Orbital apex syndrome caused by imaging-occult sphenoid sinus fungal infection and the necessity of early surgical exploration: a retrospective case series

Abstract

Objective

To characterize a highly occult subtype of sphenoid invasive fungal rhinosinusitis (IFRS) presenting as acute orbital apex syndrome (OAS) without classic invasive imaging signs, and to evaluate the efficacy of early surgical intervention.

Methods

This retrospective case series evaluated six patients (March 2021–November 2025) who developed classic OAS secondary to sphenoid IFRS despite strictly subtle imaging manifestations. All underwent navigation-guided endoscopic sinus clearance and optic nerve decompression. Pre- and postoperative neuro-ophthalmic parameters (best-corrected visual acuity [BCVA], visual field index [VFI], mean deviation [MD], and ocular motility) were analyzed. Bias-corrected effect sizes (Hedges' g) were calculated to quantify the magnitude of functional recovery.

Results

Pathological evaluation confirmed Aspergillus fumigatus (n=4), Penicillium citrinum (n=1), and Rhizopus oryzae (n=1). Four patients (66.7%) had comorbid type 2 diabetes. Following surgery, median BCVA improved from 1.91 to 0.30 LogMAR (Hedges’ g = 0.80). In the four patients completing visual field testing, median VFI improved from 11% to 91.5% (g = 1.37), and median MD improved from -28.02 to -3.33 dB (g = 1.23). Ocular motility fully recovered in two of three affected patients. Notably, the two patients lacking visual recovery presented with severe preoperative hyperglycemia (>14 mmol/L). No severe surgery-related complications occurred.

Conclusions

For occult IFRS presenting with this striking 'clinical-radiological dissociation', early navigation-guided endoscopic exploration is a crucial sight-saving strategy. Strict glycemic control is paramount to preserving this therapeutic window. Further prospective validation is warranted.

Supplementary Information

The online version contains supplementary material available at https://doi.org/10.1186/s12886-026-05170-4.

Background

Orbital apex syndrome (OAS) is a sight-threatening condition broadly defined by concurrent involvement of the optic nerve and one or more cranial nerves (the oculomotor, trochlear, abducens, and ophthalmic divisions of the trigeminal nerve) within the confined orbital apex [1]. While commonly characterized by acute or subacute visual decline, multiple extraocular muscle palsies, and a relative afferent pupillary defect (RAPD), the clinical presentation can be variable, particularly in early disease. The etiology of OAS is highly complex and represents a broad diagnostic framework, encompassing intraorbital tumors, inflammatory pseudotumors, vascular anomalies, and regional infections [2, 3]. Because the medial wall of the orbital apex is separated from the posterior ethmoid and sphenoid sinuses only by fragile bony laminae (e.g., the lamina papyracea and the lesser wing of the sphenoid bone), aggressive sinus pathologies can readily breach these anatomical barriers via direct erosion or neurovascular channels to precipitate OAS [4].

Among these rhinogenic causes, invasive fungal rhinosinusitis (IFRS) represents a fulminant and potentially fatal etiology, predominantly afflicting immunocompromised individuals or those with poorly controlled diabetes mellitus [5]. Previous literature establishes that classic IFRS typically presents with hallmark radiological features, including extensive soft tissue infiltration within the sinuses, widespread osteolytic bone destruction, and overt macroscopic invasion into the orbit or cranial cavity [6–9]. However, we recently identified a highly deceptive cohort of patients presenting with acute OAS secondary to sphenoid sinus IFRS who completely lacked these classic radiological hallmarks.

In these patients, preoperative computed tomography (CT) and magnetic resonance imaging (MRI) revealed only subtle mucosal thickening or minimal soft tissue shadows within the sphenoid sinus. Crucially, frank bone destruction was absent, and some scans were initially interpreted as strictly normal, with only hazy infiltration at the orbital apex hinting at an underlying invasive process. This striking clinical-radiological dissociation, characterized by the coexistence of profound neuro-ophthalmologic deficits and minimal sinonasal imaging changes, frequently precipitates dangerous misdiagnoses such as painful ophthalmoplegia, idiopathic optic neuritis, or Tolosa-Hunt syndrome. Therefore, this retrospective case series study of six pathologically confirmed cases aims to characterize the clinical profile of this occult IFRS subtype and to critically evaluate the therapeutic value of early, navigation-guided surgical exploration and optic nerve decompression.

Methods

This retrospective case series from March 2021 to November 2025 followed the Declaration of Helsinki and was approved by the Institutional Ethics Committee Eye Hospital of Wenzhou Medical University (No. 2026-037-K-027), which waived the requirement for informed consent. Specific diagnostic definitions were applied to ascertain cases of OAS secondary to occult sphenoid sinus fungal infections from our medical records. Case ascertainment required a preoperative clinical diagnosis of OAS, defined by the concurrent involvement of the optic nerve (evidenced by visual acuity decline or a relative afferent pupillary defect [RAPD]) and one or more cranial nerves traversing the orbital apex (manifesting as ptosis or impaired extraocular motility) [1, 5, 10]. Furthermore, patients were identified based on subtle sphenoid sinus imaging findings characterized by the absence of widespread osteolysis, focal soft tissue or mucosal changes strictly confined to the sinus cavity, or hazy apical infiltration lacking a distinct mass effect. All ascertained cases underwent intervention via navigation-guided endoscopic sinus surgery and optic nerve decompression, with definitive postoperative histopathological confirmation of fungal infection. To isolate this specific occult clinical phenotype and ensure rigorous analysis, patients were not considered for this series if they presented with non-fungal etiologies, preoperative imaging demonstrating extensive bone destruction or contiguous invasion, or confounding neuro-ophthalmic conditions (e.g., glaucoma, ischemic optic neuropathy, intracranial tumors, stroke, demyelinating diseases, or traumatic optic neuropathy), or if they lacked complete clinical and follow-up data.

All navigation-guided endoscopic explorations and optic nerve decompressions in this cohort were performed by a single senior surgeon (Y.T.). The primary surgical indications were progressive visual impairment, ophthalmoplegia, or both, secondary to a suspected fungal infection. Statistical analyses were executed utilizing GraphPad Prism version 8.0.1 and R software. To accommodate the small sample size, continuous variables are expressed as medians with interquartile ranges (IQRs), and categorical variables as frequencies (percentages). Comparisons between preoperative and postoperative parameters were conducted using paired Wilcoxon signed-rank tests. Furthermore, to quantify the magnitude of visual functional recovery and supplement traditional hypothesis testing, bias-corrected effect sizes (Hedges’ g) and their 95% confidence intervals (CIs) were calculated for paired observations. A two-tailed p < 0.05 was considered statistically significant, and a Hedges’ g ≥ 0.80 was interpreted as a large clinical effect.

Surgical procedures

Navigation-guided endoscopic sinus surgery and concurrent optic canal decompression were performed under general anesthesia. Following uncinectomy and complete ethmoidectomy to expose the sphenoethmoidal recess, the sphenoid sinus was opened to sample suspicious lesions for histopathological and microbiological diagnostics. Under real-time navigation, the medial wall and roof of the optic canal were localized and removed using a high-speed diamond drill in an anterior-to-posterior trajectory. Following complete bony removal and full visual exposure, a longitudinal incision of the optic nerve sheath was performed. The objective surgical endpoint was confirmed when normal saline irrigation yielded no further grayish white exudate from the incised sheath. Throughout the procedure, strict neural preservation was maintained by avoiding any direct mechanical manipulation of the optic nerve.

Postoperative management focused on targeted pharmacology. Empiric preoperative antifungal therapy was avoided to ensure accurate histopathological diagnosis. Postoperatively, regimens were strictly pathogen-directed: Aspergillus cases received voriconazole or liposomal amphotericin B (L-AmB) for 6–12 weeks, whereas Mucormycosis cases received intravenous L-AmB (7–14 days) followed by oral posaconazole or isavuconazole maintenance (3–6 months). Adjunctive treatments included stringent glycemic control, saline nasal irrigation, prophylactic antibiotics, and systemic mecobalamin. Therapy was maintained until complete clinical and radiological resolution, with dosages adjusted based on renal and hepatic monitoring [11–15].

Results

Clinical features

The demographic and clinical characteristics of the six patients (3 males, 3 females) are detailed in Table 1. The median age was 64.5 years (interquartile range [IQR], 45.8–69.8 years), and the median symptom duration was 1.5 months (IQR, 0.25–10.50 months). Four patients had concomitant type 2 diabetes mellitus; HbA1c levels for these patients were 7.2%, 9.4%, 5.4%, and 8.9%. The other two non-diabetic patients had HbA1c levels of 5.5% and 6.1%.

Table 1.

Clinical characteristics and treatment outcomes of patients

Case No.	Age/Sex	HbA1c (%)	Duration (months)	BCVA (LogMAR)	Radiological features		Fungal species	Residual Deficits
				Pre/Post-op	Bone Destruction	Soft Tissue Abnormality
1*	66/F	7.2	2	2.00/0.15	Focal erosion	Focal thickening	Aspergillus fumigatus	Resolved
2*	68/M	9.4	0.25	2.90/2.90	None	None	Rhizopus oryzae	Persistent NLP
3*	50/M	5.4	1	1.82/0.30	None	Focal thickening	Aspergillus fumigatus	Limited abduction
4	33/F	6.1	0.25	1.22/0.30	None	Focal thickening	Penicillium citrinum	Sluggish PLR
5*	63/F	8.9	2	2.30/2.30	Focal erosion	Focal thickening	Aspergillus fumigatus	Unchanged
6	75/M	5.5	36	0.15/0.10	None	Focal thickening	Aspergillus fumigatus	Resolved

BCVA, best-corrected visual acuity; HbA1c, glycated hemoglobin; LogMAR, logarithm of the minimum angle of resolution; PLR, pupillary light reflex. * Indicates patients with concomitant type 2 diabetes mellitus

Preoperative high-resolution CT and MRI revealed atypical paranasal sinus findings across all patients. The orbital apex predominantly exhibited hazy, patchy soft tissue shadows, conspicuously lacking the extensive soft tissue infiltration or widespread osteolysis classically associated with invasive fungal sinusitis. Specifically, two cases exhibited focal bone erosion with mucosal thickening, three demonstrated isolated focal mucosal thickening without osseous involvement, and one presented with unremarkable sinus contents and intact bony structures.

Representative case (Case 1)

A 66-year-old female with poorly controlled type 2 diabetes (HbA1c: 7.2%) presented with a 2-month history of intractable headache and progressive visual decline. Initially misdiagnosed with painful ophthalmoplegia and treated with corticosteroids, she experienced transient improvement followed by rapid deterioration. Admission BCVA was 2.00 LogMAR with positive RAPD, ptosis, and severe ophthalmoplegia (Fig. 1, column I). Perimetry showed near-total loss (VFI 0%, MD -31.59 dB). Imaging revealed localized sphenoid mucosal thickening and apical fluid signals (Fig. 1A–C).

Fig. 1.

Imaging, intraoperative, histopathological, and clinical findings of Case 1. (A, B) Preoperative axial CT indicating focal bone erosion (white arrow) and mucosal thickening. (C) T1-weighted MRI showing infiltration at the right orbital apex. (D, E) Intraoperative views: unremarkable sinus mucosa (D) and purulent fluid extrusion post-fascial incision (E). (F) Mucosal specimen. (G, H) H&E and GMS staining confirming fungal hyphae. (I, J) Gaze positions pre- (I) and 3 months postoperatively (J), showing recovery of ptosis and abduction

Emergency navigation-guided surgery was performed. Although the sphenoid mucosa appeared macroscopically normal (Fig. 1D), incising the orbital fascia released purulent fluid (Fig. 1E). Pathology confirmed Aspergillus fumigatus (Fig. 1G–H). Three months postoperatively, BCVA improved to 0.15 LogMAR with resolved headache and recovered motility (Fig. 1J). Perimetry confirmed functional salvage (VFI 83%, MD -6.74 dB), with no recurrences noted.

Surgical outcomes

All patients underwent navigation-guided endoscopic clearance of sphenoid lesions and optic nerve decompression. Intraoperative evaluation revealed no specific fungal signs (e.g., necrosis or eschar); however, in select cases, incising the apical orbital fascia released turbid intrathecal fluid, suggesting localized infectious exudation. Postoperative histopathology and special staining (PAS/GMS) identified Aspergillus fumigatus (n = 4), Penicillium citrinum (n = 1), and Rhizopus oryzae (n = 1). Intraoperative smears were positive in three cases (Cases 1, 2, and 4), while the remaining three were smear-negative despite histopathological detection of characteristic hyphae, reinforcing histopathology as the diagnostic gold standard [16–18].

Preoperative headaches in four patients resolved within one week postoperatively. The median BCVA improved from 1.91 LogMAR (IQR: 1.22–2.30) preoperatively to 0.30 LogMAR (IQR: 0.15–2.30) at final follow-up (p = 0.13). Specifically, four patients (Cases 1, 3, 4, and 6) achieved substantial visual gains, while two with profound preoperative loss (Cases 2 [NLP] and 5 [HM]) remained unchanged. In the four patients completing perimetry, the median VFI increased from 11% to 91.5% (p = 0.13), and the median MD improved from − 28.02 dB to -3.33 dB (p = 0.13), indicating partial reversibility of visual function.

Comprehensive assessments (Supplementary Table 1) revealed that postoperative recovery correlated with metabolic status: the two patients without visual improvement (Cases 2 and 5) exhibited the highest preoperative glucose levels (18.62 and 14.04 mmol/L, respectively). Conversely, patients with better glycemic control (Cases 1, 4, and 6) achieved significant recovery. Two of the three patients with preoperative ophthalmoplegia achieved complete recovery. No major perioperative complications or recurrences occurred during the median follow-up of 3.5 months (IQR: 2.25–7.75 months; range: 2–9 months).

Effect size analysis

To supplement traditional significance testing, we calculated paired bias-corrected effect sizes (Hedges’ g) with 95% confidence intervals (CIs) (Table 2). Improvement in best-corrected visual acuity (BCVA) yielded a large effect size (g = 0.80; 95% CI: -0.11 to 1.66; n = 6). Similarly, visual field parameters in the four patients with complete paired data demonstrated very large effect sizes for both the visual field index (VFI) (g = 1.37; 95% CI: 0.00 to 2.69) and mean deviation (MD) (g = 1.23; 95% CI: -0.07 to 2.46). Although limited statistical power inherent to our small sample size resulted in wide CIs crossing or touching the null value, these large effect sizes should be interpreted conservatively as exploratory signals for postoperative visual recovery, warranting further validation in larger prospective cohorts.

Table 2.

Quantitative analysis of visual functional outcomes and effect sizes

Outcome Measure	No. of Patients	Pre-op	Post-op	Effect Size (95% CI)
BCVA	6	1.91 (0.95, 2.45)	0.30 (0.14, 2.45)	0.80 (-0.11, 1.66)
VFI (%)	4	11 (2, 36)	91 (32, 100)	1.37 (0.00, 2.69)
MD (dB)	4	-28.02 (-30.81, -20.27)	-3.33 (-22.21, 0.58)	1.23 (-0.07, 2.46)

VFI, visual field index. MD, mean deviation. Effect sizes (Hedges’ g) and corresponding 95% CIs were calculated from paired observations to quantify the magnitude of clinical change, with g ≥ 0.8 interpreted as a large effect. Visual field parameters (VFI and MD) were assessed in a subset of 4 patients capable of completing the examination

Discussion

The anatomical proximity between the sphenoid sinus and the optic canal renders even minor pathologies capable of precipitating severe neurological dysfunction [4, 19]. Given the diverse etiologies of OAS, including rare structural anomalies, diagnosis heavily relies on imaging features [20, 21]. Atypical cases lacking clear radiological signs therefore pose significant diagnostic challenges, often requiring prompt endoscopic exploration [22]. Our study identifies a deceptive clinical presentation of invasive fungal rhinosinusitis (IFRS) characterized by clinical-radiological dissociation. Contrary to classic IFRS models involving extensive osteolytic destruction or macroscopic mass effects, our cohort developed rapidly progressive orbital apex syndrome (OAS) despite unremarkable or strictly subtle sinonasal imaging. These findings underscore the clinical value of early recognition and prompt multidisciplinary treatment, integrating navigation-guided endoscopic exploration with systemic antifungal and metabolic therapies, in rescuing visual function.

IFRS is a life-threatening condition traditionally categorized into distinct clinicopathological models [23–26]. However, our cohort diverges significantly from these established paradigms. Unlike the acute fulminant form characterized by storm-like tissue necrosis [23], our cohort shares certain microbiological overlaps with chronic invasive fungal rhinosinusitis (CIFRS), particularly the predominance of Aspergillus species. In this context, “occult IFRS” represents a specific clinical radiological presentation pattern rather than a novel pathological diagnosis. While classic CIFRS typically follows a highly indolent course evolving over months to years, hallmarked by dense fungal accumulation, granuloma formation, and obvious macroscopic mass effects [26], our cases present a markedly different clinical trajectory. Despite lacking macroscopic granulomas or widespread bone erosion, our patients demonstrated a unique neurovascular tropism. This cryptic microscopic invasion precipitated rapidly progressive neurological deficits far earlier than the classic progression timeline of CIFRS. Fungal lesions in this series disseminated with minimal tissue burden, spreading cryptically along perineural sheaths and microvascular spaces instead of eroding the bony framework. This microscopic invasion induced severe neurological compromise before the emergence of classic radiological signs. Given the limited sample size, this pathophysiological subtype is still hypothesis-generating, and we look forward to future cases to further validate this pathology.

We speculate that the observed clinical-radiological dissociation may be rooted in specific micro-pathological mechanisms. Based on existing literature, refractory headache could be attributed to the angiotropic and neurotropic properties of the fungi, while hazy infiltration of the orbital apex on MRI suggests perineural invasion along trigeminal nerve branches. Hyphal invasion of the microvasculature induces ischemic pain, a microscopic event that temporally precedes macroscopic bone erosion. Furthermore, the anatomical vulnerability of the confined orbital apex exacerbates the impact of inflammatory soft tissue, which can form a cuff-like encasement within the superior orbital fissure. This mechanical compression, combined with direct infiltration, precipitates multiple cranial nerve palsies [9, 27]. The absence of antecedent nasal symptoms constitutes a major diagnostic pitfall stemming from the pathogen’s propensity for submucosal spread. By disseminating microscopically along vascular and neural bundles deep within the lamina propria, the fungi preserve the macroscopic integrity of the sinonasal mucosal epithelium and maintain sinus ostium patency [28, 29]. Clinicians must therefore maintain a high index of suspicion for this occult subtype when encountering the triad of refractory headache, progressive ophthalmoplegia, and an asymptomatic nasal status.

Our study adds important nuance to conventional imaging diagnostic criteria. Paradigm models emphasized extensive bone erosion as a hallmark diagnostic indicator for IFRS, leading clinicians to prematurely exclude the diagnosis when osteolysis is absent [23]. However, our results demonstrate that lethal neural invasion frequently occurs prior to macroscopic bone destruction in the orbital apex region. This is pathologically corroborated by evidence that fungal hyphae can traverse intact Haversian canals or neural foramina without destroying the osseous structure [27]. Notably, nasoendoscopy may yield negative findings even in confirmed acute IFRS, further confounding the initial workup. For instance, isolated OAS can present without mucosal necrosis or purulent discharge, with infection ultimately proven only via optic canal biopsy [30]. Consequently, our findings suggest that the absence of frank bone destruction should not preclude an IFRS diagnosis. For orbital apex lesions of unknown etiology, clinicians might consider surgical exploration even in the presence of strictly subtle soft tissue abnormalities.

Treatment outcomes in our cohort provide crucial data for assessing early surgical intervention. While the small sample size precluded traditional statistical significance for visual acuity improvement (p = 0.13), effect size analysis revealed large improvements in BCVA (Hedges’ g = 0.80) and even larger improvements in VFI and MD (g = 1.37 and 1.23, respectively). These quantitative findings suggest a potential trend toward functional recovery. Therefore, when patients present with specific high-risk clinical features, including a metabolic background (e.g., poorly controlled diabetes), rapidly progressive neuro-ophthalmologic deficits, and a deceptive clinical-radiological dissociation, We postulate that early surgical decompression creates a critical window for recovery by alleviating localized inflammatory compression within the confined orbital apex, which, together with antifungal administration and metabolic optimization, represents the most rational, vision-rescuing strategy for highly suspicious, rapidly progressive syndromes [31].

However, this therapeutic time window is finite and governed by baseline metabolic status. Our data demonstrate that severe hyperglycemia acts as a catalyst for the rapid closure of this window, as the two patients with unrecovered vision both presented with glucose levels exceeding 14 mmol/L. This aligns with the meta-analysis by Jeong et al. [32]. The underlying mechanism involves a synergistic assault on host defenses and microvascular integrity. Hyperglycemic and acidotic conditions disrupt transferrin-iron binding, releasing free iron that certain fungi utilize to exponentially enhance their growth [33]. Simultaneously, chronic hyperglycemia exacerbates microvascular endothelial damage, which, when combined with fungal angiotropism, accelerates ischemic occlusion of the optic nerve’s vascular supply [34]. This metabolic-infection synergy irreversibly transforms reversible inflammatory edema into ischemic necrosis. Therefore, rigorous perioperative glycemic management must be elevated to a primary sight-saving therapy, equally critical to surgical decompression and antifungal administration.

Limitations

Several limitations of this study should be noted. First, the single-center retrospective design and small sample size of this rare pathology (n = 6) limit statistical power and introduce potential selection bias, precluding multivariate regression to isolate independent prognostic factors such as specific glycemic thresholds or fungal species. Second, the study lacks a non-surgical control group. Establishing such a cohort for imaging-occult IFRS is clinically and ethically unfeasible because surgical exploration is the definitive modality for histopathological confirmation in patients without gross mucosal necrosis. Withholding surgical debridement, a cornerstone of care for this lethal disease, would result in missed diagnoses and constitute an ethical violation. Therefore, a self-controlled pre- and post-operative design was the most rigorous and ethically sound methodology available to quantify the efficacy of early intervention. Third, while the median follow-up (3.5 months) limits the evaluation of long-term recurrence and final visual outcomes, it adequately captures the critical acute phase necessary to assess the immediate efficacy of early surgical decompression—the primary focus of this study. Nonetheless, future studies with extended observation periods are warranted.

Conclusions

In sphenoid-orbital apex IFRS, a deceptive subset of patients presents with a clinical-radiological paradox, developing severe orbital apex syndrome without classic imaging signs such as bone destruction. These cases are frequently misdiagnosed as noninfectious inflammatory disorders. Consequently, occult IFRS must be prioritized as a primary differential diagnosis for acute orbital apex syndrome of unknown etiology, even when sinonasal imaging appears unremarkable. Early navigation-guided endoscopic exploration is essential to secure a histopathological diagnosis and provide a critical window for functional recovery through concurrent optic nerve decompression.

Supplementary Information

Below is the link to the electronic supplementary material.

Abbreviations

BCVA

Best-corrected visual acuity

LogMAR

Logarithm of the minimum angle of resolution

HM

Hand motion

NLP

No light perception

PLR

Pupillary light reflex

RAPD

Relative afferent pupillary defect

VFI

Visual field index

MD

Mean deviation

HbA1c

Glycated hemoglobin

OAS

Orbital apex syndrome

Author contributions

Zhixing Li: Data Curation, Formal Analysis, Writing – Review & Editing. Weibin Liu, Yanchen Zhang: Investigation, Resources, Validation. Jieliang Shi, Yunhai Tu, Wencan Wu: Conceptualization, Methodology, Investigation, Writing – Review & Editing, Supervision. All authors reviewed and approved the final manuscript.

Funding

This work was supported by the National Key Research and Development Program of China (Grant No. 2021YFA1101200) and the National Natural Science Foundation of China (Grant No. 82171048).

Data availability

The datasets generated and/or analyzed during the current study are not publicly available due to patient privacy considerations but are available from the corresponding author on reasonable request.

Declarations

Ethics approval and consent to participate

This study was performed in line with the principles of the Declaration of Helsinki. Approval was granted by the Ethics Committee of the Eye Hospital of Wenzhou Medical University (Approval No. 2026-037-K-027). Due to the retrospective nature of the study, the requirement for informed consent was waived by the Ethics Committee of the Eye Hospital of Wenzhou Medical University. This decision was in accordance with the national regulations on ethical review of medical research involving human subjects.

Consent for publication

Written informed consent, which included permission for the use of clinical images and data in scientific publication, was obtained from all participants or their legal guardians.

Competing interests

The authors declare no competing interests.