Innovation in Urology: Three Dimensional Printing and Its Clinical Application

Three-dimensional (3D) printing allows rapid prototyping of novel equipment as well as the translation of medical imaging into tangible replicas of patient-specific anatomy. The technology has emerged as a versatile medium for innovation in medicine but with ever-expanding potential uses, does 3D printing represent a valuable adjunct to urological practice? We present a concise systematic review of articles on 3D printing within urology, outlining proposed benefits and the limitations in evidence supporting its utility. We review publications prior to December 2019 using guidelines outlined by the Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA) statement. Of 117 identified articles, 67 are included highlighting key areas of research as the use of patient-specific models for patient education, surgical planning, and surgical training. Further novel applications included printed surgical tools, patient-specific surgical guides, and bioprinting of graft tissues. We conclude to justify its adoption within standard practice, further research is required demonstrating that use of 3D printing can produce; direct and measurable improvements in patient experience, consistent evidence of superior surgical outcomes or simulation which surpasses existing means' both in fidelity and enhancement of surgical skills. Although exploration of 3D printing's urological applications remains nascent, the seemingly limitless scope for innovation and collaborative design afforded by the technology presents undeniable value as a resource and assures a place at the forefront of future advances.


INTRODUCTION
Stereolithography or additive manufacturing, describes forming three-dimensional (3D) objects by step-wise layered addition of material. 3D printing allows successive production of structurally distinct objects instead of the mass production of identical items typically achieved by subtractive manufacturing (1).
This underpins the primary rationale for 3D printings' increasing utilization within urology: As every patient is unique, both their surgical treatment and adjuncts to it should be similarly individualized. 3D printing can utilize data from medical imaging to produce structures customized from and for a patients' individual anatomy (2). Furthermore, products can be functionally tested then quickly adapted in the next iteration.
Three-dimensional printing has the potential to enhance patient-specificity of pre-operative counseling, surgical simulation, implantable prostheses and even transplant organs (3). Customizable surgical instruments could be shared digitally, adjusted to preference and produced as required (3,4). However, with printers used in surgical research ranging in price from $2000 to $900,000 this resource requires investment (5). Maintenance, cartridge consumables, specialist software for converting medical imaging into printable format and computer hardware further add to setup costs.
With current literature detailing diverse urological applications, distinguishing potential from proven advantage is key to guiding future practice and research alike. We therefore aim to elucidate 3D printing's confirmed benefits to patient and clinician whilst highlighting those requiring further evidence. Please refer to the Supplementary Material for details on review methodology (6). Figures 1 and 2 Illustrate products of 3D printing and a range of 3D printers.

PATIENT EDUCATION
Coupling medical 3D imaging with additive manufacturing allows accurate modeling of individual patients' anatomy by converting DICOM (digital imaging and communications in medicine) data to stereolithography (STL) format (2). Bernard created kidney models with transparent resin for renal parenchyma showing tumor, vasculature, and collecting system in patients awaiting partial nephrectomy. These reportedly improved patients understanding of kidney physiology, anatomy, tumor characteristics, and the planned procedure (7). Altahay used 3D-printed models prior to percutaneous nephrolithotomy (PCNL), reporting similar improvements in patient understanding as assessed by questionnaires before and after counseling with the 3D model (8). Teishima demonstrated superior questionnaire-assessed understanding in both patients and family members prior to partial nephrectomy when comparing explanation with 3D-printed models to using computerized tomography (CT) alone (9). Whereas Schmit reported no statistically significant improvement in patient understanding from explanation of cryoablation with 3D models compared to 2D imaging after correcting for different counseling physicians (10).
In patients undergoing partial nephrectomy or radical prostatectomy, Wake et al. compared counseling with standard imaging to imaging plus either a 3D printed model, augmented reality model or 3D computer model (11). Patient understanding was assessed via Likert scale survey with 3D-printed models and 3D computer models receiving higher scores than the control group. Only the 3D-printed model showed statistically significant improvements, including in patient-rated comfort with the surgical plan thus establishing a link between patients understanding and anxiety. A summary of reviewed articles relating to patient education is provided in Table 1.

SURGICAL PLANNING
In Zhang's study, models of patients' kidneys with T1 tumors received positive face validation scores by experienced Urologists for representation of tumor size and inter-related structures but details of renal vasculature and the collecting system were scored less favorably (12). Silberstein presented cases of T1 renal tumors where 3D-printed models were used as a realtime reference for surgeons when performing reconstruction and assessing resection in relation to the hilar vessels and collecting system (13). Surgical outcomes were reported but with no control arm for comparison. Komai reported outcomes for partial nephrectomies with nephrometry scores ≥8 where a "4D" printed model was used in surgical planning (14). The proposed 4th dimension was time as the model included a removable  tumor with a 2-5 mm margin to assess the defect in relation to renal structures but similarly no control arm was included to indicate superior outcomes. Wake et al demonstrated 3D models could (a) change experienced urologist's planned surgical approach for treatment of a renal tumor and (b) the planned approach based on a 3Dprinted model was more likely to be followed than one based on 2D imaging. (15).
Maddox used patient-specific models to simulate robotassisted partial nephrectomies prior to the actual procedure and outcomes were compared to the Tulane Urology prospectively maintained database. Cases using the surgical models had larger tumors (4. The only statistically significant difference however was a lower estimated blood loss (185.7 vs. 235.6 ml, p = 0.01) (16). Kyung et al. compared the outcomes for partial nephrectomy aided by prior inspection of a 3D-printed model against a control cohort and similarly the only significant difference was reduced estimated blood loss (17) whereas, Fan reported a significant reduction in warm ischemic time (18).
Twelve studies described using 3D-printing in prostate cancer diagnosis and management but only 2 addressed patient-tailored pre-operative planning; Shin used pre-biopsy prostatic MRI's of patients to construct translucent models of the prostate including the biopsy-proven malignant lesion and neurovascular bundles (NVB). The relationship between prostate, lesion and NVB was assessed using these before performing nerve-sparing radical robotic prostatectomy. A wider (1 mm) area of periprostatic tissue was dissected at regions high risk for extra-capsular extension based on the models. Although high risk cases (pT2c [n = 1], pT3a [n = 2], and pT3b [n = 2]), histopathology confirmed all having negative surgical margins (19). Wang used translucent prostate models with visible MRI-identified lesions to aid cognitive prostate biopsy. Standard systematic biopsies were taken, followed by 2-3 targeted biopsies. A higher positive biopsy rate was reported in the targeted biopsies and higher-grade disease was identified (20).
We identified 3 studies using 3D-printing for patientspecific aids to percutaneous nephrolithotomy (PCNL). Atalay constructed anatomically accurate models of patients with complex staghorn calculi and presented these to surgical residents before PCNL. Their understanding of the collecting system anatomy, stone location and optimal entry calyx was assessed via questionnaire with reported improvement in all categories (21). Xu created copies of patients renal anatomy, containing 3D-printed facsimiles of their stones and simulated the procedure using different puncture sites to determine the optimal approach. CT-assessed stone clearance with the models correlated well with procedures using the same puncture sites (22). Golab used a virtual 3D model of a patient with a horseshoe kidney and bilateral stones to determine the optimal puncture site and angle for PCNL, then printed a sterilisable surgical guide (23). The guide was positioned using vertebral spinous processes as markers and achieving nephroscopic access in this challenging case reportedly took 3 min.
This method of surgical guide has been utilized to implant tined leads for sacral nerve modulation with Zhang et al. reporting reduced punctures, procedural time and fluoroscopy use (24). A statistically significant reduction in effective voltage after using the surgical guide also indicated a more optimally positioned lead. An expert consensus guideline by Shaito and Ye detail potential use of 3D-printed templates for brachytherapy seed implantation (25). A summary of reviewed articles relating to surgical planning is provided in Table 2.

SURGICAL TRAINING
Gasior detailed 3D-printings use in improving understanding of complex cloacal anomalies for trainees and faculty surgeons alike (26). Tangible 3D-printed model demonstrated improved questionnaire-assessed understanding compared to 2D cloacagrams, rotatable 3D virtual models and 3D video animations.
Simulation with 3D-printed anatomical models has been a focus of research, offering patient-specific representation of pathology and tactile feedback. Simulation utilizing 3Dprinting has been described for partial nephrectomy, pyeloplasty, PCNL, robot-assisted prostatectomy, robot-assisted kidney transplant, vasectomy reversal, and transurethral resection of the prostate (TURP).
Adams created "phantom" models of cadaveric kidneys using 3D-printed molds designed from CT imaging. These were compared to the original in ultrasound and CT appearances as well as hardness, elasticity, and tensile strength. Different materials were trialed and the water-based gel, Agarose proved most similar to human kidney tissue (27). Melnyk recently created poly-vinyl alcohol (PVA) kidney phantoms using 3D-printed casts derived from a patient with a 4.2 cm exophytic renal tumor (28). Renal vasculature and collecting system were constructed from PVA and models were tested against porcine equivalents in mechanical properties & flow characteristics of the simulated vasculature. The group also compared suture tension required to approximate renal parenchymal edges and the maximum tension at which suture tension tore through parenchyma. Testing identified 7% PVA models with a three freeze-thaw cycle as the formula best replicating porcine tissue. The models were set amongst fabricated peritoneum, abdominal fat, spleen, bowel and mesentry to form an immersive simulation of robot-assisted partial nephrectomy. Similarly, Choi created prostate phantoms for simulation of TURP using PVA, agar and hollow glass powder, testing compressive & elastic properties. Models with different agar percentages were compared to normal, cancerous, and hyperplastic prostate tissue (29) with the physical properties of these simulation models being demonstrated as adjustable.
Ghazi's PVA hydrogel models of the kidney and adjacent structures allowed immersion simulation of PCNL with experts (caseload >100) and novices (caseload <20) rating it highly in similarity to the real procedure and usefulness in training. The models' realism was further supported by significant superiority from experts compared to novices in mean fluoroscopy time, number of percutaneous access attempts and stone clearance (30). Cheung et al printed molds of renal anatomy with pelvic ureteric junction obstruction and cast silicone rubber models for low-cost, reusable simulation of laparoscopic pyeloplasty (31). The simulation was validated on a 5-point Likert scale, scoring 4.75 (± 0.29), 4.50 (± 0.41), and 4.38 (± 0.48) in overall impression, realism and handling, respectively. Shee used printed casts to mold a surrogate bladder neck and urethra from silicone for an ex-vivo simulation trainer for robotic vesicourethral anastomosis (32) with an average face validity rating of 8/10 and content validity of 10/10. Experts and trainees rated the simulation superior to digital virtual reality (VR) trainers and experts performed better than residents in the simulation. Uwechue developed a 3D-printed simulation model for vascular anastomosis during robot-assisted kidney  transplant, highlighting potential training value even amongst experienced surgeons when learning new surgical techniques (33). Pinto reported an observed improvement in residents microsurgical suture time and quality on a vasectomy reversal model produced via 3D-printing (41). However, as Monda et al commented in their paper on partial nephrectomy simulation, future studies need to establish that improved performance on simulation models is associated with improvements in live surgery (42). Simulation presents a reproducible and uniform means to assess trainee performance both subjectively and with quantitative metrics. The aforementioned TURP model by Choi used materials with different ultrasound echogenicity for the central and peripheral zones of the prostate so that the resected volume of each zone could be assessed (29). Qiu incorporated tactile sensors into 3D-printed prostate models able to calculate pressure forces applied to the model (43). Witthaus et al. constructed a simulation model for robotassisted radical prostatectomy using a chemiluminescent dyeimpregnated PVA hydrogel model of the prostate and a tension wire sensor incorporated within the neurovascular bundle (44). The tension wire provided quantitative measurement of tension applied to the NVB during nerve-sparing prostatectomy and the dye allowed assessment of the surgical margin. Further metrics included the urethrovesical anastomosis leak test and task-specific times. When compared against assessment via a Global Evaluative Assessment of Robotic Skills (GEARS) and Robotic Anastomosis Competency Evaluation (RACE), higher GEARS score tallied with lower exerted force on the NVB and higher RACE scores correlated with a lower UVA leak rate. Witthaus' paper highlighted novel ways in which 3D-printing and simulation can be designed to objectively assess trainees.
Parkhomenko et al aided surgical simulation in a different way by designing a 3D-printable, portable laparoscopic trainer with a reported production cost of $26.50 and assembly time of <45 min (4). Whilst conventional trainers were scored higher by trainees, all still reported it as useful and this study demonstrated how 3D-printable designs can easily be shared across institutions for immediate production. A summary of reviewed articles relating to surgical training is provided in Table 3.

PATIENT-SPECIFIC PROSTHESES & BIOPRINTING
Outside of urology 3D-printing has been used for customized orthopedic plate sizing and molding (45) as well as for titanium and ceramic patient-specific maxillofacial implants (46). 3Dprinted implants in urology are limited by the unique mechanical and physiological functions of the genitourinary tract and further complicated by the need to be sterilisable. Patient-specific 3Dprinted extravascular stent have been used to treat posterior nutcracker syndrome (47) (48) but these do not emulate any urological structure. CT imaging was used to design a custom stent, 3D-printed from a titanium alloy and laparoscopically sited around the retro-aortic left renal vein to prevent compression.
However, substitutional grafts within the genitourinary tract remain possible via bioprinting. Organ production from native tissue over a scaffold was demonstrated in the Vacanti mouse where a chondrocyte-seeded, ear-shaped scaffold was implanted beneath its skin (49). Atala similarly used urothelial and smooth muscle cells obtained from bladder biopsy to 'grow' autologous tissue around a biodegradable bladder-shaped collagen scaffolds which were then successfully used for cystoplasty (50). Huang reported using 3D porous bacterial cellulose scaffolds seeded with rabbit lingual keratinocytes as a material for urethral reconstruction (51). These original cellulose scaffolds were not produced via 3D-printing but an integrated tissue-organ printer (ITOP) system was later used by Khang to print a porous, spiral scaffold dispersed with rabbit urothelial and smooth muscle cells within a fibrin hydrogel. This bioprinted urethra demonstrated mechanical properties equivalent to native rabbit urethra, with cells maintaining 80% viability at 7 days and demonstrating active proliferation (52). Versteegden also reported on collagen scaffolds produced via 3D-printing,   reproducing the elasticity and shape-recovery of human urethral tissue (53). Yu investigated 3D-printed polycaprolactone (PCL) scaffolds and culture of human fibroblast cells for potential use as a surrogate for tunica albuginea (54) and Oh et al successfully cultured human aortic smooth muscle and umbilical vein endothelial cells over 3D-printed PCL scaffolds as a potential tissue-engineered corpus cavernosum graft (63).
Whilst these studies demonstrate how unique mechanical properties of the urinary tract can be emulated, a more significant development toward future autologous kidney replacement is the recent bioprinting of a renal proximal convoluted tubule. This involved printing a silicone gasket within a perfusable "3D tissue chip" and seeding it with immortalized human proximal tubular cells (61) forming a polarized epithelium, functional as a barrier and damaged by known nephrotoxins. A summary of reviewed articles relating to patient-specific prostheses and bioprinting is provided in Table 4.

SURGICAL TOOLS
Park published in vitro test results for a 3D-printed anti-reflux ureteric stent (60) whilst Junco et al created ureteric stents and laparoscopic trocars via 3D-printing, tested in porcine and cadaveric models. Junco was successful in producing 9F and 12F diameter stents which were deployable via a 0.035 guidewire but the smaller 7F stent did not allow passage of a guidewire and they were unable to print a stent with a tapered end. The 3D-printed trocars maintained pneumoperitoneum and allowed instrument passage but formed larger superficial skin defects than Karl Storz and Ethicon equivalents (64). Issues of biocompatibility, sterility, durability, tensile strength, and stent encrustation were not evaluated or addressed in the study. Canvasser's pilot study of 3D-printed surgical clips proved inefficacious as they broke, failed to close and leaked more than commercially available alternatives (65). Rankin's military skin retractors 3D-printed using polylactic acid (PLA) filament were sterilisable via glutaraldehyde protocols and freshly printed retractors were sterile on bacterial testing via polymerase chain reaction (PCR). When stressed until fracture they could tolerate 13.6 kg of tangential force with no significant change after glutaraldehyde sterilization. Whilst insufficient for retraction of the abdominal wall, skin flaps or for orthopedic procedures this was fit for purpose as a skin retractor and low cost ($2.77 per retractor compared to the $23.48 stainless steel equivalent) (66). The cost of disposable equipment due to packaging, sterilization, transportation & storage could be overcome with on-site printing of tools and advantages would be greater in developing countries, where access to surgical equipment is limited by cost and transport. A summary of reviewed articles relating to surgical tools is provided in Table 5.

NOVEL USES
Morimoto described a virtual reality design interface and 3Dprinted concentric tube robot produced by this method. The surgeon-designed robot extended to curve beneath the 12th rib, into the kidney and upwards to the tumor (67). This prototyped a potential surgical instrument for tumor ablation but the study's focus was the interface which incorporated patient imaging and allowed the design and 3D-printing of patient-specific equipment by an individual surgeon.
3D-printed molds were used in two studies to standardize histological sampling of malignant renal and prostate specimens against imaging to validate the accuracy of imaging in identifying tumor location and likely histological findings (62,68,69). Antonelli used renal models to simulate PCNL and test a new "PercSac" device for capturing stones and fragments during the procedure (70).
Tse et al developed the "Endockscope" incorporating a cordless, light-emitting diode (LED) light source and a 3D-printed attachment for connecting smartphones to the flexible cystoscope eyepiece (71). Testing different smartphones against a Storz HD camera and xenon light, all were inferior to the standard system in image quality, brightness and color but combination with the Samsung S8 was uniformly rated by faculty urologists as acceptable for use.

FUTURE DIRECTIONS
Proposed applications for 3D printing within urology are expanding, with particular focus on tangible models of patient anatomy to enhance patient understanding, surgical planning and simulation training. The rapid increase in published research has been accompanied by related review articles. Smith and Dasgupta (72) reviewed applications within urological training whilst Checcuchi (73) provides a non-systematic review of both virtual and printed 3D models' utilization in robotic urological surgery. This review gives a concise overview of all reported uses of 3D printing across urology. Both Cacciamani's and Chen's systematic Explanation with patient-specific 3D-printed models can -Improve questionnaire-assessed patient understanding prior to partial nephrectomy, PCNL, and radical prostatectomy No comparison to explanation with a general renal tumor/kidney stone/prostate model. Not conclusive that patient-specific models are required to improve understanding.
-Improve patient understanding compared to routine 2D imaging Reported benefits of explanation with a 3D model confounded by duplication of the explanation process -Improve patient understanding compared to other novel models; augmented reality and 3D computer models Entailed a second explanation so benefit over standard explanation alone is confounded -Improve patient rated comfort with the planned surgical intervention More evidence required linking improved patient understanding with other measurable outcomes: e.g., patient anxiety, post-op recovery, and length of stay

SURGICAL PLANNING
Inspecting tangible patient-specific models can be used to inform and optimize surgical approach Low volume evidence that the surgical outcomes are improved significantly with many studies lacking a control arm Simulated rehearsal of procedures with patient-specific models may reduce operating time and improve surgical outcomes Evidence of improved outcomes is so far limited to estimated blood loss and warm ischemic time for partial nephrectomy. Small case series Reported higher positive targeted biopsy rate if MR identified prostate lesions are presented in 3D models Requires cost comparison of patient-specific models vs. fusion targeted biopsy Surgical guides may improve time to PCNL access The quicker nephroscopic access is unlikely to be as useful for uncomplicated cases due to the extra time required designing the guide. Automated design of guides for optimal access would be invaluable.
Using surgical guides for sacral-neuromodulation tine insertion can reduce punctures, procedure time, fluoroscopy time and optimize tine positioning.
Small cohort studies. Similarly automation of guide design would be invaluable

SURGICAL TRAINING
Understanding of complex anatomical pathology can be improved with 3D printed models Single case study The physical properties of simulation models are adjustable to approximate the behavior of different tissues Reviewed studies compared physical properties of models to porcine equivalents Phantoms of patient-specific anatomy/pathology made via 3D printed casts can be used to simulate surgery (e.g., TURP, partial nephrectomy, prostatectomy, PCNL and vasectomy) Further research is required to link practice on simulation models to improved surgical performance Simulation with replicable anatomy allows a more standardized means of trainee assessment further aided by quantifiable metrics in simulation models Requires further assessment of expert performance and tissue behaviors to interpret quantifiable metrics

SURGICAL TOOLS
Reduced cost and waste from disposable surgical tools by producing as-required tools Tools produced to date are inferior to existing alternatives in function and suitability for sterilization

PATIENT-SPECIFIC PROSTHESES AND BIOPRINTING
Future prospect of autologous grafts for urethra, corpora and even the kidney with successful generation of tissues exhibiting similar mechanical and physiological properties Very early stages of research.
reviews (74, 75) provide a comprehensive exploration of published applications with expanded analyses of setup and manufacturing costs. However, we provide a more contemporary literature search and as a mini-review, a rapidly accessible outline of both the promising existing evidence and shortcomings in linking this to an improved service; Does elevating patient understanding reduce anxiety and litigation or improve recovery? Can inspection of or simulation with a patient-specific model improve surgical outcomes? Does simulation training with tangible facsimiles improve surgical training and with superiority to existing simulation models? These answers remain unproven and should underpin future research. Regardless of how many prove durably useful, the scope of novel developments within urology showcase the freedom to innovate and rapidly prototype afforded by 3D printing. Whilst bioprinting promises major potential advances in patient-specific grafts in the future, we believe rapid prototyping coupled with network sharing and development of ideas is why 3D-printing will remain pivotal in ongoing surgical innovation. Table 6 provides an overview of all reported clinical applications and evidence limitations in the reviewed articles.

AUTHOR CONTRIBUTIONS
DM read the referenced articles and wrote the first draft of the manuscript. AB and ML reviewed and edited the manuscript and approved the final version of the manuscript.