Management of a misidentified and misnumbered intraocular glass foreign body with an endoscope in a case of endophthalmitis

Abstract

An endoscope is a useful adjunct for the retinal surgeon to overcome haze of a compromised anterior segment. It allows early surgery in trauma and infections which translates to better results. Intraocular glass foreign body is a challenging condition, demanding highly skilled surgical expertise. We present endoscopic removal of an intraocular foreign glass body in a badly traumatised and infected eye. The surgical challenge was accentuated by an imaging misdiagnosis of ‘twin metallic foreign bodies’.

Background

Ocular trauma is an unfortunate cause of visual morbidity in healthy eyes, especially in young adults. Worldwide the annual incidence of severe eye trauma was noted as 750 000, of which open globe injury (OGI) accounted for 200 000.¹ The management of OGI needs to be rapid, as there is a direct impact on the prognosis, especially in cases with endophthalmitis with intraocular foreign body (IOFB).² Ocular trauma can render the cornea permanently opaque with distorted anterior segment. Under these circumstances the visualisation of posterior segment through conventional viewing systems (operating microscope, wide-angle viewing systems) may not allow a vitreoretinal surgical intervention. In such cases delay is not an option and an endoscope proves to be a useful adjunct.³

Glass IOFB is a major challenge surgically, as magnets cannot be used, while handling transparent objects with hefty forceps can easily result in breakage within the eye. The challenge gets compounded by preoperative imaging-related difficulties, where they can be misdiagnosed.^{4 5} Herein, we discuss management of a challenging case of OGI with endophthalmitis and a glass IOFB managed using an endoscope. The glass IOFB had been reported as twin and metallic in previous imaging.

Case presentation

A 6-year-old child was brought to the clinic with reports of pain and diminution of vision in right eye for 1 day. There was alleged history of trauma to the right eye 1 day back while playing with an electric light bulb. The child had apparently lit an old incandescent bulb with a matchstick, leading to blast of the bulb. He had been noted to have penetrating OGI (corneal tear) in zone I for which an urgent corneal tear repair had been done (figure 1). At presentation, the visual acuity was restricted to perception of light with a repaired linear corneal tear involving the temporal aspects of cornea. Margin of the tear appeared to be free of infectious infiltrates, but central cornea was opaque. Anterior chamber was irregular and filled fibrinous exudates (figure 1A). There was no glow for fundus evaluation. Further examination was restricted at presentation due to non-cooperative nature of child. Left eye was unremarkable on examination.

Figure 1.

Preoperative images. (A) Image showing sutured corneal laceration and opaque cornea. No view of the posterior segment was possible through the opaque cornea. (B) B-scan ultrasound showing two highly reflective and separate echo spikes. Artefactual shadowing can be seen posterior to the spikes. (C) Non-contrast CT scan performed in the coronal section showing a single intraocular focal area of hyperdensity in the inferomedial quadrant suggestive of an intraocular foreign body. (D) Non-contrast CT scan in the axial section with inferior axial cuts showing intraocular focal area of hyperdensity in the inferomedial quadrant suggestive of intraocular foreign body. Note the gross absence of streak artefact.

Investigations

B-scan sonography was ordered with a suspicion of endophthalmitis and also to rule out presence of an IOFB. Care was taken to perform the scans very gently. The B-scan was reported to have two foreign bodies within the vitreous cavities. Mild–moderate amplitude spikes suggesting haemorrhage/exudates were also noted along with a suspicion of retinal detachment (figure 1B). CT was advised further which detected a single 8×3 mm metallic IOFB within the vitreous cavity of 2900–3100 Hounsfield unit (HU) attenuation (figure 1C, D). The CT scan was done using 2 mm intervals for axial cut sections, with sagittal and coronal reconstruction (Optima 660 128 slice MDCT scanner using 0.625 mm detector collimation).

Differential diagnosis

The patient was thus diagnosed to have a repaired OGI with endophthalmitis and a metallic IOFB. Possibility of multiple IOFBs was recorded.

Treatment

Considering presence of endophthalmitis and an unlikely significant improvement of the corneal clarity in the foreseeable future, the child was suggested to undergo an immediate endoscope-assisted pars plana vitrectomy, vitreous biopsy, IOFB removal and an intraocular antibiotic injection under general anaesthesia. The nursing staff was alerted of the situation with a requirement of intraocular magnets.

During the surgery retinal detachment with endophthalmitis was noted, and surgical field was continuously obscured due to presence of exudates. The posterior pole was obscured by blood with apparent severe injury to the macular region. Despite multiple attempts at locating a metallic IOFB, none was sighted. In a desperate attempt, the surgeon moved the intraocular magnet all over the inside of the globe while trying to locate the IOFB within the exudates, when an altered sheen was noted on the posterior pole. The substance was not responding to motion of the magnet, and thus was considered to be non-metallic and likely glass due to transparent nature of the IOFB. An intraoperative change in surgical plan was thus warranted with a switchover of instruments. First the intravitreal forceps were used which failed to grasp the IOFB due to its thickness. Later active suction was attempted with both cutter and the flute needle but both failed to lift the IOFB. Until now foreign body forceps were being avoided due to the inherent risk of glass breakage, but as there was no other foreseeable option they were employed (figure 2A, B). Care was taken to hold the glass IOFB in such a way that pressure was least on the glass and it did not break. This was challenging due to presence of vitreous exudates and retinal detachment, and the small field of the endoscope. The IOFB was then externalised through the limbal route using handshake technique (Video 1).

Figure 2.

Intraoperative images. (A) Intraoperative image showing glass foreign body grasped with forceps inside the vitreous cavity, as visualised through the limited view of the endoscope. The straight forceps can be seen in the inset image. (B) Intraoperative picture showing the successfully extracted glass foreign body.

Video 1.

bcr-2021-244569supp001.mp4 ^{(668.9KB, mp4)}

DOI: 10.1136/bcr-2021-244569.video01

Considering the previous B-scan report of twin IOFBs and the CT findings of a metallic IOFB, it was assumed that the glass IOFB had been missed on the CT scan. The surgeon continued to look for another metallic IOFB inside the eye, but none was found despite repeated rugged attempts. It was then contemplated that the IOFB had been mistaken to be metallic in the CT scan, and could be single as against the report of the B-scan. At this juncture, the surgical team measured the actual dimensions of the extracted IOFB, and found it to be 8×4 mm, matching the dimensions reported on the CT scan. Hence, it was concluded that there was a single glass IOFB inside the eye. Due to poor visual prognosis, repair of the retinal detachment and oil injection were deferred.

Outcome and follow-up

The patient was advised systemic antibiotics with supportive topical therapy in the immediate postoperative period. After discussion with radiologist, an X-ray of the orbit and B-scan ultrasound were also done to document absence of a ‘second-metallic’ foreign body (figure 3A, B). The infection resolved within 1 week, and the eye healed with phthisis without symptoms by 2 months of follow-up (figure 3C). The corneal sutures had become loose and suture removal was advised.

Figure 3.

Postoperative images. (A, B) Post-surgery X-ray of the orbits (in coronal view) and B-scan showing absence of any foreign body in right globe. (C) Post-surgery anterior segment image showing healing cornea with exudate-free anterior chamber.

Discussion

Endophthalmitis with IOFB mandates an early extraction of the IOFB due to an imminent risk of organ-loss and infection-spread.^{6 7} In the case of glass IOFBs, the challenges are accentuated due to their transparent and fragile nature causing difficulties in extraction.² In the discussed case, all these challenges increased exponentially due to an ‘impossible’ surgical view through the injured cornea, and misleading imaging reports.

The detection range on CT scans is very high for glass IOFB with size in excess of 1.5 mm (>90% detection) and for light-bulb glass (>80% detection). This is not affected by presence of ocular haemorrhage or exudates. Sensitivity is considered highest for CT scan in comparison to other imaging modalities if the IOFB is in the posterior ocular cavity (approaching 70%).^{6 8} Glass IOFB are typified by their hyperdense structure on CT scan with a HU range of 500–2700 in absence of a streak artefact.⁹ Typically, light bulb glass has a very low HU attenuation (up to 285 HU). However, the range of HU attenuation is large for glass IOFBs, for example, from as low 175 HU to as high as 2700 HU for windshield glass.¹⁰ In contrast the HU of metals typically exceed 3000 on CT scans. The HU depend on density of the substance and are thus also dependent on the quality of manufacture.¹⁰ Light bulbs typically use soda-lime type of glass and a variability in the lime content can easily lead to change in the HUs on CT scans.¹¹ In our case the range of HU for the ‘so reported metallic’ IOFB was 2900–3100, thus not typical for either metallic or glass IOFB. It is possible that due to higher lime content of the glass IOFB, the HU reached until 2900, which is actually very close to the HU for limestone (2800 HU).⁹ The other consideration is the general variability in reference range of HU for each CT machine. The HUs are calibrated for each CT machine separately.¹² The variation in HUs on MDCT scanners has been noted before and thus can explain the overlapping range seen in this case. Noticeable streak artefacts were absent in this case and thus, in retrospect, this was a soft indicator towards ‘non-metallic’ nature of the IOFB.⁹

Imaging diagnosis of glass IOFB on ultrasonography is made on the basis of their hyperechoic nature with comet-tail effect. Acoustic shadowing can be present in both metallic (except aluminium) and glass IOFBs, and this holds true for the reverberation effect related artefacts too. Both are more discrete in case of metallic IOFBs. The comet-tail artefact, as seen in our case, is a kind of a reverberation artefact and arises due to presence of two closely spaced highly reflective interfaces (figure 1B).¹³ There is not much data on accuracy of imaging modalities for detection of number of IOFBs, though ultrasonography is likely inefficient in this regard. This is because it is a highly dynamic and echographer dependent modality, and presence of vitreous spikes in a traumatised and infected eye will make this even more challenging. Arguably, an additional conundrum was generated by the partial cylinder like shape of the IOFB in our case, which may have led to two widely spaced highly hyper-reflective spikes. This may have led to misinterpretation regarding the double nature of the IOFB. It should be noted that the comet tail was seen only with respect to the posterior surface of the IOFB.

A major challenge faced during the management of this case was the extraction of the foreign body through an impossible optic media. As the child was having endophthalmitis we were left with no option but to adopt an immediate surgical approach. An endoscope is a useful tool to bypass opacified anterior segment. It allows early surgery which yields better results in the patients operated within 3 days of trauma.² It also avoids the need for donor cornea or a temporary keratoprosthesis.^{14 15} Though it proved an indispensable tool for us, we encountered multiple challenges. Due to limited field of view (up to 90 degrees in (minimally invsive vitrectomy system (MIVS)), poor stereopsis and inability to perform bimanual procedures, locating and securing the IOFB proved very difficult in our case.³ The two-dimensional view and restricted image resolution acquired through the endoscope made gripping of the transparent IOFB highly skill demanding. The introduction of endoscope in such traumatised eyes without damaging ocular structure can be challenging itself. Sometimes a false track may be formed, and the endoscope may go behind the retina/choroid. Due to retinal detachment, we had to place the endoscope far away during initial focus to prevent inadvertent retinal injury.² The tip of the fibre optic cable frequently got stained with exudates, requiring frequent cleaning. These challenges notwithstanding, the approach proved to be vital in this patient. The only other possible approach was a temporary keratoprosthesis-assisted vitrectomy followed by keratoplasty, which would have proven redundant later given the nil vision prognosis. Thus, we were able to avoid a mutilating surgery like evisceration or a redundant corneal transplant, and at the same time controlled the infection by extracting the IOFB.

Learning points.

• The shape of a large glass intraocular foreign body (IOFB) can lead to misinterpretations regarding its singularity on ultrasound.

• The Hounsfield units of glass IOFBs can vary, sometimes nearly matching that of a metallic foreign body.

• The nature of trauma should always be considered while managing IOFBs. All imaging-related information should be used together with this history while contemplating surgical planning. Information provided by ultrasonography and CT scan should be used in conjunction to judge the nature of the foreign body in difficult scenarios.

• Endoscopic surgery is a useful procedure which can be used to remove foreign body from the vitreous cavity even in presence of hazy media in emergent conditions. Glass foreign bodies pose a particularly challenging situation.

Ethics statements

Patient consent for publication

Consent obtained from parent(s)/guardian(s).