2 dB). Voice recordings were captured using Praat version 6.4.13 (2024 release) on a personal computer with a sampling frequency of 44.1 kHz. Acoustic analysis was performed with the following parameters: 1000-time steps, 250 frequency steps, a window length of 0.005 s, a Gaussian window shape, and a Fourier transform algorithm. To maintain consistency in analysis and reduce variability from voice onset and offset, only the central, stable 3-s segment of each vowel was analyzed.

Maximum phonation duration (MPD) was assessed through the sustained production of the Hindi vowel ii/iː/. Speech parameters evaluated in this study included fundamental frequency (Hz), intensity (INT) {decibel [dB]}, MPD, formant frequency, first formant (F1) [Hz], second formant (F2) [Hz], third formant (F3) [Hz], and perturbation parameters such as jitter (pitch perturbation) and shimmer (amplitude perturbation). The water swallow test was used for the swallowing assessment in both groups of participants.[8,29]

Data analysis was conducted using R, version 4.2.1 (R Foundation for Statistical Computing, Vienna, Austria). Categorical variables were described using frequencies and percentages, whereas continuous variables were reported as medians along with their interquartile ranges (IQR). The Wilcoxon rank sum test was employed for statistical comparisons, with a significance threshold set at P < 0.05. Visual representation of the results was carried out through box and whisker plots.

RESULTS

Among the 30 individuals screened, 20 participants (comprising 17 males and 3 females) fulfilled the inclusion criteria and were enrolled in the study, with an equal allocation of 10 participants per group. The ten who failed the inclusion criteria were maxillectomy defects due to the reasons other than COVID-19-associated mucormycosis, like cancer (N = 5) and bilateral total maxillectomies (N = 5) in whom conventional definitive obturators were not possible due to the lack of retention and support. The median age for Group 1 and Group 2 was 45.5 and 51.5 years, ranging from 30 to 58 years. All participants were married. Forty percent of Group 1 had education up to primary school and university, while 80% of Group 2 had primary school education. Most participants were laborers (40% in Group 1 and 60% in Group 2) and had Class 2b maxillectomy defects (infrastructural defects) [Table 1].

Table 1.

Characteristics of participants

Characteristic Group 1 - conventional (n=10)a, n (%) Group 2 - digital (n=10)a, n (%) P b
Age 45.5 (36.2–46.8) 51.5 (43.0–54.2) 0.172#
Sex
  Female 0 3 (30.0) 0.211#
  Male 10 (100.0) 7 (70.0)
Marital status
  Married 10 (100.0) 10 (100.0)
Education
  High school 0 1 (10.0) 0.125#
  Preschool 2 (20.0) 0
  Primary school 4 (40.0) 8 (80.0)
  University 4 (40.0) 1 (10.0)
Occupation
  Housewife 0 2 (20.0) 0.056#
  Labourer 4 (40.0) 6 (60.0)
  Not employed 2 (20.0) 0
  Private business 2 (20.0) 0
  Professional 2 (20.0) 0
  Retired 0 1 (10.0)
  Teacher 0 1 (10.0)
Teeth in the maxillary arch
  4 1 (10.0) 1 (10.0) 0.827#
  5–8 6 (60.0) 6 (60.0)
  9–11 3 (30.0) 3 (30.0)
Brown classification
  Class 2 b 5 (50.0) 6 (60.0) 0.700#
  Class 2 c 4 (40.0) 2 (20.0)
  Class 2 d 1 (10.0) 2 (20.0)

#Nonsignificant, aMean±SD, n (%), bWilcoxon rank sum test, Fisher’s exact test. SD: Standard deviation, n: Total number of participants

Speech analysis before rehabilitation [Table 2] showed no significant differences (P > 0.05) between the groups in fundamental frequency, F1, F2, F3, INT, and MPD. After 3 months of rehabilitation [Table 3], speech parameters remained statistically similar between groups (P > 0.05). A box and whisker plot [Figure 4] illustrates the distribution of fundamental frequency, jitter, MPD, and shimmer using Hindi vowels.

Table 2.

Comparison between the conventional and digital impression group of participants for analysis of speech using Hindi vowels (aa/aː/, ii/iː/, uu/uː/) in maxillectomy patients before rehabilitation with an obturator

Characteristics aa/aː/
ii/iː/
uu/uː/
Conventional (n=10)a Digital (n=10)a P b Conventional (n=10)a Digital (n=10)a P b Conventional (n=10)a Digital (n=10)a P b
Pre F0 (Hz) 157.1 (154.0–164.9) 167.6 (145.1–224.0) 0.489# 168.7 (162.1–178.5) 186.0 (170.3–235.2) 0.222# 165.3 (157.4–177.9) 189.8 (165.0–241.7) 0.171#
Pre F1 (Hz) 689.7 (671.6–744.3) 700.5 (634.8–852.6) >0.999# 365.1 (341.0–422.1) 405.5 (349.9–498.1) 0.340# 431.7 (401.2–456.5) 480.0 (433.6–510.5) 0.040*
Pre F2 (Hz) 1177.7 (1131.9–1857.7) 1129.8 (981.2–1176.5) 0.063# 1851.7 (1358.6–2048.0) 1440.4 (1386.6–2099.6) 0.931# 2096.6 (1267.9–2132.6) 1768.3 (950.1–2610.8) >0.999#
Pre F3 (Hz) 2604.9 (2508.1–2684.9) 2861.7 (2589.8–2925.2) 0.605# 2695.7 (2523.2–3057.1) 2649.9 (2508.9–2815.2) 0.546# 3060.2 (2603.9–3157.0) 3078.2 (2805.3–3213.8) 0.605#
Pre-INT (dB) 58.9 (56.6–61.5) 65.5 (64.4–67.3) 0.040* 61.6 (57.9–65.5) 63.5 (59.9–67.2) 0.436# 62.3 (60.3–65.1) 66.8 (64.6–69.1) 0.113#
Prejitter (%) 0.4 (0.3–1.1) 0.2 (0.2–0.3) 0.070# 0.3 (0.2–0.5) 0.2 (0.2–0.5) 0.666# 0.4 (0.2–0.7) 0.2 (0.1–0.3) 0.136#
Preshimmer (%) 6.8 (5.6–8.4) 4.0 (3.7–5.6) 0.222# 1.6 (1.6–8.3) 6.5 (3.8–8.2) 0.387# 3.7 (3.2–9.1) 3.9 (3.9–6.1) 0.931#
Pre-MPD (s) 12.0 (10.0–14.0) 12.0 (9.0–14.0) 0.789# 11.0 (11.0–12.0) 11.0 (10.0–12.0) 0.497# 12.0 (10.0–12.0) 11.0 (10.0–12.0) 0.892#

*Significant, #Nonsignificant, aMedian (IQR), bWilcoxon rank sum exact test, Wilcoxon rank sum test. F0: Fundamental frequency, F1: The first formant frequency, F2: The second formant frequency, F3: The third formant frequency, INT: Intensity, MPD: Maximum phonation duration, IQR: Interquartile range

Table 3.

Comparison between the conventional and digital impression group of participants for analysis of speech using Hindi vowels (aa/aː/, ii/iː/, uu/uː/) in maxillectomy patients after 3 months of rehabilitation with an obturator

Characteristics aa/aː/
ii/iː/
uu/uː/
Conventional (n=10)a Digital (n=10)a P b Conventional (n=10)a Digital (n=10)a P b Conventional (n=10)a Digital (n=10)a P b
Post F0 (Hz) 155.1 (139.7–157.1) 158.1 (147.5–213.0) 0.297# 160.5 (154.7–165.9) 168.4 (150.6–226.2) 0.222# 156.9 (143.2–169.4) 183.4 (154.7–239.3) 0.063#
Post F1(Hz) 726.5 (686.8–762.6) 808.6 (698.7–839.1) 0.258# 327.6 (311.6–351.5) 382.9 (306.7–456.5) 0.489# 431.3 (361.9–456.2) 506.8 (422.2–557.1) 0.387#
Post F2 (Hz) 1147.9 (1107.2–1841.6) 1236.1 (1190.7–1319.9) 0.730# 2127.7 (2105.0–2298.1) 2205.4 (2104.7–2334.5) 0.605# 2157.0 (1982.5–2258.4) 1008.7 (948.2–1092.9) 0.040*
Post F3 (Hz) 2701.2 (2517.3–3126.8) 2806.7 (2614.1–2896.7) 0.796# 2812.4 (2420.8–3141.0) 2879.9 (2760.3–3136.2) 0.546# 3077.0 (2520.9–3265.8) 2708.0 (2687.9–3059.3) 0.489#
Post I (dB) 58.6 (53.3–65.2) 62.8 (60.0–67.9) 0.258# 61.8 (59.5–63.5) 64.2 (58.6–65.9) 0.860# 59.7(57.4–60.6) 64.3 (57.4–67.1) 0.566#
Postjitter (%) 0.6 (0.4–0.7) 0.4 (0.3–0.5) 0.063# 0.3 (0.2–0.6) 0.3 (0.3–0.4) 0.796# 0.3 (0.3–0.4) 0.3 (0.2–0.5) 0.436#
Postshimmer (%) 7.5 (5.2–9.1) 5.4 (4.7–5.9) 0.340# 7.4 (1.3–12.3) 6.3 (5.3–10.5) 0.730# 6.2 (2.4–9.0) 5.2 (5.1–6.3) 0.863#
Post MPD (s) 13.0 (13.0–14.0) 13.0 (12.0–15.0) 0.928# 14.0 (14.0–15.0) 11.0 (11.0–14.0) 0.018* 13.0 (11.0–13.0) 12.0 (12.0–13.0) >0.999#

*Significant, #Nonsignificant, aMedian (IQR), bWilcoxon rank sum exact test, Wilcoxon rank sum test. F0: Fundamental frequency, F1: The first formant frequency, F2: The second formant frequency, F3: The third formant frequency, MPD: Maximum phonation duration, IQR: Interquartile range

Figure 4.

Figure 4

Box and whisker plot comparing conventional and digital impressions for fundamental frequency, Jitter, maximum phonation duration(s), and shimmer using Hindi vowels after 3 months of rehabilitation obtained from acoustic speech software Praat. Boxes show variable distribution between the 75th and 25th percentile, and the thick line represents the median. F0: Fundamental frequency, J: Jitter (%), MPD: Maximum phonation duration, S: Shimmer

For the 100 ml water swallow test, all the participants initially experienced nasal regurgitation and difficulty swallowing.[29] After rehabilitation with obturators, all showed improvement and could swallow efficiently without coughing or choking.

DISCUSSION

This exploratory study compared the conventional and digital impressions for speech analysis using Hindi vowels in maxillectomy patients after rehabilitation with an obturator. Obturators are usually preferred to close the maxillectomy defects prosthetically as they are simple, nonsurgical, and have shown significant improvement in oral health-related quality of life in patients with maxillectomy defects due to cancer and COVID-19-associated mucormycosis.[2] The impressions of the maxillectomy defects can be made conventionally using a stock metal tray and an impression material or digitally using an intraoral scanner.[20,21,22,23,24,25,26] For conventional impressions, aspiration is risky while the impression is being made. It is more technique-sensitive, and retrieving the impression becomes difficult if severe tissue undercuts or trismus exist.[20,21,22,23,24,25,26] In addition, chipping of stone models during working affects the quality of the obturator. On the other hand, making digital impressions can significantly reduce the laboratory working time, reduce the risk of aspiration of the impression material, and may benefit patients with reduced mouth opening.[20,21,22,23,24,25,26] There is still a debate regarding the feasibility of digital impressions to accurately record the maxillectomy defects. Elbashti et al. demonstrated the effectiveness of intraoral scanners in accurately digitizing models of maxillectomy defects. Similarly, Brucoli et al. and associates found that digital approaches utilizing intraoral optical scanners are reliable for recording maxillary defects after maxillectomy procedures.[25] However, Zhang et al. reported limitations, stating that digitising deeper defect sites in such patients was not feasible.[26]

For an obturator to show improvement in functions like speech and swallowing, recording accurate impressions of the maxillectomy defects becomes important. A study by Dholam et al. has shown improvement in acoustic speech parameters after the fabrication of obturators using conventional impressions.[10] Brucoli et al. in their study on the use of an optical scanner for the fabrication of maxillary obturator prostheses, showed that all participants reported 100% speech recovery.[25] In our study, speech improvements with Hindi vowels were observed in both groups following rehabilitation. Nevertheless, there were no statistically significant differences in speech outcomes between obturators made from conventional versus digital impressions (P > 0.05).

Many software programs are available for Acoustic Speech analysis, like the Praat, CSL, Dr. Speech by Tiger Electronics, etc. Dr. Speech sciences is a comprehensive software suite for voice analysis and rehabilitation in clinical settings. Compatible with windows, it offers modules for voice assessment, speech analysis, electroglottography evaluation, progress tracking, speech training, voice therapy, wave generation, and phonetogram analysis.[30] In contrast, CSL is recognised as one of the most advanced voice and speech analysis systems.[1] This high-performance commercial product includes integrated hardware and software components, such as a specialised input/output recording device for precise acoustic measurements. However, both Dr. Speech sciences and CSL are commercial tools and not freely accessible. Alternatively, Praat, a free and open-source cross-platform tool, provides scripting capabilities and is available online. Patients can also download Praat to practice therapy exercises and achieve specific therapy goals.[1] Ali et al. and colleagues conducted a study on maxillectomy patients analyzing the five Japanese vowels and found a strong correlation between acoustic measurements obtained from the commercially available CSL system and those from Praat software. This indicates that both software programs employ comparable decision strategies for formant analysis, even in the presence of pathological speech associated with maxillectomy.[1] Standard settings of the Praat software allow for effective monitoring of speech rehabilitation following prosthetic treatment, without requiring specialized equipment or extensive knowledge of speech acoustics. In this current study, voice recording and analysis of Hindi vowels were done using a computer software program, Praat version 6.4.13, before and after 3 months of rehabilitation with obturators.

Vowels are essential in speech acoustics and effectively represent general speech. The acoustic analysis of vowels is valuable for assessing speech production in maxillectomy patients and other forms of atypical speech.[1] Changes in the oral cavity structure can alter vocal frequency and intensity, which vary among individuals and denote the loudness of the voice and control over overall speech.[10] After a maxillectomy, patients often struggle with voice control and articulation.[4] An obturator is crucial in preventing nasal air escape and helps coordinate oral musculature for a proper lip seal, thereby improving speech.[9,10] It has become increasingly important for diagnosing pathological voices in vocal and acoustic analysis. Frequency is represented as pitch. Jitter and shimmer ratios, with a normal value of 1, measure the perturbations in frequency and intensity, respectively. These ratios reflect the minimal variations from one point on a sound wave to the corresponding point on the next wave.[10] Many studies have previously reported the changes in acoustic speech parameters after maxillectomy and rehabilitation with obturators using vowels in various languages.[1,5,9,12,13,14,15,16,17]

Hindi is the most commonly used official language of India. In Hindi, vowels can function as independent sounds. A vowel is coded based on the place of articulation and openness of the vocal tract. The vowels used in our study are aa/aː/, ii/iː/, and uu/uː/. The International Phonetic Association has classified Hindi vowels aa/aː/and ii/iː/as the front vowel while uu/uː/. is the back vowel. On the openness scale, ii/iː/and uu/uː/are considered narrow, while the vowel aa/aː/is considered wide.[19]

Dholam et al., in their study, reported an improvement in the maximum frequency with a definitive obturator (257.69 Hz) compared with the preoperative value (225.13 Hz) and other obturators (surgical obturator, 229.69 Hz; interim obturator, 229.69 Hz); however, it was not statistically significant (P = .288).[10] This study found no statistically significant differences (P > 0.05) between obturators made from conventional and digital impressions when comparing median (IQR) values of fundamental frequency, F1, F2, F3, intensity (INT), and MPD during speech analysis of Hindi vowels.

For the 100 ml water swallow test, patients were asked to drink 100 ml of water as quickly and comfortably as possible. All patients in both groups complained of nasal regurgitation and difficulty swallowing before rehabilitation. After rehabilitation with obturators in both groups, all patients showed improvement and could swallow 100 ml of water efficiently without coughing or choking. Similarly, in a study by Grover et al., speech intelligibility, nasal resonance, and swallowing ability improved with an obturator.[8]

To the best of our knowledge, this is the first study to compare the conventional and digital impressions for speech analysis using Hindi vowels in maxillectomy patients after rehabilitation with an obturator. Our study’s novelty lies in combining digital dentistry, maxillofacial prosthetics, and linguistically-relevant speech analysis, offering practical insights for both clinical prosthodontists and speech therapists working in diverse linguistic settings. The findings of this preliminary study suggest that obturators created from digital impressions perform similarly to those made with conventional impressions in enhancing speech for patients with infrastructural maxillectomy defects. The intra-oral scanner (Medit i600) adequately scanned both the defect and nondefect sides to fabricate the obturator, and hence, digital impressions can be used for making maxillary obturators. Our experience showed that patients rehabilitated with obturators fabricated using the digital impression technique demonstrated superior fit, faster adaptation, improved comfort, minimal need for adjustments, and equally good speech outcomes compared to those fitted with conventionally fabricated prostheses. A key contributor to these improved outcomes is the accuracy of the impression technique. Conventional tray impressions are susceptible to distortion due to anatomical undercuts, saliva, and patient movement, often resulting in a poor palatal seal. This suboptimal fit can lead to increased nasal air emission, hypernasality, and reduced intraoral pressure, negatively impacting speech clarity. In contrast, digital impressions provide greater precision and consistency, overcoming many limitations associated with physical impression materials. Patients also reported greater comfort with digital scanning, particularly those experiencing postsurgical sensitivity or anxiety, as it avoids the discomfort and gagging often associated with conventional trays. The intraoral scanner used in our clinical setting has a standard tip size of 22.2 mm × 15.9 mm and a scan area of 15 mm × 13 mm. A smaller 19 mm tip was also available for capturing smaller defects or for patients with limited mouth opening. Most patients in our study presented were Brown’s Class II defects (infrastructural defects), which allowed for adequate visualisation and accurate digital recording of the defect area. However, given the scanner’s limited field of view, digital impressions may not be suitable for all maxillary defects, especially in Brown’s Class III and IV defects.

A major limitation of this study was the small sample size, similar to other studies on acoustic speech parameters in maxillectomy patients. This exploratory study was conducted with a logistically feasible sample size, including 20 participants (17 men, 3 women) out of 30 screened, with 10 in each group, with most patients in both groups having Brown’s Class II defects (infrastructural defects). Due to the recent emergence of COVID-19-associated mucormycosis, a rare pathology, developing a formal hypothesis with a larger sample was not feasible. The low incidence of this condition further restricted recruitment from a single institute. Future studies with larger sample sizes, other sections of the Browns classification, and longer follow-ups will help provide more accurate results. A larger, adequately powered study can be planned to statistically compare the conventional and digital techniques based on the preliminary trends observed here.

CONCLUSIONS

Within the aforementioned limitations, this exploratory study on QOL and speech parameters in maxillectomy patients undergoing obturator rehabilitation concluded that for infrastructural defects, voice recording, and acoustic analysis using Praat provided valuable insights into pitch, intensity, and formants for comprehensive speech evaluation. No significant differences (P > 0.05) were observed between obturators fabricated using conventional and digital impressions when analyzing Hindi vowels based on the fundamental frequency, formants (F1, F2, and F3), intensity (INT), and MPD. These findings indicate that conventional and digital impression techniques effectively fabricate obturators, allowing flexibility in clinical practice for maxillectomy patients.

Conflicts of interest

There are no conflicts of interest.

Funding Statement

This research received an Intramural Grant from All India Institute of Medical Sciences Bhopal (No. RC/AIIMS-BPL/IM-LOI/AP/2023/332).

REFERENCES

  • 1.Ali IE, Sumita Y, Wakabayashi N. Comparison of Praat and computerized speech lab for formant analysis of five Japanese vowels in maxillectomy patients. Front Neurosci. 2023;17:1098197. doi: 10.3389/fnins.2023.1098197. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 2.Chouksey G, Dugad J, Gupta V, Dholam K, Goel P, Choure R. Oral health-related quality of life after rehabilitation with maxillary obturators: A comparison of patients with maxillectomy defects associated with cancer and post-COVID-19 mucormycosis. J Prosthet Dent. 2023 doi: 10.1016/j.prosdent.2023.11.009. S0022-3913(23)00757-6. [DOI] [PubMed] [Google Scholar]
  • 3.Beumer J, Curtis TA. Speech, palatopharyngeal function, and restoration of soft palate defects. In: Beumer J, Curtis TA, Firtell DN, editors. Maxillofacial Rehabilitation: Prosthodontic and Surgical Considerations. St. Louis: Mosby; 1979. pp. 244–91. [Google Scholar]
  • 4.Beumer J, Curtis TA. Restoration of acquired hard palate defects: Etiology, disability, and rehabilitation. In: Beumer J, Curtis TA, Firtell DN, editors. Maxillofacial Rehabilitation: Prosthodontic and Surgical Considerations. St. Louis: Mosby; 1979. pp. 188–243. [Google Scholar]
  • 5.Sumita YI, Hattori M, Murase M, Elbashti ME, Taniguchi H. Digitised evaluation of speech intelligibility using vowels in maxillectomy patients. J Oral Rehabil. 2018;45:216–21. doi: 10.1111/joor.12595. [DOI] [PubMed] [Google Scholar]
  • 6.Fadhil SM, Mumcu E. Rehabilitation of a patient with palatal defect-A case report. J Surg Surg Res. 2019;5:93–6. [Google Scholar]
  • 7.Babu MR, Premalatha BS, Halkud R. A cross-sectional study on articulatory working space area in mandibulectomy or maxillectomy. Int J Health Sci Res. 2020;10:125–33. [Google Scholar]
  • 8.Grover R, Jurel SK, Agarwal B, Rao J, Kapoor S, Mishra N, et al. Speech intelligibility, nasal resonance, and swallowing ability of maxillectomy patients with customized obturator: A non randomized controlled study. J Indian Prosthodont Soc. 2021;21:249–55. doi: 10.4103/jips.jips_98_21. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 9.Tobey EA, Lincks J. Acoustic analyses of speech changes after maxillectomy and prosthodontic management. J Prosthet Dent. 1989;62:449–55. doi: 10.1016/0022-3913(89)90180-7. [DOI] [PubMed] [Google Scholar]
  • 10.Dholam KP, Bachher G, Gurav SV. Changes in the quality of life and acoustic speech parameters of patients in various stages of prosthetic rehabilitation with an obturator after maxillectomy. J Prosthet Dent. 2020;123:355–63. doi: 10.1016/j.prosdent.2019.03.005. [DOI] [PubMed] [Google Scholar]
  • 11.Cler GJ, Perkell JS, Stepp CE. Oral configurations during vowel nasalization in English. Speech Commun. 2021;129:17–24. doi: 10.1016/j.specom.2021.02.005. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 12.Yoshida H, Furuya Y, Shimodaira K, Kanazawa T, Kataoka R, Takahashi K. Spectral characteristics of hypernasality in maxillectomy patients. J Oral Rehabil. 2000;27:723–30. doi: 10.1046/j.1365-2842.2000.00537.x. [DOI] [PubMed] [Google Scholar]
  • 13.Lee AS, Ciocca V, Whitehill TL. Acoustic correlates of hypernasality. Clin Linguist Phon. 2003;17:259–64. doi: 10.1080/0269920031000080091. [DOI] [PubMed] [Google Scholar]
  • 14.Sumita YI, Ozawa S, Mukohyama H, Ueno T, Ohyama T, Taniguchi H. Digital acoustic analysis of five vowels in maxillectomy patients. J Oral Rehabil. 2002;29:649–56. doi: 10.1046/j.1365-2842.2002.00911.x. [DOI] [PubMed] [Google Scholar]
  • 15.Kwon HB, Chang SW, Lee SH. The effect of obturator bulb height on speech in maxillectomy patients. J Oral Rehabil. 2011;38:185–95. doi: 10.1111/j.1365-2842.2010.02139.x. [DOI] [PubMed] [Google Scholar]
  • 16.Wu S, Huang X, Wang J, Hong N, Li Y. Evaluation of speech improvement following obturator prostheses for patients with palatal defect. J Acoust Soc Am. 2018;143:202. doi: 10.1121/1.5020781. [DOI] [PubMed] [Google Scholar]
  • 17.Elbashti ME, Sumita YI, Hattori M, Aswehlee AM, Taniguchi H. Digitized speech characteristics in patients with maxillectomy defects. J Prosthodont. 2019;28:649–55. doi: 10.1111/jopr.12705. [DOI] [PubMed] [Google Scholar]
  • 18.Wang Y, Hattori M, Liu R, Sumita YI. Digital acoustic analysis of the first three formant frequencies in patients with a prosthesis after maxillectomy. J Prosthet Dent. 2024;132:1082–7. doi: 10.1016/j.prosdent.2022.10.010. [DOI] [PubMed] [Google Scholar]
  • 19.Shosted R, Carignan C, Rong P. Managing the distinctiveness of phonemic nasal vowels: Articulatory evidence from Hindi. J Acoust Soc Am. 2012;131:455–65. doi: 10.1121/1.3665998. [DOI] [PubMed] [Google Scholar]
  • 20.Alanezi A, Aljanahi M, Moharamzadeh K, Ghoneima A, Tawfik AR, Khamis AH, et al. Development and comparison of conventional and 3D-printed laboratory models of maxillary defects. Dent J (Basel) 2023;11:115. doi: 10.3390/dj11050115. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 21.Elbashti ME, Hattori M, Patzelt SB, Schulze D, Sumita YI, Taniguchi H. Feasibility and accuracy of digitizing edentulous maxillectomy defects: A comparative study. Int J Prosthodont. 2017;30:147–9. doi: 10.11607/ijp.5095. [DOI] [PubMed] [Google Scholar]
  • 22.Elbashti ME, Hattori M, Patzelt SB, Aswehlee AM, Sumita YI, Taniguchi H. Precision and trueness of computerized optical impressions in maxillectomy defects: An in vitro 3D comparison. Int J Prosthodont. 2019;32:289–92. doi: 10.11607/ijp.6077. [DOI] [PubMed] [Google Scholar]
  • 23.Elbashti ME, Sumita YI, Kelimu S, Aswehlee AM, Awuti S, Hattori M, et al. Application of digital technologies in maxillofacial prosthetics literature: A 10-year observation of five selected prosthodontics journals. Int J Prosthodont. 2019;32:45–50. doi: 10.11607/ijp.5932. [DOI] [PubMed] [Google Scholar]
  • 24.Srivastava G, Padhiary SK, Mohanty N, Patil PG, Panda S, Cobo-Vazquez C, et al. Digital workflow feasibility for the fabrication of intraoral maxillofacial prosthetics after surgical resection: A systematic literature review. Acta Odontol Scand. 2024;83:392–403. doi: 10.2340/aos.v83.40870. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 25.Brucoli M, Boffano P, Pezzana A, Corio C, Benech A. The use of optical scanner for the fabrication of maxillary obturator prostheses. Oral Maxillofac Surg. 2020;24:157–61. doi: 10.1007/s10006-020-00836-9. [DOI] [PubMed] [Google Scholar]
  • 26.Zhang M, Hattori M, Elbashti ME, Sumita YI. Feasibility of intraoral scanning for data acquisition of maxillectomy defects. Int J Prosthodont. 2020;33:452–6. doi: 10.11607/ijp.6763. [DOI] [PubMed] [Google Scholar]
  • 27.Brown JS, Rogers SN, McNally DN, Boyle M. A modified classification for the maxillectomy defect. Head Neck. 2000;22:17–26. doi: 10.1002/(sici)1097-0347(200001)22:1<17::aid-hed4>3.0.co;2-2. [DOI] [PubMed] [Google Scholar]
  • 28.Sahoo AK, Sahoo PK, Gupta V, Behera G, Sidam S, Mishra UP, et al. Assessment of changes in the quality of voice in post-thyroidectomy patients with intact recurrent and superior laryngeal nerve function. Cureus. 2024;16:e60873. doi: 10.7759/cureus.60873. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 29.Patterson JM, Hildreth A, McColl E, Carding PN, Hamilton D, Wilson JA. The clinical application of the 100mL water swallow test in head and neck cancer. Oral Oncol. 2011;47:180–4. doi: 10.1016/j.oraloncology.2010.11.020. [DOI] [PubMed] [Google Scholar]
  • 30.Núñez Batalla F, González Márquez R, Peláez González MB, González Laborda I, Fernández Fernández M, Morato Galán M. Acoustic voice analysis using the Praat program: Comparative study with the Dr. Speech program. Acta Otorrinolaringol Esp. 2014;65:170–6. doi: 10.1016/j.otorri.2013.12.004. [DOI] [PubMed] [Google Scholar]

Articles from The Journal of the Indian Prosthodontic Society are provided here courtesy of Wolters Kluwer -- Medknow Publications

RESOURCES