Diffusion Tensor Imaging of the Auditory Pathway in Prelingual Deaf Children in Comparison to Normal Hearing Children in the 1 to 7 Years of Age Group

Abstract

Objective This article aims to determine the microstructural alterations in the auditory pathway in prelingual deaf children using diffusion tensor imaging (DTI)-derived parameters—fractional anisotropy (FA) and apparent diffusion coefficient (ADC), and secondarily to evaluate these changes in rubella and cytomegalovirus (CMV) positive cases.

Materials and Methods A consecutive series of consenting deaf and normal children between 1 and 7 years of age, forming the case and control groups, respectively, underwent DTI, audiological tests, and testing for rubella, CMV, and toxoplasma infections. FA and ADC were measured at four locations bilaterally: lateral lemniscus (LL), inferior colliculus, medial geniculate body, and auditory cortex (AC).

Result The mean ADC values were higher and the mean FA values were lower in cases (19 males, 21 females, mean age 2.65 years) than the controls (21 males, 19 females, mean age 4.63 years) at all eight sites. Sixteen (40%), 17 (42.5%), and 7 (17.5%) cases had severe, severe to profound, and profound hearing loss, respectively, the FA and ADC values being significantly different for LL. For rubella and CMV immunoglobulin G, 20/40and 17/40 cases were positive, respectively, 11 for both, and none for toxoplasma. Significant decrease in FA was seen at LL and AC in rubella/CMV positive cases.

Conclusion Microstructural changes are seen throughout the auditory pathway in prelingual deaf children, especially with rubella and/or CMV positive status. Further studies may pave the path to segregate out patient groups potentially more responsive to cochlear implant.

Keywords

diffusion tensor imaging - prelingual deafness - auditory pathway - fractional anisotropy - apparent diffusion coefficient

Introduction

Prelingual deafness (PD) refers to hearing loss that has occurred prior to language acquisition. It can be congenital or acquired due to ototoxic drugs, infection, trauma, etc. that occur during the initial few years of life.

Cochlear implant (CI), one of the best options to rehabilitate profoundly deaf patients, aims at bypassing the damaged hair cells of cochlea and transmitting the auditory stimuli directly to the nerve fibers through electrical stimulation. It is best done within 4 years of age, when the central nervous system has maximum plasticity for auditory stimulation.[1] In PD children, sensory deprivation may alter the structure and organization of the auditory pathway, which in turn can affect the outcome of CI in them.[2] An accurate imaging workup is required prior to CI, to diagnose abnormalities of inner ear, 7th and 8th nerve complex, brain, and other associated structures to enable appropriate patient selection, side of implantation, and the choice of the device.

Diffusion tensor imaging (DTI) is an advanced magnetic resonance imaging (MRI) technique that studies diffusion properties of water molecules within a tissue and helps in understanding its microarchitecture. Its two main parameters, fractional anisotropy (FA) and apparent diffusion coefficient (ADC), can provide valuable information regarding the structural orientation and integrity of various white matter tracts including the auditory pathway.[3]

It means that comparison of these parameters in young children with and without PD can determine the levels and extent of microarchitectural alterations within the auditory pathways. Further, comparison can be made in PD cases with and without TORCH infection to determine possible suitability and prognosis of CI.

Materials and Methods

After approval from the Institute Ethics Committee, the study was conducted, enrolling 40 cases and controls each, between January 2021 to May 2022. Cases consisted of PD children aged 1 to 7 years being evaluated for possible CI at our institute, while controls were derived from normal hearing children who were already undergoing MRI for other reasons such as isolated peripheral nonsyndromic low-flow vascular malformations or limb trauma and without any systemic complaints or neurological issues, after following the due consent process. Cases with severe neurological disorders such as seizures, brain tumors, diffuse white matter diseases, significant brain atrophy, or hydrocephalus were not included.

The scans of brain and inner ear were performed on a 3.0 T MRI machine (GE DISCOVERY 750W) using a 32-channel head coil under intravenous sedation/general anesthesia ([Table 1]).

Table 1 Parameters used in the MRI sequences of cases and controls

Sequences	TR (ms)	TE (ms)	ST (mm)	NEX	Spacing (mm)	FOV (cm × cm)	Matrix	Acquisition time
Axial T1 FSE	2,186	10	4.0	2	1.5	20 × 16	320 × 256	2 min 35 s
Axial T2 PROPELLER FSE	5,753	120	4.0	3	1.5	20 × 20	352 × 352	1 min 44 s
Coronal T2 FRFSE	4,175	142	4.0	2	1.5	18 × 14	320 × 256	1 min 16 s
Sagittal T2 FRFSE	3,045	139	4.0	2	0.5	20 × 20	320 × 256	1 min 50 s
Axial T2 FLAIR	9,500	93	4.0	1	1.5	20 × 16	288 × 192	2 min 42 s
Axial 3D FIESTA-C[a]	5.8	2.3	0.8	2	0.4	16 × 16	288 × 288	5 min 47 s
Axial DTI	5,834	112	4.0	2	1.5	20 × 20	128 × 140	2 min 14 s
Axial 3D SWAN	42.5	23	1.6	0.7	0	22 × 15	260 × 256	3 min 39 s

Abbreviations: 3D, three-dimensional; DTI, diffusion tensor imaging; FIESTA, fast imaging employing steady-state acquisition; FLAIR, fluid-attenuated inversion recovery; FOV, field of view; FRFSE, fast relaxation fast spin echo; FSE, fast spin echo; MRI, magnetic resonance imaging; NEX, number of excitations; ST, slice thickness; SWAN, susceptibility-weighted angiography; TE, echo time; TR, repetition time.

^a Included posterior fossa and inner ear (all other sequences were of whole brain).

The images were then transferred to AW 3.2 software workstation and postprocessed. Color encoded FA and ADC maps were generated. FA and ADC were estimated using region of interest (ROI) analysis at four locations along the auditory pathway bilaterally, which included lateral lemniscus (LL), inferior colliculus (IC), medial geniculate body (MGB), and the auditory cortex ((AC); [Fig. 1]). ROI placement was guided by color-encoded maps and anatomic images. At LL a square ROI of area 4 mm² was used in all cases. Similarly, a square ROI of 6 mm² and a rectangular ROI of 9.6 mm² were used at IC and MGB, respectively. At AC a freehand ROI was drawn around the margins of Heschl's gyrus (Brodmann areas 41 and 42) in axial images. ROI drawing was performed by two radiologists (J.K.J. and S.K.) at joint sessions by common consensus.

Fig. 1 Region of interest (ROI) placement technique on axial T1-weighted and diffusion tensor imaging-derived fractional anisotropy (FA) and apparent diffusion coefficient (ADC) images. (A) Color-encoded FA map showing ROI placement at bilateral lemnisci.. (B) Axial FA map showing average FA estimated at lateral lemnisci. (C) Axial ADC map showing ROI placed at lateral lemnisci with average ADC displayed. (D) Axial T1 image denoting anatomic location of lateral lemniscus. (E) Color-encoded FA map showing ROI placement at bilateral inferior colliculi. (F) Axial FA map showing average FA estimated at inferior colliculi. (G) Axial ADC map showing ROI placed at inferior colliculi with average ADC displayed. (H) Axial T1 image denoting anatomic location of inferior colliculi. (I) Color-encoded FA map showing ROI placement at bilateral medial geniculate body (MGB). (J) Axial FA map showing average FA estimated at MGB. (K) Axial ADC map showing ROI placed at bilateral MBG with average ADC displayed. (L) Axial T1 image denoting anatomic location of MGB. (M) Color-encoded FA map showing ROI placement at bilateral Heschl's gyrus (auditory cortex). (N) Axial FA map showing average FA estimated at bilateral Heschl's gyrus. (O) Axial ADC map showing ROI placed at bilateral Heschl's gyrus with average ADC displayed. (P) Axial T1 image denoting anatomic location of Heschl's gyrus. Note that the color-coded FA maps (A, E, I, M) denote direction of white matter fibers using 3 colors—red (left-right), blue (craniocaudal), and green (anteroposterior), using conventional FA map coding scheme.

All the children underwent audiological testing at the institute's audiometry facility by means of three tests prior to the MRI, brainstem evoked response audiometry (BERA), otoacoustic emissions, and behavioral response audiometry. Depending upon the BERA findings, the cases were grouped into three categories of hearing loss based on the severity: severe hearing loss, severe to profound hearing loss, and profound hearing loss. Controls did not undergo these tests, but were clinically examined to rule out deafness and to possess normal language functions.

Cases also underwent a TORCH screening using serum antibodies (immunoglobulin [Ig] M and IgG) against toxoplasma (TX), rubella (RB), and cytomegalovirus (CMV).

Statistical analysis was done using IBM SPSS Statistics (release 22.0; IBM, Somers, New York, United States). Descriptive analysis was done to obtain mean or median values of FA and ADC at all four locations bilaterally. Normality was assessed using Shapiro–Wilk's test and Q-Q plot.

To compare the mean FA and ADC between cases and controls, independent samples t-test was used except for median ADC at bilateral AC, for which Mann–Whitney U test was used. Sidewise comparison of mean FA was done using paired t-test. Sidewise comparison of ADC was done using paired t-test (LL and IC) or Wilcoxon signed rank test (MGB and AC).

Correlation between FA and ADC was assessed using Pearson's correlation or Spearman's rho (right MGB and bilateral AC).

The paired (right and left) FA and ADC values were averaged at each of the four locations. The average of two sides was then used to compare between degrees of hearing loss using one-way analysis of variance (ANOVA). Tukey post hoc analysis was also employed.

The average FA was further compared between TORCH reactive and nonreactive cases, as well as for cases with and without cochlear nerve abnormality/absence, using independent samples t-test.

Results

The study group consisted of 40 cases with mean age of 2.65 years (19 males and 21 females), and an equal number of controls with mean age of 4.63 years (21 males and 19 females) between 1 and 7 years of age. The cases had bilateral hearing loss of similar degree on both sides, classified as severe (n = 16, 40%), severe to profound (n = 17, 42.5%), and profound (n = 7, 17.5%). Fifty percent (n = 20) and 42.5% (n = 17) cases showed IgG reactivity to RB and CMV, respectively, with 11 being reactive to both, thus 65% cases were positive for RB, CMV, or both, while none were reactive for TX IgG.

Five children had bilateral inner ear malformation as confirmed by high-resolution computed tomography of temporal bone and MRI and included two cases of bilateral incomplete partition of cochlea (type II) with enlarged vestibular aqueduct, two cases of bilateral cochlear hypoplasia with hypoplastic cochlear nerves, and one case of bilateral rudimentary otocyst with cochlear nerve aplasia.

Comparison of FA and ADC at the four pairs of anatomic locations along the auditory pathway between cases and controls demonstrated statistically significant decrease in FA and increase in ADC at all locations in the former ([Table 2]).

Table 2 Mean fractional anisotropy (FA) and apparent diffusion coefficient (ADC) of cases and controls

	Fractional anisotropy (FA)			Apparent diffusion coefficient (ADC)
Anatomic location	Case (n = 40) Mean (SD)	Control (n = 40) Mean (SD)	p-Value[a] (independent t-test)	Case (n = 40) Mean (SD) [×10−6 mm2/s]	Control (n = 40) Mean (SD) [×10−6 mm2/s]	p-Value[a] (independent t-test)
Right LL	0.508 (0.024)	0.670 (0.019)	<0.001	780.8 (30.0)	694.7 (18.5)	< 0.001
Left LL	0.510 (0.029)	0.670 (0.024)	<0.001	792.1 (31.1)	698.6 (16.0)	< 0.001
Right IC	0.675 (0.043)	0.756 (0.022)	<0.001	792.0 (31.9)	718.1 (16.1)	< 0.001
Left IC	0.673 (0.032)	0.755 (0.020)	<0.001	794.40 (37.5)	719.0 (11.5)	< 0.001
Right MGB	0.311 (0.017)	0.365 (0.019)	<0.001	878.4 (33.9)	797.6 (11.5)	< 0.001
Left MGB	0.322 (0.022)	0.364 (0.017)	<0.001	890.0 (38.6)	799.8 (11.5)	< 0.001
Right AC[b]	0.196 (0.018)	0.225 (0.015)	<0.001	966.5 (38)	915.5 (18)	< 0.001
Left AC[b]	0.185 (0.018)	0.218 (0.014)	<0.001	994.5 (43)	915.0 (21)	< 0.001

Abbreviations: AC, auditory cortex; IC, inferior colliculus; IQR, interquartile range; LL, lateral lemniscus; MGB, medial geniculate body; SD, standard deviation.

^a p-Value < 0.05 is significant.

^b For AC, the ADC values were nonnormal distribution, hence were evaluated using Mann–Whitney U test and the values are median (IQR), as against t-test and mean for all other values in the table.

Comparison of FA and ADC between two sides at the four places of examination in the cases demonstrated mean difference ranging from 0.010 to –0.011 and 28 to –11.32 × 10⁻⁶ mm²/s, respectively.

Comparison of mean FA and degree of hearing loss showed statistically significant changes at LL only between the three categories on ANOVA (p = 0.021), with post hoc Tukey analysis further revealing that this effect was majorly due to decreased FA in the children with profound hearing loss ([Table 3]). Similar observations were made for ADC too, where ANOVA demonstrated significant differences between the groups for LL only (p = 0.001125), and the effect on post hoc analysis was due to differences in between the severe and severe to profound and the severe and profound groups ([Table 3]).

Table 3 Comparison of extent of FA change across three categories of hearing loss, viz., severe, severe to profound, and profound

Mean FA and ADC (average of two sides) across categories of hearing loss severity
	Severe (n = 16) FA (SD) ADC (SD) [×10−6 mm2/s]	Severe to profound (n = 17) FA (SD) ADC (SD) [×10−6 mm2/s]	Profound (n = 7) FA (SD) ADC (SD) [×10−6 mm2/s]	p-Value (ANOVA)
Lateral lemniscus	0.522 (0.021) 771.843 (30.874)	0.503 (0.022) 793.529 (27.701)	0.496 (0.024) 802.714 (25.664)	0.021[a] 0.001125[a]
Inferior colliculus	0.683 (0.024) 785.062 (33.363)	0.660 (0.030) 804.029 (32.298)	0.686 (0.052) 785.642 (38.297)	0.087 0.054
Medial geniculate body	0.318 (0.018) 878.093 (34.827)	0.316 (0.015) 883.970 (30.423)	0.315 (0.018) 899 (50.670)	0.886 0.205
Auditory cortex	0.196 (0.017) 1005.312 (128.530)	0.186 (0.012) 979.205 (33.527)	0.186 (0.016) 1012.928 (64.843)	0.109 0.357
Comparison of pairs of hearing loss with respect to mean difference in FA at lateral lemniscus (Tukey post hoc analysis)
Comparison pair		Mean difference in FA at lateral lemniscus		p-Value
Severe–severe to profound		0.018		0.062
Severe–profound		0.026		0.038[a]
Severe to profound–profound		0.007		0.727
Comparison of pairs of hearing loss with respect to mean difference in ADC at lateral lemniscus (Tukey post hoc analysis)
Comparison pair		Mean difference in ADC [×10−6 mm2/s] at lateral lemniscus		p-Value
Severe–severe to profound		21.69		0.03409[a]
Severe–profound		30.87		0.00149[a]
Severe to profound–profound		9.18		0.53031

Abbreviations: ADC, apparent diffusion coefficient; FA, fractional anisotropy; SD, standard deviation.

^a p-Value < 0.05 is significant.

TORCH positive status demonstrated significantly decreased FA at LL (p = 0.042) and AC (p = 0.016), but not at IC (p = 0.18) and MGB (p = 0.27), and the three cases with abnormalities in cochlear nerve did not show statistically significant difference from those with normal nerves.

Discussion

In our study, we utilized DTI to estimate the diffusion parameters of FA and ADC, at four locations along bilateral auditory pathway locations, viz., LL, IC, MGB, and AC, and demonstrated significantly reduced FA and increased ADC at all locations in cases as compared with controls. The decrease in FA is similar to the results obtained in previous studies conducted by Wu et al,[4] Miao et al,[5] Huang et al,[6] Chang et al,[7] Lin et al,[8] Jiang et al,[9] and Park et al.[10] Estimation of ADC along auditory pathway of children with hearing loss has not been reported before. The alterations are suggestive of structural alterations in the auditory pathway which may be due to reduction in the number of axons, changes related to axonal or myelin integrity, and abnormalities with fascicular structure and organization. These structural changes may be the result of longstanding sensory deprivation that could induce alterations at microstructural level. These changes can occur all the way from cochlea to the AC.[11]

In the present study, 26 out of 40 cases had a positive TORCH serology (IgG reactivity), and the study could demonstrate some effect in the form of lower FA in LL and AC in cases with TORCH positive versus negative status. To the best of our knowledge, DTI findings of auditory pathway have not been documented in children with hearing loss having positive TORCH serology. As we conducted the serological testing after 1 year of age, it is assumed that the IgG reactivity is either due to a past infection (in utero or during infancy) or a recent infection instead of it being due to vertical transmission of antibodies because the latter tend to disappear by 12 months of age.[12] [13] [14] [15] Nijman et al conducted a DTI-based study in preterm infants who acquired CMV infection postnatally, comparing 21 infected with 61 noninfected neonates, and found decreased FA in the occipital white matter in the former group, however, noting that this difference did not result in unfavorable short-term neurodevelopmental status.[16] van der Voorn et al also demonstrated raised ADC and decreased FA in white matter in CMV infection, and found it to be similar to that of periventricular leukomalacia, noting that both insult occur during the same developmental period harboring immature oligodendrocytes.[17]

CMV is the most common infective cause of sensorineural hearing loss and the proposed mechanisms by which it induces sensorineural hearing loss includes immune-mediated injury, apoptosis of neurons in spiral ganglion of cochlea, and even by inducing vascular abnormalities in the cochlea.[18] RB infection is also a cause of encephalitis and various neurological sequelae including hearing loss.[19] Ours is the first article studying this patient group. Comparison of CMV, RB, and CMV + RB cases did not reveal significant difference, further strengthening our hypothesis that it is the lack of sensory stimulus resulting in the alteration of DTI parameters, not the parenchymal insult and neither the type of insult. However, in cases with extensive parenchymal changes, which are hypothesized to be due to vasculitis phenomenon and infarcts along with encephalitis, they would obviously result in further and possibly more severe changes. Such cases have not been included in our study, and also these are not considered candidates for CI in our institutional practice. Of note, we observed statistically significant changes in FA at LL and AC in TORCH positive cases, not in IC or MGB, and the reasons are difficult to fathom; however, it can be possibly guessed from the ROI placement technique. Both LL and AC included white matter bundles, the subarcuate fibers included in the latter's ROI, while IC and MGB are predominantly deep gray matter nuclei, and possibly lesser affected. However, this remains a conjecture and larger studies are needed for validation.

Somewhat analogous to the sensory stimulus deficit causing the changes in postinfective cases would be the patients with cochlear nerve abnormalities/deficiency. A DTI study in 12 patients with cochlear nerve deficiency has demonstrated a significant reduction in the FA of bilateral LL and IC, in comparison to normal hearing control.[4] Our study had only three such patients, and while the mean FA in these was lower than control, it was not statistically significant, possibly necessitating a larger sample size.

We also examined if one or more side showed more changes, such that the “better off” side could be a prospective better candidate for CI, and while statistical tests did show significance at couple of locations for both FA and ADC, the absolute values of the differences are not significant to be of significance in real life.

One significant drawback of our study is the discrepancy in the mean age of cases and controls. It is known that brain volume as well as white matter maturation progressively increases, with the early childhood period a rather exuberant phase for neurodevelopment such that a four times increase in volume is seen up to 5 years of age. It is thus not unfathomable, that some of the differences in our two groups may be due to the different ages of maturation.

Conclusion

Reduced FA and increased ADC were consistent findings throughout the auditory pathway in PD cases, remarkably so in cases with CMV or RB IgG positivity in LL and AC. Both these diffusion parameters may be routinely calculated as part of the imaging studies prior to CI and also in follow-up imaging after CI surgery to correlate with the clinical outcomes so to establish further guidelines. Further studies can also focus on determining the laterality to decide upon suitability of CI.

Conflict of Interest

None declared.

Acknowledgments

The authors acknowledge Dr. Alok Ranjan, Biostatistics faculty, Department of Community and Family Medicine, All India Institute of Medical Sciences Patna for the statistics related to the study, and Mr. Sanjeev Kumar, Radiographer Grade I, for the data keeping and liasioning with various cases and controls.

Data Availability

The essential data pertaining to the study are attached in the [Supplementary File S1] (available in the online version).

Ethical Approval

This study was performed according to the Declaration of Helsinki, and after obtaining approval from the Institute Ethics Committee.

• References

• 1 Kral A, Sharma A. Developmental neuroplasticity after cochlear implantation. Trends Neurosci 2012; 35 (02) 111-122
• 2 Campbell R, MacSweeney M, Woll B. Cochlear implantation (CI) for prelingual deafness: the relevance of studies of brain organization and the role of first language acquisition in considering outcome success. Front Hum Neurosci 2014; 8: 834
• 3 Shaikh S, Kumar A, Bansal A. Diffusion tensor imaging: an overview. Neurol India 2018; 66 (06) 1603-1611
• 4 Wu CM, Ng SH, Wang JJ, Liu TC. Diffusion tensor imaging of the subcortical auditory tract in subjects with congenital cochlear nerve deficiency. AJNR Am J Neuroradiol 2009; 30 (09) 1773-1777
• 5 Miao W, Li J, Tang M. et al. Altered white matter integrity in adolescents with prelingual deafness: a high-resolution tract-based spatial statistics imaging study. AJNR Am J Neuroradiol 2013; 34 (06) 1264-1270
• 6 Huang L, Zheng W, Wu C. et al. Diffusion tensor imaging of the auditory neural pathway for clinical outcome of cochlear implantation in pediatric congenital sensorineural hearing loss patients. PLoS One 2015; 10 (10) e0140643
• 7 Chang Y, Lee SH, Lee YJ. et al. Auditory neural pathway evaluation on sensorineural hearing loss using diffusion tensor imaging. Neuroreport 2004; 15 (11) 1699-1703
• 8 Lin Y, Wang J, Wu C, Wai Y, Yu J, Ng S. Diffusion tensor imaging of the auditory pathway in sensorineural hearing loss: changes in radial diffusivity and diffusion anisotropy. J Magn Reson Imaging 2008; 28 (03) 598-603
• 9 Jiang M, Wen Z, Long L. et al. Assessing cerebral white matter microstructure in children with congenital sensorineural hearing loss: a tract-based spatial statistics study. Front Neurosci 2019; 13: 597
• 10 Park KH, Chung WH, Kwon H, Lee JM. Evaluation of cerebral white matter in prelingually deaf children using diffusion tensor imaging. BioMed Res Int 2018; 2018: 6795397
• 11 Shepherd RK, Hardie NA. Deafness-induced changes in the auditory pathway: implications for cochlear implants. Audiol Neurootol 2001; 6 (06) 305-318
• 12 Leuridan E, Van Damme P. Passive transmission and persistence of naturally acquired or vaccine-induced maternal antibodies against measles in newborns. Vaccine 2007; 25 (34) 6296-6304
• 13 Leuridan E, Hens N, Hutse V, Aerts M, Van Damme P. Kinetics of maternal antibodies against rubella and varicella in infants. Vaccine 2011; 29 (11) 2222-2226
• 14 Kiliç A, Altinkaynak S, Ertekin V, Inandi T. The duration of maternal measles antibodies in children. J Trop Pediatr 2003; 49 (05) 302-305
• 15 Niewiesk S. Maternal antibodies: clinical significance, mechanism of interference with immune responses, and possible vaccination strategies. Front Immunol 2014; 5: 446
• 16 Nijman J, Gunkel J, de Vries LS. et al. Reduced occipital fractional anisotropy on cerebral diffusion tensor imaging in preterm infants with postnatally acquired cytomegalovirus infection. Neonatology 2013; 104 (02) 143-150
• 17 van der Voorn JP, Pouwels PJ, Vermeulen RJ, Barkhof F, van der Knaap MS. Quantitative MR imaging and spectroscopy in congenital cytomegalovirus infection and periventricular leukomalacia suggests a comparable neuropathological substrate of the cerebral white matter lesions. Neuropediatrics 2009; 40 (04) 168-173
• 18 Xia W, Yan H, Zhang Y. et al. Congenital human cytomegalovirus infection inducing sensorineural hearing loss. Front Microbiol 2021; 12: 649690
• 19 Cohen BE, Durstenfeld A, Roehm PC. Viral causes of hearing loss: a review for hearing health professionals. Trends Hear 2014; 18: 2331216514541361

Publication History

Article published online:
24 October 2024