Accuracy of biomicroscopy, ultrasonography and spectral-domain OCT in detection of complete posterior vitreous detachment

This was a prospective, consecutive, interventional case series based at the tertiary care Department of Ophthalmology, University Clinical Center Tuzla, Bosnia and Herzegovina. The study included all patients admitted for surgical treatment of epiretinal membrane (ERM) and macular hole (MH) in the period from January 1, 2015, to December 31, 2019. Exclusion criteria were age younger than 18, dense cataract or opaque ocular media that precluded fundus visualization, vitreous haze or bleeding, history of ocular trauma, and previous vitrectomy. The current study was approved by the University Clinical Center Tuzla Ethics Committee. Written informed consent was obtained from all patients after receiving an explanation of the investigative nature and intent of the study and tenets of the Helsinki Declaration were followed. As a part of preoperative preparation, all patients underwent a complete systemic evaluation which included general laboratory findings and a complete systemic examination to identify systemic diseases and drug use. In order to minimize the risk of spontaneous PVD all eligible patients underwent a complete preoperative ophthalmological assessment (BM, US, and SD-OCT) one day before the surgery. The first independent investigator performed slit-lamp BM with fully dilated pupils using a 78D lens (Volk Optical Inc., Enterprise Drive Mentor, OH, USA). A PVD was identified by the presence of a Weiss ring and or the definitively detached visible posterior hyaloid membrane. It was classified as presence or absence of PVD on BM. The second independent investigator performed ocular US using UD-800 (Tomey, Nagoya, Japan). Both vertical and horizontal views were used and the mobility of the posterior vitreous was examined during saccadic eye movements with a high gain (90 dB), real-time, through-the-lid contact technique. A PVD status was considered when posterior vitreous cortex was well defined and completely separated from the retina situated posterior to the equator and at the optic nerve head. It was classified as presence or absence of PVD in US. Spectral-domain OCT images were obtained with fully dilated pupils using Cirrus HD-OCT (Carl Zeiss Meditec, Inc., Dublin, CA) with the Macular Cube Scans (200x200) accompanying automatic centering at the fovea. An acceptable image quality was 8 and above, and adequate scan position was if the top of the scan was at least two “nasal retina thicknesses” above the retinal pigment epithelium at the foveal center.10 Both investigators interpreted OCT scans independently in a masked fashion. Posterior vitreous detachment was confirmed on OCT when a hyperreflective linear signal was clearly separated from the neuroretina. Disagreement regarding PVD status was resolved by joint review of the macular scans. It was classified as presence or absence of PVD in OCT. All results of preoperative examinations were compared with the findings of triamcinolone-assisted 23 gauge 3-port pars plana vitrectomy (PPV) performed on the following day. All surgeries were performed using Constellation (Alcon Laboratories, TX, USA) and recorded with a recording camera. After initial core vitrectomy triamcinolone was injected and posterior vitreous cortex was evaluated. Attached vitreous was considered when firm vitreous attachment that had to be removed by vitreous cutter suction was noted during the surgery. After complete vitrectomy all patients underwent internal limiting membrane (ILM) peeling and the gas tamponade based on surgical indication. The evaluation of PVD status was performed during the surgery and on the surgical video by both examiners. It was classified as presence or absence of PVD during PPV. Binary and categorical variables are reported as absolute numbers and percentages, and were tested for differences between groups defined by the PPV-confirmed PVD status using the chi-squared test. Sensitivity, specificity, positive predictive value, and negative predictive value were calculated for the BM, US and SD-OCT findings, using a PPV as a gold standard. Univariable and multivariable receiver operating characteristics (ROC) to obtain the values of the area under the curve (AUC) and corresponding 95% confidence intervals (95% CI) were based on logistic regression models predicting the PPV-confirmed PPD result. Multivariable models were adjusted for sex and age. AUCs were compared using deLong test. All the analyses were performed using Stata version 15.1 (StataCorp, College Station, Texas, USA). The significance level for hypothesis testing was set to 0.05.