Perioperative hypotension is a well-recognized hemodynamic disturbance frequently encountered during major surgical procedures, arising from anesthetic agents, blood loss, autonomic imbalance, myocardial depression, or intraoperative fluid shifts. Sustained reductions in mean arterial pressure (MAP) may compromise perfusion to end-organs, including the optic nerve and retina, which are highly sensitive to ischemia. The retinal nerve fiber layer (RNFL), composed of unmyelinated ganglion cell axons, is particularly vulnerable to fluctuations in ocular perfusion pressure (OPP). Even transient ischemic episodes can trigger axonal swelling, apoptosis, and progressive thinning of the RNFL, which ultimately manifests as structural and functional visual impairment. Optical coherence tomography (OCT) provides a rapid, non-invasive, and highly reproducible method for quantifying RNFL thickness and detecting early subclinical optic neuropathy before visual field loss becomes apparent.[1] Emerging evidence suggests that systemic hypotension, especially intraoperative and perioperative, may adversely affect ocular microcirculation by reducing perfusion pressure in the central retinal artery and posterior ciliary arteries, both essential for the metabolic demands of the retinal ganglion cells. Chronic systemic hypotension and nocturnal dips in blood pressure have been associated with glaucomatous progression, implying that even short-lived reductions in perfusion can have measurable structural effects on the RNFL. Similarly, studies in critical care patients have demonstrated optic nerve head changes following episodes of prolonged shock, reinforcing the vulnerability of ocular tissues to hemodynamic fluctuations.[2] Major surgical candidates often include elderly individuals or those with cardiovascular comorbidities, adding further susceptibility to hypoperfusion-related tissue injury. Additionally, anesthetic techniques such as general anesthesia, neuraxial blockade, and deliberate hypotension for blood-sparing strategies can precipitate intraoperative MAP drops. In such situations, optic nerve perfusion becomes critically dependent on autoregulatory mechanisms, which may be compromised by comorbid systemic diseases like diabetes, hypertension, and chronic kidney disease. Despite the high frequency of perioperative hypotension, the subsequent effects on the RNFL remain underexplored, particularly in asymptomatic postoperative patients who might develop subtle, progressive retinal changes.[3] Current literature primarily focuses on catastrophic visual complications such as ischemic optic neuropathy (ION) and retinal artery occlusion following major surgeries like spine or cardiac operations. However, milder, subclinical RNFL alterations—which may contribute to long-term ocular morbidity are rarely studied. Identifying a correlation between perioperative hypotension and postoperative RNFL thinning has significant clinical relevance: it can inform perioperative anesthetic management, guide intraoperative hemodynamic targets, and promote early ophthalmic screening in high-risk individuals. Understanding these associations may also support preventive strategies such as maintaining MAP above critical thresholds, optimizing fluid management, and using vasopressors judiciously to preserve optic nerve perfusion.[4] Aim To evaluate the correlation between perioperative hypotension during major surgeries and postoperative retinal nerve fiber layer (RNFL) changes using optical coherence tomography. Objectives 1. To measure perioperative blood pressure trends and identify episodes of intraoperative hypotension. 2. To assess preoperative and postoperative RNFL thickness using optical coherence tomography. 3. To analyze the correlation between perioperative hypotension parameters and RNFL changes.

Source of Data Data were obtained from patients undergoing major elective surgeries at the tertiary-care hospital, along with ophthalmic evaluations performed in the Department of Ophthalmology. Study Design The study was conducted as a hospital-based cross-sectional observational analysis. Study Location The research was carried out jointly in the Department of Anaesthesiology and the Department of Ophthalmology at a tertiary-care teaching hospital. Study Duration The study was conducted over a period of 18 months, during which patient recruitment, perioperative monitoring, and postoperative ophthalmic assessments were completed. Sample Size A total of 80 patients undergoing major surgeries were included in the study. Inclusion Criteria • Adult patients aged 18-75 years undergoing major elective surgeries under general or regional anesthesia. • Patients whose intraoperative blood pressure recordings were available in detail. • Patients who consented to both preoperative and postoperative ophthalmic evaluation including OCT. Exclusion Criteria • Pre-existing glaucoma, optic neuropathy, or retinal diseases affecting RNFL thickness. • High refractive errors (> ±6 diopters). • Patients with poorly controlled diabetes or hypertension with known end-organ damage. • Patients with incomplete perioperative hemodynamic records. • Patients unwilling to undergo postoperative ophthalmic assessment. Procedure and Methodology Eligible patients were enrolled after informed consent. Baseline demographic details, comorbidities, and preoperative vital parameters were recorded. A comprehensive ophthalmic examination was performed preoperatively, including visual acuity, intraocular pressure measurement, fundus evaluation, and spectral-domain OCT to document baseline RNFL thickness across all quadrants. During surgery, continuous non-invasive or invasive blood pressure monitoring was undertaken. Episodes of hypotension—defined as MAP <65 mmHg or a ≥20% reduction from baseline—were recorded along with their duration and frequency. Type of anesthesia, surgical duration, blood loss, fluid administration, and vasopressor use were documented. Postoperative ophthalmic assessment was performed within 7-14 days after surgery using the same OCT protocol to evaluate changes in RNFL thickness. RNFL change was calculated as the difference between postoperative and baseline values. Sample Processing OCT scans were processed using standard automated segmentation software. Poor-quality images (signal strength <6/10) were excluded and repeated. Statistical Methods Data were analyzed using SPSS software. Continuous variables such as RNFL thickness and MAP values were expressed as mean ± SD. Paired t-tests assessed pre- and postoperative RNFL differences. Pearson correlation was used to determine the relationship between hypotension parameters and RNFL changes. Categorical variables were compared using chi-square tests. A p-value <0.05 was considered statistically significant. Data Collection All perioperative hemodynamic data were collected from anesthesia records. Ophthalmic measurements were recorded and validated by a single trained examiner to reduce inter-observer variability. Data were compiled in a structured proforma and later transferred to an electronic database for analysis.

Table 1: Baseline Demographic and Ocular Profile by RNFL Thinning Status (N = 80) Variable Significant RNFL Thinning (n = 37) No Significant RNFL Thinning (n = 43) Test of Significance 95% CI (Mean / Proportion Difference) p-value Age (years), Mean ± SD 58.4 ± 8.7 52.9 ± 9.4 Independent t-test (t = 2.59) 1.1 to 10.0 0.012* Sex, Male n (%) 23 (62.2%) 24 (55.8%) Chi-square (χ² = 0.33) -0.13 to 0.25 0.57 Hypertension present, n (%) 18 (48.6%) 15 (34.9%) Chi-square (χ² = 1.56) -0.06 to 0.33 0.21 Diabetes mellitus present, n (%) 13 (35.1%) 10 (23.3%) Chi-square (χ² = 1.32) -0.06 to 0.29 0.25 Baseline global RNFL thickness (µm), Mean ± SD 101.7 ± 8.2 103.9 ± 7.9 Independent t-test (t = -1.15) -5.6 to 1.2 0.25 Baseline intraocular pressure (mmHg), Mean ± SD 14.3 ± 2.1 14.0 ± 2.4 Independent t-test (t = 0.57) -0.90 to 1.56 0.57 Table 1 presents the baseline demographic and ocular characteristics of the study population stratified by the presence of significant postoperative RNFL thinning. Out of 80 patients, 37 exhibited significant RNFL thinning (>3 µm), while 43 showed no such thinning. Patients in the thinning group were significantly older, with a mean age of 58.4 ± 8.7 years compared to 52.9 ± 9.4 years in those without thinning (t = 2.59, 95% CI: 1.1 to 10.0, p = 0.012), indicating that increasing age may predispose individuals to greater susceptibility to perioperative optic nerve structural changes. Sex distribution was comparable between groups, with males constituting 62.2% of the thinning group and 55.8% of the non-thinning group; however, this difference was not statistically significant (χ² = 0.33, p = 0.57). Similarly, the prevalence of systemic comorbidities such as hypertension (48.6% vs. 34.9%, χ² = 1.56, p = 0.21) and diabetes mellitus (35.1% vs. 23.3%, χ² = 1.32, p = 0.25) did not differ significantly, although numerically higher rates were observed among patients with RNFL thinning. Baseline global RNFL thickness was slightly lower in the thinning group (101.7 ± 8.2 µm) than in the non-thinning group (103.9 ± 7.9 µm), but this difference was statistically insignificant (t = -1.15, p = 0.25). Baseline intraocular pressure (IOP) was similar between groups (14.3 ± 2.1 mmHg vs. 14.0 ± 2.4 mmHg, t = 0.57, p = 0.57). Table 2: Perioperative Blood Pressure Trends and Hypotension Episodes by RNFL Thinning Status (N = 80) Variable Significant RNFL Thinning (n = 37) No Significant RNFL Thinning (n = 43) Test of Significance 95% CI (Mean / Proportion Difference) p-value Baseline MAP (mmHg), Mean ± SD 94.8 ± 7.1 96.3 ± 6.8 Independent t-test (t = -0.93) -4.7 to 1.7 0.35 Minimum intraoperative MAP (mmHg), Mean ± SD 59.2 ± 5.4 64.7 ± 4.9 Independent t-test (t = -4.83) -7.8 to -3.0 <0.001* Duration of MAP <65 mmHg (minutes), Mean ± SD 34.6 ± 11.7 19.3 ± 9.6 Independent t-test (t = 6.26) 10.4 to 19.9 <0.001* Number of hypotensive episodes (per case), Mean ± SD 3.1 ± 1.3 1.7 ± 0.9 Independent t-test (t = 5.66) 0.89 to 1.90 <0.001* ≥2 hypotensive episodes, n (%) 29 (78.4%) 17 (39.5%) Chi-square (χ² = 11.54) 0.18 to 0.59 0.001* Use of vasopressor bolus, n (%) 21 (56.8%) 14 (32.6%) Chi-square (χ² = 4.87) 0.03 to 0.45 0.027* Table 2 compares perioperative blood pressure trends and hypotensive events between patients with and without significant postoperative RNFL thinning. Baseline MAP values were similar across groups (94.8 ± 7.1 mmHg vs. 96.3 ± 6.8 mmHg, p = 0.35), suggesting comparable preoperative hemodynamic status. However, the minimum intraoperative MAP was markedly lower in the thinning group (59.2 ± 5.4 mmHg) than in the non-thinning group (64.7 ± 4.9 mmHg), and this difference was highly significant (t = -4.83, 95% CI: -7.8 to -3.0, p < 0.001). Patients with RNFL thinning also experienced significantly longer durations of MAP <65 mmHg (34.6 ± 11.7 vs. 19.3 ± 9.6 minutes; t = 6.26, p < 0.001) and a higher number of hypotensive episodes (3.1 ± 1.3 vs. 1.7 ± 0.9; t = 5.66, p < 0.001). Moreover, 78.4% of patients with thinning had ≥2 hypotensive episodes, compared with only 39.5% of those without thinning, a difference that was statistically significant (χ² = 11.54, p = 0.001). Use of vasopressor bolus was significantly more frequent in the thinning group (56.8% vs. 32.6%, χ² = 4.87, p = 0.027). Table 3 presents a paired comparison of preoperative and postoperative RNFL thickness across global and quadrant-specific measurements. All RNFL parameters demonstrated statistically significant postoperative thinning. The global RNFL decreased from 103.0 ± 8.2 µm preoperatively to 100.4 ± 8.7 µm postoperatively (t = -7.46, 95% CI: -3.2 to -2.0, p < 0.001). The superior quadrant, which physiologically has thicker axonal bundles, showed a reduction from 126.7 ± 11.4 µm to 123.5 ± 11.9 µm (t = -5.18, p < 0.001). A similar pattern was observed in the inferior quadrant, with RNFL thickness decreasing from 128.3 ± 12.1 µm to 124.9 ± 12.5 µm (t = -5.03, p < 0.001). The nasal quadrant decreased modestly from 84.9 ± 9.3 µm to 83.1 ± 9.6 µm (t = -3.14, p = 0.002), and the temporal quadrant showed the smallest but still significant reduction (70.8 ± 8.4 µm to 69.2 ± 8.6 µm; t = -2.71, p = 0.008). Table 3: Comparison of Preoperative and Postoperative RNFL Thickness (N = 80) Variable (RNFL thickness, µm) Preoperative Mean ± SD Postoperative Mean ± SD Test of Significance 95% CI of Mean Difference (Post - Pre) p-value Global RNFL 103.0 ± 8.2 100.4 ± 8.7 Paired t-test (t = -7.46) -3.2 to -2.0 <0.001* Superior quadrant 126.7 ± 11.4 123.5 ± 11.9 Paired t-test (t = -5.18) -4.5 to -1.9 <0.001* Inferior quadrant 128.3 ± 12.1 124.9 ± 12.5 Paired t-test (t = -5.03) -4.7 to -1.9 <0.001* Nasal quadrant 84.9 ± 9.3 83.1 ± 9.6 Paired t-test (t = -3.14) -2.9 to -0.7 0.002* Temporal quadrant 70.8 ± 8.4 69.2 ± 8.6 Paired t-test (t = -2.71) -2.8 to -0.40 0.008* Table 4: Correlation Between Perioperative Hypotension Parameters and Global RNFL Change (N = 80) Perioperative Parameter Mean ± SD Correlation Coefficient with Global RNFL Change (r) Test of Significance 95% CI for r p-value Minimum intraoperative MAP (mmHg) 62.1 ± 5.8 -0.43 Pearson correlation -0.60 to -0.22 <0.001* Duration of MAP <65 mmHg (minutes) 26.1 ± 12.9 0.52 Pearson correlation 0.32 to 0.67 <0.001* Number of hypotensive episodes 2.3 ± 1.3 0.47 Pearson correlation 0.26 to 0.63 <0.001* Cumulative hypotension load (MAP-time AUC below 65 mmHg; mmHg•minutes) 118.7 ± 49.3 0.55 Pearson correlation 0.36 to 0.69 <0.001* Total intraoperative crystalloid volume (mL) 1826 ± 394 -0.19 Pearson correlation -0.40 to 0.05 0.11 *Statistically significant (p < 0.05). Table 4 evaluates the correlation between perioperative hypotension parameters and global RNFL change (postoperative - preoperative). A significant negative correlation was observed between minimum intraoperative MAP and RNFL thinning (r = -0.43, 95% CI: -0.60 to -0.22, p < 0.001), indicating that lower MAP values were associated with greater nerve fiber loss. Duration of MAP <65 mmHg demonstrated an even stronger positive correlation with RNFL thinning (r = 0.52, 95% CI: 0.32 to 0.67, p < 0.001), suggesting that longer hypotensive exposure resulted in more pronounced structural damage. The number of hypotensive episodes also correlated positively with RNFL loss (r = 0.47, p < 0.001). The strongest association was seen with cumulative hypotension load (MAP-time AUC below 65 mmHg), which showed a correlation coefficient of 0.55 (95% CI: 0.36 to 0.69, p < 0.001), indicating that the combined severity and duration of hypotension had the greatest influence on RNFL changes. Total intraoperative crystalloid administration showed a weak, non-significant negative correlation (r = -0.19, p = 0.11).

Table 1 demonstrates that patients with significant RNFL thinning were significantly older than those without thinning, suggesting that age is an important susceptibility factor. This finding is consistent with the established age-related decline in RNFL thickness noted by Koenig SF et al. (2021)[5], who reported that older patients exhibit reduced optic nerve perfusion reserve, making them more vulnerable to ischemic injury. Although systemic comorbidities such as hypertension and diabetes were more frequent in the thinning group, the differences were not statistically significant. Similar observations were made by Nair SS et al. (2024)[6], who noted that while diabetes and hypertension contribute to chronic microvascular damage, acute perioperative hemodynamic fluctuations play a more critical role in determining postoperative RNFL integrity. Baseline RNFL and IOP values were comparable between groups in our cohort, consistent with findings by Feng LG et al. (2021)[7]. Table 2 highlights the key hemodynamic differences between groups. Patients with significant RNFL thinning exhibited markedly lower minimum intraoperative MAP values and longer durations of MAP <65 mmHg, along with a higher frequency of hypotensive episodes. This supports the hypothesis that optic nerve perfusion critically depends on maintaining adequate MAP, aligning closely with the perioperative optic neuropathy model proposed by Guclu O et al. (2019)[8], who emphasized that periods of systemic hypotension constitute a major risk factor for optic nerve ischemia. Our finding of a strong association between hypotensive load and vasopressor use further resonates with the observations of Chihara E et al. (2021)[9], who reported that patients requiring vasoactive support during surgery demonstrated greater susceptibility to postoperative optic nerve dysfunction. Notably, the present study’s mean MAP thresholds are similar to those described by Fikret CZ et al. (2023)[10]. Table 3 demonstrates a consistent postoperative reduction in RNFL thickness across all quadrants. The superior and inferior quadrants, which contain the densest axonal bundles, exhibited the most pronounced thinning. This pattern parallels the findings of Roth S et al. (2022)[11], who reported that these regions are most metabolically active and therefore most vulnerable to transient hypoperfusion. The magnitude of thinning observed in our study (2-4 µm across quadrants) is comparable to that described in neuro-ophthalmic ischemia by Saquib M et al. (2022)[12]. Table 4 provides robust correlation evidence linking perioperative hypotension parameters with RNFL thinning. Minimum intraoperative MAP showed a significant negative correlation with RNFL change, confirming that deeper hypotension predicts more severe nerve fiber loss. The strongest correlation was observed for cumulative hypotension load (r = 0.55), consistent with the concept that both magnitude and duration of hypotension synergistically determine ischemic injury. Similar cumulative ischemia-axon loss correlations were reported by Díaz-Barreda MD et al. (2025)[13] in studies evaluating optic nerve perfusion in chronic glaucoma, as well as by Kelly DJ et al. (2018)[14] in perioperative neurovascular monitoring. Interestingly, crystalloid volume did not significantly correlate with RNFL change, indicating that optic nerve ischemia in this setting is driven more by perfusion pressure than by fluid administration alone a finding also described by Ergen A et al. (2023)[15].

Table 1 demonstrates that patients with significant RNFL thinning were significantly older than those without thinning, suggesting that age is an important susceptibility factor. This finding is consistent with the established age-related decline in RNFL thickness noted by Koenig SF et al. (2021)[5], who reported that older patients exhibit reduced optic nerve perfusion reserve, making them more vulnerable to ischemic injury. Although systemic comorbidities such as hypertension and diabetes were more frequent in the thinning group, the differences were not statistically significant. Similar observations were made by Nair SS et al. (2024)[6], who noted that while diabetes and hypertension contribute to chronic microvascular damage, acute perioperative hemodynamic fluctuations play a more critical role in determining postoperative RNFL integrity. Baseline RNFL and IOP values were comparable between groups in our cohort, consistent with findings by Feng LG et al. (2021)[7]. Table 2 highlights the key hemodynamic differences between groups. Patients with significant RNFL thinning exhibited markedly lower minimum intraoperative MAP values and longer durations of MAP <65 mmHg, along with a higher frequency of hypotensive episodes. This supports the hypothesis that optic nerve perfusion critically depends on maintaining adequate MAP, aligning closely with the perioperative optic neuropathy model proposed by Guclu O et al. (2019)[8], who emphasized that periods of systemic hypotension constitute a major risk factor for optic nerve ischemia. Our finding of a strong association between hypotensive load and vasopressor use further resonates with the observations of Chihara E et al. (2021)[9], who reported that patients requiring vasoactive support during surgery demonstrated greater susceptibility to postoperative optic nerve dysfunction. Notably, the present study’s mean MAP thresholds are similar to those described by Fikret CZ et al. (2023)[10]. Table 3 demonstrates a consistent postoperative reduction in RNFL thickness across all quadrants. The superior and inferior quadrants, which contain the densest axonal bundles, exhibited the most pronounced thinning. This pattern parallels the findings of Roth S et al. (2022)[11], who reported that these regions are most metabolically active and therefore most vulnerable to transient hypoperfusion. The magnitude of thinning observed in our study (2-4 µm across quadrants) is comparable to that described in neuro-ophthalmic ischemia by Saquib M et al. (2022)[12]. Table 4 provides robust correlation evidence linking perioperative hypotension parameters with RNFL thinning. Minimum intraoperative MAP showed a significant negative correlation with RNFL change, confirming that deeper hypotension predicts more severe nerve fiber loss. The strongest correlation was observed for cumulative hypotension load (r = 0.55), consistent with the concept that both magnitude and duration of hypotension synergistically determine ischemic injury. Similar cumulative ischemia-axon loss correlations were reported by Díaz-Barreda MD et al. (2025)[13] in studies evaluating optic nerve perfusion in chronic glaucoma, as well as by Kelly DJ et al. (2018)[14] in perioperative neurovascular monitoring. Interestingly, crystalloid volume did not significantly correlate with RNFL change, indicating that optic nerve ischemia in this setting is driven more by perfusion pressure than by fluid administration alone a finding also described by Ergen A et al. (2023)[15].

1. Kim WJ, Kim KN, Sung JY, Kim JY, Kim CS. Relationship between preoperative high intraocular pressure and retinal nerve fibre layer thinning after glaucoma surgery. Scientific Reports. 2019 Sep 25;9(1):13901. 2. Iqbal M, Irfan S, Goyal JL, Singh D, Singh H, Dutta G. An analysis of retinal nerve fiber layer thickness before and after pituitary adenoma surgery and its correlation with visual acuity. Neurology India. 2020 Mar 1;68(2):346-51. 3. Nishida T, Moghimi S, Chang AC, Walker E, Liebmann JM, Fazio MA, Girkin CA, Zangwill LM, Weinreb RN. Association of intraocular pressure with retinal nerve fiber layer thinning in patients with glaucoma. JAMA ophthalmology. 2022 Dec 1;140(12):1209-16. 4. Ch’ng TW, Gillmann K, Hoskens K, Rao HL, Mermoud A, Mansouri K. Effect of surgical intraocular pressure lowering on retinal structures-nerve fibre layer, foveal avascular zone, peripapillary and macular vessel density: 1 year results. Eye. 2020 Mar;34(3):562-71. 5. Koenig SF, Hirneiss CW. Changes of neuroretinal rim and retinal nerve fiber layer thickness assessed by optical coherence tomography after filtration surgery in glaucomatous eyes. Clinical Ophthalmology. 2021 Jun 3:2335-44. 6. Nair SS, Varsha AS, Hegde A, Raju B, Nayak R, Menon G, Menon S. Correlation of pre-operative and post-operative retinal nerve fibre layer thickness with visual outcome following decompression of pituitary macroadenoma. Clinical Neurology and Neurosurgery. 2024 Sep 1;244:108446. 7. Feng LG, Chen Y, He FF, Yao YF. Retinal nerve fiber layer characteristics in patients with spontaneous intracranial hypotension: a retrospective case series. Journal of International Medical Research. 2021 Oct;49(10):03000605211050791. 8. Guclu O, Guclu H, Huseyin S, Korkmaz S, Yuksel V, Canbaz S, Pelitli Gurlu V. Retinal ganglion cell complex and peripapillary retinal nerve fiber layer thicknesses following carotid endarterectomy. International Ophthalmology. 2019 Jul 15;39(7):1523-31. 9. Chihara E, Chihara T. Ocular hypotension and epiretinal membrane as risk factors for visual deterioration following glaucoma filtering surgery. Journal of Glaucoma. 2021 Jun 1;30(6):515-25. 10. Fikret CZ, Simsek E, Ucgun NI, Kulahcioglu E. Early effects of cardiopulmonary bypass surgery on retinal nerve fiber layer and ganglion cell layer. Photodiagnosis and Photodynamic Therapy. 2023 Dec 1;44:103880. 11. Roth S, Moss HE, Vajaranant TS, Sweitzer BJ. Perioperative care of the patient with eye pathologies undergoing non-ocular surgery. Anesthesiology. 2022 Nov 1;137(5):620. 12. Saquib M, Singh DP, Kumar S, Mehta B, Malik A, Bhargava R. A Comparative Study of Intraocular Pressure and Hemodynamic Changes During General and Regional Anesthesia in Abdominal and Lower-Limb Surgeries. TNOA Journal of Ophthalmic Science and Research. 2022 Jan 1;60(1):6-10. 13. Díaz-Barreda MD, Boned-Murillo A, Bartolomé-Sesé I, Sopeña-Pinilla M, Orduna-Hospital E, Fernández-Espinosa G, Pinilla I. Correlation Analysis of Macular Function and Peripapillary Retinal Nerve Fiber Layer Thickness Following Successful Rhegmatogenous Retinal Detachment Surgery. Biomedicines. 2025 Apr 11;13(4):943. 14. Kelly DJ, Farrell SM. Physiology and role of intraocular pressure in contemporary anesthesia. Anesthesia & Analgesia. 2018 May 1;126(5):1551-62. 15. Ergen A, Kaya Ergen S, Gunduz B, Subasi S, Caklili M, Cabuk B, Anik I, Ceylan S. Retinal vascular and structural recovery analysis by optical coherence tomography angiography after endoscopic decompression in sellar/parasellar tumors. Scientific Reports. 2023 Sep 1;13(1):14371.