C horoidal melanoma is the most common primary intraocular malignancy in adults and remains a significant cause of visual morbidity and mortality. The Collaborative Ocular Melanoma Study (COMS) and subsequent American Joint Committee on Cancer (AJCC) classification have established tumor size as a dominant prognostic factor, with incremental risk of metastasis linked to increasing thickness and largest basal diameter. ^,, In addition, subretinal fluid (SRF) and macular involvement are considered important for visual prognosis in benign intraocular tumors. ^,, Therefore, accurate, reproducible measurement of tumor dimensions-specifically apical thickness and largest basal diameter (LBD) with minimal patient burden is essential for staging, prognostication, and treatment planning in ocular oncology practice.

Traditionally, B-scan ultrasonography (US) has served as the reference standard for measuring choroidal tumor dimensions, particularly thickness. Although widely available, US has limitations, including operator dependence, lower axial resolution, longer exam duration, and a tendency to overestimate tumor height relative to optical coherence tomography (OCT). Enhanced depth imaging OCT (EDI-OCT) and spectral-domain OCT (SD-OCT) have improved visualization of small, posteriorly located tumors. However, their restricted field-of-view (typically 30°-55°) and limited penetration into the choroid hinder consistent evaluation of larger or peripheral lesions. A previous study demonstrated that ultrasonography overestimated thickness by approximately 55% compared with EDI-OCT for small melanomas, highlighting the discrepancy between modalities and the need for improved depth-resolved imaging. This discrepancy has critical implications for treatment planning, particularly in determining eligibility for brachytherapy vs enucleation based on COMS size criteria, where a 1-2 mm difference in measured thickness can alter management decisions. Magnetic resonance imaging (MRI) captures choroidal melanoma well but provides limited spatial resolution for intraocular and retinal details.

Various benign and secondary choroidal tumors also require precise imaging. Choroidal hemangioma, a benign vascular hamartoma, often causes exudative retinal detachment and requires differentiation from amelanotic melanoma. Choroidal osteoma, a rare ossifying tumor usually affecting young women, is characterized by calcified plaques in the juxtapapillary or macular region that can develop choroidal neovascularization and progressive visual decline. Metastatic choroidal tumor represents the most common intraocular malignancy overall, frequently originating from breast and lung cancer and presents as creamy, plateau-like lesions and may be multifocal or bilateral. For these entities, as with melanoma, precise measurements of tumor dimensions and assessment of secondary complications such as SRF and macular involvement are critical for management.

Ultrawide-field (UWF) imaging and optical coherence tomography (OCT) have advanced imaging technologies in terms of tumor detection, following tumor size after treatment, recurrence monitoring, and surgical planning. ^, Recent advances in swept-source OCT (SS-OCT), operating at longer wavelengths (∼1060 nm), allow for deeper choroidal penetration and visualization of the choroidoscleral interface.

The recently developed ultra-widefield swept-source OCT (Xephilio OCT-S1; Canon Medical Systems, Tokyo, Japan) extends this capability further by achieving single-acquisition coverage of up to 23 × 20 mm, corresponding to approximately 80° retinal field of view with 5.3 mm penetration depth. This ultrawide-field enables seamless visualization from the macula to the peripheral retina, encompassing the choroid-scleral interface, without the need for montage acquisition protocols.

UWF SS-OCT offers distinct clinical advantages in ocular oncological applications, wherein accurate characterization of tumor morphology, dimensional parameters, and choroidal structural relationships constitutes a fundamental prerequisite for optimal therapeutic decision-making. This technological advancement proves particularly advantageous for longitudinal surveillance protocols, where examination reproducibility, patient compliance, and imaging tolerability are critical.

Despite these technical advantages, there is limited clinical data evaluating the performance of ultra-widefield OCT in the context of ocular oncology. While 1 study has shown UWF-OCT can capture peripheral retinal degenerations and breaks easily, its applicability across tumor sizes and morphologies, its concordance with established modalities such as US and MRI, and factors to consider when applying this specific modality in ocular oncology remain poorly reported.

In this study, we investigated patients with choroidal melanoma, hemangioma, osteoma, and metastatic choroidal tumor using UWF swept-source OCT to determine its applicability across tumor types and size categories, to cf quantitative measurements with US, SD-OCT, and MRI, and identify predictors of UWF-OCT measurement quality.

This was a retrospective diagnostic comparison of patients diagnosed with choroidal tumors between January 2023 and August 2025 at a high-volume, referral-based tertiary hospital: Gangnam Severance Hospital, affiliated with Yonsei University College of Medicine, Seoul, Korea. Institutional Review Board (IRB) approval was obtained from Gangnam Severance Hospital IRB (No. 3-2025-0211), and the study adhered to the tenets of the Declaration of Helsinki. The requirement for informed consent was waived by the IRB due to the retrospective nature of the study and the absence of identifiable patient information.

Consecutive patients with a clinical diagnosis of choroidal tumor (including choroidal melanoma, hemangioma, osteoma, and metastatic choroidal tumor) who underwent multimodal imaging during the study period were screened. Eyes with inadequate image quality due to media opacity or incomplete records for primary outcomes were excluded to minimize information bias. Clinical variables collected included demographics, tumor type, location, tumor features, and staging when applicable, with tumor dimensions assessed according to Collaborative Ocular Melanoma Study (COMS) criteria using largest basal diameter and apical thickness. Sex was defined as the biological sex recorded in the patients’ medical records and was categorized as male or female.

Clinical diagnosis of choroidal melanoma was established based on characteristic multimodal imaging findings including tumor dimensions, acoustic hollowness on B-scan ultrasonography, optical shadowing on OCT, and associated features including subretinal fluid and orange pigment. All diagnoses were confirmed by experienced ocular oncologists at a high-volume tertiary referral center. TFSOM-related features were considered assisting risk-assessment parameters within the overall multimodal clinical evaluation and were not used as sole diagnostic criteria for choroidal melanoma. For diagnostic classification and TFSOM-related height thresholds, conventional ultrasonographic measurements rather than UWF-OCT-derived thickness measurements were used. For small lesions where clinical findings were insufficient for definitive diagnosis of either choroidal melanoma or benign nevus, the lesion was classified as an indeterminate choroidal melanocytic lesion (ICML). In cases where the diagnosis remained uncertain despite multimodal imaging assessment, fine-needle aspiration biopsy (FNAB) was performed.

All eyes underwent ultra-widefield swept-source OCT imaging using the Xephilio OCT-S1 (Canon Medical Systems Corporation, Tochigi, Japan). A single-capture 23 × 20 mm raster volume of 807 segmented images was acquired per eye with UWF-OCT. Additional imaging included spectral-domain OCT (SD-OCT, Heidelberg Spectralis HRA + OCT, Heidelberg Engineering, Heidelberg, Germany), B-scan ultrasonography (US) (Ellex Eye Cubed ultrasound system, Ellex Medical, Adelaide, Australia), and orbital MRI (Ingenia Elition X system, Philips Healthcare, Best, the Netherlands). Wide-field fundus images were captured with the Optos P200DTx California system (Optos, Dunfermline, UK) to document tumor location and extent.

Figure 1 illustrates the tumor segmentation, interpolation, and measurement protocol. The tumor’s anterior margin (curve A) was drawn by auto-segmentation of UWF-OCT at the base of the RPE. The tumor’s posterior margin (curve B) was produced by connecting points at the tumor base of the hyperreflective junction of the inner sclera at the choroido-scleral interface (CSI). If applicable, UWF-OCT’s auto-segmentation tool of curve A and curve B was used. In cases with dense posterior shadowing where the CSI was obscured, Curve B was generated using the Canon RX OCT viewer’s built-in drawing tool, which requires 3 reference inputs: 2 longitudinal base elevation points placed at the lateral margins of the tumor base, and one central point. After the 2 lateral base elevation points were identified at the tumor base margins—confirmed using invert, angiogram, Fit Scale, and High-Resolution display modes to ensure objective boundary identification and minimize reliance on standard grayscale contrast—the line connecting these 2 points was drawn and measured, and its geometric midpoint was identified. The central reference point was then placed posterior to this midpoint along the perpendicular bisector of the basal chord, at the expected continuation of the adjacent visible choroidoscleral interface curvature, verified in Real Scale mode. As the central point was moved, the software automatically generated the corresponding circular eyeball contour and CSI layer. The central point was therefore constrained by the 2 lateral base elevation points and adjacent ocular curvature, rather than being freely adjusted to alter tumor height. Because the software generates a circular arc from these 3 predefined noncollinear reference points, the resulting Curve B is software-generated CSI contour rather than a free-hand drawing.

UWF-OCT–based method for measuring choroidal tumor dimensions. (A) Fit Scale view with RGB fundus photograph inset. Curve A (anterior margin) is auto-segmented at the base of the retinal pigment epithelium; Curve B (posterior margin) follows the choroidoscleral interface (CSI). Tumor thickness (2088 µm) is the maximal perpendicular distance between Curves A and B; LBD (9619 µm) is the maximal chord along the CSI. (B) In cases where posterior shadowing obscures the CSI, 3 reference points (arrows) are placed at the 2 lateral base margins and one central point, positioned posterior to the midpoint of the basal chord as described in the Methods. (C) The Canon RX OCT viewer automatically generates a smooth circular eyeball contour from the 3 reference points, defining the interpolated Curve B. (D) Real Scale mode confirming true anatomical proportionality and geometric perpendicularity of measurement lines (thickness 2088 µm, LBD 9619 µm); all measurements were performed in this mode. (E) Angiogram and invert mode demonstrating objective delineation of the RPE boundary (Curve A) and lateral base elevation points, used for consensus verification of correct measurement placement. (F) Invert mode alone showing clear CSI layer visualization and lateral base elevation onset, confirming the reliability of boundary identification using the Canon RX OCT viewer.

The physiological basis for this approach rests on the established observation that choroidal tumors originate within the choroid and characteristically grow inward toward the vitreous cavity, with the rigid sclera acting as a natural mechanical barrier to outward extension. ^,, All patients underwent routine orbital MRI, which confirmed absent extra-scleral extension and intact scleral curvature in all measurable cases, validating the use of circular eyeball contour interpolation as a reliable posterior tumor boundary reference.

Tumor thickness was measured as the maximum perpendicular distance between curves A and B, following the protocol established by Shields et al for EDI-OCT measurement of small choroidal melanomas. Largest basal diameter (LBD) was defined as the maximum linear basal chord across the base of the tumor measured at CSI layer. In ultrasonography, LBD of the tumor is usually measured by placing calipers from one border of the tumor to another border. Following the methodology originally described for ultrasonographic LBD measurement by Franca et al, we adapted this approach for UWF-OCT–based measurement at the CSI layer: LBD was measured using 2 straight-line segments for tumors ≥12 mm and a single straight-line segment for tumors <12 mm. In mushroom-shaped choroidal melanomas, LBD was defined as the maximum linear chord at the choroidoscleral interface at the base of the tumor stalk, consistent with COMS basal diameter definition and ultrasonographic caliper placement methodology. The Canon RX OCT viewer software offers 2 display modes: “Real Scale” displays images in true anatomical proportions without distortion, while “Fit Scale” optimizes images to fill the viewing window. All measurements were performed using the Canon RX OCT viewer’s “Real Scale” mode, which displays images in true anatomical proportions without scaling distortion, ensuring accurate dimensional assessment and rechecked for clear boundary in “Fit scale” and “High Resolution” mode. To objectively verify lateral base elevation point placement and RPE boundary identification, the Canon RX OCT viewer’s angiogram and invert display modes were additionally utilized where needed. The angiogram mode overlays flow information and inverts signal intensity, enhancing visualization of the RPE boundary and choroidal elevation onset independent of grayscale contrast. The invert mode alone provides superior delineation of the CSI layer and lateral base margins, particularly in cases where standard grayscale contrast is insufficient for confident boundary identification. ( Figure 1 E and 1 F).

All UWF-OCT measurements were independently performed by 2 experienced observers (M.K., a vitreoretinal specialist, and T.Y.K., a vitreoretinal fellow) who were masked to measurements from other imaging modalities. Discrepancies exceeding 10% were adjudicated through consensus review. Measurements from ultrasonography and MRI were extracted from clinical reports generated by certified ultrasonographers and radiologists, respectively.

Thirty-nine eyes from 39 patients were included, with a mean age of 59 years (median 61, range 19-94) of whom 18 (46.2%) were male. Mean follow-up was 8.5 months (median 7, range 2.5-53.4) from initial diagnosis. Eight patients (20.5%) had prior cancer history, most commonly lung cancer. Tumors included choroidal melanoma (n = 27, 69.2%), metastatic choroidal tumor (5, 12.8%), choroidal hemangioma (4, 10.3%), choroidal osteoma (2, 5.1%), and indeterminate choroidal melanocytic lesion (1, 2.6%). Tumors were located within the macula in 20 eyes (51.2%), extra-macula in 19 eyes (48.7%), and specifically peripapillary in 8 eyes (20.5%).

By COMS classification, 11 tumors (28.2%) were small, 20 (51.3%) medium, and 8 (20.5%) large. Among melanomas, 13 (48.1%) were AJCC Stage I, 5 (18.5%) Stage IIA, and 4 (14.8%) Stage IIB. Twenty-nine tumors (74.4%) were dome-shaped and 7 (17.9%) mushroom-shaped. Subretinal fluid was present in 32 eyes (80.0%). Baseline characteristics are summarized in Table 1 .

TABLE 1

Baseline Characteristics for Patients With Choroidal Tumor (n = 39, 39 Eyes).

Demographic Features
Age, years
Mean ± SD	59.2 ± 16.9
Median (range)	61 (19-94)
Sex , n (%)
Male	18 (46.2%)
Female	21 (53.8%)
Hypertension, n (%)	18 (46.2%)
Diabetes, n (%)	7 (17.9%)
Cancer history, n (%)	8 (20.5%)
Lung cancer	3 (7.7%)
Breast cancer	2 (5.1%)
Thyroid cancer	2 (5.1%)
Renal cancer	1 (2.6%)
Eyes, n (%)
Right	18 (46.2%)
Left	21 (53.8%)
Clinical features
Color, n (%)
Pigmentation	22 (56.4%)
Nonpigmented	11 (28.2%)
Mixed	6 (15.4%)
Location of Tumor, n (%)
Macula	20 (51.2%)
Extra-macula	19 (48.7%)
Peripapillary	8 (20.5%)
Tumor shape, n (%)
Dome	29 (74.4%)
Mushroom	7 (17.9%)
Others	3 (7.7%)
Ultrasonography measurements
Tumor thickness, mm	4.9 ± 2.9 (1.3-12.9)
Largest basal diameter, mm	10.0 ± 3.6 (4.0-21.1)
Tumor type, n (%)
Choroidal Melanoma	27 (69.2%)
Metastatic choroidal tumor	5 (12.8%)
Choroidal hemangioma	4 (10.3%)
Choroidal Osteoma	2(5.1%)
Indeterminate choroidal melanocytic lesion	1 (2.6%)
Choroidal tumor size (COMS staging), n (%)
Small	11(28.2%)
Medium	20(51.3%)
Large	8 (20.5%)
Melanoma (AJCC staging), n (%) (n = 27)
Stage I	13 (48.1%)
Stage IIA	5 (18.5%)
Stage IIB	4 (14.8%)
Stage IIIA	1 (3.7%)
Stage IIIB	1 (3.7%)
Stage IV	3 (11.1%)
Subretinal fluid, n (%)	32 (80.0%)
Subretinal hemorrhage, n (%)	3 (7.7%)
Retinal detachment, n (%)	4 (11.4%)
Brachytherapy, n (%)	34 (87.2%)
Transpupillary thermotherapy, n (%)	23 (59.0%)
Follow up duration (months), Mean ± SD (range)	8.5 ± 8.9 months (2.5-53.4)

Only gold members can continue reading. Log In or Register to continue

Share this:

• X
• Facebook

Related

Related posts:

1. Zeb1 Mediates Injury-Induced Mesenchymal Transition in the Mouse Corneal Endothelium In Vivo
2. Associations Between Routine Vaccinations and the Risk of New-Onset Idiopathic Uveitis
3. Second-Generation ELZA-sub400 Protocol: Individualized High-Fluence Cross-Linking for Ultra-Thin Keratoconus Corneas
4. Comment on: Visual Field Progression in Patients With Glaucoma With Delayed Follow-Up

Stay updated, free articles. Join our Telegram channel

Join