Introduction

Foetal ventriculomegaly, defined as dilation of the cerebral ventricles (atrial diameter ≥ 10 mm), is a common cerebral anomaly on prenatal ultrasound, with an estimated prevalence of 0.3–1.5 per 1000 births1. Foetal ventriculomegaly is typically categorized as mild (10–12 mm), moderate (13–15 mm), or severe (> 15 mm) ventriculomegaly2. Foetal ventriculomegaly can also be classified as unilateral or bilateral ventriculomegaly as well as isolated or non-isolated ventriculomegaly3.

Foetuses with severe ventriculomegaly are known to have a poor prognosis in accordance with survival and neurodevelopmental outcome4. However, the prognosis for infants with mild-to-moderate ventriculomegaly is widely variable, which makes genetic counselling quite challenging in clinical practice. The prognosis depends on whether it is combined with structural abnormalities, especially central nervous system (CNS) anomalies, chromosomal aberrations, congenital infections of cytomegalovirus (CMV) or toxoplasmosis (TOX), and the progression of the ventricular dilation5.

Therefore, in this retrospective study, which comprehensively focused on aetiology, we systematically investigated the chromosomal aberrations of foetal ventriculomegaly by single nucleotide polymorphism (SNP) array, the additional CNS anomalies by magnetic resonance imaging (MRI), and the status of congenital foetal CMV infections. Furthermore, pregnancy outcomes were evaluated based on the follow-up information on foetuses with different types of ventriculomegaly.

Results

We analysed a total of 548 foetuses with ventriculomegaly by SNP array. The maternal age was between 18 and 46 years (median = 28.68 ± 4.49 years); 10.58% (58/548) of the gravidas were of advanced maternal age (35 years or older at delivery); 60.95% (334/548) of the gravidas were nulliparous; and the gestational age (GA) ranged from 22 to 34+5 weeks (median = 27.57 ± 5.78 weeks). The demographic and clinical characteristics of the gravidas is shown in Supplementary Table S1.

There were 509 (92.88%) gravidas who underwent foetal MRI of the CNS after 26 weeks GA. Eighteen (3.54%, 18/509) cases with additional CNS anomalies were detected, including 6 with agenesis of the corpus callosum, 5 with arachnoid cysts, 3 with hypoplasia of the corpus callosum, 2 with Dandy–Walker malformation, 1 with migrational abnormality, and another with holoprosencephaly.

In the 3 gravidas with serous positive IgM and low IgG avidity of CMV, CMV infection was confirmed in 2 cases by polymerase chain reaction (PCR) testing of amniotic fluid. As regards the 2 foetuses, one had moderate bilateral ventriculomegaly, hyperechogenic bowel, and hepatic calcifications, whereas the other had moderate bilateral ventriculomegaly and pericardial effusion.

Prevalence of chromosomal aberrations in foetal ventriculomegaly

The overall prevalence of chromosomal aberrations in foetuses with ventriculomegaly was 7.30% (40/548) (Table 1). Among all cases, numerical chromosomal aberrations were identified in 17 (3.10%), including 1 triploid and 3 with mosaic trisomy (Table 2). In addition, 23 (4.20%) foetuses were identified with pathogenic/likely pathogenic (P/LP) copy number variants (CNVs) and 2 (0.36%) with uncertain clinical significance (VUS).

Table 1 Summary of 750 K SNP array results in 548 fetuses with ventriculomegaly.
Table 2 Characteristics of fetuses with aneuploidies/polyploidy detected by 750 K SNP array among the 548 fetuses with ventriculomegaly.

In the 23 foetuses with P/LP CNVs, 14 were identified with deletions/duplications larger than 5 Mb, and these were also detected with karyotyping (Table 3). Karyotyping detected subtelomeric aberrations in 5 cases, including 3 confirmed cases (foetuses 30, 31, and 32) and 2 missed cases (foetuses 33 and 37). Furthermore, the 3 confirmed cases (foetuses 30, 31, and 32) were confirmed to be inherited from normal parents with balanced translocations. Among the other 9 cases subjected to karyotyping, 1 was missed (foetus 35) and 3 (foetuses 21, 26, and 40) with GA exceeding 32 weeks had cell culture failure.

Table 3 Characteristics of fetuses with P/LP CNVs detected by 750 K SNP array among the 548 fetuses with ventriculomegaly.

Thirteen types of microdeletion/microduplication syndromes were identified in 14 cases, including 1q21.1 recurrent microduplication, 3q29 microduplication syndrome, Cri-du-chat syndrome, 7q11.23 duplication syndrome, 8p23.1 duplication syndrome, 12p13.33 microdeletion syndrome, 15q26 overgrowth syndrome, 16p13.11 recurrent microduplication, 16p11.2 microduplication syndrome, Miller–Dieker syndrome (MDS), 17q12 recurrent deletion syndrome, 22q11 duplication syndrome, and steroid sulphatase deficiency (STS). Additionally, 3 cases were further confirmed by karyotyping among 5 cases with deletions/duplications > 5 Mb (Table 3).

Incidence of chromosomal aberrations in unilateral vs bilateral ventriculomegaly

The incidences of chromosomal aberrations in foetuses with unilateral ventriculomegaly and bilateral ventriculomegaly were 5.71% (21/368) and 10.56% (19/180), respectively (Table 1). The incidence in foetuses with unilateral ventriculomegaly was significantly lower than in those with bilateral ventriculomegaly (P = 0.040) (Fig. 1a). However, the incidences of P/LP CNVs were not significantly different between these groups [3.26% (12/368) vs. 6.11% (11/180), P = 0.118] (Fig. 1b).

Figure 1
figure 1

The incidences of chromosomal aberrations in subgroups of foetuses with ventriculomegaly. (a) The incidence of chromosomal aberrations in the unilateral ventriculomegaly group [5.71% (21/368)] was significantly lower than in the bilateral ventriculomegaly group [10.56% (19/180)] (P = 0.040); (b) the incidences of P/LP CNVs were not significantly different between the unilateral ventriculomegaly group [3.26% (12/368)] and bilateral ventriculomegaly group [6.11% (11/180)] (P = 0.118); (c) the incidence of chromosomal aberrations in the isolated ventriculomegaly group [2.38% (7/294)] was significantly lower than in the non-isolated ventriculomegaly group [12.99% (33/254)] (P < 0.0001); (d) the incidence of P/LP CNVs in the isolated ventriculomegaly group [1.36% (4/294)] was significantly lower than in the non-isolated ventriculomegaly group [7.48% (19/254)] (P < 0.0001); (e) The incidence of chromosomal aberrations in the mild ventriculomegaly group [5.92% (26/439)] was significantly lower than in the severe ventriculomegaly group [23.08% (6/26)] (P = 0.003), whereas the incidence in the mild-to-moderate ventriculomegaly group [6.51% (34/522)] was significantly lower than that in the severe ventriculomegaly group [23.08% (6/26)] (P = 0.005); (f) The incidences of P/LP CNVs were not significantly different among the mild ventriculomegaly [3.87% (17/439)], moderate ventriculomegaly [3.61% (3/83)], and severe ventriculomegaly [11.54% (3/26)] groups (mild vs. moderate ventriculomegaly, P = 1.000; mild ventriculomegaly vs. severe ventriculomegaly, P = 0.169; and moderate ventriculomegaly vs. severe ventriculomegaly, P = 0.146).

Incidence of chromosomal aberrations in isolated vs non-isolated ventriculomegaly

The incidences of chromosomal aberrations were 2.38% (7/294) and 12.99% (33/254) in foetuses with isolated and non-isolated ventriculomegaly, respectively (Table 1). The incidence in foetuses with isolated ventriculomegaly was significantly lower than in those with non-isolated ventriculomegaly (P < 0.0001) (Fig. 1c). Likewise, the incidence of P/LP CNVs in foetuses with isolated ventriculomegaly was significantly lower than in those with non-isolated ventriculomegaly [1.36% (4/294) vs.7.48% (19/254), P < 0.0001] (Fig. 1d).

Subgroup analysis of different severities of foetal ventriculomegaly

The incidences of chromosomal aberrations were 5.92% (26/439), 9.64% (8/83), and 23.08% (6/26) in foetuses with mild, moderate, and severe ventriculomegaly, respectively (Table 1). The incidence in foetuses with mild ventriculomegaly was significantly lower than in those with severe ventriculomegaly (5.92% vs. 23.08%, P = 0.003), whereas that in foetuses with mild-to-moderate ventriculomegaly was significantly lower than in those with severe ventriculomegaly [6.51% (34/522) vs. 23.08% (6/26 ), P = 0.005]. However, statistical analysis showed that these incidences were not significantly different between foetuses with mild ventriculomegaly and those with moderate ventriculomegaly (5.92% vs. 9.64%, P = 0.208) or between foetuses with moderate ventriculomegaly and those with severe ventriculomegaly (9.64% vs. 23.08%, P = 0.147) (Fig. 1e).

The incidences of P/LP CNVs were 3.87% (17/439), 3.61% (3/83), and 11.54% (3/26) in foetuses with mild, moderate, and severe ventriculomegaly, respectively. However, statistical analysis showed that these incidences were not significantly different from each other (mild vs. moderate ventriculomegaly, P = 1.000; mild vs. severe ventriculomegaly, P = 0.169; and moderate vs. severe ventriculomegaly, P = 0.146) (Fig. 1f).

The clinical follow-up assessments

Clinical follow-up assessments were performed on 495 (90.33%) cases (Table 4). Except for the foetus (foetus 23) confirmed with 16p13.11 recurrent microduplication (neurocognitive disorder susceptibility locus) inherited from normal mother and whose mild bilateral ventriculomegaly was resolved at 33 GA resulting in a favourable outcome at birth, the other 39 foetuses with chromosomal aberrations all underwent termination of pregnancy (TOP). The ventriculomegaly of one foetus with VUS progressing from mild to severe underwent TOP at 34 GA.

Table 4 Clinical follow-up assessment of the 548 fetuses with ventriculomegaly.

The rate of normal infant was 81.24% (368/453) in foetuses without chromosomal aberrations, except those lost to follow-up. The 2 cases confirmed with CMV infection underwent TOP. The rates of normal infant in isolated and non-isolated ventriculomegaly group were 95.76% (226/236) and 91.61% (142/155), respectively. The outcomes of foetuses were not significantly different between these groups (P = 0.088). In foetuses with isolated ventriculomegaly, the most common reason for TOP was progressive ventriculomegaly. Additional CNS anomalies detected by MRI included agenesis/hypoplasia of the corpus callosum, Dandy–Walker malformation, migrational abnormality, and holoprosencephaly. Meanwhile, as all the 20 foetuses with severe ventriculomegaly underwent TOP, the rates of normal infant in the mild and moderate ventriculomegaly groups were 95.60% (326/341) and 84.00% (42/50), respectively. The outcome of foetuses with mild ventriculomegaly was significantly better than that of those with moderate ventriculomegaly (P = 0.003).

Additionally, for foetuses without chromosomal aberrations, except those that underwent TOP or intrauterine foetal death, the rate of foetuses born with clinical defect was 5.88% (23/391). The clinical manifestations included developmental delay, intellectual disability, delayed speech and language development, autism, epilepsy, hyperactivity disorder, hydrocephalus, etc. All these infants were diagnosed with mild-to-moderate ventriculomegaly that never aggravated to severe ventriculomegaly. The rate of clinical defects after birth in the mild ventriculomegaly group was significantly lower than in the moderate ventriculomegaly group [4.40% (15/341) vs. 16.00% (8/50), P = 0.003]; however, it was not significantly different between the isolated and non-isolated groups [4.24% (10/236) vs. 8.39% (13/155), P = 0.088].

Discussion

The Society for Maternal-Foetal Medicine (SMFM) has recommended that once ventriculomegaly is detected, a thorough evaluation should be performed, including detailed sonographic evaluation of foetal anatomy, amniocentesis for the assessment of chromosomal aberrations, and a workup for foetal infection5. However, few studies have comprehensively explored all the causes of ventriculomegaly. In this study, we systemically investigated the aetiological factors, including chromosomal aberrations by SNP array, foetal CNS structural abnormalities by MRI, and foetal CMV infections by PCR in order to improve prenatal diagnosis and clinical counselling.

The prevalence of MRI-detected additional CNS anomalies was 3.54% (18/509), which was lower than the results from previous studies that ranged from 5 to 10%1, 6, 7. While more than 80% of cases included in our study had mild ventriculomegaly (10–12 mm), selection bias may explain the difference. In our study, the most common abnormality missed by ultrasound but detected by MRI was agenesis/hypoplasia of the corpus callosum (9/18), which may induce poor outcome. Moreover, another foetus was identified with migrational abnormality, which has rarely been detected by ultrasound. On account of the MRI-detected additional CNS anomalies associated with adverse outcome, pregnancies were terminated, except for the 5 cases with arachnoid cysts. Thus, foetal MRI for CNS has been proven as a useful tool to identify additional structural anomalies, a quality that may lead to modification of pregnancy management.

The overall prevalence of chromosomal aberrations was 7.30% (40/548), comprising 42.50% chromosomal numerical aberrations and 57.50% P/LP CNVs. Further, karyotyping was performed on 14 cases with P/LP CNVs exceeding 5 Mb, including 5 with microdeletion/microduplication syndromes. However, 3 cases, including 2 with subtelomeric aberrations, were missed (foetuses 33, 35, and 37) and 3 at third-trimester had cell culture failure. Combined with those P/LP CNVs < 5 Mb, we estimated that 15 (2.73%) cases with chromosomal aberrations would be misdiagnosed if karyotyping alone was used in our study. Thus, CMA should be preferred to karyotyping as a first-tier test for prenatal diagnosis of foetal ventriculomegaly.

Until recently, several studies have focused on the relationship between foetal ventriculomegaly and chromosomal aberrations and have reported prevalence of chromosomal aberrations ranging from 5.1 to 17.4%8,9,10,11,12,13,14. We speculate that such a wide range could be due to different distribution of the ventriculomegaly degrees or combined ultrasound abnormalities among different studies. Further subgroup analysis was performed in our study.

Consistent with data presented by Chang et al.8, the incidences of chromosomal aberrations in foetuses with bilateral ventriculomegaly (10.56%) was significantly higher than in those with unilateral ventriculomegaly (5.71%). The incidences of both chromosomal aberrations and P/LP CNVs in foetuses with non-isolated ventriculomegaly (12.99% and 7.48%, respectively) were significantly higher than in those with isolated ventriculomegaly (2.38% and 1.02%, respectively). Wang et al.15 conducted a review and revealed that the incidence of chromosomal aberrations in foetuses with ventriculomegaly associated with other structural anomalies (9.5–36%) was higher than in those with isolated ventriculomegaly (1.5–12%)16,17,18, with the similar results for the incidence of P/LP CNVs (6.6–37.9% vs. 4–9.5%)9, 10, 19, 20. Thus, CMA should be highly recommended for prenatal diagnosis of foetal ventriculomegaly, especially in foetuses with bilateral or non-isolated ventriculomegaly.

Notably, the incidences of chromosomal aberrations in different degrees of foetal ventriculomegaly were still discordant. Several studies reported that the incidence was higher in mild-to-moderate ventriculomegaly than in severe ventriculomegaly10, 19; however, contradicting results were reported by others20, 21. In our study, the incidence of both chromosomal aberrations and P/LP CNVs in foetuses with severe ventriculomegaly (23.08% and 11.54%, respectively) was higher than in those with mild-to-moderate ventriculomegaly (5.92% and 3.47%, respectively); however, no significant difference was in the P/LP CNV incidences between these two groups. As the number of cases with severe ventriculomegaly was small in our study, large-scale studies are still required to determine the relationship between the incidence of chromosomal aberrations and the degrees of foetal ventriculomegaly.

Several CNVs have been reported to be associated with foetal ventriculomegaly, for example 22q11.2 microdeletion, 15q11.2 microdeletion, 16p13.11 microdeletion, 16p11.2 microdeletion, and 1q21.1 microdeletion12, 13, 22, 23. In our study, 13 types of microdeletion/microduplication syndromes were identified in 14 cases. Ten have been recognized as significant contributors to poor neurodevelopmental outcome, including 1q21.1 recurrent microduplication, 3q29 microduplication syndrome, Cri-du-chat syndrome, 7q11.23 duplication syndrome, 8p23.1 duplication syndrome, 12p13.33 microdeletion syndrome, 16p13.11 recurrent microduplication, 16p11.2 microduplication syndrome, Miller–Dieker syndrome (MDS), and 22q11 duplication syndrome. However, whereas 16p13.11 recurrent microduplication has low penetrance24, feotus 23 that inherited this CNV form normal mother was born with favourable outcomes. Additionally, a de novo 12q24.2 microdeletion of 392 kb was detected in foetus 36 with unilateral moderate ventriculomegaly and talipes equinovarus, encompassing exons 24–31 of mediator of RNA polymerase II transcription subunit 13-like (MED13L) gene (OMIM: 608771). As a haploinsufficient gene, MED13L is linked to mental retardation and distinctive facial features, with or without cardiac defects (MIM: 616789). However, it was not sufficiently powered to address the association of specific CNVs with foetal ventriculomegaly.

Postnatal follow-up from 6 months to 3 years after birth was performed in 495 (90.33%) cases in our study. TOP was opted for all the cases with severe ventriculomegaly due to poor prognosis both in terms of survival and neurodevelopmental disorders3, 4. The rate of abnormal clinical outcome in foetuses with mild-to-moderate ventriculomegaly was 5.88%, which concurred with previous studies that had rates ranging from 4 to 11%1, 7, 16, 25, 26. The main clinical manifestation was neurodevelopmental delay. Furthermore, we compared the clinical outcome in foetuses with mild ventriculomegaly to that in those with moderate ventriculomegaly. Consistent with previous studies7, 16, 18, 27, the rate of normal outcome in mild ventriculomegaly (95.60%) was significantly higher than in moderate ventriculomegaly (84.00%). Our results corroborated a previous submission5 that the outcome of foetuses with moderate ventriculomegaly is favourable; however, an increased risk of neurodevelopmental disabilities does exist.

This study still has a few limitations. First, we only detected CMV in the amniotic fluid of gravidas with low IgG avidity and positive IgM for CMV; therefore, other cases with CMV infection might have been missed in our study. Congenital TOX infection has also been confirmed to be associated with foetal ventriculomegaly28, 29. According to SMFM recommendations5, the tests for CMV and TOX should be performed in further studies, regardless of the history of exposure or the presence of symptoms. Second, a number of single-gene disorders have been associated with ventriculomegaly, such as L1 cell adhesion molecule (L1CAM)-associated hydrocephalus, Pettigrew syndrome, Walker–Warburg syndrome, Meckel syndrome, Joubert syndrome, etc.30, 31. These conditions have always been associated with severe ventriculomegaly and mostly untreatable; thus, genetic diagnoses are crucial for the parents to assess the risk of recurrence and to facilitate prenatal or preimplantation diagnosis when planning the next pregnancy. Gene sequencing, as a useful approach to remedy the shortage under this condition, should be implemented in future studies.

Conclusion

When a foetal ventriculomegaly is confirmed by ultrasound, CMA should be regarded as the first-tier test for prenatal diagnosis, especially in foetuses with bilateral ventriculomegaly and non-isolated ventriculomegaly. Tests for infection such as CMV with amniotic fluid samples should be considered in combination with prenatal CMA. MRI is a vital tool in identifying additional CNS anomalies that are associated with adverse outcome. The outcome of foetuses with mild ventriculomegaly is favourable; nonetheless, there is an increased risk of neurodevelopmental disabilities in foetuses with moderate ventriculomegaly.

Methods

Subjects

Between September 2014 and December 2018, a total of 548 singleton pregnancies with foetal ventriculomegaly detected by ultrasound and recruited from the Prenatal Diagnosis Center of West China Second University Hospital, Sichuan University, were included in this study. Pre-test counselling was conducted by trained clinical geneticists. All foetal samples were obtained through amniocentesis for CMA testing at a GA from 22 to 34+5 weeks, which was calculated according to crown-rump length in the first trimester. Clear amniotic fluid samples were tested directly, whereas blood-stained amniotic fluid samples were cultured before CMA testing. Additionally, the peripheral blood samples of the parents were obtained to confirm the inherited or de novo CNVs of the foetuses and determine the clinical significance. The flow chat is shown in Fig. 2.

Figure 2
figure 2

The flow chart of the study.

Lateral ventricle atria were measured on an axial plane at the level of the thalamic nuclei, just below the standard image, to measure the biparietal diameter (BPD). When foetal ventriculomegaly were suspected, ultrasonography and echocardiography were performed by two foetal sonographers (Voluson E8, GE Medical Systems, Zipf, Austria) to confirm the diagnosis in the hospital, as per the guidelines of the International Society of Ultrasound Obstetrics and Gynaecology5, 32. Furthermore, foetal MRI of the CNS was recommended after 26 weeks GA5, 33.

There were 368 (67.15%) foetuses with unilateral ventriculomegaly and 180 (32.85%) with bilateral ventriculomegaly. The foetuses were categorized into 2 groups: isolated ventriculomegaly (n = 294) and non-isolated ventriculomegaly (n = 254). Non-isolated ventriculomegaly anomalies included structural anomalies in other system(s) or sonographic soft markers. The soft markers included intracardiac echogenic focus, increased nuchal fold, hyperechogenic bowel, mild pyelectasis, short femur, short humerus, aberrant right subclavian artery, absent or hypoplastic nasal bone, single umbilical artery, choroid plexus cysts, and enlarged cisterna magna34, 35. Concomitantly, the foetuses were also categorized into the following 3 groups according to the atrial diameter of the lateral ventricle: mild ventriculomegaly (10–12 mm) (n = 439), moderate ventriculomegaly (13–15 mm) (n = 83), and severe ventriculomegaly (> 15 mm) (n = 26).

Written informed consent was obtained from all gravidas before testing. The study was approved by the Medical Ethics Committee of West China Second University Hospital, Sichuan University. The research was conducted in accordance with the relevant guidelines and clinical norms.

Chromosomal microarray analysis

Genomic DNA of foetuses and parents was extracted from amniotic fluid and peripheral blood samples using the QIAamp DNA Blood Mini Kit (Qiagen, Valencia, CA, USA) and subsequently subjected to SNP array analysis using the CytoScan 750 K Array (Thermo Fisher Scientific, Santa Clara, CA, USA). The procedures were conducted according to the manufacturers’ instructions with 250 ng of gDNA, as described in our previous publication36. The molecular karyotype analysis was performed using Chromosome Analysis Suite (ChAS) v4.1 (ThermoFisher Scientific). The GRCh37 (hg19) genome was used for annotation. CNVs larger than 100 kb or that affected more than 50 contiguous probes were considered.

The detected CNVs were systematically evaluated for clinical significance based on scientific literature review and the following public databases: PubMed (https://www.ncbi.nlm.nih.gov/pubmed/), UCSC Genome Browser (https://genome.ucsc.edu/), Online Mendelian Inheritance in Man (OMIM) (https://omim.org/), Database of Genomic Variants (https://dgv.tcag.ca/dgv/app/home), DECIPHER (https://decipher.sanger.ac.uk/browser), GeneReviews (https://www.ncbi.nlm.nih.gov/books/NBK1116/), ClinVar (https://www.ncbi.nlm.nih.gov/clinvar/), and ClinGen (https://www.ncbi.nlm.nih.gov/projects/dbvar/clingen/). The clinical significances of detected CNVs were classified into 5 categories: pathogenic CNVs (P CNVs), likely pathogenic CNVs (LP CNVs), VUS, likely benign CNVs, and benign CNVs. This classification was conducted according to the guidelines of the American College of Medical Genetics and Genomics as follows: (1) initial assessment of genomic content, (2) established triplosensitive, haploinsufficient, or benign genes or genomic regions, (3) gene number, (4) cases from published literature, public databases, and/or internal lab data, and (5) inheritance patterns/family history for patient37. The CNVs defined as likely benign or benign were not reported.

Fluorescence in situ hybridization analysis (FISH)

This procedure was also described in our previous study36.

Chromosomal karyotyping

This procedure was also described in our previous study36.

lgM and lgG antibodies of TORCH agents detection

This procedure was described in Supplementary Methods.

Polymerase chain reaction (PCR) test for CMV-DNA

This procedure was also described in Supplementary Methods.

Clinical follow-up assessments

Clinical follow-up assessments via telephone were scheduled and conducted from 6 months to 3 years after birth to evaluate pregnancy outcomes and circumstances after birth, including postnatal imaging and developmental details, etc.

Statistical analysis

Statistical analysis was performed using SPSS software, version 24.0. Continuous data were presented as the means ± standard deviations. Comparison of continuous variables between groups was performed using the two-tailed Student’s t test and analysis of variance using Levene’s test. If the distribution was extremely skewed, a nonparametric two–tailed Mann–Whitney U test was used. Comparison of categorical variables between groups was performed using the Χ2 or Fisher exact test. A value of P < 0.05 was considered to indicate statistical significance.