Comparison of Optical Coherence Tomography With Fundus Photographs, Fluorescein Angiography, and Histopathologic Analysis in Assessing Coats Disease

Key Points

Question  How is optical coherence tomography (OCT) useful in assessing Coats disease?

Findings  In this cohort study of 29 eyes from 28 children with Coats disease, the OCT analysis revealed the axial location of fluid, exudates, and atrophy that were not visible with photographs and fluorescein angiograms. Distinct OCT morphologies, when compared with histologic findings from biorepository eyes with Coats disease, were consistent with retinal anatomic changes and intraretinal and subretinal cells and depositions.

Meaning  Optical coherence tomography augments photographs and angiography to reveal transient and permanent effects of Coats disease on the retina, and these may be useful in monitoring disease activity and responses to therapy.

Importance  Coats disease is a rare pediatric vitreoretinopathy that can cause devastating visual and anatomic outcomes.

Objective  To compare optical coherence tomography (OCT) with fundus photographs, fluorescein angiography (FA), and histopathologic findings in Coats disease.

Design, Setting, and Participants  This retrospective cohort study was conducted in a single tertiary institution (Duke Eye Center) and identified 28 children with Coats disease through a review of medical records from December 2002 to January 2018. Four eyes were obtained from a biorepository for histopathologic analysis.

Main Outcomes and Measures  Macular OCT, fundus photographs, and FA results were reviewed and compared for morphological changes. These were compared with retinal histopathological findings.

Results  The mean (SD) age was 9.5 (5.5) years for the 28 children (and 29 eyes) with clinical imaging results, and 24 (86%) were boys. A comparison between imaging modalities revealed OCT features that were not visible in photographs or FA, including exudates in multiple retinal layers (23 [82.1%]), small pockets of subretinal fluid (4 [14.3%]), an outer retinal atrophy overlying fibrotic nodules (7 [25.0%]), and small preretinal hyperreflective OCT dots (25 [89.3%]). Next, a comparison with light micrographs introduced an association of OCT findings with possible pathological features, including hyperreflective linear structures on OCT that appeared consistent with cholesterol crystals, small hyperreflective dots with macrophages, outer retinal tubulations with rosettes, and analogous OCT histopathology features such as intraretinal vessels entering fibrotic nodules and retinal pigment epithelium excrescences under the subretinal fluid. An OCT analysis revealed intraretinal cystoid spaces in 19 eyes, but in 9 of 19 (47.4) this was not associated with cystoid macular leakage; rather, fluorescein leakage was observed from peripheral telangiectatic vessels. Additionally, exudates were intraretinal only (6 [21.4%]) or both intraretinal and subretinal (17 [60.7%]); none were subretinal only. In eyes with follow-up results, new fibrosis developed in 8 of 17 eyes (47.1%). Fibrosis developed in 5 of 5 eyes (100%) with baseline subretinal fluid vs 3 of 12 without (25%; 95% CI, 22%-92%) and in 7 of 9 eyes (77.8%) with subretinal exudates vs 1 of 8 (12.5%) without (95% CI, 16%-89%).

Conclusions and Relevance  Optical coherence tomography may show the transient and permanent effects of Coats disease on the retina. These results suggest that exudates and fluid in the macular subretinal space appear later in the disease and may result in fibrosis formation. Further studies are needed to confirm if early treatment could prevent vision-threatening macular fibrosis.

The advent of wide-field imaging, including the use of wide-field fundus photography (FP) and fluorescein angiography (FA) under anesthesia, has improved the visualization of peripheral vascular findings in Coats disease.^1-3 However, information about cross-sectional ultrastructural changes in Coats disease remains lacking in earlier stages of the disease because histopathological specimens usually include eyes that were enucleated for advanced disease.^4-9 Microscopic examinations of enucleated eyes have shown lipid laden macrophages and cholesterol clefts in the sensory retina and subretinal fluid (SRF).^4-9

Optical coherence tomography (OCT) provides an opportunity for a cross-sectional ultrastructural analysis of the retina in patients. Because the resolution of OCT is at the micrometer level, OCT findings are arguably comparable with histology results. This provides clinicians with a powerful tool to examine Coats disease, especially in its early stages. To our knowledge, there have been several publications on OCT findings in Coats disease to date^10-14 but none have combined an OCT and histopathologic analysis.^15,16 The objective of this study was therefore to compare the appearance of retinal and subretinal findings in Coats disease on OCT with FP, FA, and relevant histopathologic analyses. Using these findings, we also examined the pathogenesis of macular fibrosis, which causes irreversible vision loss in Coats disease.¹⁷

This was a retrospective cohort study of all patients younger than 21 years who received a diagnosis of Coats disease, presented before receiving treatment to the Duke Eye Center between December 1, 2002, and January 31, 2018, and underwent OCT, FP, and FA imaging. The study was conducted between July 26, 2017, and April 29, 2018, and was designed in accordance with the tenets of the Declaration of Helsinki. The study was approved by the institutional review board at Duke University and was Health Insurance Portability and Accountability Act–compliant. Consent was waived because of the retrospective nature of the study.

A diagnosis of Coats disease included fulfilling the clinical criteria of having idiopathic retinal telangiectasia with or without exudation and/or exudative retinal detachments. Demographic information and the staging of Coats disease were documented. Affected eyes were divided based on clinical examination, FP, or FA results into 5 stages of disease as described by Shields et al.¹⁸

Fundus photography (color, red-free) and FA images were obtained using tabletop systems (Zeiss FF450; Carl Zeiss Meditec), ultra-widefield (Optos 200Tx; Optos, Inc), or during examination under anesthesia (EUA; RetCam 2; Clarity Medical Systems). The OCT images were attained using tabletop time-domain or spectral-domain OCT (Stratus OCT; Carl Zeiss Meditec or Spectralis HRA-OCT; Heidelberg Engineering) or portable handheld spectral-domain OCT (Envisu C2300; Leica Microsystems).

Paraffin embedded tissues from 4 enucleated eyes were obtained from the Duke Biorepository and Precision Pathology Center. Using light microscopy, these histologic sections were examined and compared with invivo OCT findings. The statistical analysis was performed using JMP Pro, version 13.0 (SAS Institute Inc; 95% CIs for percentage difference between groups were calculated and the statistical significance was represented by 95% CIs that did not contain 0.

Twenty-eight patients and 29 eyes met the entry criteria. The mean (SD, range) age at presentation was 9.5 (5.5, 1.1–18.7) years and 24 (86%) were boys. Of the participants, 11 (39.3%) were white, 7 (25.0%) were African American, 4 (14.3%) were Asian, 1 (3.6%) was Hispanic, and 5 (17.9%) were unreported. At presentation, the stages of disease were 1 (1 [3.6%]), 2A (4 [14.3%]), 2B (12 [42.9%]), 3A1 (8 [28.6%]), and 3A2 (4 [14.3%]). Snellen or equivalent visual acuities were recorded at presentation for 26 eyes and ranged from 20/20 to 20/8000 (median, 20/200 [interquartile range, 1.20 (logMAR)]).

Clinical information from 4 biorepository eyes was reviewed. Two patients were boys. These were younger patients (mean [SD, range] age was 1.3 [0.7, 0.4-1.9] years) with a more severe disease, and the enucleation was because of concern for retinoblastoma in 3 eyes with total retinal detachment or uncontrolled glaucoma in 1 blind and painful eye. None of the enucleated eyes had received any vitreoretinal surgery. One eye was treated with laser cyclophotocoagulation 1 month before enucleation.

Color or red-free photography was performed in the clinic in 21 of 29 eyes (72.4%; 5 tabletop: age range, 8.5-18.7 years; 16 ultra-widefield: age range, 4.0-18.4 years) and during EUA in 8 of 29 (27.6%; age range, 1.1-14.5 years). Fluorescein angiography imaging, which was not performed in 4 eyes, was obtained in the clinic in 13 of 25 eyes (52%; 3 tabletop: age range, 15.6-18.4 years; 10 ultra-widefield: age range, 8.2-18.7 years) and during EUA in 12 of 25 (48%; age range, 1.1-14.5 years). Twenty-five of 29 eyes (86.2%) underwent OCT in the clinic (6 [24%] time-domain and 19 [76%] spectral-domain; age range, 4.0-18.7 years) while 4 of 29 (13.8%) underwent OCT with a handheld device during EUA (spectral-domain: age range, 1.1-4.4 years).

Twenty-eight of 29 eyes (96.6%) had exudates on photographs; 23 (82.1%) were within the macula and 20 (71.4%) were within the fovea (Figures 1 and 2; eFigure 1 in the Supplement). On OCT imaging, the exudates appeared as bright hyperreflective foci or broad sheets and were observed in the following retinal layers: nerve fiber (7 [25.0%]), ganglion cell (GCL) (8 [28.6%]), inner plexiform (15 [53.6%]), inner nuclear (INL) (19 [67.9%]), outer plexiform (19 [67.9%]), outer nuclear (ONL) (23 [82.1%]), and in the subretinal space (17 [60.7%]). Two eyes had highly hyperreflective linear structures (within subretinal exudates and SRF respectively) on OCT that corresponded to refractile cholesterol crystals on photographs. On light micrographs, eosinophilic exudates were also observed in every retinal layer and subretinally and cholesterol clefts were seen in the SRF.

On OCT, the highest density of exudates were observed in the upper half of the ONL (correlate for histological Henle fiber layer, which is part of the outer plexiform¹⁹) in 13 eyes (46.4%) (9 [69.2%] of these were star shaped on photographs), and in the subretinal space in 10 eyes (35.7%). While exudates were found intraretinally only in 6 eyes (21.4%), exudates were observed both intraretinally and subretinally in 17 eyes (60.7%). Notably, no eye had subretinal exudates in the absence of intraretinal exudates.

Retinal detachment was apparent on photographs in 12 eyes (41.4%) and involved the macula and foveal center in 6 (20.7%) and 4 eyes (13.8%), respectively (Figure 2). Optical coherence tomography also demonstrated macular SRF in these 6 eyes. In an additional 4 eyes, OCT illustrated a small pocket of macular SRF that was not visible in photographs (Figure 1B). The OCT results further showed small hyperreflective dots in the SRF and retina in 3 of 10 eyes (30.0%). Importantly, the small hyper-reflective dots were observed in areas without exudates in photographs and enface infrared images.

On light micrographs, the retina surrounding SRF was thickened and diffusely infiltrated by macrophages. Light micrographs also illustrated cholesterol crystals and macrophages in the eosinophilic SRF and retinal pigment epithelium (RPE) excrescences with RPE thinning underlying the SRF. On OCT, these RPE excrescences were not well defined because of shadowing from exudates, but in 2 eyes a few hyperreflective RPE excrescences were seen under SRF in areas without exudates.

Seven of 29 eyes (24.1%) presented with macular fibrosis (Figures 3 and 4, eFigure 2 in the Supplement). In photographs, the fibrosis appeared foveal-involving, nodular, pigmented, and surrounded by dense exudation. Optical coherence tomography also showed that macular fibrosis was nodular, intermediately hyperreflective with areas of intense hyperreflectivitiy because of the pigment, and surrounded by intraretinal and subretinal exudates. The OCT results further showed the presence of hyporeflective cystoid spaces (in the GCL (3 [42.9%]), INL (7 [100%]), and ONL (3 [42.9%]) and overlying outer retinal atrophy in 7 of 7 eyes (100%). In 3 of these 7 eyes (42.9%), the hyperreflective material associated with the fibrotic nodule transversed through all layers of the retina and created a full-thickness macular hole. On light micrographs, the outer retina overlying all fibrotic nodules were also observed to be atrophic.

On histologic sectioning, 4 and 1 fibrotic nodules were observed in 2 eyes, respectively. All fibrotic nodules contained cholesterol clefts that were intermixed with macrophages, pigment granules, and fibrinous material. Multinucleated giant cells, lymphocytes, plasma cells, and capillaries were also observed. One of these nodules also had a fibrous capsule and 2 nodules were accompanied by rosette formation in the overlying retina. Eyes with fibrotic nodules on OCT were examined and 3 of 7 eyes (42.9%) had a corresponding hyperreflective tubule with a hyporeflective core in the overlying and surrounding retina. One eye without fibrosis but with extensive macular atrophy was also observed to have these tubular structures on OCT (eFigure 3 in the Supplement).

Photographs demonstrated a prominent retinal vessel that appeared to dive toward the nodule in 6 of 7 eyes (85.7%), and this was accompanied by a few spots of hemorrhage in 1 eye (14.3%). Fluorescein angiography, which was available for 6 eyes with macular fibrosis, showed staining with areas of leakage (3 [42.9%]), staining (2 [28.6%]), or diffuse macular leakage (1 [14.3%]). On OCT, a hyperreflective elongated structure with posterior shadowing that was continuous with inner retinal vasculature and traveled toward the nodule was observed in 4 of 7 eyes (57.1%). On histologic sectioning, we observed a vessel from the inner retina approaching a fibrotic nodule found in the macula. Because of histologic processing, the neurosensory retinas were artificially detached from the RPE and choroid. All 5 fibrotic nodules were observed to be attached to the undersurface of and partially encircled by the neurosensory retina without any connection to the RPE/choroid. No nodule was accompanied by a hemorrhage under the distant RPE or a break in the Bruch membrane.

Optical coherence tomography showed small preretinal hyperreflective OCT dots in 25 eyes (86.2%), epiretinal membrane in 11 eyes (37.9%), and inner surface wrinkling in 12 eyes (41.4%) (Figure 2B). Optical coherence tomography further demonstrated intraretinal hyporeflective cystoid spaces in 19 eyes (65.5%). Seven (36.8%) of these eyes had coexisting macular fibrosis, as previously described. One eye had coexisting macular atrophy without fibrosis; the cystoid spaces were observed in the INL and FA demonstrated a window defect (eFigure 3 in the Supplement). Of the remaining 11 eyes, 2 (18.2%) demonstrated cystoid macular leakage on FA with intraretinal fluid observed in the INL and ONL (eFigure 4 in the Supplement) whereas 9 (81.8%) had no macular hyperfluorescence on FA but had intraretinal cystoid spaces between peripheral leaking aneurysms and the fovea and cross-sectionally in the GCL (1 [11.1%]), INL (7 [77.8%]), and ONL (9 [100%]) (Figure 5).

Nineteen eyes (65.5%) had a 6-month or longer follow-up after receiving treatment (average [SD], 41 [42] months). Seven of these eyes (36.8%) had laser and/or cryopexy, 9 (47.4%) had this treatment plus anti–vascular endothelial growth factor injections, 2 (10.5%) had vitreoretinal surgery, and 1 (5.3%) had vitreoretinal surgery with anti–vascular endothelial growth factor injections.

Macular exudates and SRF improved or resolved after treatment in all 14 of 14 eyes with macular exudates and 5 of 5 eyes with macular SRF at baseline. No eye developed new macular exudates or SRF. In comparison, of the 17 eyes with no macular fibrosis at baseline, 8 (47.1%) developed new macular fibrosis (6 [35.3%] were fovea-involving). Of the 16 eyes with no previous macular atrophy, 9 (56.3%) developed new macular atrophy (5 [31.3%] were foveal-involving). No eye had a resolution of macular fibrosis or atrophy.

The likelihood of developing new macular fibrosis at the final follow-up was 75% higher when macular SRF was present (5/5 eyes) vs absent (3/12 eyes [25%]) (95% CI, 22%-92%) and 65% higher when macular subretinal exudates were present (7/9 eyes [77.8%]) vs absent (1/8 eyes [12.5%]) (95% CI, 16%-89%) on OCT at baseline. Moreover, the development of macular atrophy was associated with the development of macular fibrosis; the proportion of eyes that developed macular atrophy was 63% higher in eyes that developed macular fibrosis (7/8 [87.5%]) vs those with no new macular fibrosis (2/8 [25%]) (95% CI, 12%-88%). eFigure 5 in the Supplement illustrates an eye that developed macular and extramacular fibrosis posttreatment.

To our knowledge, this is the first study of pediatric Coats disease to compare OCT with FP, FA, and relevant light micrographs. A literature search on PubMed using the terms optical coherence tomography, Coats disease, histology, and pathology between January 1, 2007, and April 29, 2018, revealed no prior published articles on this topic. Of note, imaging in clinics was possible in children as young as 4 years with ultra-widefield retinal imaging and OCT. This may have been in part because both systems lack the white light illumination that can be bothersome to young children.

This study showed multiple advantages in adding OCT to traditional imaging with photography and FA in Coats disease. First, OCT documented exudates in all retinal layers, which could not be appreciated in photographs. Second, OCT explained the macular star pattern of exudation in some eyes with Coats disease. These exudates for the most part accumulated in the upper half of the ONL; previous reports had shown that Henle fiber layer corresponds with the usually hyporeflective ONL.¹⁹ Third, while photographs showed macular detachment in 6 eyes, OCT demonstrated SRF in 10 eyes. Subtle SRF may be missed on clinical examination or with non–OCT-based imaging. Fourth, OCT demonstrated the presence of extensive outer retinal atrophy above fibrotic nodules, which is not appreciated in photographs. Fifth, OCT is the only clinical imaging modality capable of showing small preretinal hyperreflective OCT dots, which may represent lipid, inflammatory cells, or red blood cells.

Furthermore, this study revealed several observations on OCT that may be associated with features on a histopathologic analysis. Cholesterol clefts and macrophages are hallmark features of Coats disease and were observed in the enucleated specimens. Hyperreflective linear structures that resembled cholesterol crystals were also observed in 2 eyes on OCT, both on the cross-sectional image and enface infrared image. Additionally, small hyperreflective dots were observed in 3 of 10 eyes (30.0%) with SRF in areas without exudates and we speculate that they represent the macrophages that were observed on light micrographs. These hyperreflective subretinal and intraretinal dots have been previously observed in central serous chorioretinopathy and were theorized to represent macrophages.²⁰ Light micrographs showed macrophages within exudates as well, but these would be difficult to identify on OCT because exudates are also hyper-reflective.

Other findings found on both light microscopy and OCT include RPE excrescences under SRF, retinal rosettes, and intraretinal vessels approaching a fibrotic nodule. Retinal pigment epithelium excrescences were observed under extensive SRF on both light micrographs and OCT. On light micrographs, retinal rosettes overlying 2 fibrotic nodules were observed. A corresponding hyperreflective tubule in the outer retina was also observed on OCT in 3 of 7 eyes (42.9%) with fibrotic nodules and 1 of 1 eye (100%) without fibrosis but with extensive atrophy. These tubules have a hyperreflective wall that surrounds a hyporeflective core and resemble “outer retinal tubulations” previously described in macular degeneration, Bietti crystalline dystrophy, and choroidemia.²¹ It has been theorized that the appearance of outer retinal tubulations is due to the arrangement of degenerating photoreceptors in a tubular fashion,²¹ while rosettes in Coats disease are postulated to represent secondary changes due to retinal degeneration or detachment.²² To our knowledge, this study is the first to describe outer retinal tubulations on OCT in Coats disease and we suspect that the rosettes that were previously described in Coats disease represent outer retinal tubulations because both are lined by an external limiting membrane with photoreceptor segments either lining the wall or projecting into the lumen.^22,23

An intraretinal vessel approaching a fibrotic nodule was observed on light micrographs in 1 biorepository eye. On OCT, corresponding hyperreflective elongated structures from the inner retinal vasculature entering the nodules were observed in 4 of 7 eyes (57.1%). Additional findings of retinal vessels diving down into the lesion (6/7 eyes [85.7%]) with adjacent dots of blood (1/7 eyes [14.3%]) in photographs, leakage (4/6 eyes [66.7%]) on FA, and cystoid intraretinal spaces (7/7 eyes [100%]) on OCT possibly due to leakage from neovascularization (vs atrophy) all support the hypothesis that type 3 neovascularization can occur in fibrotic nodules. This study therefore supports the existing literature that reported the presence of neovascularization in fibrotic nodules in Coats disease.^24,25 However, we did not observe any hemorrhage under RPE or Bruch membrane disruption that would suggest a connection between the subretinal nodule and the choroid. Chang and colleagues⁵ analyzed 13 eyes with macular fibrous nodules and similarly did not observe hemorrhage below RPE, but others had observed that the Bruch membrane can be disrupted with blood vessels extending from the choroid to the bottom of a fibrotic mound.^22,26

Consistent with prior reports, this study showed that exudative fluid and material may diffuse posteriorly to the macula and from the intraretinal tissue to the subretinal space.^26,27 This is supported by findings of cystoid spaces without leakage between extramacular lesions and the fovea and absent on the other side of the fovea. Furthermore, while some eyes only have intraretinal exudates, all eyes with subretinal exudates also have intraretinal exudates. These findings suggest that fluid or exudates in the macular subretinal space are seen later during the disease.

We also found that fluid or exudates in the macular subretinal space were associated with new macular fibrosis formation. Subretinal exudates have previously been postulated to incite the recruitment of phagocytic inflammatory cells that lead to fibrosis.^28,29 Manschot and de Bruijn²⁸ demonstrated on electron microscopy that phagocytic cells in an eye with Coats disease and subretinal fibrosis had empty lipid vacuoles and (1) pigment granules within vacuoles or (2) native pigment granules that were embedded directly in cytoplasm. They also observed (1) pigment epithelium cells that had taken on a fiber producing ability and (2) typical fibroblasts.²⁸ Because native pigment granules that are embedded in cytoplasm are found in RPE cells whereas pigment granules in vacuoles are found in macrophages, the authors concluded that subretinal exudates cause the recruitment of RPE cells that differentiate into phagocytic cells with the ability to produce fiber, as well as macrophages and fibroblasts that migrate from the retina. We similarly observed the presence of phagocytic cells and multinucleated giant cells in fibrotic nodules on light micrographs. However, without electron microscopy, we could not determine if the pigment granules were in vacuoles or embedded in the cytoplasm. Using light micrographs, we further observed the presence of an intraretinal vessel approaching the fibrotic nodule and proliferating capillaries within the nodule. Therefore, the inflammatory cascade that causes fibrosis in Coats disease may also recruit endothelial cells that proliferate and form new blood vessels.

The limitations of this study include its retrospective nature, nonstandardized imaging, and small sample size, which restrict the power to detect differences between subgroups. The comparison between the OCT findings and the histopathology was also limited given the more severe disease seen in enucleated eyes.

Fluid and exudates appear in the macular subretinal space later in the disease course and are associated with new macular fibrosis. We therefore postulate that early treatment before exudation into the macular subretinal space may prevent macular fibrosis, although this cross-sectional study cannot confirm this association. This study also highlights the importance of OCT in multimodal imaging. Optical coherence tomography augments photographs and FA and demonstrates features that are comparable with histopathologic findings.

Accepted for Publication: September 20, 2018.

Corresponding Author: Cynthia A. Toth, MD, Department of Ophthalmology, Duke University Eye Center, 2351 Erwin Rd, Durham, NC 27710 (cynthia.toth@duke.edu).

Published Online: November 21, 2018. doi:10.1001/jamaophthalmol.2018.5654

Author Contributions: Dr Toth had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Study concept and design: Ong, Mruthyunjaya, Toth.

Acquisition, analysis, or interpretation of data: All authors.

Drafting of the manuscript: Ong.

Critical revision of the manuscript for important intellectual content: All authors.

Statistical analysis: Ong.

Obtained funding: Toth.

Administrative, technical, or material support: Ong, Cummings, Toth.

Study supervision: Toth.

Conflict of Interest Disclosures: All authors have completed and submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Dr Ong receives grant support from the International Retinal Research Foundation. Dr Vajzovic reports receiving grant support from the Research to Prevent Blindness Unrestricted Grant to Duke Eye Center, Alcon, Roche, and Second Sight and personal fees from Alcon, DORC, Genentech, Jannsen, and Alimera Sciences. Dr Mruthyunjaya receives consulting fees from Optos. Dr Toth receives grant support from the National Institutes of Health (NIH) and royalties from Alcon. No other disclosures are reported.

Funding/Support: This article was supported by The International Retinal Research Foundation, The Hartwell Foundation, The Andrew Family Charitable Foundation, the NIH (grants 1RO1 EY025009 and P30 EY005722), and grant UL1 RR024128 (Pilot project grant) from Duke Translational Research Institute and NIH Roadmap for Medical Research.

Role of the Funder/Sponsor: The funding organizations had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.

Disclaimer: The contents of this article are solely the responsibility of the authors and do not necessarily represent the official view of the NIH.

Additional Contributions: Du Tran-Viet, BS, Vincent Tai, MS, and Katrina Winter, BS, (Duke University) provided assistance with image processing and analysis, and Sandra Stinnett, DrPH, completed statistical analysis for this study. They were compensated for their contributions.