Use of 3-dimensional printing in surgical exploration of a nasolacrimal duct obstruction in a dog

Abstract

A 1.5-year-old spayed female English setter dog was presented because of mucopurulent discharge emanating from the left medial canthal region of 8-months duration despite medical management and repeated nasolacrimal flushing. Dacryocystorhinography demonstrated obstruction at the level of the lacrimal sac. Three-dimensional (3-D) modelling software was used to print a 3-D construct of the facial bones and a drill guide over the region of obstruction. The 3-D prints were sterilized and utilized during surgery to facilitate access to the lacrimal sac. The left lacrimal sac was identified, explored, and flushed. Patency was re-established, and the dog was asymptomatic 7 months after surgery.

Nasolacrimal duct (NLD) disorders can occur due to obstruction of the nasolacrimal system (1). Epiphora (excessive tearing) is the most common clinical sign or presenting complaint (2). However, mucopurulent punctal, conjunctival, and/or nasal discharge; swelling or draining of a fistula in the inferior medial canthal region; and/or punctal foreign bodies may also occur (1).

The nasolacrimal system has 2 openings (puncta) located 1 to 2 mm inside the medial margin of the superior and inferior eyelids. These puncta connect via short ducts (canaliculi) and join to become the lacrimal sac, which is encased within the lacrimal bone. The NLD exits the lacrimal sac and courses rostrally through a bony canal to exit the ventromedial aspect of the nasal cavity (3). Extraluminal compression as well as intraluminal obstruction of the NLD can lead to a loss of patency. Nasolacrimal duct obstructions occur in various species, with the most frequent area of obstruction or retention of foreign debris (grass awns, cellular debris) occurring as the duct passes through the lacrimal bone (4–9). Another common cause of intraluminal NLD obstruction is mucosal inflammation decreasing luminal size or completely obstructing the NLD (10). Less common causes include tissue trauma, fibrosis, or neoplasia which compresses the duct externally (10). In cases of obstruction, flushing of the NLD or surgical intervention may be necessary to re-establish patency and eliminate clinical signs (1,9).

Although treatment of dacryocystitis and NLD obstruction with medication and irrigation has been well-documented in dogs, this is not always effective for the resolution of clinical signs (1,3,4,11,12). There is a limited number of reports outlining the surgical treatment of dacryocystitis, with only 1 report describing foreign material as the underlying cause (9,10,13–16). The surgical approach to access the NLD is more complex than that for other periocular surgeries and may require a great deal of microsurgical dexterity (17). The anatomic location and minute size of the NLD make access to and cannulation of the duct challenging; removal of bone is often required to access the duct. As such, the use of a 3-D printed model specific to the patient can facilitate surgical planning, visualization of the optimal surgical approach, and an overall enhanced surgical outcome through a more precise anatomical approach. To the authors’ knowledge, this is the first case report detailing the use of 3-D modelling as an aid in the surgical planning and treatment of recurrent dacryocystitis secondary to lacrimal sac obstruction in a dog.

Case description

A healthy 1.5-year-old spayed female English setter dog weighing 17.0 kg was presented to Michigan State University Veterinary Teaching Hospital (MSU VTH) for evaluation of mucopurulent discharge emanating from the left medial canthal region of 8 mo duration (Figure 1A). Immediately prior to the onset of discharge, the owner reported removing a Canada wild rye grass seed (Elymus canadensis) from the dog’s left conjunctival fornix. The dog was evaluated by the referring veterinarian and treated with oxytetracycline and polymyxin B ophthalmic ointment (Terramycin; Pfizer Animal Health, New York, New York, USA), q8h, OS for 5 d, and concurrent tobramycin 0.3% ophthalmic solution (Tobramycin ophthalmic solution USP; Akorn, Decatur, Illinois, USA), q8h, OS for 7 d, followed 10 d later by neomycin polymyxin B dexamethasone 0.1% ophthalmic suspension (Neomycin and polymyxin B and dexamethasone ophthalmic suspension USP; Bausch and Lomb, Bridgewater, New Jersey, USA), q12h OS for 1 mo. No other ophthalmic abnormalities were noted. During the initial course of treatment (1 mo duration), the mucopurulent discharge subsided. However, it returned shortly after discontinuation of medications. The NLD was successfully flushed normograde twice by the referring veterinarian using saline solution under sedation, but the discharge persisted. The owner reported 1 incidence of blood-tinged discharge emanating from the left medial canthus, occurring after treatment and flushing of the NLD and before presentation at MSU VTH.

Figure 1.

A — Close-up view of mucoid discharge emanating from the patient’s left inferior lacrimal puncta on initial presentation. B — Transverse CT scan of the skull with contrast dye highlighting catheterization of the left inferior lacrimal puncta (arrowhead) and a filling defect, where the dye pools at the level of the left nasolacrimal sac (arrow). C — A 3-D model was generated based on contrast CT. The nasolacrimal sac is modeled in a blue color and the simulated surgical trajectory of the burr (magenta) from the maxilla to the nasolacrimal sac and the lesion is depicted. D — A patient-specific surgical guide (blue) to delineate the area of bone needing to be burred is designed based on the simulated path of the burr and the contour of the maxilla.

Findings from a full physical examination, including oral examination, were within normal limits. Complete ophthalmic and neuro-ophthalmic examinations, including retropulsion, revealed no abnormalities aside from mild mucopurulent discharge emanating from the left inferior punctum. Schirmer tear test values were 18 mm of wetting per min OD and 17 mm OS. Intraocular pressures were 15 mmHg OD and 18 mmHg OS, as measured by rebound tonometry (Tonovet: iCare, Vantaa, Finland). No fluorescein dye retention was noted on either cornea, and no dye was observed emanating from either nostril (Jones test negative). Normal airflow was present in both nostrils. Slit lamp biomicroscopy (Kowa SL-17; Kowa Company, Tokyo, Japan) and indirect ophthalmoscopy (Keeler All Pupil II; Keeler Instruments, Broomall, Pennsylvania, USA) were unremarkable.

Differential diagnoses for mucopurulent ocular discharge included infectious (i.e., distemper) and inflammatory/immune-mediated (i.e., chronic superficial keratitis) causes, anatomic eyelid abnormalities (entropion), abnormally located hairs (distichia, ectopic cilia), ocular trauma, qualitative and quantitative tear film abnormalities, foreign material, allergies, neoplasia, and obstruction of the NLD and/or lacrimal sac. Based on the historical removal of foreign material from the fornix and clinical signs, foreign material resulting in unresolved dacryocystitis secondary to lacrimal sac obstruction was suspected. Additional diagnostic tests included a complete blood (cell) count (CBC), serum biochemistry panel, and aerobic culture and antibiotic susceptibility test of the mucopurulent discharge. Anaerobic culture was not pursued at the time of initial presentation due to financial limitations. Blood analysis was unremarkable. The culture results revealed Staphylococcus pseudintermedius and rare Actinomyces canis organisms susceptible to all antimicrobials tested, except ampicillin and penicillin.

The dog was routinely pre-medicated and anesthetized. Pre-medication for the CT scan included butorphanol tartrate (Torbugesic; Zoetis, Parsippany, New Jersey, USA), 0.3 mg/kg body weight (BW), IM, and acepromazine maleate (PromAce; Boehringer Ingelheim Vetmedica, St. Joseph, Montana, USA), 0.03 mg/kg BW, IM. The dog was maintained under general anesthesia using isoflurane (Isoflurane; Isothesia, Henry Schein, Melville, New York, USA) in oxygen at a concentration of 1.5% to 3%.

Both nasolacrimal ducts were flushed with dye in a normograde direction using a 25-gauge IV catheter. There was no resistance and good flow through the right NLD; however, there was significant resistance and lack of passage through the left NLD. Computed tomography of the head was performed (CT Scanner: Revolution EVO; GE Healthcare, Chicago, Illinois, USA) with transverse 0.625-mm slices at 120 kVP and 200 mAs. Pre-contrast imaging showed soft tissue thickening immediately adjacent to the medial aspect of the left lacrimal sac. Contrast dacrocystorhinography (Omnipaque; iohexol injection 52%; Marconi Medical Systems Canada, Brampton, Ontario) revealed normal dye passage and anatomy of the right NLD. A filling defect involving the left NLD was observed at the level of the opening to the lacrimal sac with no passage of dye rostral to the defect (Figure 1B). With contrast administered normograde, a focal dilation of contrast and a small amount of gas were present caudally at the level of the nasolacrimal fossa, immediately preceding entry to the lacrimal canal. Retrograde contrast administration revealed an area of stricture bordered by the inferior rim of the orbital bone and caudo-superior aspect of the maxillary sinus. A diagnosis of NLD obstruction was made, presumed secondary to foreign material (surmised Canada wild rye grass seed) based on clinical signs and history. Surgical exploration of the left nasolacrimal system was discussed, and an initial decision was made by the owner to assess response to medical therapy. The dog was discharged with neomycin B polymyxin dexamethasone 0.1% ophthalmic suspension (Neomycin and polymyxin B and dexamethasone ophthalmic suspension USP; Bausch and Lomb), q8–12h OS until recheck, carprofen (Rimadyl; Pfizer Animal Health, New York, New York, USA), 2.2 mg/kg BW, PO, q12h, for 2 wk, and amoxicillin trihydrate/ clavulanate potassium (Clavamox; Pfizer Animal Health), 14.5 mg/kg BW, PO, q12h for 2 wk.

Although the volume of mucopurulent discharge decreased after flushing the left NLD, intermittent drainage from the left medial canthus persisted despite ongoing medical therapy. The owner elected to pursue exploratory surgery 4 wk after initial evaluation at MSU VTH.

3-D prints of the maxillary and lacrimal bones and their associated foramina overlying the region of obstruction were made from the CT scan images using 3-D modeling software (3-D Modeling Program; Mimics 19.0, materialize NV, Leuven, Belgium) (Figure 1C). Computer-aided design software (3-D Printing Material: Meshmixer; Autodesk, San Rafael, California) was used to generate the necessary computer files from the original CT images to produce a 3-D rendering of the regions of interest. A 3-D Printer (Form 2; Formlabs, Somerville, Massachusetts, USA) using white resin “ink” was employed to make the 3-D prints. A surgical drill guide mirroring the contour of the facial bones was designed and printed to ensure accurate entry to the obstructed region (Figure 1D). Printed materials were gas sterilized.

The dog was routinely anesthetized as described for the CT scan. Maintenance fluids (LRS, Lactated Ringer’s Injection Rx, USP; Hospira, Lake Forest, Illinois, USA), 5 mL/kg BW per hour, IV, and cefazolin sodium (Cefazolin: ANCEF; West-Ward Pharmaceuticals, Cherry Hill, New Jersey, USA), 22 mg/kg BW, IV, on induction and q90min during surgery were administered under anesthesia. The fur was clipped from the left inferior eyelid extending 10 cm distal from the eyelid margin. The left conjunctival sac, globe surface, and periorbital skin were routinely prepared for surgery using diluted 2% povidone iodine solution. The patient was positioned in sternal recumbency with the head in an elevated position, exposing the left eye and adnexal region. The left periorbital area was routinely draped and a 2-cm vertical skin incision was made using a #15 scalpel blade (Bard-Parker; Aspen Surgical, Caledonia, Michigan, USA) beginning 3 mm rostral to the infraorbital foramen (the foramen is located within the maxillary bone, approximately 5 to 8 mm below the center of the inferior orbital rim) and ending 5 mm distal to the central aspect of the inferior eyelid margin. A self-retaining adjustable retractor (Lone Star Retractor; Cooper Surgical, Trumbull, Connecticut, USA) was placed within the skin incision to provide adequate exposure. The subcutaneous tissues were dissected using blunt and sharp dissection with Steven’s tenotomy scissors (Sklar Instruments, West Chester, Philadelphia, USA), with careful attention to avoid the angularis oculi vein located dorsomedial to the surgical incision. The nasolabialis muscle was transected and the inferior aspect of the orbicularis oculi muscle was isolated and separated from the underlying bone using a Freer periosteal elevator. The position of the underlying lacrimal sac was identified using the printed surgical guide (Figure 2A) and a 4-mm diameter round burr and air drill (Air Drill; Air Pen Drive, DePuy Synthes Power Tools, Palm Beach Gardens, Florida, USA) were used to remove a portion of the lacrimal and caudal maxillary bones (Figure 2B), measuring approximately 5 mm in diameter, overlying the osseous lacrimal canal exposing the underlying lacrimal sac and NLD. Hemostasis was achieved using bone wax, phenylephrine 1% ophthalmic solution (Akorn, Decatur, Illinois, USA), and digital compression. The lacrimal sac was entered using a #15 scalpel blade (Bard-Parker; Aspen Surgical), explored, and flushed. No foreign material was identified.

Figure 2.

A — Placement of the 3-D printed guide marking the location of bone to be burred. B — Burring of the lacrimal bone using a 4-mm diameter burr and an air drill. C — Post-operative side view of the dog, 1-month post-surgery demonstrating cannulation of the left nasolacrimal duct using 4-0 Ethilon which was secured to the patient’s nose using 1/2 inch white hospital tape and 4-0 nylon suture. D — Photograph of the left eye 2 mo after surgery demonstrating the absence of discharge and renewed patency of the left nasolacrimal duct.

A catheter (Tom Cat Catheter: Sovereign; Kendall international, Mansfield, Massachusetts, USA) was passed normograde through the opening of the lacrimal sac rostrally. A 4-0 nylon suture (Ethilon; Ethicon, Cincinnati, Ohio, USA) was then threaded through the catheter to cannulate the distal nasolacrimal duct. The suture was clamped with mosquito hemostats at the opening of the NLD and the catheter was removed.

A 25-gauge catheter was passed normograde from the superior lacrimal punctum into the lacrimal sac. The same piece of 4-0 nylon suture (Ethilon; Ethicon) described previously was threaded up the 25-gauge catheter to exit the superior lacrimal punctum; the catheter was removed. The superior punctum was used due to swelling involving the inferior punctum at the time of surgery. The incision into the lacrimal sac and the remaining bony defect were left to heal by second intention. The subcutaneous tissues and nasolabial muscle were closed using 6-0 polyglactin 910 (Polyglactin 910: Vicryl; Ethicon, Cincinnati, Ohio) suture in 3 layers using a simple continuous suture pattern. The skin was closed intradermally with 6-0 polyglactin 910 (Polyglactin 910: Vicryl; Ethicon) suture. The rostral and caudal aspects of the 4-0 nylon (Ethilon; Ethicon) used to cannulate the left NLD were secured to the skin using 1/2 inch white medical tape and 4 simple interrupted sutures using 4-0 nylon (Ethilon; Ethicon) (Figure 2C).

The dog recovered uneventfully from anesthesia. An Elizabethan collar was kept in place for the entire suture retention time of 1 mo to prevent self-induced trauma to the surgical site. The dog was treated with ofloxacin 0.3% ophthalmic solution (Ofloxacin ophthalmic solution, USP 0.3%; Akorn) and OptixCare ophthalmic lubricant (OptixCare Eye Lube; CLC Medica, Waterdown, Ontario), q6h OS for 2 wk, carprofen (Rimadyl; Pfizer Animal Health), 2.2 mg/kg BW, PO, q12h, and amoxicillin trihydrate/clavulanate potassium (Clavamox; Pfizer Animal Health), 14.5 mg/kg BW, PO, q12h, for 2 wk, and tramadol (Tramadol; Sun Pharmaceutical Industries, Cranbury, New Jersey, USA), 4.4 mg/kg BW, PO, q8h for 10 d.

A decrease in mucopurulent ocular discharge was noted within the first 48 h after surgery. The dog tolerated the nasolacrimal stent and was discharged without complication.

The referring veterinarian re-examined the patient 2 wk after surgery. The nasolacrimal stent was securely in place and the dog was comfortable. Follow-up examination at Michigan State University 1 mo after surgery demonstrated resolution of mucopurulent discharge and healing of the surgical site. The left NLD was patent (Jones positive) and the 4-0 nylon (Ethilon: Ethicon) suture was removed without complication. Seven months after surgery, the dog remained asymptomatic with no recurrence of nasolacrimal discharge (Figure 2D). Repeat contrast dacryocystorhinography could have been considered to verify patency of the NLD; this was not pursued due to financial constraints of the owner.

Discussion

Surgical approaches to the lacrimal sac similar to the one described here have been reported (9,13,14), along with the use of a high-speed burr, air drill, and stent material (13,14). In cases of NLD obstructions, dacryocystorhinography is a valuable imaging tool to identify the location of the obstruction (8,10,11,14–16,18,19). However, the 3-D printed model used in this case was a unique, useful tactile tool that may facilitate a minimally invasive surgical approach. Disadvantages of this modelling technology include the costs associated with acquiring a 3-D printer and necessary computer-aided design (CAD) editing software, in addition to familiarity with the software itself. Alternatively, lacrimoscopy could have been considered herein. The advantage of this technique includes its minimal invasive nature, potentially avoiding the need for cutting/ disrupting bone to access the site of obstruction. However, disadvantages include the high instrumentation costs and the necessary technical skill set required by the user to adequately perform this technique. 3-D printing has become increasingly common in physician-based medicine as a clinical, research, and hands-on teaching tool (20). 3-D printing is also rapidly gaining popularity in veterinary medicine, particularly for orthopedic procedures (19–22). The creation of custom prosthetics and patient-specific guides (23) has facilitated rehearsal of surgeries prior to performing the procedure on a live patient (24), enabled pre-contouring of plates for orthopedic procedures (25), and allowed for printing of complex biomodels, such as cerebral aneurysms (26). This has decreased surgical and operating room times in 46% of human studies (20,24). Access to 3-D printed implements has also enhanced the cosmetic surgical outcomes in 72% of studies (24,27).

Dacryocystorhinography, in combination with CT imaging, has permitted evaluation of the bone surrounding the NLD and the development of a specific treatment plan for the patient. However, it does not provide a tangible object to view, handle, and/or manipulate. The 3-D model provided a full-scale, intricately detailed, anatomical representation of the skull, which facilitated localization of the NLD, including the area of obstruction (19,28). This model also allowed for a simulated trajectory of the drill to be generated, permitting access into the nasolacrimal fossa, prior to surgery. Measurements could be made directly from the model and in the operating room, which were applied during surgery to mark the drill site and locate the lacrimal sac (29). Based on this, a patient-specific surgical guide was made and access to the duct and fossa was obtained without complication. The 3-D model allowed the surgeon to pre-plan the surgical approach, perform a mock surgery, and identify the necessary surgical instruments required for surgery.

In contrast to previous veterinary reports using 3-D models (30), this study evaluated the utility of printing only the portion of the skull containing the NLD and lacrimal sac. This significantly reduced the printing time and the cost of materials. Prolonged printing time has been considered a limitation of using 3-D printed models (21). In addition, 33% of human-based studies noted that 3-D printing was costly (20). The 3-D printer used in this report (Form 2 Formlabs) costs approximately $3000 USD. When used to print a small surgical model for a dog weighing 50 kg, the cost of consumables alone is relatively inexpensive at approximately $30 USD.

Although no foreign material was identified before, during, or after surgery, the lacrimal sac was inflamed and difficult to cannulate. This correlated with the filling defect observed on imaging. The presence of chronic inflammation likely contributed to obstruction of the NLD; therefore, nylon suture was used as a stent in lieu of tubing due to the amount of swelling present and the small size of the NLD. The lacrimal sac was left open to heal by second intention as there was concern that suture used to close the sac might propagate an increased inflammatory reaction and lead to permanent obstruction of the nasolacrimal system. Similarly, urethral openings left to heal by granulation were associated with less inflammation and no post-operative urethral stricture formation (31). Leaving the nasolacrimal sac to heal via second intention and more importantly, long-term cannulation of the nasolacrimal system, resulted in patency and resolution of presenting clinical signs in this case.

Soft tissue structures were closed superficial to the bony defect, which was also allowed to heal by granulation. Ongoing remodeling and rapid revascularization have been shown to be equally effective and less time-consuming than more invasive surgical techniques used to bridge gaps in bone (32).

This case highlights the benefits of CT imaging in conjunction with 3-D modeling to aid in planning a surgical approach to relieve NLD obstruction. The secondary signs of epiphora and mucopurulent discharge from the lacrimal puncta were resolved.