Design and Performance Testing of a Novel In Vivo Laparoscope Lens Cleaning Device

A common tool for diagnosis and treatment of gastrointestinal, gynecologic, and other anatomical pathologies is a form of minimally invasive surgery (MIS) known as laparoscopy. There are roughly 4 × 10⁶ laparoscopic surgeries performed in the U.S. every year, with an estimated 15 × 10⁶ performed across the world [1–3]. The popularity of laparoscopy is rapidly rising globally due more favorable quality of life outcomes when compared to open surgery [1,4–15]. Additionally, this trend toward laparoscopy has been shown to demonstrate benefits of decreased length of inpatient hospital stays as well as decreased operative patient morbidity [16].

During a laparoscopic surgery, a surgeon makes one or more small incisions in the body, and places a port, known as a trocar, in each incision. The trocar acts as a tunnel by which the surgeon can access the inside of the body (Fig. 1) [17]. Typically, inert CO₂ gas travels through a tube that attaches to a port on a trocar. The CO₂ insufflates the abdominal cavity generating a more spacious surgical environment, allowing the surgeon to safely visualize and operate. Small instruments, including laparoscopes, fit through these trocars and allow the surgeon to operate on the patient without having to perform more invasive open-surgery methods. The laparoscope is a long slender tool that allows the surgeon to see inside the patient's body by illuminating the interior space and collecting optical images/video through a lens at the distal tip. Laparoscopes are generally categorized as straight or angled. Straight laparoscopes provide visualization directly in front of the main axis with a lens angle of 0 deg, while angled laparoscopes provide visualization off-axis often at a 30 deg or 45 deg angle.

Often during laparoscopic surgeries, lens clarity may become impaired via (1) condensation or (2) smearing of bodily fluids and tissues (i.e., blood and adipose) (Fig. 2) [18–23]. Such an event may be referred to as obstructed lens visualization (OLV). The current gold standard (GS) to clean the laparoscope is to physically remove it from the body, wipe the lens with a fluid and cloth combination, and then reinsert it into the body. This typically takes anywhere from ∼15 to 60 s. Laparoscopes typically require cleaning up to 30 times during a surgery depending on surgical complexity. This can lead to increased time of surgery and anesthesia, which correlates to higher incidence rates for surgical complications and postsurgical infections, both of which can result in increased hospital re-admissions and general operating room (OR) inefficiencies [24–28]. In summary, OLV causes medical and monetary pains for patients, hospitals, and third-party payers, as well as an unwanted distraction to the surgeon that poses safety concerns.

When laparoscope visualization is improved, surgeons are able to perform more accurate and safe surgery, ultimately overcoming medical and monetary issues associated with longer surgeries such as complication rates, operational delays, and increased financial costs [19]. The fiscal impact for implementing a low-cost device relative to time saved in surgery could effectively decrease economic burdens on hospitals, improving overall value of care on a per surgery basis [29]. For example, a study shows that a commercially available device known as the FloShield saved an average of $620 per gastric bypass procedure (∼10 min/surgery) by addressing lens fogging events in vivo [30,31]. Although there is another version of the device, “FloShield Plus” that offers additional features for addressing blood and fat tissue obstructions on the lens.

The proposed novel in vivo laparoscope lens cleaner device provides a simple and effective method, with clearing times comparable to that of the FloShield product. However, as the device is capable of cleaning all major types of lens contaminants in vivo, it stands to potentially achieve even greater cost savings. It may additionally help to optimize OR workflow by increasing OR availability with the accumulated time saved per surgery [31], if a single OR room saves enough time in one day. Furthermore, such time savings might actually increase throughput regarding the number of surgeries performed, ultimately increasing hospital profit by completing more surgeries in less time than previous methods produced. In terms of medical outcomes, the time saved, and decreased time spent with obstructed lens visual, has the potential to decrease risk to the patient and improve surgeon-focused outcomes as well. As previously suggested, decreased time under anesthesia stands to improve patient outcomes regarding postsurgical site infection rates and surgical complication rates and possibly reducing the associated readmission rates and accompanying penalties, benefiting multiple stakeholders in this healthcare ecosystem [32–35]. Furthermore, decreasing intraoperative interruptions stands to decrease fatigue and stress placed for the surgical team while simultaneously improving surgical team efficiency and patient outcomes [24–29,36,37]. More so, the time saved may also decrease monetary pains placed on hospitals, as OR time costs ∼$60 per minute [29,38]. In summary, our novel in vivo laparoscope lens cleaner stands to improve the value of care by increasing quality and improving safety, while lessening medical and financial stresses on patients, hospitals, and third party payers.

Our technological innovation addresses the major healthcare ecosystem needs while overcoming some of the shortcomings of the current products available. Through background research and the NSF Innovation Corps customer discovery process (117 stakeholder interviews), engineering specifications were developed from customer needs. The most important specifications were that the device would: (1) not require the laparoscope to be removed from the body, (2) operate effectively, easily, and quickly, to remove condensation, blood, and fat from the lens surface and (3) be compatible with available laparoscopes and trocars. Figure 3 displays expected device performance relative to a number of established competitors [39–44]. The proposed design stands to address major shortcomings from current technologies. Compared to ex-vivo cleaners, the proposed technology works in vivo and saves time since the need to remove the scope from the body is mitigated. Compared to other in vivo lens cleaning technologies, it clears all debris types effectively while maintaining a simple interface that is easy for a user to manage during a laparoscopic procedure. While the proposed design is compatible with current OR processes and equipment, the first prototype may not maintain absolute compatibility across the vast spectrum of trocar ports on the market.

After multiple iterations of conceptualizing, prototyping, and testing, a final design was determined to meet the aforementioned customer needs [45]. The design consists of a long sheath that surrounds the laparoscope from base to tip (Fig. 4(a)) that includes a handle with an actuation knob and scope alignment features (Fig. 4(b)), in addition to a mechanical wiper in the tip region (Fig. 4(c)) controlled by the actuation knob. Once the scope is placed inside the device (Fig. 4(d)), the mechanical wiper can be oriented in line with the laparoscope to allow for insertion through a standard trocar. Once inside the body cavity, the wiper allows for laparoscope lens cleaning when OLV events occur, in essence analogous to a windshield wiper on a car. The sheath is intended to stay stationary relative to the laparoscope during use. This is especially beneficial for orientation and usability with angled lens laparoscopes. The wiper is connected to a long shaft and actuation knob allowing for user-actuation during OLV events. The base end of the sheath includes a user-friendly interface allowing for implementation of either a rotation or a push-pull mechanism for effective actuation and cleaning of the laparoscope by rotating or pushing and pulling the wiper component. Rotating the wiper is the primary cleaning motion for a 0 deg laparoscope. It is necessary to push-pull the wiper into contact for cleaning and out of view postcleaning. Conversely, rotation is important for aligning the wiper on an angled laparoscope, although the primary cleaning motion is a push-pull mechanism. The base also prevents rotation of the device around the laparoscope, and consequentially the wiper component, during use. This is important, as in order to clean the lens surface on a 30 deg laparoscope, the wiper component must be oriented correctly. The current design maintains viability and cleaning performance despite the fact that different methods of actuation remain dependent on the angle of the laparoscope lens, suggesting reasonable versatility for the current device design.

Two experimental procedures were designed and executed. The first experiment focused on quantifying lens cleaning efficacy, utilizing a custom-fabricated “warm-body” environment to simulate the various lens debris encountered during real surgery. The second experiment focused on potential mechanical durability (i.e., safety) of the proposed device design.

A warm body simulator (WBS) chamber was needed to emulate conditions encountered in a human body during a laparoscopic procedure to test the lens cleaning device. We designed and built a WBS to introduce thermal- and humidity-related characteristics while aiding in the simulated generation of the three major types of OLV debris that may be encountered inside the body during surgery: condensation, blood, and adipose.

A commercially available foam chamber was selected to simulate a real surgical cavity due to insulating thermal properties, ease of manufacturability, and low cost. Thermal regulation was accomplished using a circulating water bath (Thermo Scientific NESLAB RTE-7 Circulating Water Bath; Waltham, MA), producing a reliable and repeatable temperature level of 98.8±1.7 °F over ∼400 min of operation, after taking ∼100 min to heat from room temperature.

The WBS (Fig. 5) included mobile debris depositors and two imaging stations. The debris depositors were long-handled cups that held simulated blood (VATA Anatomical Healthcare Models: Simulated Blood, Canby, OR) with comparable viscosity to real blood and fresh pork fat from a local supermarket to simulate human fat tissue. Debris deposition on the laparoscope lens for blood and fat tissue residue was accomplished with simple contact between the laparoscope tip and the type of debris, as the debris depositor was moved toward the lens surface from its original position. In order to generate fogging/condensation events, water was added to the bottom of the WBS before chamber warming commenced. Once the WBS temperature was stable, the scope was removed from the chamber and chilled in a cooled container, and then reinserted into the WBS—the relative humidity and temperature differences generating a reliable and consistent fogging event. Regardless of debris type, an OLV event was deemed acceptable if more than 75% of the image appeared distorted, as this threshold (albeit subjective in nature) was notably more extreme than those referenced in literature studying suboptimal surgical view in the OR. Such a threshold generated a conservative debris threshold, and as such suggest that strong lens cleaning performance would ideally translate to more common OLVs cleaning scenarios [22].

Although surgical end-users would be the ultimate judge of a device in the OR regarding efficacy, this benchtop experiments sought objective quantification of lens cleaning efficacy. After much research and deliberation, a custom “Ronchi Ruling grid” imaging target was selected to aid in the process, as it would have a set resolution with consistently defined spacing between vertical and horizontal lines (henceforth referred to as “the grid”), seen on the far right of Fig. 5.

Additionally, the lens cleaning device docked in fixtures for 0 deg and 30 deg laparoscopes, which we referred to as a collective imaging station (IS). The IS was designed and implemented into the WBS in an attempt to generate repeatable and reliable image outputs (Fig. 6). This IS helped improve image reliability by (1) holding parallel orientation between a laparoscope lens surface and the corresponding grid surface to reduce image distortion from 0 deg and 30 deg laparoscopes and their associated targets and (2) physically constraining the camera, laparoscope, and imaging target, as not all of the computational methods explored for these experiments fully compensated for small shifts/variations in the physical real-world structures that are amplified in digital image pixels. Alongside these IS controls, other controlled parameters include the laparoscope light intensity maintained at 30% of its maximum brightness (Stryker Endoscopy 988 3 CHIP, Stryker X8000 Light Source, Stryker SDC Ultra, Kalamazoo, MI), and ambient lighting controlled by the closed environment of the WBS. This version of this imaging base accommodates the functional prototype, coupled with 5 mm Stryker laparoscopes (502-555-030 30 deg laparoscope, 502-55-010 0 deg laparoscope, 988-210-122 camera; Kalamazoo, MI).

An experimental protocol was established with the intent to allow for full reproducibility of the experimental results. Five test methods (Fig. 7) were conducted using a single device prototype for lens cleaning on 0 deg and 30 deg laparoscopes. The five types of tests were (1) null, (2) gold-standard (GS), (3) fog, (4) simulated blood (comparable viscosity to real blood), and (5) simulated fat tissue (fresh pork fat). Each test method consisted of 20 trials, for a total of 100 lens cleaning trials per laparoscope (200 total trials). The lens cleaning device was placed in the IS base fixture and not removed until all testing for a single laparoscope type was completed.

The Null test was intended to validate the reliability and repeatability of the IS-WBS system. Therefore, no debris deposition or cleaning process was implemented in the null test. The laparoscope was placed in the IS and an image was acquired. It was then completely removed from the device and reinserted for another image acquisition event.

The gold-standard (GS) test was similar to the null approach, except implementation of the current GS cleaning process was added. When the laparoscope was removed from the device, it was wiped on a sponge-FRED combination (current lens cleaning GS) and reinserted for acquisition of the second image (Medtronic, Dublin, Ireland) [41]. No debris deposition occurred, generating a “best case scenario” for the GS.

The fog test was accomplished in a similar manner, with the addition of a few steps. An original clean image (OI) was acquired. The laparoscope was then removed from the IS. The laparoscope shaft was inserted into a plastic tube-channel, held within an ice bucket, for 10 s of cooling. Upon reinsertion into the IS, a fog event occurred and a postdebris image (PDI) was taken if it met criteria for an adequate OLV (<75% distortion). The device was then actuated for lens cleaning and a postcleaning image (PCI) was taken. Simulated blood and fat tests were accomplished in an almost identical manner: OI to PDI to PCI, except without laparoscope removal from the WBS. This was accomplished by using the long-handled cups within the WBS to deposit debris for the PDI image without removing the scope from the IS. Example images of fog, simulated blood, and simulated fat acquired images can be seen in Fig. 8. Additional data gathered outside of image acquisition included the number of actuations required to result in a visually satisfying image, apparently free of debris.

Extensive research has been performed in the realm of image processing, and in particular, numerous methods have been explored at quantifying image clarity and acuity [46–56]. Some techniques used for image comparison are “full-reference” techniques, where a reference image is known. Others are “no-reference” techniques, where the reference image is not available, and the quality assessment is blind to the ideal image [47]. These experiments employ a combination of no-reference and full-reference techniques. The reference image(s) used in these experiments were often subjected to less controllable noise variables such as glare, small shifts or rotations of the camera or camera focus, etc. Such noise variables cause slight shifts in pixel value, location, or intensity in the image array. Attempts to mitigate these issues were employed in the imaging hardware, protocol, and software analysis as previously described.

The analyses were ultimately used to determine the difference between two sets of two images: (1) OI versus PDI and (2) OI versus PCI. All methods were explored and analyzed using a custom matlab script (MathWorks, Natick, MA). All metrics were reported on a 0:1 scale, where 0 represents low image agreement and 1 represents ideal image agreement (i.e., poor cleaning efficacy and strong cleaning efficacy, respectively). Metrics used to quantify image clarity between the paired image combinations included each image's histogram correlation coefficient (HCC), and ratios from each paired image's Euler number and image quality measurement (IQM). The HCC observed relationship between binned grayscale image pixel values. A nonblurred image would hold more bin values representing black and white colors while a smudged/blurred image would result in more gray-colored pixels due to blurring of the black and white grid. For Euler numbers, after converting pixels from grayscale values to binary 0 or 1 values, the bweulermatlab function separated the image into (1) “objects” (connected white pixels) and (2) “holes” (black pixels inside of white pixel “objects”). The function then computes the total number of objects minus holes and provides the output Euler number. This number can be positive or negative, depending on the image. The black and white grid target conveniently generates a distinguishable objects and holes due to the nature of its design. The IQM employed an image sharpness measure utilizing a frequency domain approach and finds that a lower IQM score correlating to higher blur levels in an image. For additional details into this method, please consult the aforementioned literature reference for specifics, presented by the original authors [55].

As some of these scores ranged widely on the 0:1 scale (namely, the HCC), an overall “effective cleaning score” (an average of the HCC, Euler ratio, and IQM ratio), was also calculated to obtain a metric by which a general comparison could be obtained.

After testing for cleaning efficacy of the initial functional prototype design, a refined overmolded wiper design was generated for testing the mechanical durability of a wiper configuration anticipated to be more robust, reliable, and manufacturable (Fig. 9). Exploratory testing was performed to determine potential safety levels relative to wiper adhesion resiliency. Each specimen was manufactured by overmolding silicone rubber over a stainless steel wire using a custom wiper overmold. The wire design combined with wiper material and aligned manufacturing process is intended to form a reliable and repeatable bonded wiper to the wire, allowing the bottom wiper edge to maintain a strong bond while still allowing for effectively lens cleaning efficacy. Wiper specimens (N = 33) were manufactured for early stage testing and underwent mechanical tensile testing until adhesion failure was reached.

Mechanical testing was accomplished combining test specimen with custom-fabricated fixtures that would isolate the desired tensile forces, primarily by compensating for and/or limiting unwanted deformation of the silicone rubber material while still subjecting the wire-wiper to shearing forces that may otherwise try to dislodge it during surgical use. A “base” fixture was designed to seat and constrain the wire inside of a hole that used a set screw to lock the wire in place, while a “claw” fixture was designed to seat and constrain the wiper in a notch that allowed for applicable force isolation for the desired failure testing (Fig. 10). Before each specimen was testing, the horizontal part of the wire inside the overmolded silicone was carefully aligned parallel with the wiper seat notch to ensure wire forces were not measured directly. Doing so removed its potential noise variable of a wire-silicone composite measurement, and instead focused measurements on wire-silicone adhesion forces (albeit still relative to the wire as its geometry interacts with the silicone overmolded wiper.

An Instron Electro-Puls E1000 mechanical tensile testing machine (Norwood, MA) was used for these experiments. Once the base and claw fixtures were in place, specimens were loaded into the base fixture, the claw fixture was adjusted for (a) the proper height to ensure no preload was imparted onto the specimen and (b) the overmolded wire was properly aligned with the claw wire-wiper notch as previously mentioned. The wire was then pulled at a rate of 0.3 mm/s to failure of a zero point slope threshold, demonstrating a catastrophic failure level. Time-dependent tensile strain and load were observed over the course of each specimen trial.

During interviews with clinicians, they repeatedly suggested a primary concern that needed to be addressed—the wiper component could not fall off inside the patient. Such an event could possibly develop from a misuse case scenario. Further questions were posed to clinicians around how the device might be misused during use. Numerous interviews with clinicians suggested that an expected misuse case would involve removal of the wiper in an orientation that would interfere with the trocar port. Therefore, a misuse case test was also executed to establish threshold targets for silicone tear/break-away forces that the design must meet for adequate safety performance and risk mitigation. A custom fixture was attached to a force gage (Chatillon DFE-025, AMETEK; Berwyn, PA) and constrained a 5 mm laparoscope, with an attached prototype device (Figs. 11(a) and 11(b)), while being pulled through a commercially available trocar (Kii Sleeve 5 mm × 100 mm Trocar, Applied Medical, Rancho Santa Margarita, CA). Before being pulled through the trocar, the wiper component of the novel device was oriented as it would be in a misuse case (Fig. 11(c)), where the wiper would catch on the lip of the trocar cannula during misuse removal (Fig. 11(d)). Average mean force observed was 1.6±0.5 kgf, with a maximum of 2.7 kgf. This data, combined with the Empirical Rule, suggests that 3.1 kgf suffices as a safety threshold for wiper retention force based on expected misuse case data. As such, the wiper component needed to have attachment strengths ≥3.0 kgf in areas of concern regarding mechanical failure.

As seen in Figs. 12 and 13, both 0 deg and 30 deg laparoscopes produced notably higher metric scores across the board for PCI-OI when compared to PDI-OI metrics, showing that debris deposition did indeed generate obscured views compared to the original clean images. In both figures, the table cells below bar graphs show the average score for each performance metric, and error bars are included on the graph itself, representing the standard deviation between the 20 trials for each metric. Table cells highlighted in green are values within 5% of the GS average, while cells highlighted in yellow fell between 5 and 15%, and cells highlighted in red fell below the 15% threshold. The GS represented a conservative target threshold for satisfactory cleaning as that specific test did not include any debris deposition on the lens (i.e., cleaned an already clean scope).

Performance metrics for both the 0 deg and 30 deg laparoscopes show consistently high scores for clearing fog debris from the lens surfaces. Clearing simulated blood and fat from the lens surfaces held more variation in lens-clearing scores, suggesting additional difficulty in clearing these types of debris from a lens surface. For the 0 deg laparoscope had IQM scores for the simulated blood and fat within the 5–15% range. For the 0 deg laparoscope, the IQM for simulated blood PCI versus PDI was at a 55% improved metric, and the IQM and Effective Cleaning Score for simulated fat tissue averaged a 26% metric improvement.

Wiper-wire adhesion data show encouraging results. Measurements for each specimen were averaged and standard deviation (StDev) was obtained. Additionally, relative standard deviation (RSD) was calculated for each category to ascertain a relative degree of error measurement in the performance data, where a low RSD represents consistent and repeatable metric would show a tight StDev relative to the sample mean. Summary data shows fairly consistent and repeatable data in Table 1.

The maximum load is essentially the reported threshold before catastrophic failure for the proposed wire-wiper design, while the tensile strain at this point shows the extension of the wiper at the point before catastrophic failure. Data show an average maximum load of roughly 4.5 kgf (9.9 lbf) with >90% repeatability and reliability expressed by the corresponding RSD. Furthermore, the small RSD seen for the tensile strain suggests an extremely reliable test setup. The maximum load experienced before failure was 2.5 kgf.

The load at first yield (i.e., the first point at which the slope of the stress–strain curve is zero) was slightly lower than the maximum load value, averaging 3.4 kgf (∼7.5 lbf) (example stress–strain curve for a specimen can be seen in Fig. 14). This figure shows a particularly interesting “double-peak” relationship. This event was seen in all samples, and may actually suggest multiple yield points levels or perhaps multiple mechanisms of failure. Images of wipers postfailure can be seen in Fig. 15. It should be noted that wipers remained connected to the wire and did not completely dislodge from the part.

The prototype device demonstrated high efficacy for cleaning, as results show high image agreement scores between OI versus PCI trials. The slight variations seen in the data regarding debris-specific and lens angle-specific factors may suggest that such aspects may effect lens cleaning performance. These variations could also potentially be related to usability features given the difference in actuation on a 0 deg versus a 30 deg laparoscope for the proposed lens cleaning device design.

The prototype device demonstrates strong lens cleaning performance, and has the potential to substantially decrease the time required to clean a laparoscope during a surgical procedure given its efficacious ability to allow for in vivo lens cleaning, coupled with the alleviated need to remove the laparoscope from the body cavity. Even though time to clean the laparoscope was not extensively or thoroughly measured for each trial, this prototype device appeared to take an estimated ∼1–5 s to achieve a clean lens, depending on number of actuations required. Yet, studies have shown that total time spent cleaning a laparoscope lens during surgery average roughly 31 s/cleaning event [22]. This information combined with general observations during the experiment suggests that even the higher number of actuations for the proposed prototype device used in this study are still an improvement upon the current methods used in laparoscope cleaning.

Mechanical testing data showed promising wire-wiper adhesion forces (relative to those expected to be experienced in a minimally invasive surgical environment). The average retention force of the wiper before failure was 4.5 kgf compared the safety threshold from a misuse case scenario of 3.1 kgf, suggesting a safety factor of ∼1.5. Interesting to note is the fact that all measured wiper failure points not only fell below the misuse case scenario threshold, but also that no wiper fully disconnected from the wire component (confirmed after testing). This may be due to material compliance, the geometry of the wire offering a mechanical “catch,” or perhaps a combination of such features. These forces suggest a robust design capable of handling mechanical uncertainties in a laparoscopic surgery. Another interesting output was the double-peak phenomena as shown in the sample stress–strain curve of Fig. 14. Although more work is needed for confirmation, it is currently assumed that this is relative to an initial release point of the adhered wire that then mechanically catches on the hook-like geometry of the wire. The observation might suggest a potential safety net built into the current design due to wire-wiper configuration and interaction. This may be reaffirmed by the fact that during testing, no wiper was ever completely dislodged from the wire part, providing strong data supporting potential safety performance of the innovative design, as even if the device were to fail during surgical use and no longer be capable of cleaning, the device could be safely removed with the failed wiper still constrained to the wire. While mechanical testing data was performed under an exploratory experiment to determine current mechanical integrity thresholds of the refined wiper design, more research is needed to understand force levels that the wiper may experience in a typical surgery. In particular, additional focus and priority for such work may consider misuse scenarios above others. The prototype was also used in a cadaveric model and demonstrated strong lens cleaning performance. While objective data was not gathered during this feasibility testing, user comments suggested validation of the concept with regard to lens cleaning performance and usability. During use in the cadaver, the device was subjected to tissue solid residue, condensation, and a variety of bodily fluids. It cleaned effectively for over 70 OLV events despite build-up of contaminant on the wiper. Future testing will include live animal studies.

The general trend in both the lens cleaning efficacy and mechanical durability experiments and data (both objective metrics and subjective observation), support the notion of the innovative design resulting in a functional and high-performance device that stands to adequately address the issue of in vivo laparoscopic lens cleaning.

• NSF (No. 1844732; Funder ID: 10.13039/100000001).

Authors Idelson, Uecker, and Rylander all disclose the following financial conflicts of interest:

• They are listed as inventors on patent pending submissions related to technology discussed in this paper, and may afforded royalties for relative patents.

• They are cofounders and equity holders in the small business entity known as ClearCam Inc.

Furthermore, Idelson is a salaried employee at ClearCam Inc.

There are no other conflicts of interest to disclose.