Choroidal thickness, ganglion cell complex, and photoreceptor outer segment length evaluation in patients receiving tamoxifen therapy by spectral domain optical coherence tomography

Background To evaluate choroidal thickness, ganglion cell complex (GCC) and photoreceptor outer segment length were measured in patients with breast cancer undergoing tamoxifen therapy, using spectral-domain optical coherence tomography (SD-OCT); results were compared with those for normal eyes. Methods Forty-four patients with breast cancer, undergoing tamoxifen therapy, and 41 healthy controls were included in this prospective, comparative study. All participants underwent a complete ophthalmologic evaluation and SD-OCT. Subfoveal, nasal (nasal distance to fovea 500, 1000, 1500 μm), and temporal (temporal distance to fovea 500, 1000, 1500 μm) choroidal thickness measurements were performed using the enhanced depth imaging mode of SD-OCT. Using an Early Treatment Diagnostic Retinopathy Study (ETDRS) circle at the macular level, the automated retinal segmentation software was applied to determine the thickness of the GCC. The photoreceptor outer segment (PROS) length was determined manually, as the distance from the inner surface of the ellipsoid zone to the inner surface of retina pigment epithelium. tamoxifen therapy. Increased choroidal thickness, PPE and thinning GCC were detected in tamoxifen users. These OCT findings may be an early indicator of retinal toxicity for patients undergoing tamoxifen therapy in the follow-up period.

tamoxifen therapy. Increased choroidal thickness, PPE and thinning GCC were detected in tamoxifen users. These OCT findings may be an early indicator of retinal toxicity for patients undergoing tamoxifen therapy in the follow-up period.
Keywords tamoxifen retinopathy, choroidal thickness, ganglion cell complex, PROS, spectral-domain optical coherence tomography Background Tamoxifen, a selective estrogen-receptor modulator, is recommended to treat early and advanced-stage breast cancer in premenopausal and postmenopausal women [1]. It reduces the available estrogen to cancer cells by competitively inhibiting the binding of estrogen to the estrogen receptors on breast tissues [2].
We aimed to evaluate structural changes by spectral-domain optical coherence tomography (SD-OCT) before development of pathological retinal findings.

Methods
This prospective, comparative clinical study was carried out between September 2018 and January 2019 in Okmeydani Training and Research Hospital, Istanbul, Turkey. Informed consent was obtained from all participants, and the study was carried out in agreement with the Declaration of Helsinki for research involving human subjects. The first group comprised patients with breast cancer, undergoing tamoxifen therapy, following up at Okmeydani Training and Research Hospital Radiation oncology clinic; the control group comprised age-and gender-matched healthy subjects who were examined at the Eye Clinic of the same hospital. Patients who had been receiving tamoxifen therapy for less than one year were not included. All patients were receiving tamoxifen 20 mg per a day.
Retinal diseases (such as epiretinal membrane, vein occlusion, macular hole, vitreomacular traction syndrome, or diabetic retinopathy), any systemic disorders, previous ocular surgery or trauma, congenital malformations of the eye, glaucoma and ocular hypertension, significant media opacities precluding fundus examination and/or imaging, best corrected visual acuity (BCVA) worse than 20/20, spherical equivalent refractive errors of more than ± 3 dioptre, any medication use within the last three months, history of smoking or alcohol intake, and pregnancy were evaluated as exclusion criteria. All participants' haemoglobin, blood vitamin D, vitamin B12, iron, thyroid functional tests (T3, T4, and TSH levels), and systolic and diastolic blood pressure values were within the normal range.
A complete ophthalmic examination, including measurement of BCVA using Snellen charts, Goldmann applanation tonometry, slit-lamp biomicroscopy, and dilated fundus examination, was conducted. Axial length (AL) was evaluated using an AL-Scan optical biometer (Nidek Co., Gamagori, Japan). Choroidal thickness was measured using the enhanced depth imaging (EDI) mode of an SD-OCT (Spectralis HRA+OCT; Heidelberg Engineering Inc., Heidelberg, Germany). Twenty-five, each comprised of 40 averaged scans, were obtained in a 10° × 20° rectangle centred on the fovea. The choroid was measured from the outer portion of the hyperreflective line corresponding to the retinal pigment epithelium-Bruch's membrane complex to the inner surface of the sclera/choroidal junction (manually drawn by the examiner). All SD-OCT measurements were performed between 10:00 and 11:00 am. Subfoveal, nasal, and temporal measurements of choroidal thickness were conducted manually. Nasal and temporal choroidal thickness measurements were carried out at 500 μm intervals from the fovea centralis (500, 1000 and 1500 μm distance to the fovea centralis; Figure 1).
Choroidal thicknesses of eyes with pachychoroid phenotype that were greater than 300 μm, with and without retina pigment epithelium abnormalities, were evaluated as PPE and uncomplicated pachychoroid epitheliopathy (UCP), respectively [12].
Using an Early Treatment Diagnostic Retinopathy Study (ETDRS) circle at the macular level, the automated retinal segmentation software was applied to determine thicknesses of the ganglion cell complex (GCC) by adding the parameters of the macular retinal nerve fibre, macular ganglion cell, and macular internal plexiform layers ( Figure 2).
The photoreceptor outer segment (PROS) length was determined manually, as the distance from the ellipsoid zone to the inner surface of the retina pigment epithelium (RPE) after automatic retinal segmentation ( Figure 1).
Only the right eye of each participant was evaluated for statistical analysis. Choroidal thickness and PROS length were measured manually by two independent graders without prior information about the subjects, and the averages of their measurements were used in the statistical analysis. The reliability of measurements between the two graders was evaluated using the intraclass correlation coefficient (Table 1).
Statistical analyses were performed using IBM SPSS Statistics version 21. The variables were investigated using histograms and analytical methods to test whether they are normally distributed or not. Student's t-test and the chi-squared test were used to compare the variables between groups. The effects of age and axial length were adjusted using ANCOVA test. A 5% type I error level was used to infer statistical significance.

Results
A total of 85 eyes were enrolled in this prospective case-control study. Forty-four patients constituted the tamoxifen group, whereas 41 age and sex-matched subjects constituted the control group. All participants were female. Patients' mean age was 51.6 ± 7.4 years in the tamoxifen group and 52.0 ± 5.5 years in the control group (p = 0.786). The mean duration of tamoxifen use in the study group was 3.95 ± 1.91 (1-9) years. The groups had comparable ALs (22.7 ± 0.4 vs 22.7 ± 0.3; p = 0.713).
The mean choroidal thickness was statistically greater in the tamoxifen group than the control group in all quadrants (p < 0.001 for all quadrants). The results are shown in Table   2. Figure Table 3).
No statistically significant difference was found between the groups in mean subfoveal PROS length, mean central macular thickness (CMT), or GCC thickness in the rest of the outer rings (p = 0.640, 0.144, 0.122, 0.233, and 0.468, respectively; Tables 4 and 3).

Discussion
The incidence of tamoxifen-associated ocular toxicity is estimated to be between 0.9% and 12% [13]. Generally, ocular toxicity occurs with high-dose therapy (cumulative dose > 100 g) [14]. There are no controlled studies in the literature that evaluated structural optical coherence tomography (OCT) changes under tamoxifen therapy. Reported OCT findings associated with tamoxifen depend on case series; these findings are ellipsoid zone defect, thinning of the inner retinal layers, cystic foveal cavitation, and crystalline deposits in the inner retina [15].
In our study, the mean choroidal thickness measurements in the subfoveal, nasal, and temporal quadrants were all grater in tamoxifen users than in the control group. UCP was detected in 75% of all tamoxifen users. PPE was detected in five patients (11.3%). The  [10,16]. In our study, the five patients with PPE in one eye had UCP in the fellow eye. We detected CSCR findings in a patient receiving tamoxifen therapy. The risk factors for CSCR are well documented; these are exogenous cortisol therapies, endogenous hypercortisolism, male sex, exogenous testosterone therapy, and polycystic ovary syndrome [17,18,19]. Furthermore, several studies have shown that choroidal thickness increases in various inflammatory diseases such as Vogt -Koyanagi Harada disease and Familial Mediterranean Fewer disease [20,21].
Estrogen receptors (ER) in the retina were first demonstrated in rats [22], and then both types of ER receptors (ER α and ER β) [23] were identified in human retinas. Tamoxifen is a partial agonist, which activates ER α with lower efficacy compared to estrogen [24].
Reduced expression of ER α is associated with CSCR [25]. In addition, serum testosterone, cortisol, and cortisone levels increase during tamoxifen therapy [26]. Testosterone has been shown to increase choroidal vascular permeability and vascular dilatation [27].
These mechanisms may explain increasing choroidal thickness, development of PPE and CSCR in tamoxifen users.
In vitro studies have shown that tamoxifen is toxic to human RPE and photoreceptors [28].
Maculopathy associated with tamoxifen was demonstrated in previous reports. Case series in the literature include crystalline maculopathy [7], macular hole [8], and pseudocystic foveolar cavitation [9]. However, there are no previous reports evaluating macula, GCC thickness, and PROS length.
This study may be important in demonstrating the relationship between the use of tamoxifen and reduction in GCC thickness. Tamoxifen-associated optic neuropathy has been previously reported [6]. Retinal ganglion cell (RGC) layer thinning occurs earlier than retinal nerve fibre layer (RNFL) thinning after acute optic neuropathies, and quantification of the macular RGC layer thickness may provide more valuable information on neuronal damage than RNFL thickness does [29]. Toxic optic neuropathies are primarily due to injury of the RGC layer [30]. This suggests that ganglion cell toxicity plays an important role in the development of toxic optic neuropathy. We found that tamoxifen users had lower GCC thicknesses in all inner rings and the nasal outer ring of the ETDRS inlay. In

Funding
No financial support was received for this submission.

Availability of data and materials
The data that support the findings of this study are available from the corresponding author upon reasonable request.

Authors' contributions
SB conceived the idea. SB and OKG participated in the design of the study. OKG and GK collected the data. AC and BE performed the statistical analysis. SB, OKG, AC, BE, and GK drafted and revised the manuscript. All authors were involved in manuscript preparation and further revisions. All authors read and approved the final manuscript for publication and agreed to be personally accountable for their contributions.

Ethics approval and consent to participate
All procedures performed in studies involving human participants were in accordance with the ethical standards of the institutional and/or national research committee and with the 1964 Helsinki Declaration and its later amendments or comparable ethical standards.
Ethics committee approval for ophthalmological examination and to use the results of research and data analysis in breast cancer patients undergoing tamoxifen therapy was obtained from the Marmara University Medical Faculty Ethic Committee, Turkey (approval number 09.2018.427). The research adhered to the tenets of the Declaration of Helsinki.
The written informed consent to participate in the study and relevant section of the manuscript was obtained from all individuals enrolled in the study.

Consent for publication
Not applicable.