Currently, the field of ophthalmology is experiencing immense technological advancements. One of the most recent and exciting innovations is the three-dimensional (3-D) digitally assisted vitreoretinal surgery (DAVS) system for performing posterior-segment surgeries. This system allows the surgeon to operate in “heads-up” position, alleviating the physical stress sustained by his or her back and neck due to the use of binoculars of a microscope.1–7
The few prior studies done on the clinical application of DAVS have focused only on the superior ergonomics of these systems.1–7 Such systems are said to provide the surgeon a very high-definition image at high magnifications, enabling the surgeon to maintain a more physiological position throughout the surgery and allowing for easy training of the assistant(s). However, mass acceptance of any new system depends on its ability to improve the performance of surgeries. A direct comparison of surgical performance of posterior-segment surgeries using DAVS and an analog microscope (AM) has never been done.
The study was done to compare the surgical performance of internal limiting membrane (ILM) peeling during vitrectomy for idiopathic full-thickness macular hole (FTMH) using a DAVS with an AM.
Patients and Methods
This was a prospective trial conducted at Aravind Eye Hospital, Madurai, India, after receiving approval from the institutional review board. The study adheres to the tenets of the Declaration of Helsinki.
The DAVS used during the study was NGENUITY 3D Visualization System (Alcon, Forth Worth, TX) in collaboration with the TrueVision Visualization System (Santa Barbara, CA). The eye-piece of the microscope is replaced by a high-dynamic-range camera. The camera is wired to a central processing unit, which processes two overlapping stereo images from the live surgery and displays the image on a 55-inch LED monitor. The surgeon views the screen from a distance of about 5 feet with the help of a passive polarized 3-D glasses (Figure 1). The surgical microscope used was OPMI Lumera T (Carl Zeiss Meditec, Jena, Germany), and the wide-angle viewing system used was RESIGHT 500 (Zeiss, Oberkochen, Germany). The vitrectomy machine used was Constellation Vitreoretinal Surgical System (Alcon Laboratories, Fort Worth, TX). “Pinch-and-grab” technique was used to perform ILM peeling. The forceps used was a 25-gauge Grieshaber DSP ILM peeling forceps (Alcon Laboratories; Fort Worth, TX).
The surgeon (NB) wearing polarized three-dimensional glasses to operate on a patient with idiopathic full-thickness macular hole using the NGENUITY 3D Visualization System.
All the surgeries were performed by a single surgeon (NB), who has more than 15 years of post-fellowship experience of performing high-volume vitreoretinal surgeries. The surgeon used this system for about 6 months before the start of the study to overcome the learning curve. The surgeries done in this first 6-month period included FTMH, epiretinal membrane (ERM), rhegmatogenous retinal detachment (RRD), macular tractional retinal detachments, combined mechanism retinal detachment, vitreous hemorrhage, endophthalmitis, dropped nucleus, scleral-fixated intraocular lens, and silicon oil removal.
The study included patients undergoing 25-gauge pars plana vitrectomy and internal limiting membrane (ILM) peeling for idiopathic FTMH. All of the patients were pseudophakic at the time of vitrectomy. Patients with traumatic macular hole (MH); myopic MH; MH associated with ERM; history of prior vitreoretinal surgery; or presence of any media opacity like pterygium, corneal opacity, posterior capsular opacity, vitreous hemorrhage, or asteroid hyalosis were excluded from the study.
Forty consecutive patients were included in the study and randomized into two groups. Group A patients were operated on with the help of an AM, whereas Group B patients were operated on with the help of a DAVS. As per the usual protocol followed at our institution, two or more surgeons perform the various surgeries planned for that day, going between each system. Similarly, Dr. Babu also performed various different types of surgeries on either system, depending on its availability. In both the groups, the surgical steps were similar. First core vitrectomy was done, followed by posterior vitreous detachment (PVD) induction. Intravitreal triamcinolone (Aurocort; Aurolab, Madurai, Tamil Nadu, India) was used to stain the vitreous, so as to ensure its complete removal. ILM was then stained with 0.05% solution of heavy brilliant blue G dye (HBBG) (Ocublue Plus; Aurolab, Madurai, Tamil Nadu, India) mixed with 10% dextrose in 1:2 proportions. HBBG was injected slowly under saline and allowed to stay for about 2 minutes before washing it. The stained ILM was grasped and lifted with the help of a 25-gauge ILM-peeling forceps to create a break in the ILM, constructing a flap. The edge of the ILM flap was then grasped with the forceps and peeled circumferentially around the MH for 2 disc diameters (DDs) from the edges of the hole.
All surgical videos included in the study were collected and converted into a similar two-dimensional format (ie, mp4) by the audio-visual department of the hospital. All of the videos were independently observed by two fellows (PK, ORN). Neither of these two fellows were present in the operating room while the surgeries were performed. The videos that were not recorded completely or that had too poor a quality for grading were also excluded from the study.
The data analyzed included the total surgical time required for ILM peeling; number of attempts in creating the ILM flap; number of attempts to complete ILM peeling; and intraoperative complications, including peeling-related hemorrhage(s), retinal break(s), and retinal touch. Surgical time required for ILM peeling was defined as the time between the initial contact and the last contact of the ILM peeling forceps with ILM. Number of attempts in creating the ILM flap included the number of misses, final attempt to pinch the ILM, and the attempts to create the flap. Number of attempts to complete ILM peeling was defined as the number of attempts taken to repeatedly grasp the ILM flap and circumferentially extend ILM peeling to 2 DD around the edges of the hole. The small hemorrhages that appeared during ILM peeling stopped spontaneously and disappeared within 24 hours were taken to be “peeling-related hemorrhage(s).” These are related to the nature of ILM. Sometimes while trying to pinch the ILM, by mistake retina was pinched. This caused hemorrhages, which stopped only after raising the intraoperative intraocular pressure, left a permanent bruise on the retinal surface, and could be seen postoperatively on ocular coherence tomography (OCT) as a dimple in the inner retinal layers. This was termed as “retinal touch.” “Retinal break” was defined as a discontinuity of all the layers of the retina. This can also occur if by mistake retina is pinched.
Statistical analysis was performed with STATA statistical software, Version 14.0 (StataCorp, College Station, TX). Continuous variables were expressed as mean (± standard deviation). Normality of the data was verified using Histogram plot and Shapiro-Wilk test. Independent student's t-test, Mann-Whitney U test, or Wilcoxon sign-rank test was used to find out the difference between two continuous variables. P value less than .05 was considered statistically significant.
Videos of 20 cases were included in each group (Table 1). All the cases were surgically similar. Consecutive patients were included, and none of the patients were excluded.
Comparison of Surgical Performance of ILM Peeling Performed Under the Analog Microscope and the Digitally Assisted Vitreoretinal Surgery System
The average surgical time required to complete the ILM peeling was 123.05 seconds ± 42.23 seconds in the AM group and 142.35 seconds ± 31.49 seconds in the DAVS group (P = .109). The mean number of surgical attempts required to create the ILM flap were 1.05 ± 0.22 in the AM group and 1.70 ± 1.22 in the DAVS group (P = .008). The mean number of surgical attempts required to complete the ILM peeling were 22.85 ± 9.95 in the AM group and 27.20 ± 7.16 in the DAVS group (P = .121).
The mean number of peeling-related hemorrhages that occurred in the AM group and the DAVS group were 3.35 ± 3.75 and 2.20 ± 1.47, respectively (P = .794). Intraoperative retinal breaks were not created in any of the patients in either of the groups. Retinal touch was noted in one patient in the AM group and three patients in the DAVS group (P = .534).
All the MHs in both the groups closed after the surgery. The improvement of best-corrected visual acuity in both the groups was similar.
ILM peeling is one of the most delicate, technically challenging, and potentially dangerous maneuvers that a vitreoretinal surgeon needs to perform.8 The underlying retina can be damaged during either the creation of the initial incision in the ILM, during finding and grasping the edge of the initial incision, or during creation and extension of the flap. Most surgeons prefer to peel at least 2 DD of ILM around the edge of the FTMH. For adequate peeling, it is important to differentiate between the area where the ILM has been peeled and where it needs to be peeled.8 Hence, it is critical that the new visualization systems should provide a high-resolution image with excellent contrast and depth perception.
Eckardt et al. found that the resolution of the eyepiece of AM was about twice that obtained with DAVS, whereas the depth of field obtained at medium-to-high magnification was identical with both the systems.2 They found that due to the lower resolution obtained with DAVS, frequent focus adjustments had to be made, leading to an increase in the duration of surgery. However, how these theoretical variables affect the surgical performance is largely unknown.
Our study results show that the mean number of flap creation attempts were significantly higher in DAVS group compared to the AM group. This may be due to the lower resolution obtained with DAVS. As the surgeon used forceps to create the ILM flap, he preferred to remain superficial and miss the ILM, rather than going too deep and grabbing the retina and injuring it. However, there was no statistically significant difference in the mean number of flap extension attempts between the two groups. The creation of the ILM flap was difficult using DAVS compared to AM; however, the surgeon did not find this as a clinically significant disadvantage. Also, once the flap was made, completing the ILM peeling was similar with both the techniques. Similarly, the mean surgical time was also statistically similar in both the groups. This was because majority of time was spent in flap extension and not flap creation. We performed ILM peeling with the help of the “pinch-and-grab” technique, and do not have much experience of using a diamond-dusted membrane scrapper or a loop. But we feel that the creation of ILM flap with these techniques would also be more challenging while using DAVS compared to AM.
Intraoperative retinal breaks were not created in any of the patients in our study, the rate being similar to our usual experience. The mean peeling-related hemorrhages were also statistically similar in both the groups. Although not statistically significant, the incidence of retinal touch during surgery was three times higher with DAVS than AM. These touches occurred mainly during the creation of the ILM flap, which was more difficult to achieve with DAVS. Although experimental studies have proved that the depth of field with DAVS is similar to that provided by AM, the results of our study showed that DAVS can limit the surgical performance of some very fine maneuvers. However, the final clinical outcome (ie, closure of the holes) was similar with both the imaging modalities.
Few previous studies have compared the difference in surgical time between DAVS and AM. Eckardt et al. compared the performance of 20 volunteers in performing meticulous tasks like placing black and transparent sequins onto needles located on a biconvex polystyrene disk and stacking nails in a Lincoln Log fashion.2 They found that the mean duration to perform these tasks was higher when performed with DAVS than AM; however, this difference did not reach statistical significance. They suggested that it took longer to complete the tasks as frequent focusing was required to improve the resolution of the 3-D image. In the study done by Romano et al., a single experienced surgeon performed a number of surgeries like vitrectomy for RRD, TRD, ERM peeling, and FTMH. It took longer to perform the surgeries using DAVS than AM; however, this difference was also not statistically significant.3 They attributed this difference to the learning curve of the surgeon. Coppola et al. found that the RRD surgery took less time with DAVS than AM; however, this difference was also not statistically significant.5 They argued that this improvement in speed with DAVS was because certain steps like the use of triamcinolone for vitreous staining could be avoided due to the magnified high-resolution image that enabled the surgeon to visualize the vitreous remnants optimally. However, the main limitation of their study was the inhomogeneous nature of the sample in both the groups.
In our experience, the learning curve with this system is not very long. Dr. Babu became proficient with the machine in fewer than 10 surgeries. At our center, the less-experienced surgeons and even the fellows operate using the DAVS. All are comfortable with the system, and none of them find it difficult in adjusting to the 3-D viewing system, even if they have to perform surgery alternatively on either system. As reported in literature, we also found DAVS to have better ergonomics than AM. While operating, the surgeons at our center have experienced much less stress on their neck and back. Performing surgeries was also easier with DAVS due to the higher magnification provided by the 3-D system compared to AM. Concerns have been raised about surgeons experiencing “3-D asthenopia” while operating.9 However, none of the surgeons at our center have experienced these symptoms.
To the best of our knowledge, there have been no previous studies comparing surgical performance of any surgical maneuver performed using the two imaging modalities. The strengths of our study were the use of a metric system for the evaluation of surgical performance of various surgical maneuvers, homogeneity of the sample in both the groups, careful exclusion of the surgeon's learning curve, and performance of all the surgeries by a single experienced surgeon who followed the same protocol during all the surgeries. The limitations of our study include a small size, evaluation of experience of only a single surgeon and only a single surgery.
This is the first prospective report that directly compares the use of AM and DAVS in terms of surgical performance of a posterior-segment surgery. In our small initial series, we found that DAVS provides almost identical surgical performance as AM, however it can limit the surgical performance of some extremely fine maneuvers like creation of the ILM flap.
- Bhadri PR, Rowley AP, Khurana RN, et al. Evaluation of a stereoscopic camera-based three-dimensional viewing workstation for ophthalmic surgery. Am J Ophthalmol. 2007;143(5):891–892. doi:10.1016/j.ajo.2006.12.032 [CrossRef]
- Eckardt C, Paulo EB. Heads-up surgery for vitreoretinal procedures: An experimental and clinical study. Retina. 2016;36(1):137–147. doi:10.1097/IAE.0000000000000689 [CrossRef]
- Romano MR, Cennamo G, Comune C, et al. Evaluation of 3D heads-up vitrectomy: Outcomes of psychometric skills testing and surgeon satisfaction. Eye (Lond). 2018;32(6):1093–1098. doi:10.1038/s41433-018-0027-1 [CrossRef]
- Kunikata H, Abe T, Nakazawa T. Heads-up macular surgery with a 27-gauge microincision vitrectomy system and minimal illumination. Case Rep Ophthalmol. 2016;7(3):265–269. doi:10.1159/000452993 [CrossRef]
- Coppola M, La Spina C, Rabiolo A, et al. Heads-up 3D vision system for retinal detachment surgery. Int J Retina Vitreous. 2017;3:46. doi:10.1186/s40942-017-0099-2 [CrossRef]
- Coppola M, La Spina C, Querques G, Bandello F. Correspondence. Retina. 2017;37(5):e57. doi:10.1097/IAE.0000000000001606 [CrossRef]
- Figueroa MS. 3D vitrectomy. Is it really useful?Arch Soc Esp Oftalmol. 2017;92(6):249–250. doi:10.1016/j.oftal.2017.02.001 [CrossRef]
- Li K, Wong D, Hiscott P, Stanga P, Groenewald C, McGalliard J. Trypan blue staining of internal limiting membrane and epiretinal membrane during vitrectomy: Visual results and histopathological findings. Br J Ophthalmol. 2003;87(2):216–219. doi:10.1136/bjo.87.2.216 [CrossRef]
- Kim SH, Suh YW, Song JS, et al. Clinical research on the ophthalmic factors affecting 3D asthenopia. J Pediatr Ophthalmol Strabismus. 2012;49(4):248–253. doi:10.3928/01913913-20120207-03 [CrossRef]
Comparison of Surgical Performance of ILM Peeling Performed Under the Analog Microscope and the Digitally Assisted Vitreoretinal Surgery System
|Criterion||Group A (AM)||Group B (DAVS)||P Value|
|Mean surgical time for completing ILM peeling||123.05 ± 42.23 seconds||142.35 ± 31.49 seconds||.109|
|Mean number of surgical attempts required to initiate ILM flap||1.05 ± 0.22||1.70 ± 1.22||.008|
|Mean number of surgical attempts required to complete ILM peeling||22.85 ± 9.95||27.20±7.16||.121|
|Mean number of peeling-related hemorrhages||3.35 ± 3.75||2.20 ± 1.47||.794|
|Patients with intraoperative retinal breaks||0||0|
|Patients with intraoperative retinal touch||1||3||.534|