Introduction

Cerebral hemispheres have interhemispheric connection served by corpus callosum (CC). CC reflects the topographic organization, connectivity of the fiber tracts and fiber tract sizes. Electrophysiology studies have elucidated CC and cortical lobe representation.1 Hence variation in tract connectivity and volumes is reflected in the CC. Anatomically prefrontal cortex fibers are represented in the genu and anterior part of the body. Premotor, motor, and sensory cortical connections are located in the three segments of the body of CC. Posterior parietal, temporal, occipital cortical connections are represented in the splenium of CC2 3 (Fig. 1). A recent diffuse tensor imaging-based study has observed an anteroposterior topography of interhemispheric tracts within the CC, consistent with neuroanatomical studies.4

Close

Fig. 1 (A) Subdivisions of corpus callosum (CC) and white matter tracks—(a) rostrum and genu, (b) body, (c) splenium. (B) Gross lobar representation of CC (a) frontal, (b) parietal, (c) temporal, and (d) occipital lobe. (C) Corpus callosum fibers and their various cortical representation.

Export to PPT

Fig. 1 (A) Subdivisions of corpus callosum (CC) and white matter tracks—(a) rostrum and genu, (b) body, (c) splenium. (B) Gross lobar representation of CC (a) frontal, (b) parietal, (c) temporal, and (d) occipital lobe. (C) Corpus callosum fibers and their various cortical representation.

The pyramidal neurons in layer III of cortical regions have axonal projections to the contralateral cerebral hemisphere that traverses through the CC. In neurodegenerative conditions, subsequent to neuronal cell death these axons undergo Wallerian degeneration and cause morphological/volumetric changes in CC.5 6 In support of above there were reports on CC segmental volume atrophy in mild cognitive impairment due to Alzheimer’ dementia (AD), frontotemporal dementia (FTD), Parkinson’s plus syndromes, and vascular dementia.7 8 9 10 11 In this way, CC anatomical maps with volume analysis has the potential to aid in the study of neurodegenerative and cognitive disorders as CC reflects the effects of neurodegeneration to specific cortical regions. Though the volume of each segment of CC gives more accurate information reflecting the cortical degenerative process, it has few limitations. These include being time-intensive, lack of normative data on CC segmental volume, variation in CC volumes in normal individuals, and more importantly unavailability of quantitative volumetric measurements in routine clinical practice.

In this scenario, the shape profile of CC can serve as an alternative to CC volume as a proxy marker for regional cortical degenerative process and thus help in phenotyping of dementia. CC shape profile of specific segments can be studied using more easily accessible imaging modality to clinicians such as magnetic resonance imaging/computed tomography-brain. However, there is paucity of studies on CC shape profile in various phenotypes of degenerative dementia.12 The objective of the current study is to investigate the profile of CC shape in different phenotypes of dementia.

Materials and Methods

Our study included patients who underwent advanced neuroimaging in our facility as part of clinical evaluation for dementia referred from geriatric clinic during 2017 to 2018. The imaging details were retrieved from the imaging hospital database using a unique identification number allotted to each patient. In all these cases the diagnosis was reached after a detailed evaluation and on the consensus of two psychiatrists specialized in geriatric psychiatry. The clinical diagnosis is based on DSM-5 (Diagnostic and Statistical Manual of Mental Disorders) criteria and appropriate specific recent consensus criteria for each phenotype.13 14 15 16 17 Patients with atypical features of dementia and mixed etiology of dementia were excluded from the study. The current study is a part of a larger retrospective study on patients registered under geriatric clinic and services which has the approval of the Institutional Review Board. Anonymity of the patients was maintained using an alphanumeric code.

Results

Sample included 24 Alzheimer’s patients in our study. Among these there was atrophy of genu (25%), body (anterior: 25%, middle: 12.5%, posterior: 4.1%), and splenium (three divisions-79.2, 79.2, and 0%) (percent indicates proportion of patients). However, these are not exclusive with few patients noted to have overlapping features. On analysis, CC pattern atrophy with cortical regions in AD cases represented involvement of posterior parietal, temporal with few having prefrontal, premotor, and primary motor, primary somatosensory as part of generalized atrophy. In the seven PCA-AD, all cases had splenium atrophy in anterior third (100%) and 57.1% had atrophy of posterior part of body of CC. The segmental volume loss represented posterior parietal and few had primary somatosensory lobe as shown in Table 1.

We have included three FTD subtypes in our study. Among them, we analyzed 17 Bv-FTD, 11 SD, and four PNFA. In the Bv-FTD, there was atrophy of genu (58.8%), entire body (58.8%), and splenium (three divisions: 23.5, 23.5, and 0%). The CC segmental atrophy represented atrophy of prefrontal, premotor, primary motor, and primary somatosensory lobe. In the SD, there was atrophy of body of CC (anterior: 81.8%, middle: 81.8%, posterior: 63.6%) and splenium (three divisions: 63.6, 18.2, and 0%) with no atrophy of genu in any case. The involved segments represented premotor, primary motor, primary somatosensory, posterior parietal, and temporal lobe atrophy. In PNFA both the genu and splenium were spared with predominant involvement of the body of CC (anterior: 100%, middle: 75%, and posterior: 0%) (Table 1). The segments involved represented premotor, primary motor, and in few cases posterior parietal cortical regions.

In other neurodegenerative conditions such as PD and other parkinsonian disorders, we evaluated five cases of PDD, seven cases of DLBD, three cases of PSP, and three cases of CBD. The pattern of atrophy of CC involved genu (20%) body (anterior: 60%, middle: 60%, posterior: 40%) and anterior third of splenium (40%). The segmental involvement in the PDD represented prefrontal, premotor, primary motor, and primary somatosensory and posterior parietal involvement. In DLBD there was atrophy of genu (28.6%), anterior half of body (85.7%), and splenium (three divisions: 57.1, 14.2, and 0%). The segmental pattern of CC atrophy in these cases represented premotor, posterior parietal, prefrontal, and temporal/occipital lobe involvement. In the PSP there was atrophy of genu (100%), body (anterior: 100%, middle: 66.7% and posterior: 0%), and anterior third of splenium (33.3%). CC segmental atrophy represented prefrontal, premotor and primary motor, and posterior parietal atrophy in PSP. In CBD, CC the pattern of atrophy involved genu (33.3%), body (anterior: 100%, middle: 0%, and posterior: 66.7%), and anterior third of splenium (100%) suggesting premotor, posterior parietal, primary somatosensory atrophy, and prefrontal atrophy (Table 1). The CC for each sample phenotype is highlighted as shown in Fig. 2.

Close

Fig. 2 CC profile in AD, PCA, Bv-FTD, PNFA, SD, PSP, PD, DLBD, and CBD. CC is highlighted in each figure. AD, Alzheimer’s dementia; Bv-FTD, behavioral variant of frontotemporal dementia; CBD, corticobasal degeneration; CC, corpus callosum; DLBD, diffuse Lewy body dementia; PCA, posterior cortical atrophy; PD, Parkinson’s disease dementia; PNFA, progressive nonfluent aphasia; PSP, progressive supranuclear palsy; SD, semantic variant of frontotemporal dementia.

Export to PPT

Fig. 2 CC profile in AD, PCA, Bv-FTD, PNFA, SD, PSP, PD, DLBD, and CBD. CC is highlighted in each figure. AD, Alzheimer’s dementia; Bv-FTD, behavioral variant of frontotemporal dementia; CBD, corticobasal degeneration; CC, corpus callosum; DLBD, diffuse Lewy body dementia; PCA, posterior cortical atrophy; PD, Parkinson’s disease dementia; PNFA, progressive nonfluent aphasia; PSP, progressive supranuclear palsy; SD, semantic variant of frontotemporal dementia.

Discussion

We felt the need for the study as we observed that in a clinical setting diagnosing phenotyping dementia is associated with certain challenges to a clinician.18 Phenotype of dementia is based on the historical temporal evolution of clinical symptoms, as it gives information on specific cortical regions involved and helps in phenotyping dementia. However, in many cases due to recall bias or lack of informants, it is often difficult to phenotype dementia clinically. Few others include cerebral comorbidity in degenerative dementia, systemic diseases, autoimmune conditions, and psychiatric conditions which can potentially confuse the clinical picture and phenotyping.19 20 21 22 23 In this situation, we felt CC shape profile can help as an easily available visual aid in the phenotyping dementia.

When we did an overall profile analysis, normal CC has an “eel-like” tubular appearance with a large bulbous Genu and a smaller bulbous splenium. The genu has a rostrum as a subunit which gives it a “beak-like” appearance. Though there are multiple overlaps of anatomical areas the gradient of severity of segmental involvement in the clinical context may act as a clue for underlying pathology.

Majority of earlier studies on AD investigated the volume and thickness of CC. These studies reported predominant involvement of posterior part of CC. Studies also found atrophy of posterior part of CC in MCI due to AD and progression in CC atrophy with the severity of AD.24 25 26 Our study was in agreement with previous studies with predominant involvement of the posterior part of CC (splenium). The involvement of specific CC segments in AD could be understood from the Wallerian degeneration of temporoparietal fibers which pass through the splenium.27 28 In comparison with AD versus PCA, there was more severe involvement of the body of CC in PCA compared with AD and helped to differentiate the phenotypes. Studies by Walterfang et al and Kaufer et al compared CC morphology among FTD and AD. These studies reported more consistent posterior involvement of CC in AD and anterior involvement of CC in FTD.12 29 Our findings were similar to the previous study on CC shape profile which helped in differentiating AD and FTD phenotypes. To the contrary, few studies did not find any difference in CC between AD and FTD.30

Fewer studies have looked at CC profile among the FTD subtypes. These studies have reported more diffuse involvement of CC in Bv-FTD, anterior part of body of CC in PNFA, whereas in SD there was minimal involvement of inferior genu compared controls.11 12 Our finding on CC profile in Bv-FTD is in agreement with above mentioned studies. The diffuse involvement of CC in Bv-FTD can be explained by the degenerative process of medial frontal and orbitofrontal cortex affecting anterior part of CC, degenerative process of superior and inferior temporal regions affecting the posterior part of CC.31 In PNFA, the involvement of left inferior frontal and insula reflected as atrophy of anterior and mid part of CC.32 Our finding in PNFA with sparing of genu and splenium and predominant involvement of body of CC was in agreement with previous studies. In SD our study reported involvement of body and anterior half of splenium in contrast to the inferior genu in previous studies.12 The minimal involvement of CC in PNFA could be explained by neuroanatomy of left anterior temporal poles with many of the axonal fibers traversing through the anterior commissure.33 However, overall, the subtle shape differences in CC could help in differentiating FTD subtypes along with other imaging markers.

In other neurodegenerative conditions such as PDD and other parkinsonian plus disorders, CC profiling may help to guide clinical diagnosis. The literature on CC in movements disorders is mostly on PD and PSP and very little on DLBD and CBS. Study by Goldman et al reported involvement on the mid-anterior and central part of CC in PD with similar picture noted in our PDD sample.8 Whereas, DLBD is a diffuse condition with bilateral involvement. The CC cannot differentiate unilateral involvement from bilateral involvement of cerebral hemisphere. Among the posterior dementia DLBD and AD are common. Apart from the clinical profile, CC also helps differentiation between the two. In AD there is atrophy of the entire body and entire splenial part, whereas in DLBD there is more pronounced atrophy of the first segment of body and splenium. The bulbosity of splenium is preserved in both cases. The sensorimotor fibers are relatively preserved in DLBD. However, there were few studies which did not find any difference in CC profile in PD, PDD, and DLBD compared with that in healthy controls.34 35

In our analysis in PSP, we found an overall thinning of body of CC giving an appearance of dumbbell as genu and splenia were spared. This is similar to the previous studies by Yamauchi et al which reported atrophy of the middle anterior part of CC.9 11 36 This helps in differentiating PSP from FTD and AD. Studies on CBD showed a similar pattern as in PSP with involvement of the middle part of CC.37 There is an overlap in terms of CC involvement in PD, PDD, DLBD, PSP, and CBD with subtle differences among them. Though CC helps in differentiating PD, PD plus syndromes from AD and FTD cannot differentiate PD and PD plus syndromes from each other. This is due to inherent limitations in CC as structural biomarker which will not identify unilateral or bilateral pathology and brainstem involvement. In these cases, clinical profile, imaging markers such as mid-brain will help in differentiating PD phenotypes.38

The basis for atrophy or narrowing of CC in CNS degenerative conditions is due to loss of gray matter neurons leading to secondary axonal loss.6 CC is also now hypothesized to be a pathway for abnormal proteins to spread across hemispheres such as amyotrophic lateral sclerosis.39 Few initial studies on neurodegenerative condition reported white matter microstructural changes prior to gray matter atrophy.40 However, it is difficult to study white matter atrophy in the clinical setting. Knowledge about the segmental anatomy of the CC and what it represents may help subtype these subtle tracts and cortical gray matter atrophy better.4

Certain confounders that can contribute to the variation in shape and thickness of CC include gender, handedness, and age-related white matter change.41 42 43 It is possible that some of the morphological changes of CC in our patients could have exaggerated or attenuated by above confounding factors which is a limitation. It was established that the degree of atrophy of CC depends on the duration and the severity of neurodegenerative condition.44 This was not taken into consideration which is a limitation of our study. Lack of control group in our study is another limitation.

To summarize in posterior dementias such as AD, AD-PCA corresponded to the atrophy of posterior part of CC. In Bv-FTD anterior part of CC was atrophied. In conditions such as PNFA, PDD, PSP, CBD, there is atrophy of body of CC. These findings are in agreement with the pathology of CNS regions. The atrophy pattern of SD is not in agreement with previous reports which need further study.

Conclusion

Corpus callosum is a white matter bundle connecting both cerebral hemispheres. Any pathology to cortical gray matter results in changes in white matter and reflects in CC. In the current study, we explored the shape profile of CC in various phenotypes of dementia. Our visual analysis and interpretation of CC shown in different phenotypes of dementia have a unique shape profile of CC with few overlapping features. Our preliminary study highlights that CC shape profile along with comprehensive clinical history and examination can be useful in phenotyping dementia.