Transorbital point-of-care ultrasound versus fundoscopic papilledema to support treatment indication for potentially elevated intracranial pressure in children

Purpose To compare transorbital point-of-care ultrasound techniques —optic nerve sheath diameter (US-ONSD) and optic disc elevation (US-ODE)— with fundoscopic papilledema to detect potentially raised intracranial pressure (ICP) with treatment indication in children. Methods In a prospective study, 72 symptomatic children were included, 50 with later proven disease associated with raised ICP (e.g. pseudotumour cerebri, brain tumour, hydrocephalus) and 22 with pathology excluded. Bilateral US-ONSD and US-ODE were quantified by US using a 12-MHz-linear-array transducer. This was compared to fundoscopic optic disc findings (existence of papilledema) and, in 28 cases, invasively measured ICP values. Results The sensitivity and specificity of a cut-off value of US-ONSD (5.73 mm) to detect treatment indication for diseases associated with increased ICP was 92% and 86.4%, respectively, compared to US-ODE (0.43 mm) with sensitivity: 72%, specificity: 77.3%. Fundoscopic papilledema had a sensitivity of 46% and a specificity of 100% in this context. Repeatability and observer-reliability of US-ODE examination was eminent (Cronbach’s α = 0.978–0.989). Papilledema was detected fundoscopically only when US-ODE was > 0.67 mm; a US-ODE > 0.43 mm had a positive predictive value of 90% for potentially increased ICP. Conclusion In our cohort, transorbital point-of-care US-ONSD and US-ODE detected potentially elevated ICP requiring treatment in children more reliably than fundoscopy. US-ONSD and US-ODE indicated the decrease in ICP after treatment earlier and more reliably than fundoscopy. The established cut-off values for US-ONSD and US-ODE and a newly developed US-based grading of ODE can be used as an ideal first-line screening tool to detect or exclude conditions with potentially elevated ICP in children.

Point-of-care (PoC) ultrasound of the optic nerve sheath diameter (US-ONSD) provides an alternative non-invasive, rapid and reliable method to estimate ICP in children [14][15][16], demonstrating a good correlation to invasively measured ICP [17][18][19] and a high intra-and inter-observer reliability [15].Transorbital ultrasound can also be used to detect an elevation of the optic disc (US-ODE) [20], and this technique has been described as a useful adjunct to papilledema evaluation to diagnose potentially elevated ICP [21].
This study compares the diagnostic accuracy of US-ONSD, US-ODE, and fundoscopic papilledema in children with symptoms compatible with raised ICP to detect intracranial pathology with treatment indication.
In addition, we investigated the repeatability and reliability of US-ODE measurement and sought to define US-ONSD and US-ODE cut-off values associated with potentially elevated ICP and the presence of fundoscopic papilledema.

Study design
Paediatric patients were included in this prospective observational study (2018)(2019)(2020)(2021) according to the following criteria: All patients presented with clinical symptoms suggestive for raised ICP (headache, vomiting, irritability, visual disturbance, e.g.blurred/double vision).Some of them appeared for the first time, some were already known and pre-treated, and all required further investigation due to the duration or severity of symptoms.All patients underwent fundoscopy by independent ophthalmologists (inpatient/ outpatient).Additionally, all patients underwent transorbital ultrasound of ONSD and ODE by an investigator blinded to fundoscopy.US findings were not included in decision making regarding further diagnostics and treatment.Only patients tolerating US investigations without sedation were included.The maximum time-interval between fundoscopy and transorbital US was 24 h, with no ICP-affecting therapy in between.Depending on kind, duration and severity of symptoms and fundoscopic findings, further imaging diagnostics were performed.
In 28 patients, an additional invasive ICP measurement (by intraparenchymal ICP probe) was performed, either because the diagnostic work-up required it, e.g. in PTC [22], or the preceding diagnostics did not yield clear findings (e.g.arachnoid cyst or clinical suspicion of craniostenosis in absence of papilledema).
According to findings and treatment decisions, patients were stratified into 2 groups: treatment group (TG) and nontreatment group (NTG).
Inclusion criteria for TG included the following in addition to the existing symptoms as described above: (1) imaging evidence of disease/disease-recurrence associated with increased ICP (hydrocephalus, space-occupying arachnoid cyst, brain tumour, craniosynostosis with copper-beaten skull or involvement of multiple sutures) and/or (2) evidence of raised intracranial pressure ≥ 18.5 mmHg (≥ 20.5 mmHg if child was sedated and/or obese) and/or (3) diagnosis of PTC according to the Friedman criteria [22].
Inclusion criteria for NTG encompassed the following: (1) absence of fundoscopic papilledema and absence of radiological evidence of a disease associated with increased ICP or (2) absence of fundoscopic papilledema, absence of imaging findings and intracranial pressure < 18.5 mmHg.
Informed consent was obtained from parents and children old enough to understand the study prior to US-investigation.According to the written ophthalmologic reports, fundoscopic findings were classified into (A) papilledema (in any manifestation) and (B) no papilledema.
All procedures were in accordance with the Code of Ethics of the World Medical Association (Declaration of Helsinki).The study protocol was approved by the institutional ethics committee (180/2018BO2).The study meets the STROBE guidelines for reporting observational studies.
In 40/50 patients of the TG, fundoscopy, US-ONSD and US-ODE were acquired before and after ICP decreasing therapy.In 24 of those 40 patients, at least one further follow-up investigation was performed (maximum follow-up interval 50 weeks).

PoC ultrasound
One examiner (SRK), highly experienced in transorbital ultrasound, performed all US-investigations.US was performed in supine position, head straight and not elevated.A 12-MHz linear transducer (Philips, Epiq 5G) was placed to the closed eyelid.US-ONSD determination was performed 3 mm behind the bulb, 90° to the optic nerve in axial plane (Fig. 4e).From three measurements per eye, mean ONSD and resulting mean binocular US-ONSD were calculated as previously described [23].Additionally, the optic disc elevation was measured three times per side (Fig. 4e) and mean binocular US-ODE was calculated.

Sample size calculation
The sample size estimation for comparison of US-ONSD means was calculated at 16 patients per group to achieve 99% power with an alpha of 0.1 (p < 0.001).The sample size estimate for comparison of US-ODE means was calculated at 22 patients per group to achieve 95% power with an alpha of 1 (p < 0.01).With an estimated loss of 10% due to exclusion criteria and data loss, the study aimed to recruit a minimum of 50 patients.

Statistical analysis
SPSS (PASW Statistics 27, IBM) statistical software was used.Data were tested for normal distribution using the Kolmogorov-Smirnov or Shapiro-Wilk test.Parametric data were reported as means ± standard deviation (sd).Depend- ing on normality of distribution, correlation of parametric variables was tested using Pearson's or Spearman's correlation; point-biserial correlation r was used for comparing parametric to nominal variables.Receiver operating characteristic (ROC) curves were created to define US-ONSD and US-ODE cut-off values according to Youden-Index calculation.Analysis of variance (ANOVA) was used to determine intra-observer variability of US-ODE measurement.Cronbach's alpha (α), a model of internal consistency, was used to test repeatability and observer reliability of US-ODE determination.Repeatability and reliability are considered reasonable when Cronbach's alpha is > 0.7, strong if > 0.8 and excellent if > 0.9.The Mann-Whitney U test and dependent t test were used for comparing mean values or paired samples, respectively.Chi-Quadrat-test was applied for comparing qualitative parameters.Bonferroni's correction was used to correct multiple comparisons.Statistical significance was set at p < 0.05: p < 0.05*, p < 0.01**, p < 0.001***.

Repeatability and intra-observer reliability of US-ODE measurement
Repeatability of US-ODE measurements was excellent with Cronbach's α = 0.978 (left) and 0.989 (right).A scanrescan reproducibility analysis of two US-ODE measures on each side yielded a mean bias of 0.006 mm (range − 0.24 to 0.4 mm) and 0.001 mm (range − 0.3 to 0.2 mm) left and right side, respectively.
To test for intra-observer variability, sequential measurements of US-ODE were performed in 20 random patients at two different timepoints, without therapy affecting ICP in between.The interval between investigations was 5 days to 6 months.The examiner was blinded to results obtained at the first timepoint.Mean US-ODE was 0.22 ± 0.33 mm at first and 0.19 ± 0.32 mm at second investigation.Testing for intra-observer reliability demonstrated eminent correlation (Cronbach's α = 0.985).

US-ODE cut-off value to detect papilledema by fundoscopy
We sought to determine the minimal US-ODE that manifests as fundoscopic papilledema.ROC analysis identified 0.675 mm as cut-off value to detect papilledema by fundoscopy with excellent diagnostic accuracy (Fig. 3a).

ROC analysis of US-ONSD, fundoscopic papilledema and US-ODE to detect potentially raised ICP
The optimal US-ONSD cut-off value to detect potentially raised ICP was 5.73 mm with excellent diagnostic accuracy (Fig. 3b).Fundoscopic papilledema showed a specificity of 100% to detect potentially raised ICP, with a sensitivity of 46%.The ideal US-ODE cut-off value was 0.43 mm with good diagnostic accuracy, better than papilledema on fundoscopy but inferior to US-ONSD.The combination of US-ONSD and US-ODE (AUROC: 0.745, sensitivity: 70%, specificity: 78.9%) could not increase diagnostic accuracy compared to US-ODE alone.Based on the results from "US-ODE cut-off value to detect papilledema by fundoscopy" and "ROC-analysis of US-ONSD, fundoscopic papilledema and US-ODE to detect potentially raised ICP" Sections, a US-based grading system of optic disc elevation was established (Fig. 4).
At end of follow-up, four patients had persistent papilledema (2 patients each with initial tumour and PTC; time intervals were 2, 8, 12, and 24 weeks after therapy).However, US-ONSD significantly decreased to normal values (5.29 ± 0.4 mm) in all.US-ODE significantly decreased as well (0.28 ± 0.25 mm, p < 0.001), but the four children with persistent papilledema had significantly elevated US-ODE (0.64 ± 0.14 mm) compared to those without (0.19 ± 0.2 mm).

Discussion
This study is the first to compare the diagnostic validity of US-ONSD, US-ODE, and fundoscopic papilledema to detect raised ICP with treatment indication in children.It is the largest prospective paediatric cohort used to define US-ONSD and US-ODE cut-off values to identify elevated ICP in need of treatment.Regression dynamics for all parameters after therapy were analysed.Finally, a US-based classification of the optic disc elevation as additional first-line screening tool for raised ICP in children was developed.

US-ONSD is an excellent non-invasive screening tool for detecting potentially elevated ICP requiring treatment
In children, symptoms like headache and vomiting are common and unspecific [24] but can indicate a significant ICP increase.The standard non-invasive, diagnostic tool is fundoscopy.The misconception that absence of papilledema is diagnostic to rule out raised ICP is still very common [25], although it is known that increased ICP may occur without papilledema in 40-60% of cases [8,[11][12][13].In an infant study, only 1/56 children with raised ICP had fundoscopic papilledema [10].Part of the pathophysiology of papilledema is the translation of ICP to the peri-optic subarachnoid space inside the dural optic nerve sheath (ONS) [26].A pressure rise within the ONS results in a widening of the elastic ONS and increase of ONSD.The response time of the ONS to pressure changes is prompt due to the elasticity-coefficient of the dura [27].As one result of persistently raised pressure within the ONS, an optic disc swelling can occur as a secondary late event [28].US-ONSD, which measures the primary event in the pathophysiological chain that may lead to papilledema, is a reliable tool for assessing ICP in children with excellent intra-individual correlation of ONSD and ICP and a rapid decline in ONSD after ICP reduction [29].Furthermore, in presence of optic nerve atrophy, papilledema is unlikely to occur at all, whereas an increase of ONSD will still occur in response to raising ICP [28].
The use of ONSD measured via US or MRI/CT [30] has been described in children [31][32][33], but a prospective comparison between ONSD, ODE and papilledema in symptomatic patients with/without potentially elevated ICP has not yet been made.
We demonstrate that children with confirmed underlying diseases associated to increased ICP had significant higher US-ONSD values (mean 6.5 ± 0.63 mm) compared to those without disease (mean 5.31 ± 0.38 mm, comparable to reference ONSD cut-off values of 5.2/5.3mm for ICP < 10 mmHg [17,18]).Most importantly, and in agreement with literature [34], more than 50% of the patients of the TG had no fundoscopic papilledema.These patients had -accordingly -significantly lower US-ODE values.
A correlation between US-ODE and fundoscopic papilledema is existent, and a highly diagnostic cut-off value for US-ODE was identified above which papilledema was found fundoscopically.The correlation analysis between US-ONSD, US-ODE and papilledema with invasively measured ICP showed convincing results for US-ONSD, but not for US-ODE or fundoscopic papilledema.Figure 2 shows that individual children with ICP values between 25 and 45 mmHg had no papilledema and would have been missed using fundoscopy as only screening tool.The ROC analysis revealed that the diagnostic accuracy of an US-ONSD cut-off value of 5.73 mm is clearly superior compared to the use of US-ODE or fundoscopy.This US-ONSD cut-off value is concurrent to a previously published cut-off value of 5.75 mm to detect ICP > 20 mmHg in children [17,18].

US-ONSD is an outstanding tool to monitor treatment outcome
Confirmation of success after treatment of elevated ICP is of clinical importance.Several studies described a relevant reduction of ONSD after treatment [29,31,35], and an immediate ONSD decrease after ICP lowering is known [29].The mean time-interval for papilledema development is 7-10 days [36], indicating that an acute ICP elevation is not rapidly followed by papilledema.The resolution time of papilledema is even longer with 6-10 weeks [28,37,38].
In our cohort, elevated US-ONSD values decreased significantly after therapy, but were still above normal 1-14 days after therapy.Of this cohort, 47.5% had no initial papilledema.Papilledema persisted in 19/21 patients at first investigation after therapy.After a follow-up of 2-50 weeks, US-ONSD values normalized (5.29 ± 0.4 mm), whereas 4 patients had persistent papilledema (2-24 weeks after therapy).Most importantly, US-ONSD values did not differ between patients with/without papilledema.These results suggest that in contrary to papilledema, treatment success is immediately reflected by an ONSD decrease.The regression of elevated ONSD values shows a hysteresis effect, since a longer-lasting ONS distention leads to a (transient) loss of dural elasticity in terms of a memory effect, which precludes a fast return to normal ONSD values.Hysteresis probably depends on extent and duration of ICP increase [39] and type of pathology [29].
In summary, US-ONSD can, unlike papilledema, reliably demonstrate a treatment success in very early stages, and it can be quantified in numbers.

ODE can be measured easily and reliably with US
A few recent studies reported on US-ODE examination as a tool to detect papilledema and increased ICP in adults [40][41][42] and children [20].Tessaro et al. published a retrospective pilot-study on 40 children and compared US-ODE and US-ONSD values with invasively measured ICP [43].Twenty-six children had elevated ICP, and the optimal ODE cut-off to detect raised ICP was 0.66 mm (sensitivity-96%, specificity-93%).In our study indications for treatment were not only based on invasively measured ICP values but also on clear radiologic findings such as massive hydrocephalus, space-occupying tumours or secondary craniostenosis with massive copper beaten skull appearance in X-ray, known to result in elevated ICP [44,45].We found the same US-ODE cut-off value of 0.675 mm associated to papilledema, suggesting that a relevant ICP increase seems to manifest on fundoscopy only if ODE exceeds 0.7 mm.
However, the US-ODE cut-off value with best diagnostic accuracy to detect potentially raised ICP requiring treatment was lower at 0.43 mm with a PPV of 90%.This suggests that patients with an US-ODE grade II (0.43-0.67 mm) are very likely to carry a significantly raised ICP, however will not be detected by fundoscopy.The diagnostic sensitivity of fundoscopic papilledema is much lower and inferior to US-ODE using our grading system.US-ODE values are furthermore quantitative numbers, which can reveal even smaller differences and thus carry more information than purely qualitative fundoscopic findings.US-ODE showed a high repeatability and intra-observer reliability, similar to another recent study that additionally demonstrated an excellent interrater reliability [46].

Limitations
The major limitation of this study is, despite its prospective nature, the relatively small cohort especially in the subgroups.Another limitation is that fundoscopic findings were dichotomized only (papilledema vs. no papilledema) without consideration of the papilledema severity.However, in clinical routine, especially at initial presentation, the relevant and often only message in the written report is "presence or absence" of papilledema, due to the high specificity of papilledema.The Frisén-rating of severity of papilledema [47,48] is very much examiner dependent and its reproducibility and discriminative ability is limited even among ophthalmologists [49].

Conclusions
US-ONSD and US-ODE are a reliable, non-invasive, diagnostic method for detecting potentially elevated ICP superior to fundoscopy.In addition, US-ONSD and US-ODE can indicate treatment success earlier and more reliably.US-ONSD and US-ODE cut-off values can be used in screening for potential and treatment-relevant ICP elevation in children.The proposed US-based grading scale of ODE is a further adjunct to support diagnosis.

Fig. 1 Fig. 2
Fig. 1 Initial US-ONSD and US-ODE in symptomatic patients considering fundoscopic ODE findings and indication for treatment.a US-ONSD and b US-ODE at initial presentation of 72 children with signs and symptoms of raised ICP.***p < 0.001

Fig. 3
Fig. 3 ROC analysis of fundoscopic findings, US-ONSD and US-ODE.a AUROC curve for minimal US-ODE cut-off value to detect papilledema fundoscopically.b AUROC curve of US-ONSD and US-ODE cut-off value to detect raised ICP requiring treatment in chil-