Purpose
To assess the additive value of foveal swept-source optical coherence tomography (OCT)–based biometry to the preoperative fundus examinations for diagnosing macular abnormalities in patients scheduled for cataract surgery.
Design
Diagnostic testing evaluation.
Methods
Consecutive patients 50 years of age and older planned for cataract surgery from one institution were retrospectively enrolled. All patients underwent foveal swept-source OCT, and macular spectral domain (SD) OCT scans before pupil dilation as well as dilated fundus biomicroscopy examination. The effectiveness of fundus biomicroscopy examinations, foveal swept-source OCT scans, and the combination of both in identifying macular diseases was analyzed with macular spectral-domain OCT scans as reference.
Results
Seventy-eight of the eligible 442 eyes (442 patients) were excluded because of noninterpretable macular spectral-domain OCT OCT scans or foveal swept-source OCT scans. The remaining 364 eyes of 364 patients (mean age 73.59±9.26 years [range 49-96], 172 males) formed the study group. Fundus biomicroscopy alone vs fundus biomicroscopy with the addition of foveal swept-source OCT yielded 36% vs 63% sensitivity, 94% vs 72% specificity, 79% vs 58% positive predictive value (PPV), and 71% vs 76% negative predictive value (NPV), respectively. This diagnostic improvement was significant compared with fundus biomicroscopy alone ( P = 2.98 –8 ).
Conclusion
Combined fundus biomicroscopy and foveal swept-source OCT scans improved the detection of macular abnormalities prior to cataract surgery but it was inferior to macular spectral-domain OCT scans. Additional studies to assess the cost-effectiveness of adding foveal swept-source OCT scan in comparison to macular spectral-domain OCT scan to the preoperative cataract evaluation are required.
C ataract surgery is currently the most commonly performed surgical procedure worldwide. With the rapid evolution and improvement of the technique, the expectations for a superb uncorrected visual acuity outcome are increasing. Among other factors, the visual outcome following surgery depends on the state of ocular health. A thorough preoperative evaluation is critical to facilitate proper patient management and to enable the surgeon to meet patient expectations. Missed macular pathologies might lead to unexpectedly low postoperative best corrected visual acuity and occasionally to worsening of existing macular pathology. However, preoperative identification of macular pathology might aid in adjustment of the treatment plan. Specifically, a combined surgery can be offered, premium intraocular lenses (IOLs) may be waived, retina specialist consultation may be considered, and patient expectations can be adjusted. , The current standard of care for evaluating the macula prior to surgery in most institutions is a dilated fundus biomicroscopy examination. However, factors such as poor pupil dilation, photophobia, advanced cataract, and other opacities of the ocular media can obscure macular pathology. The rates of macular pathologies existing before cataract surgery that are missed by fundus biomicroscopy examination range from 4.9% to 25.0%. , , Recent studies have suggested adding a macular spectral-domain optical coherence tomography (OCT) examination to the preoperative routine to detect macular pathologies prior to cataract surgery. , However, the cost-effectiveness of a routine preoperative macular spectral-domain OCT examination before standard cataract extraction has not yet been established.
The IOLMaster-700 biometry device uses foveal swept-source OCT as part of fixation verification during biometry measures. The foveal fixation is based on a lower-quality imaging than that of a dedicated macular OCT. Such foveal imaging, which is integrated into the standard biometry examination, has the potential to improve the detection of macular pathologies in patients before cataract surgery and may also reduce the need for a dedicated macular OCT examination. , To our knowledge, only 2 studies have evaluated the effectiveness of foveal swept-source OCT in macular pathologies detection. , However, neither of them assessed the additive value of routine foveal swept-source OCT scans to the standard dilated fundus biomicroscopy examination in the detection of macular pathologies.
The aim of this study was to evaluate the additive diagnostic value of foveal swept-source OCT scans with the use of macular spectral-domain OCT as reference to detect macular pathologies as part of the preoperative evaluation process for cataract surgery.
Methods
Participants and Ethics
This retrospective study included consecutive patients aged >50 years who were referred for a standard cataract extraction due to senile cataract between November 2017 to January 2018. All patients underwent an swept-source OCT biometry examination and macular spectral-domain OCT imaging for their routine preoperative evaluation process. Patients with either noninterpretable macular spectral-domain OCT or foveal swept-source OCT scans due to low quality were excluded. The study was approved by the institutional review board (IRB) of Shaare Zedek Medical Center, Jerusalem, Israel, and adhered to the tenets of the Declaration of Helsinki.
Study Design
Patient data were retrieved retrospectively from their medical records. All preoperative evaluation examinations were performed on the same day and included visual acuity, intraocular pressure measurement (Goldmann tonometer), ocular biometry (IOLMaster-700, version 1.70; Carl Zeiss Meditec AG, Jena, Germany), a complete ophthalmic biomicroscopic evaluation by a consultant ophthalmologist (including a slit-lamp fundus biomicroscopy examination using 78- and 90-diopter lenses after pharmacological pupil dilation), and a macular spectral-domain OCT (Heidelberg Spectralis, software version 6.9.4.0, Heidelberg Engineering, Germany). Both macular spectral-domain OCT and foveal swept-source OCT scans were performed before pharmacologic pupil dilation.
The foveal swept-source OCT scans of the IOLMaster-700 biometer are based on detection of the foveal pit in 1 × 1-mm central macular imaging (44-mm scan depth with 22-µm resolution in tissue). The foveal imaging is one single horizontal scan that is constructed from an average of 4 B-scans. Those scans were reviewed by a single consultant ophthalmologist (Y.W.) who was blinded to the patient’s detailed fundus biomicroscopy examination, macular spectral-domain OCT scans, and medical records. The foveal swept-source OCT scans were categorized as either normal or pathological, without any attempt to diagnose a specific pathology.
Macular spectral-domain OCT scans were performed by means of the Heidelberg Spectralis device (software version 6.9.4.0; Heidelberg Engineering, Germany) following the same protocol consisting of a dense macular scan, high-speed mode, >15 images averaged, scan angle 20 degrees, 5.9 × 5.9 mm, X scaling 11.47 µm/pixel, and Z scaling 3.87µm/pixel. All macular spectral-domain OCT scans were analyzed independently by a retina specialist (J.H.), who was blinded to the patient’s detailed fundus biomicroscopy examination, foveal swept-source OCT scans, and medical records. The presence of macular pathology and the type of pathology were documented. A macular spectral-domain OCT examination was considered pathological if it was expected to be clinically significant. This was based on the clinical judgment of the retina specialist who conducted the examination. The findings on macular spectral-domain OCT scans were used as reference to assess the false positive rate (FPR), false negative rate (FNR), sensitivity, specificity, accuracy, positive predictive value (PPV), and negative predictive value (NPV) of the fundus biomicroscopy examination, the foveal swept-source OCT scans, and the combination of both methods. To test for an additive diagnostic value of the foveal swept-source OCT scans to the fundus biomicroscopy examinations, the detection ability of the fundus biomicroscopy alone was compared to the combined results of the fundus biomicroscopy examination with those of the foveal swept-source OCT. For the combined fundus biomicroscopy and foveal swept-source OCT results, a macula was defined as “pathological” if one of the examinations diagnosed a macular pathology. Alternately, a macula was considered “normal” if both examinations indicated that the macula was within normal limits.
Statistical Analysis
The statistical analysis of the fundus biomicroscopy examination alone and the combination of the patient’s fundus biomicroscopy examination and foveal swept-source OCT scans findings was based on a confusion matrix that was built according to the description in the study design section. McNemar test was performed with continuity correction to determine whether there was a significant difference of macular pathology detection ability between the fundus examination alone and the combination of the fundus examination and the foveal swept-source OCT findings. A P value of 5% or less was considered statistically significant. Also calculated were multiple classification metrics (FPR, FNR, sensitivity, specificity, accuracy, PPV, and NPV) for each method. The analyses were performed by means of Python 3 with packages scikit-learn (0.22.1), stats models (0.11.1), and pandas (1.0.0).
Results
Four hundred forty-two eyes of 442 consecutive patients were eligible for study entry. Forty (9.05%) patients were excluded because of noninterpretable macular spectral-domain OCT scans, and 38 of the remaining 402 patients (8.60%) were excluded because of noninterpretable foveal swept-source OCT scans. The final study group was composed of 364 patients. The flowchart of the study participants and the diagnostic results of the fundus biomicroscopy examination alone and in combination with the foveal swept-source OCT are presented in Figure 1 . The study patients’ demographics are summarized in Table 1 .
| Parameter | Total |
|---|---|
| Number of eyes (number of patients, right/left eye) | 364 (364, 181/183) |
| Mean age (range) | 73.59 ± 9.26 (49-96) |
| Sex, n (%) | |
| Male | 172 (47.25) |
| Female | 192 (52.75) |
One hundred thirty-eight eyes (37.91%) demonstrated macular pathologies on macular spectral-domain OCT, and 12 eyes had more than 1 pathology (yielding a total of 153 macular pathologies). Figure 2 lists the types of macular pathologies detected by the 3 modalities (spectral-domain OCT, fundus biomicroscopy, foveal swept-source OCT, and fundus biomicroscopy + foveal swept-source OCT).
