Endoscopic sinus surgery (ESS) transformed the field of otolaryngology. Although the endoscope enhanced the surgical management of sinus disease, it offered a two-dimensional (2D) view of the surgical field compared to direct vision. Endoscopes did not provide ‘natural’ depth perception, a factor widely regarded as critical for efficient and accurate surgical movements. Instead, otolaryngologists relied on monocular cues, anatomic knowledge, and haptic feedback to perceive spatial impression. Despite these compensatory mechanisms, 2D visualization did not match stereoscopic cues.
Three-dimensional (3D) visualization has been pursued since the introduction of the endoscope. Multiple technologies to achieve stereopsis have been utilized, which include dual-camera video, dual chip-on-the-tip, shutter mechanism, polarization, and stereoscopic displays. Although surgeons have universally remarked on the improved depth perception with 3D endoscopy, conflicting results with regard to speed, error rate, and training have been reported when compared to conventional endoscopy. Resolution degradation, bulky technology, increased cost, and user side-effects are some of the reasons why stereoscopic technology has not been widely adopted, particularly for ESS.
Recent advances in stereoscopic technology, including robotics as well as introduction of novel stereoscopic endoscopes and displays, have reinvigorated the field of 3D minimally invasive surgery. Evolving stereoscopic technology promises to revolutionize endoscopic surgery in many surgical disciplines including endoscopic sinus and skull-base surgery.
ENDOSCOPIC SINUS SURGERY
Introduced in the 1970s by Wigand and Draf, ESS revolutionized the surgical management of sinus disease. At the center of the transformation was the endoscope, a tool that allowed superior illumination and panoramic views of the paranasal sinuses. Although nasal endoscopy had been introduced in the early 20th century by Hirschmann using a modified cystoscope, it was not until the 1960s when advances in optics led to the development of the modern rigid endoscope . The endoscope was helpful not only for the diagnosis of sinonasal pathology but also for the development of minimally invasive surgical techniques. Prior to ESS, the majority of sinonasal surgery was performed with the headlight using trans-facial, coronal, or gingivobuccal incisions. Over the past few decades, ESS has expanded to treat sinonasal tumors, orbital pathology, and anterior skull-base tumors. Angled endoscopes, improved instrumentation, and advances in imaging have allowed surgical management of a wide range of pathology with subsequent improvement in outcomes.
LIMITATIONS OF ENDOSCOPY
Although the endoscope provided many advantages, it lacked the ability to provide binocular vision. Monocular endoscopes and displays created a 2D image of the operative field, thereby impairing depth perception, hand-eye coordination, and size estimation [2,3]. Human kinematic studies have demonstrated that monocular cues require longer movement times and a tendency to underestimate distance between surgical landmarks . Depth perception is widely regarded as a critical parameter for efficient movements during surgery and reducing surgical errors. The disadvantages of 2D visualization were illustrated in a landmark study in 2003. The authors conducted a performance analysis of 252 bile duct injuries, where they reported that visual perceptual illusions were the primary cause of errors in 97% of the cases. They demonstrated that visual misperceptions rather than ‘errors of skill, knowledge or judgment resulted in the vast majority of laparoscopic bile duct injuries’ .
Despite these limitations, endoscopic surgeons learned to create 3D constructs from monocular cues. These cues included relative structure/size, color/texture gradients, and motion parallax. Surgeons also relied on haptic cues and surgical movements to infer spatial relationships . These visual and tactile cues enabled experienced surgeons to perform increasing complex maneuvers and tasks during ESS. Nonetheless, 2D visualization has limited the growth of some fields such as neuroendoscopy.
Although technical improvements have enhanced monocular cues in endoscopy, 2D visualization is inherently unable to match the depth perception gained from binocular cues such as convergence, stereopsis, and vertical disparities. The human visual cortex superimposes two subtly different retinal images obtained from different viewing angles to achieve stereopsis. Similarly, most stereoscopic systems produce two slightly different images, which are then displayed separately to each eye, thereby rendering a 3D representation of a surgical field [2,6].
Until recently, two major image generation strategies have been employed for stereopsis, namely the dual-channel and shutter mechanism technologies. Dual-channel technology utilizes two distinct images generated from either a camera or video chip to render a 3D image. Differences in brightness, color, and sharpness can result in user side-effects such as headaches, nausea, or ocular fatigue. In addition, the small ‘inter-pupillary’ distance between the endoscopes/chips creates a weak 3D effect. Shutter mechanism technology relies on temporal shifts of a single camera to generate slightly different, rapidly alternating images to enable stereopsis [6,7]. Unpredictable movements of the camera may result in images with low disparity and weak stereopsis in addition to result in user side-effects.
EXPERIMENTAL STUDIES OF STEREOSCOPIC SYSTEMS
Over the past two decades, the literature has examined the role of stereo-endoscopy in many surgical disciplines (Table 1[3,8–13]). However, none of the early literature explored the role of 3D technology in ESS due to technical limitations. These stereoscopic technologies included first and second-generation dual channel, and shutter mechanism systems, time-parallel and time-multiplexed systems, as well as various 3D displays. These technologies were compared to available 2D systems in clinical and laboratory settings. Outcomes included subjective (depth perception, efficiency, confidence, dexterity, and user side-effects) and objective (operative time, error rate) measures. The role of stereo-endoscopy in surgical education was a central theme in the literature. Varying results have been noted in the literature due to dissimilarities in technology, inter-patient variability, and lack of standardized measures.
First-generation laparoscopic systems
A majority of surgeons in early studies noted improved depth perception with first-generation 3D endoscopic systems. Using shutter glasses to perform laparoscopic surgery, Becker et al. reported that surgeons were ‘enthusiastic about spatial impression’ generated by the technology. In a 1995 study, using a dual-camera endoscope, authors reported that 20/32 laparoscopic surgeons with varying experience reported superior depth perception, but 15/32 also noted no improvement in surgical maneuvers. Forty percent of surgeons also reported worse illumination and image resolution compared to 2D . Other authors in clinical and nonclinical settings noted similar results [9,12,14,15]. Furthermore, a small number of surgeons in multiple studies noted user side-effects such as dizziness, eyestrain, fatigue, and headaches [8,12,14–16].
Conflicting results regarding objective measures such as task performance, operative time, and error rate were reported in the literature. Early studies using first-generation dual-camera and shutter mechanism technologies did not report any significant improvement in clinical or nonclinical task performance (dexterity, positioning, accuracy), time, or error rate. Nonclinical tasks included placing beads on a suture , bean placement into dishes , knot-tying , grasping and displacing objects , or determining relative position of sticks . Clinical tasks included laparoscopic surgery such as cholecystectomies . Laparoscopic dissection (suturing and knot-tying) by experienced surgeons in animals did not reveal any difference in completion time or performance using shutter glasses compared to 2D visualization . Even when faster performance was noted on certain tasks with stereovision, the significance was eliminated after repetition. This was illustrated in a nonclinical study where suturing and knot-tying were performed 12% (P = 0.06) faster in 3D by novice and experienced surgeons; however, after two repetitions this was no longer statistically significant .
Later studies using evolving first-generation stereoscopic technology for laparoscopy, however, noted subtle improvements in objective measures as well as decreased user side-effects. Nonclinical studies using shutter glasses noted improvement in task performance and time [19–21]. For example, van Bergen et al. noted a statistically significant faster performance (26%) and decrease in error rate (43%) with 60 students performing five standardized tasks. Other clinical studies also noted improvement in operative time and performance [19,21,22].
Second-generation laparoscopic systems
A landmark study comparing direct vision, a second-generation 3D endoscope, and 2D endoscope found impaired task performance by 35–100% with 2D when compared to direct vision. Stereoscopic visualization reduced this handicap by 41–53% (P < 0.03). Novices using 2D endoscopy had highest levels of impairment (more correctional movements, longer task completion), followed by surgeons using 2D, novices using 3D, and experienced surgeons using 3D endoscopy . Since then, multiple studies using standardized tasks in nonclinical settings have compared second-generation 3D systems with 2D systems (standard and high definition). Many have found an improvement in performance times and error rates. One such study found no difference in time, but decreased error rate in novice surgeons using a novel 3D camera system when compared to a standard 2D system . Statistically significant improvements in time and reduction in error rates were noted with novel stereoscopic endoscope , and passive polarizing stereoscopic display  when compared to a 2D system. In addition, decreased  or absent  side-effects were noted with second-generation stereoscopic systems.
Robotic stereoscopic systems
Recently introduced robotic systems provided an alternative to traditional stereoscopic surgery. Advantages included endo-wristed surgery, and high-definition binocular vision resulting in improved depth perception and surgical efficiency. However, significant cost and size requirements, as well as restrictions for multiquadrant surgery counted among its disadvantages [25–28]. Randomized blinded studies examining robotic systems in clinical and nonclinical settings concluded that 3D visualization improved task performance, decreased operative times, and reduced surgical errors when compared to the 2D robotic mode [25,26]. In a nonclinical study, complex tasks (knot-tying, needle threading) were performed better (even by experienced surgeons) with stereoscopic vision, demonstrating an average decrease in error rate (P = 0.004) and operative time (P < 0.05) . Finally, two studies comparing stereoscopic laparoscopic and robotic systems found that task performance was better with 3D visualization regardless of surgical tool/method used [27,29].
LIMITED CLINICAL UTILIZATION OF STEREOSCOPIC TECHNOLOGY
Although surgeons noted improved spatial perception with stereoscopic technology, its widespread clinical use in all surgical disciplines, including ESS, has been limited for many reasons. These included poor image resolution of early 3D endoscopes, which often used 1-chip-camera technology . In addition, bulky endoscopes, eyewear, and head-mounted displays added to surgeons’ reluctance to adopt this technology. Subtle differences in image quality resulted in user side-effects including headaches, nausea, and ocular fatigue. Although subsequent 3D endoscopes were far better at image generation, limitations of scaling and viewing angles limited their use among many endoscopic surgeons. In addition, 30% of individuals, despite passing conventional stereovision testing, are unable to appreciate adequate depth perception with most 3D systems secondary to a variety of ophthalmologic conditions [2,7]. The cost of these systems also prevented surgeons and hospitals from acquiring this technology.
Endoscopic sinus surgeons were particularly hesitant about adopting stereoscopic endoscopy. Technical limitations of endoscopes made them impracticable for sinus surgery. The narrow, dark nasal passages required endoscopes with small diameters, which could provide superior illumination and resolution. In addition, the lack of angled endoscopes made current stereo-endoscopes difficult to use for frontal recess surgery. One of the few studies in ESS using 3D technology was published by White et al. in 1997. The authors looked at performance outcomes in novices using direct, 2D, and 3D displays in endoscopic sinus surgery. They found that endoscopic performance tasks were performed better in a sheep cadaveric model using a 3D display than a 2D display (P < 0.05). Furthermore, they noted that diagnostic and therapeutic performance scores (execution time divided by tasks accomplished) were statistically better (P < 0.05) using 2D displays (mean scores 56.8 and 41.1) over direct endoscopy (mean scores 94.1 and 74.1) .
NOVEL THREE-DIMENSIONAL ENDOSCOPE
Although many recent developments have occurred in 3D technology, the most significant for ESS has been the introduction of a novel 3D stereo-endoscope (Fig. 1). Unlike previous 3D endoscopes, this endoscope mimics the compound eye of arthropods. Using a microscopic array of lenses positioned over a single video chip, this endoscope generates multiple images that are processed to reconstruct a 3D image. The variance in the image formed by the lens array allows stereopsis. Surgeons view the image on a stereoscopic monitor with lightweight polarized glasses (Fig. 2) [4,31,32▪,33–35].
The first clinical study was conducted in laparoscopic gynecological surgery where novice and experienced surgeons were randomly assigned to a 2D or 3D visualization group. In a multivariate analysis, the authors found the duration of surgery to be shorter in the 3D group (P = 0.04). In addition, longer procedures (>60 min) of similar complexity, regardless of surgical expertise, were higher in the 2D group (36 vs. 20%). All surgeons reported superior depth perception, anatomic understanding, and surgical efficiency with the endoscope. The authors believed that this endoscope improved the learning curve for the novices and helped experienced surgeons perform complex tasks with ease. In addition, no surgeon experienced any side-effects which have tempered enthusiasm for 3D endoscopy in the past .
Additional studies, largely published in endoscopic sinus and skull-base surgery, have noted promising results for 3D endoscopy (Table 2[4,31,32▪,33–38]). In a clinical study of 12 patients undergoing endoscopic approaches to the skull-base, the authors noted an improvement in anatomical recognition and depth perception, particularly near the carotid artery and optic nerve in sphenoid sinus. Larger clinical studies have noted favorable subjective (e.g. depth perception) and objective results with this technology. In 2009, Tabaee et al. published a retrospective series of 13 patients undergoing 3D endoscopic pituitary surgery compared to a matched group treated with a 2D endoscope. The authors noted no difference in operative time, hospital length of stay, extent of resection, or complication rate . A subsequent larger series of 90 patients (58 in 3D group, 32 in 2D group) concurred with the findings and in addition found no difference in estimated blood loss or readmission rate [32▪].
Other authors reported more complete ESS as a result of stereoscopy. In one study, additional ethmoid partitions were removed in two patients to achieve an ethmoidectomy, and residual tumor was identified and resected in one patient. Some limitations, namely the loss of orientation due to overmagnification and inability to visualize a frontal cell, were, however, noted . Similarly, other authors have noted decreased image clarity and resolution with this stereo-endoscope compared to high-definition 2D endoscopes. Surgeons have also noted a smaller field of view, central darkness, inadequate light in narrow sinonasal clefts, and poor visualization with minimal debris on the scope [34,35]. None of the surgeons in all the studies experienced any side-effects as noted with dual-camera and shutter stereoscopy. However, a few surgeons did note that a short acclimation to the system was necessary [4,33,35].
Cadaveric dissections performed of the sinuses, orbital apex, and midline skull-base targets have revealed superior spatial orientation with this novel 3D endoscope when compared to the 2D endoscopy [36,38]. In addition, surgeons performing standardized endoscopic tasks in a training model with 2D and 3D endoscopes noted improvement in performance with 3D endoscopy. A decrease in time (37 to 33 s; P = 0.56) and error rate (2.27 to 1.33; P = 0.26) was noted with simpler tasks. A decreased error rate was noted for a more complex task (8/15 surgeons failed with 2D, 3/15 failed with 3D) . In another simulator experiment, 33 neurosurgeons and otolaryngologists were randomized to complete two runs of an endoscopic task in either 2D first/3D second, 3D first/2D second, or 3D first/3D second. Surgeons using 3D in the second run showed a significantly higher efficiency (P = 0.04) in both the 2D/3D or 3D/3D scenarios .
LEARNING AND EDUCATION
Multiple studies have noted the significant impact that stereoscopic technology plays in the education of novice users. In 1999, Taffinder et al. noted that task impairment was greatest in novice users using a 2D endoscopy and least in experienced users using 3D technology. This finding fueled subsequent studies, which found that stereoscopy improved speed and/or decreased the error rate in novice users [23,24,27,35]. Many studies noted similar performance improvement in experienced users [3,13,25,26,37]. One study even suggested that experienced surgeons benefited even more than novices when using 3D visualization .
Other authors, however, reported that anatomical education and experience played a greater role in depth perception rather than the method of visualization [9,39]. This was illustrated in study where 3D depth maps were constructed based on a viewer's interpretation of 2D image. The authors found that novices randomized to receive a short anatomical lesson created depth maps similar to those of experienced surgeons. Another study noted that inexperienced laparoscopic surgeons who were initially trained in 3D endoscopy performed the same tasks with greater speed and decreased error rate in 2D (P < 0.05) even 3 months later .
Three-dimensional endoscopic sinus surgery is in its infancy. Although stereo-endoscopy has been available for decades, technical limitations have obviated its use in sinus surgery. Recent advances have produced a stereo-endoscope largely used for endonasal approaches to the skull-base. Continued developments promise to deliver a 3D endoscope that will provide high-definition, 3D visualization for endoscopic sinus surgery and beyond (Singh A, personal communication).
Conflicts of interest
There are no conflicts of interest.
REFERENCES AND RECOMMENDED READING
Papers of particular interest, published within the annual period of review, have been highlighted as:
▪ of special interest
▪▪ of outstanding interest
Additional references related to this topic can also be found in the Current World Literature section in this issue (p. 87).
1. Govindaraj S, Adappa ND, Kennedy DW. Endoscopic sinus surgery: evolution and technical innovations. J Laryngol Otol 2010; 124:242–250.
2. Hofmeister J, Frank T, Cuschieri A, Wade N. Perceptual aspects of two-dimensional and stereoscopic display techniques in endoscopic surgery: review and current problems. Semin Laprosc Surg 2001; 8:12–24.
3. Taffinder N, Smith SGT, Huber J, et al. The effect of second generation 3D endoscope on laparoscopic precision of novices and experienced surgeons. Surg Endosc 1999; 13:1087–1092.
4. Tabaee A, Anand VK, Fraser JF, et al. Three dimensional endoscopic pituitary surgery neurosurgery. Operative Neurosurg 2009; 64:288–295.
5. Way LW, Stewart L, Gantert W, et al. Causes and prevention of laparoscopic bile duct injuries: analysis of 252 cases from a human factors and cognitive psychology perspective. Ann Surg 2003; 237:460–469.
6. Szold A. Seeing is believing: visualization system in endoscopic surgery (video, HDTV, stereoscopy, and beyond). Surgical Endosc 2005; 19:730–733.
7. Radermacher K, Pichler CV, Fisher S, Rau G. 3-D. Visualisation in surgery. Helmholtz-Institute Aachen; 1998.
8. Chan AC, Chung SC, Yim AP, et al. Comparison of two-dimensional vs. three dimensional camera systems in laparoscopic surgery. Surg Endosc 1997; 11:438–440.
9. Jones DB, Brewer JD, Soper, NJ. The influence of three-dimensional video systems on laparoscopic task performance. Surg Laparosc Endosc 1996; 6:191–197.
10. Votanopoulos K, Brunicardi FC, Thornby J, Bellows CF. Impact of three-dimensional vision in laparoscopic training. World J Surg 2008; 32:110–118.
11. Becker H, Melzer A, Schurr MO, Buess G. 3-D video techniques in endoscopic surgery. 1993; 1:40–46.
12. Hanna GB, Shimi SM, Cushieri A. Randomized study of influence of two-dimensional versus three-dimensional imagining on performance of laparoscopic cholecystectomy. Lancet 1998; 351:348–351.
13. Storz P, Buess G, Wolfgang K, Kirshinak A. 3D HD versus 2D HD: surgical task efficiency in standardized phantom tasks. Surg Endosc 2012; 26:1454–1460.
14. Dion YM, Gaillard F. Visual integration of data and basic motor skills under laparoscopy. Influence of 2-D and 3-D vide-camera systems. Surg Endosc 1997; 11:995–1000.
15. McDougall EM, Soble JJ, Wolf JS, et al. Comparison of three-dimensional and two dimensional laparoscopic video system. Endoscopy 1996; 10:371–374.
16. Birkett DH, Josephs LG, Este-McDonlad JSO. A new 3-D laparoscope in gastrointestinal surgery. Surg Endosc 1994; 8:1448–1451.
17. Crosthwiate G, Chung T, Dunkley P, et al. Comparison of direct vision and electronic two- and three-dimensional display systems on surgical task efficacy in endoscopic surgery. Br J Surg 1995; 82:849–851.
18. Muller MD, Camartin C, Dreher E, et al. Three-dimensional laparoscopy. Gadget or Process? A randomized trail on the efficacy of three-dimensional laparoscopy. Surg Endosc 1999; 13:469–472.
19. Van Bergen P, Kunert W, Bessell J, et al. Comparative study of two-D and three-D vision systems for minimally invasive surgery. Surg Endosc 1998; 12:948–954.
20. Pietrabissa A, Sarcella E, Carobbi A, Mosca F. Three-dimensional versus two-dimensional video system for the trained endoscopic surgeons and the beginner. Endosc Surg 1994; 2:315–317.
21. Von Pichler C, Rademacher K, Rau G. The state of 3-D technology and evaluation. Minim Invasive Ther Allied Technol 1996; 5:419–426.
22. Wenzl R, Lehner R, Vry U, et al. Three-dimensional video endoscopy: clinical use and gynaecological laparoscopy. Lancet 1994; 344:1621–1622.
23. Kong SH, Oh BM, Yoon H, et al. Comparison of two- and three-dimensional camera systems in laparoscopic performance: a novel 3D system with one camera. Surg Endosc 2010; 24:1132–1143.
24. Smith R, Day A, Rockall T, et al. Advanced stereoscopic projection technology significantly improves novice performance of minimally invasive surgical skills. Surg Endosc 2012; 26:1522–1527.
25. Bandani KK, Bhandari A, Tewari A, Menon M. Comparison of two dimensional and three- dimensional suturing: is there a difference in a robotic surgery setting? J Endourol 2005; 19:1212–1215.
26. Byrn JC, Schluender S, Divino CM, et al. Three-dimensional imaging improves surgical performance for both novice and experienced operators using the da Vinci Robot System. Am J Surg 2007; 193:519–522.
27. Blavier A, Gaudissart Q, Cadiere G, Nyssen AS. Comparison of learning curves and skill transfer between classical robotic laparoscopy according to the viewing conditions: implications of training. Am J Surg 2007; 194:115–121.
28. Jourdan IC, Dutson E, Garcia T, et al. Stereoscopic vision provides a significant advantage for precision robotic laparoscopy. Br J Surg 2004; 91:879–885.
29. Wagner OJ, Hagen M, Kuramann A, et al. Three-dimensional vision enhances task performance independently of surgical method. Surg Endosc 2012; 26:2961–2968.
30. White PS, Frizelle FA, Hanna GB, et al. Comparison of direct monocular endoscopic, two- and three-dimensional display systems on surgical task performance in functional endoscopic sinus surgery. Clin Otolaryngol 1997; 22:65–67.
31. Kaufman Y, Sharon A, Klein O, et al. The three-dimensional ‘insect eye’ laparoscopic imagining system- a prospective randomized study 2007; 4:31–34.
32▪. Kari E, Oyeaiku N, Dadashev V, Wise S. Comparison of traditional 2-dimensional endoscopic pituitary surgery with new 3-dimensional endoscopic technology: intraoperative and early postoperative factors. Int Forum Allergy Rhinol 2012; 2:2–8.
This is a large series comparing two-dimensional to three-dimensional endoscopy for endoscopic approaches to the sella and pituitary surgery.
33. Manes RP, Barnett S, Batra PS. Utility of novel 3-dimensional stereoscopic vision system for endoscopic sinonasal skull base surgery. Int Forum Allergy Rhinol 2011; 1:191–197.
34. Brown SM, Tabaee A, Singh A, et al. Three-dimensional endoscopic sinus surgery: feasibility and technical aspects. Otolaryngol Head Neck Surg 2008; 138:400–402.
35. Shah RN, Leight D, Patel MRA, et al. A controlled laboratory and clinical evaluation of three-dimensional endoscope for endonasal sinus and skull base surgery. Am J Rhinol Allergy 2011; 25:141–144.
36. Roth J, Fraser J, Singh A, et al. Surgical approaches to orbital apex: comparison of endoscopic, endonasal and transcranial approaches using novel 3D endoscope. Orbit 2011; 30:43–48.
37. Fraser JF, Allen B, Anand VK, Schwartz TH. Three-dimensional neurosteroendoscopy: subjective and objective comparison to 2D. Minim Invas Neurosurg 2009; 52:25–31.
38. Roth J, Singh A, Gurston N, et al. Three-dimensional and 2-dimensional endoscopic exposure of midline cranial base targets using expanded endonasal and transcranial approaches. Neurosurgery 2009; 116–130.
39. Sindu RS, Tompa D, Jang R, et al. Interpretation of three-dimensional structure from two-dimensional endovascular images: implications for educators in vascular surgery. J Vasc Surg 2004; 39:1305–1311.
© 2013 Lippincott Williams & Wilkins, Inc.