Abstract
BACKGROUND AND PURPOSE: In fetuses with lateral ventriculomegaly and normal posterior fossa cerebrospinal spaces, third ventricle distention is a compelling clue that supports a diagnosis of aqueductal stenosis. However, this association assumes normal ventricular anatomy. Structural constraints can impair pressure-induced compliance. We aimed to determine how thalamic massa intermedia size alterations may impact the size of the third ventricle in the setting of congenital aqueductal stenosis.
MATERIALS AND METHODS: This retrospective study was performed at a single academic pediatric hospital after institutional review board approval. We searched our brain MRI reports for all examinations describing aqueductal stenosis and included all the patients who had both fetal and postnatal examinations. Patients with interhypothalamic adhesions and hydrocephalus unrelated to congenital aqueductal stenosis were excluded from this study. We evaluated all the MRIs for the presence of thalamic massa intermedia and documented third ventricle diameters (supraoptic recess, central and suprapineal recesses) and the thalamic massa intermedia circumference. The Spearman correlation was used to identify the potential relationship between the thalamic massa intermedia circumference and third ventricle size in fetal and postnatal MRIs. Patients were also stratified into 2 groups based on the presence or absence of thalamic massa intermedia. Mann-Whitney U tests were used to compare third ventricle diameters between these groups.
RESULTS: The study included both fetal and postnatal studies from 59 patients. The overall third ventricle diameter was inversely proportional to the circumference of the thalamic massa intermedia in both groups (fetal: P = .001, ρ = −0.422; [95% CI, −0.628 to −0.181]; postnatal: P < .001, ρ = −0.653; [95% CI, −0.782 to −0.479]). Nonetheless, dilation of anterior and posterior recesses still occurred when the mid third ventricle was nondilated or less severely dilated in patients with an enlarged thalamic massa intermedia. Third ventricle dilation was most severe in patients lacking a thalamic massa intermedia compared with patients with a thalamic massa intermedia (P < .001).
CONCLUSIONS: In patients with suspected congenital aqueductal stenosis, lack of marked third ventriculomegaly as conventionally measured can sometimes be explained by thickening of the thalamic massa intermedia. In this circumstance, it is important to evaluate the extreme recesses of the third ventricle for evidence of dilation on fetal MRI.
ABBREVIATIONS:
- CAS
- congenital aqueductal stenosis
- IQR
- interquartile range
- TMI
- thalamic massa intermedia
SUMMARY
PREVIOUS LITERATURE:
Third ventricle dilation is a compelling clue that supports a diagnosis of aqueductal stenosis in fetuses with normal ventricular and posterior fossa anatomy. The thalamic massa intermedia is a bridgelike structure across the third ventricle. Previous studies comparing third ventricle diameter and thalamic massa intermedia size in the general population reported controversial results.
KEY FINDINGS:
The third ventricle diameter was inversely proportional to the circumference of the thalamic massa intermedia. Nonetheless, dilation of the extreme recesses still occurred when the mid third ventricle was nondilated or less severely dilated in patients with an enlarged thalamic massa intermedia. Third ventricle dilation was most severe in patients lacking a thalamic massa intermedia.
KNOWLEDGE ADVANCEMENT:
In patients with suspected congenital aqueductal stenosis, lack of significant third ventriculomegaly as conventionally measured can sometimes be explained by thickening of the thalamic massa intermedia. In this circumstance, it is important to evaluate the extreme recesses of the third ventricle for dilation on fetal MRI.
Hydrocephalus is a disorder characterized by an imbalance between the production and resorption of CSF that results in pressurized distention of the ventricular system.1 The estimated prevalence of infantile hydrocephalus is 1.1 in 1000 infants, though this varies in different parts of the world.2 On imaging, hydrocephalus manifests as generalized ventriculomegaly or uneven expansion of ventricular system components with morphologic changes indicative of elevated intraluminal pressure with or without interstitial edema, depending on the underlying cause, severity, and duration.1 Therefore, it is prudent to evaluate each ventricle separately after a broader categorization into intraventricular or extraventricular hydrocephalus.
Congenital aqueductal stenosis (CAS) is one of the most common causes of infantile hydrocephalus, particularly in patients with nonsyndromic congenital hydrocephalus.1,2 Stenosis of the cerebral aqueduct may develop from intrinsic or, less commonly, extrinsic mechanisms and may result in increased supratentorial ventricular volume and pressure.2,3 Anatomic changes consequent to aqueductal stenosis may vary on the basis of the location of the luminal narrowing.3 In more chronic CAS processes, possible presentations are focal ventricular enlargement, pulsion diverticula, subependymal dissection, and spontaneous ventriculocisternostomies.3,4 Downward bulging of the dilated third ventricle into the interpeduncular cistern with or without pineal recess enlargement is the typical presentation of third ventricle distention in patients with aqueductal stenosis.3
Early detection of CAS is important because it might result in severe hydrocephalus-related complications.4,5 MRI is the preferred study for fetal diagnosis, with a reported sensitivity and specificity of 95% and 65%, respectively, in patients presenting with severe ventriculomegaly.5⇓-7 While postnatal MRI usually shows direct stenosis or funneling of the aqueduct, fetal diagnosis is mostly based on secondary changes. Third ventricle enlargement is an important secondary finding that can be diagnosed when the diameter exceeds 4 mm in the coronal plane at any age on fetal MRI or is >2 SDs above the mean for gestational age. Furthermore, the degree of third ventricle enlargement predicts the severity of CAS.
The thalamic massa intermedia (TMI) is a normal anatomic connection between the 2 parts of the thalamus.8,9 In children, the TMI is usually centered in the anterior‐superior third ventricle.8 Although TMI absence is a common variation in healthy individuals, the prevalence is higher in patients with midline malformations.8,9 The cause of TMI size variability and absence is still not well‐known. Previously, a few studies compared TMI size and third ventricle diameter in the general population, and different authors reported controversial results.10⇓-12 We have clinically observed an inverse relationship between TMI thickness and third ventricle size and, notably, cases of CAS without significant enlargement of the third ventricle transverse diameter. In this study, we evaluated the correlation between the size of the third ventricle and the TMI in the setting of CAS on fetal and postnatal MRI examinations.
MATERIALS AND METHODS
Study Design and Inclusion Criteria
This work resulted from a retrospective study conducted at Children’s Hospital of Philadelphia. The institutional review board of the hospital approved this study and waived the need for informed consent due to its retrospective nature. We included all patients who had at least 1 fetal and 1 postnatal MRI who had a reported aqueductal stenosis in 1 of the MRI reports. To obtain these data, we searched the radiology information system for the phrase “aqueductal stenosis” using Illuminate InSight (Illuminate). After obtaining the primary data, a pediatric neuroradiologist with 13 years of postfellowship experience (M.T.W.) and a radiologist research fellow (O.S.) re-evaluated all cases in consensus to confirm the presence of aqueductal stenosis by direct inspection of the aqueduct on postnatal imaging. The last fetal MRI and first postnatal MRI were chosen for the evaluation and measurements. Exclusion criteria were the existence of interhypothalamic adhesions, external lesions that could alter the diameter of the third ventricle, encephalocele, and holoprosencephaly. Interhypothalamic adhesions are bridgelike structures across the third ventricle, similar to the TMI;13 therefore, we excluded these cases under the presumption that they may be confounding by preventing or reducing the degree of third ventricle distention in patients with CAS. We did not find any instances of incorrectly reported aqueductal stenosis. The patient selection process is depicted in Fig 1.
Selection process of patients.
Imaging Data Acquisition
All fetal MRI examinations were performed at 1.5T while postnatal examinations were performed on either 1.5T or 3T MRI scanners (Siemens, Philips, GE Healthcare). We obtained measurements on the T2-weighted series of each MRI. The ranges of acquisition parameters for fetal MRIs were TE = 64–99 ms, TR = 1.1–1400 ms, slice thickness = 3–4 mm, and interslice gap = 3 mm. The ranges of acquisition parameters for postnatal MRIs were TE = 68–406 ms, TR = 3200–6400 ms, slice thickness = 1–2 mm, and interslice gap = 0–2 mm.
Massa Intermedia and Third Ventricle Measurements
We evaluated both fetal and postnatal MRIs to determine whether there was a visible TMI. All MR imaging examinations were evaluated and measured using NilRead (Version 22; Hyland Software). When present, we measured craniocaudal and anterior-posterior TMI diameters on sagittal T2-weighted series (Fig 2A). Circumference length and TMI area were calculated automatically on the basis of manually drawn ROIs on sagittal T2-weighted images. Measurements were recorded as 0 when the TMI was absent. Third ventricle measurements, including the central third ventricle diameter, supraoptic recess diameter, and suprapineal recess diameter, were obtained on coronal T2-weighted series in both groups (Fig 2B, -C, and -D). Specifically, the maximum transverse diameter of the central third ventricle (henceforward “central third ventricle diameter”) was measured in the coronal plane at or about the level of the hypothalamic sulcus in accordance with prior literature.14,15 The suprapineal recess was measured at its maximum diameter at or about the level of the pineal gland. The supraoptic recess was measured at its maximum diameter at or just anterior to the level of the optic chiasm.
Measurement methods in a 34-weeks’ gestational age female fetus with aqueductal stenosis. A, Sagittal T2WI shows a third ventricle with enlarged supraoptic (black arrowhead) and suprapineal recesses (white arrowhead). The black ellipse shows the measurement method of massa intermedia circumference and area; the long axis of the ellipse represents the craniocaudal length, and the short axis represents the anterior-posterior length. B, C, and D, Coronal T2WI of the same patient shows the supraoptic recess (white dotted line, B), suprapineal recess (white dotted line, C), and the central third ventricle diameter measurements (white dotted line, D).
Fetal MRI of a 22-weeks’-gestation male fetus with aqueductal stenosis (A, B, C, and D). A, Sagittal T2WI shows a thick TMI (star) and aqueductal stenosis with hydrocephalus. Coronal T2WI shows a small central third ventricle diameter (white arrow, B) but an enlarged supraoptic recess (black arrowhead, C) and a large supraspinal recess (curved white arrow, D). Postnatal 1-day-old sagittal T2WI (E) of the same patient confirms thickening of the TMI (star) and CAS, with milder central third ventricle dilation. Fetal MRI of a 21-weeks’-gestation female fetus with aqueductal stenosis (F, G, H, and I). F, Sagittal T2WI shows aqueductal stenosis and marked dilation of the third ventricle with no visible TMI. Coronal T2WI shows an enlarged central third ventricle (white arrow, G), an enlarged supraoptic recess (black arrowhead on H), and an enlarged suprapineal recess (curved white arrow, I). Postnatal 2-day-old sagittal T2WI (J) of the same patient confirms CAS with massive supratentorial ventriculomegaly and hydrocephalus.
Statistical Analysis
Nominal variables are presented as percentages, and quantitative data, as median and interquartile range (IQR). The Spearman rank correlation was used to assess the correlation between the TMI measurements and the central third ventricle diameter, suprapineal recess diameter, and supraoptic recess diameter. Mann-Whitney U tests were used to compare quantitative measurements between groups negative and positive for TMI. We used SPSS Statistics for Windows (Version 29.0; IBM) to analyze the data of this study. P < .05 was considered significant.
RESULTS
Fifty-nine patients and 118 MRI examinations were included in this study. Twenty-nine (49.2%) patients were male. The patients’ gestational age range at the time of fetal MRI was between 19 and 35 weeks with a median of 23 weeks (IQR, 21–27 weeks); at postnatal MRI, the median age was 7 days (IQR, 2–28 days). The TMI was present in 33 (55.9%) fetal and 31 (52.5%) postnatal cases. The medians of the central diameter of the third ventricle on fetal and postnatal MRIs were 3.9 mm (IQR, 2.8–6 mm) and 7.9 mm (IQR, 5.9–10.9 mm), respectively (P < .001) (Tables 1 and 2). The median area of TMI on fetal MRI was 18.4 mm2 (IQR, 9.2–29.75 mm2); on postnatal MRI, it was 37 mm2 (IQR, 15.7–69.4) mm2) (P < .001) (Tables 1 and 2).
TMI sizes in patients with CAS during the fetal and childhood periodsa
Third ventricle sizes in patients with CAS during the fetal and childhood periods
All TMI measurements had a significant moderate inverse correlation with the central third ventricle diameter. The highest correlation on fetal MRIs was between the anterior-posterior diameter of the TMI and the central third ventricle diameter (ρ = –0.456, P < .001); and on postnatal MRIs, it was between the circumference of the TMI and the central third ventricle diameter (ρ = –0.653, P < .001). Table 3 shows the correlation coefficients of different TMI measurements with the central diameter of the third ventricle. TMI measurements and supraoptic and suprapineal recess measurements showed no significant correlation in fetal MRIs. The correlation in the postnatal period was significant but weak, ranging from −0.313 to −0.371. Table 3 shows the correlation coefficients of different TMI measurements and supraoptic and suprapineal recesses.
Spearman correlation coefficients of the third ventricle measurements and TMI measurements for all MRIs, fetal subgroups, and postnatal subgroups
Comparison of third ventricle diameters between patients with and without TMI
The central diameter of the third ventricle was significantly larger in patients without a TMI (5.95 mm; range, 3.68–7.8 mm) compared with patients with a TMI (3.4 mm; range, 2.8–5.05 mm) on fetal MRI (P = .002), while the supraoptic (P = .463) and suprapineal (P = .171) recess diameters were not significantly different (Table 4). On postnatal MRIs, all third ventricle measurements were significantly larger in the TMI absent group (Online Supplemental Data).
DISCUSSION
Congenital aqueductal stenosis is an important cause of fetal and infantile hydrocephalus. Third ventricle enlargement is a helpful imaging finding to support the diagnosis of CAS, especially in cases in which the aqueduct is unevaluable or suboptimally assessed.3⇓-5,16 However, a substantial third ventriculomegaly does not occur in all cases of CAS, and dilation of the extreme recesses of the third ventricle does not always require an enormously enlarged third ventricle. We hypothesized that interindividual anatomic differences of the TMI could be associated with differences in the size and shape of the third ventricle and that TMI thickening would tend to constrain the adjacent third ventricle regions, impeding dilation. We found an inverse relationship between the size of the central third ventricle diameter and the TMI, even though there was similar third ventricle recess distention in fetuses with CAS supporting this notion.
Previous literature has similarly demonstrated the importance of third ventricle recess evaluation in patients with CAS. Heaphy-Henault et al5 compared the fetal imaging findings between CAS and non-CAS-associated hydrocephalus and concluded that third ventricle recess dilation is significantly more common in CAS than in other hydrocephalus etiologies. Together with our results, this finding suggests that third ventricle recesses deserve special scrutiny in patients with suspected CAS because recess dilation is relatively impervious to TMI thickening compared with the more conventionally measured transverse third ventricle diameter on fetal MRI. Therefore, lack of enlargement of the central transverse diameter of the third ventricle should not be used as an exclusion criterion for CAS, particularly in fetuses with prominent TMI. When CAS is suspected, all components of the third ventricle should be carefully evaluated in 3 planes for abnormal dilation. Earlier-onset third ventriculomegaly is sometimes appreciable in the sagittal plane by visibility of the anterior-inferior recesses that are typically absent or MR occult until later in the second trimester due to normal hypothalamic glioepithelial abutmen.17
Thinning of the TMI was commonplace in our cohort of patients with CAS. Moreover, nearly one-half lacked a TMI, significantly more frequently than the <20% prevalence reported in healthy children.8,9 Because our cohort included solely patients with CAS, these findings suggest that CAS and TMI absence are correlated, potentially due to developmental abnormalities of midline brain structures and/or pressure-induced TMI disruptions.
TMI was larger on postnatal than on fetal MRI examinations, potentially related to interval patient growth and development between studies and/or inherent limitations in fetal measurement acquisitions. However, third ventricle enlargement generally increased between the pre- and postnatal MRI examinations, probably reflecting progressive hydrocephalus superimposed on age-related differences in third ventricle size. Progressive ventricle distention disproportionate to TMI enlargement may account for the development of the inverse relationship between third ventricle recess dilation and TMI size seen only after birth. Alternatively, or additionally, third ventricle recess size discrepancies between the pre- and postnatal groups could be partially explained by inherent differences in slice thickness.
Cases of CAS with interhypothalamic adhesions, a potentially confounding variable, were eliminated from further review at the outset under the assumption that these may constrain the third ventricle from dilation in a manner similar to that of an interthalamic adhesion. Future studies would be useful to evaluate the effect of interhypothalamic adhesions on third ventricle size in patients with CAS.
This study has a few limitations. Our cohort included solely patients already identified as having CAS. Given that third ventricle dilation is a well-established, commonly used diagnostic criterion for aqueductal stenosis, some patients with CAS without third ventriculomegaly may have been excluded unintentionally. Such exclusions would be expected to cause the magnitude of the inverse relationship between TMI and third ventricle size reported here to be an underestimation, assuming TMI thickening as the culprit in patients with CAS without third ventriculomegaly. Nonetheless, cases of CAS without significant dilation of the central third ventricle were still represented in our data set, some of which were diagnosed prenatally and others postnatally. Additionally, our cohort included only fetuses between gestational weeks 19 and 35; therefore, future studies would be required to determine the validity of our findings in younger fetuses.
The current fetal MR imaging technique does not allow highly accurate or precise quantitative assessment of the TMI and third ventricle, particularly in younger/smaller developing brains and when motion artifacts are present. While we acknowledge that third ventricle recess measurement errors could result in millimetric differences among patients, we believe that the largest diameter measurements are a more important representation of disease state than measurements obtained at focal anatomic points. Thus, we targeted the measurement location description into 1–2 coronal slices but refrained from providing an exact anatomic point of measure because the more clinically relevant largest diameter does not always match an exact anatomic point. Therefore, we avoid suggesting specific thresholds or definitive values, instead emphasizing the significance of observing a generally enlarged or absent TMI in patients with CAS and recommending careful attention to the extreme recesses of the third ventricle.
CONCLUSIONS
In patients with suspected aqueductal stenosis, lack of significant third ventriculomegaly as conventionally measured can sometimes be explained by thickening of the TMI; thus, a normal transverse diameter of the central third ventricle should not be used as an exclusion criterion. In this circumstance, evaluating the extreme recesses of the third ventricle is important to determine if there is evidence of abnormal dilation, especially during fetal imaging.
Footnotes
Disclosure forms provided by the authors are available with the full text and PDF of this article at www.ajnr.org.
References
- Received March 15, 2024.
- Accepted after revision April 24, 2024.
- © 2024 by American Journal of Neuroradiology