Closed-Circuit Dual-Port Injector System for Fully Automated Contrast Delivery in Diagnostic Cerebral Angiography
Article information
Abstract
Purpose
During diagnostic cerebral angiography, catheter navigation requires manual contrast “puff” injections, while subsequent 2D/3D runs often use automated power injectors. Using power injectors for navigation puffs has not been described. We present a closed-circuit dual-port injector system (Nemoto Press Duo Elite) that integrates both navigation puff delivery and diagnostic run injection into a single automated platform, eliminating all manual tableside contrast handling. A foot pedal interface enables operator-controlled puff timing, potentially reducing contrast waste and air embolism risk while improving single-operator ergonomics with future remote robotic implications.
Materials and Methods
This retrospective comparative cohort study compared 19 consecutive patients undergoing diagnostic cerebral angiography with foot pedal-controlled puff injections (June–July 2023) to 19 historical controls using manual hand injections (May 2021). Both groups used 90% contrast concentration. Fluoroscopy time, radiation dose, contrast utilization, and safety outcomes were compared.
Results
Groups were demographically matched (mean age 52.1±14.2 vs. 50.2±12.9 years; 73.7% female). All 38 procedures achieved diagnostic adequacy with no complications. The foot pedal group demonstrated significantly shorter procedure time per vessel (11.5±4.4 vs. 18.9±10.5 min/vessel, P=0.010) with no significant differences in fluoroscopy time (P=0.171), radiation dose (P=0.690), or contrast delivered (88.7±30.9 vs. 88.2±42.5 mL, P=0.966). A trend toward improved contrast efficiency was observed (23.4±9.4 vs. 27.4±10.4 mL/vessel, P=0.226). Despite undergoing significantly more 3D rotational runs (1.3±1.0 vs. 0.6±0.7, P=0.030), the foot pedal group maintained comparable safety metrics, strengthening the non-inferiority findings.
Conclusion
A closed-circuit dual-port injector system integrating automated navigation puff delivery with diagnostic run injection demonstrates non-inferiority to manual injection for diagnostic cerebral angiography, with shorter procedure time per vessel (39% reduction, P=0.010), though interpretation is limited by differences in indication distribution. By eliminating manual tableside contrast handling, this approach enables precise digital contrast accounting and reduces air embolism risk, establishing a foundation for remote and robotic angiography applications.
INTRODUCTION
Tableside operator presence is currently necessary for many steps of neuro-angiography, including catheter and wire manipulation, device deployment, fluoroscopy table positioning, and contrast administration. Automated, hands-free approaches are increasingly emphasized in interventional neuroradiology for ergonomic and workflow benefits [1,2]. Fully automated procedures would require remote control of the angiography table, X-ray system, and contrast injection [1,2]. While commercial solutions exist for table positioning and X-ray control, complete automation of contrast administration has remained a persistent barrier to complete procedural automation.
Automated contrast injectors have become standard for pre-programmed diagnostic angiographic runs, where fixed injection parameters enable reproducible, high-quality subtracted imaging [3,4]. The Nemoto Press Duo Elite (Nemoto Kyorindo Co., Ltd.) is a significant advancement as the first U.S. Food and Drug Administration (FDA)-approved dual-port system, with Port A storing iodinated contrast and Port B providing normal saline for precise digital control of contrast concentration [5]. This capability is particularly valuable when imaging metallic hardware such as stents, coils, and flow diversion devices, where pre-programmed parameters can be adjusted from 10% to 100% contrast immediately depending on artifact density [6]. However, a critical gap remains: during vessel catheterization, operators still require manual tableside hand injections to deliver small-volume “puff” injections for navigation and vessel selection [7,8]. These manual injections result in substantial contrast waste from syringe dead space, residual manifold volumes, and aspiration procedures, while also increasing the risk of air embolism during syringe handling [9,10]. This requirement for manual navigation puffs prevents the realization of fully hands-free contrast administration during procedural workflows.
An injector linked foot pedal addresses this gap by enabling automated delivery of navigation puffs through the injector system. While power injectors are routinely used for diagnostic runs at fixed rates, using them to deliver navigation puffs for vessel catheterization has not been previously described. The pedal delivers operator-controlled, fixed-volume injections on demand, combining the flexibility and timing control of manual hand injections with the consistency and precision of power injection [11,12]. Unlike a hand switch alternative, which would require sterile preparation and negate ergonomic advantages, the foot pedal preserves handsfree control while maintaining sterile field integrity. This method eliminates all manual tableside contrast injections, creating a fully closed system that achieves complete procedural automation and establishes a platform for potential future remote-control capabilities over fiber optic connections [13,14].
This study presents the first clinical experience with foot pedal-controlled puff injections for diagnostic cerebral angiography, focusing on the ergonomics, ease of use, technical performance, and safety of the system. The goal is to establish proof-of-concept for this closed-system approach as a foundation for discussing future remote injection control capabilities [15].
MATERIALS AND METHODS
Study Design
This retrospective comparative cohort study was conducted at a single center. The foot pedal group consisted of 19 consecutive patients who underwent diagnostic cerebral angiography using the foot pedal-controlled puff injection technique (June 14 to July 12, 2023). The control group consisted of 19 consecutive patients who underwent diagnostic cerebral angiography using traditional manual hand injection techniques (May 5 to 26, 2021). All procedures were performed by a single operator. The historical control period was selected to represent standard institutional practice prior to adoption of the foot pedal system.
Equipment Description and Technical Configuration
The Nemoto Press Duo Elite injector system was used (Fig. 1). The system features a digital dilution platform with dual ports: Port A contains iodinated contrast medium, while Port B contains normal saline solution. This configuration enables real-time digital mixing at precisely operator-specified concentration ratios. The system was configured to maintain a 90% contrast dilution ratio (0.9) for all study cases, providing optimal image quality while minimizing iodinated contrast exposure. The digital platform automatically calculates and delivers precise contrast-saline mixture ratios without manual pre-mixing.
Illustration demonstrating setup for a closed-system cerebral angiography employing complete injector-controlled contrast administration with the Nemoto Press Duo Elite injector. (A) The components in the angiography room and control room are demonstrated. The angiography machine components include the biplane fluoroscopy setup and contrast injector with their respective foot pedals. The control room houses the injector control module. The inset shows the injector setup in a closed-system fashion with Y tubing connected to a two-gang manifold (Merit Medical). The heparinized saline infusion runs at baseline, and manifold stopcocks can be modified to handle injector contrast versus manual contrast mode (not employed in the study). (B) The Nemoto Press Duo Elite in-room unit, with separate syringes for contrast and saline and a digital display showing total volume remaining of each. (C) The setup screen for the injector module in the control room, configured for Trace Shot (puff injections). After completing the Air Check and while on the Accept Bolus screen, pressing the foot pedal activates the Trace Shot/puff injection (seen on top right). For this study, the flow rate was typically set at 1 mL/sec and volume at 2 mL each time the foot pedal was pressed. CCA, common carotid artery; CTAP, cone beam CTA; CA, internal carotid artery. CCA, CTAP, and CA are arbitrary labels that can be used to save injection settings for each vessel per user preference; the labels shown here are examples and can be customized as needed.
The foot pedal interface delivers pre-programmed “puff” injections with vessel-specific flow rates and volumes. When activated by operator foot pressure, the pedal immediately triggers delivery of predetermined injection protocols. The system can operate in 2 manifold states: (1) heparin flush state for continuous anticoagulation (baseline state) and (2) injector state for contrast delivery for puffs and subtracted runs (contrast state). This dual-mode capability eliminates the need for manual contrast syringes at the tableside throughout the procedure.
The Trace Shot function, which delivers puff injections each time the foot switch is pressed, is configured via the setup screen and can be used with both ANGIO Protocol and Microcatheter Protocol modes. The programming ranges allow flow rates of 0.1–3.0 mL/sec (0.1 mL/sec increments), volumes of 1.0–6.0 mL (0.1 mL increments), and a fixed pressure limit of 300 psi. The Trace Shot can only be performed when the device is on the Accept Bolus screen, which is displayed after completion of the Air Check. Injection occurs only while the foot switch start button is pressed, with the foot switch connected to the Main Unit via a wired connection. For repeated injections, an injection can be activated via the foot switch after the bolus injection is accepted.
The injector system has a total capacity of 300 mL (150 mL each of contrast and saline in the dual ports). Using 90% dilution, this capacity proved sufficient for all cases in this series; no case ran out of contrast during the procedure. Injection parameters were programmed once at the beginning of each case based on planned vessel selection. The current version uses a wired connection from the foot pedal to the injector console, with the cable long enough to reach tableside. Future iterations will feature wireless pedals to eliminate cable constraints. System setup time was not significantly different from traditional manual injection preparation.
Patient Selection Criteria and Vascular Access Approach
Inclusion criteria were patients scheduled for diagnostic cerebral angiography for standard clinical indications, including aneurysm evaluation, post-treatment follow-up imaging, pulsatile tinnitus workup, arteriovenous malformation assessments, or other cerebrovascular anomaly investigations. Exclusion criteria included documented contrast allergies requiring special preparation protocols, severe renal insufficiency (estimated glomerular filtration rate <30 mL/min/1.73m2), and emergency cases where standard injection protocols could not be safely followed.
Transradial approach was standard unless contraindicated. All vascular access procedures followed standard institutional protocols for site selection, preparation, and catheter insertion techniques. Standard biplane digital subtraction angiography (DSA) was performed for all vessels. Three-dimensional rotational angiography was added when clinically indicated for detailed vessel characterization.
The only manual contrast injection required during the entire procedure was a single 5 mL injection through the vascular access site at procedure initiation (angiographic runthrough). This volume should be added to the actual injector contrast numbers when calculating total contrast exposure. All subsequent injections, including both navigation puffs and diagnostic runs, were delivered exclusively via the foot pedal-controlled injector system.
Injection Protocols and Parameter Specifications
Vessel-specific injection parameters were pre-programmed into the Nemoto system based on established angiographic standards and institutional protocols. For diagnostic subtraction runs, the following parameters were used: common carotid artery (CCA) at 7 mL/sec for 11 mL total volume, internal carotid artery (ICA) at 6 mL/sec for 9 mL volume, external carotid artery (ECA) at 4 mL/sec for 4 mL volume, vertebral artery (VA) at 6 mL/sec for 6 mL volume, and subclavian artery (SA) at 10 mL/sec for 15 mL volume.
For navigation and vessel catheterization, a modified puff protocol delivered 2 mL of contrast over 2 seconds. This injection rate and volume provided sufficient visualization to safely catheterize vessels while minimizing contrast exposure. In cases requiring CCA roadmaps for selective ICA or ECA catheterization, a 6 mL injection at 6 mL/sec (6/6 modified puff mode) was used to provide adequate opacification for guidewire navigation.
The standardized operational technique sequence involved: (1) initial catheter positioning in the target vessel using conventional fluoroscopic guidance with navigation puffs as needed, (2) confirmation of optimal catheter position, and (3) activation of the coupled biplane fluoroscopy foot pedal to deliver the angiographic injection and complete acquisition of subtracted angiographic images. The subtracted diagnostic runs were completed as usual through the subtraction fluoroscopy pedal, which was coupled and timed to the injector delivery. The foot pedal was operated by the operator’s left foot when the right foot controlled the X-ray pedal, or by the right foot when an assistant or technologist operated the radiation pedal.
The foot pedal-controlled puff system and the conventional injector-angio system served distinct roles at different procedural stages. At vascular access initiation, a single 5 mL manual hand injection was performed (the only manual injection in the entire procedure). During catheter navigation, all puff injections were delivered via the foot pedal-controlled Trace Shot function (2 mL at 1 mL/sec). For common carotid roadmap acquisition, a modified puff protocol (6 mL at 6 mL/sec) was used. During diagnostic runs (2D DSA and 3D rotational angiography), the conventional coupled injector-angio system delivered pre-programmed vessel-specific parameters via the fluoroscopy foot pedal.
Data Collection Methodology and Statistical Analysis
Parameters prospectively collected included patient demographics, procedural access sites, total case times, fluoroscopy exposure times, total radiation doses, number of vessels selected, total contrast volumes, puff injection counts, diagnostic adequacy, and operator satisfaction. Diagnostic adequacy was defined as achievement of sufficient image quality to answer the clinical question prompting the angiographic study. Operator satisfaction was assessed using a standardized 3-point scale (1=poor, 2=adequate, 3=excellent).
Contrast measurements, a key aspect of this closed system, differed methodologically between groups. In the control group, contrast use was estimated by technologists at the end of each case based on contrast bottles opened, tableside manual contrast syringes used, and total power injector runs completed, which is standard practice in angiographic suites. In the foot pedal group, actual contrast volumes were recorded directly from the digital display on the injector console. Any procedural complications, technical system failures, or device malfunctions were also recorded.
Descriptive statistics were calculated using R software. Continuous variables are presented as mean±standard deviation with ranges, and categorical variables as counts and percentages.
RESULTS
The study included 38 patients: 19 in the foot pedal group and 19 in the control group. Groups were well-matched for baseline characteristics. Mean age was 52.1±14.2 years (range 22–74) in the foot pedal group versus 50.2±12.9 years (range 26–66) in controls (P=0.661). Both groups had identical sex distribution with 14 females (73.7%) and 5 males (26.3%) (P>0.999). Race/ethnicity was similar between groups. Clinical indications differed significantly between groups, with pulsatile tinnitus more common in the foot pedal group (36.8% vs. 5.3%) and subarachnoid hemorrhage/hemorrhage workup more common in controls (47.4% vs. 5.3%, P=0.035) (Table 1).
Patient demographics, baseline characteristics, and procedural parameters
Transradial access was utilized in 17 cases (89.5%) in the foot pedal group versus 18 cases (94.7%) in controls (P>0.999). Two patients in the control group required access site modifications due to radial artery tortuosity or inadequate blood return; no access site modifications were needed in the foot pedal group (P=0.486) (Table 2).
Contrast utilization, injection parameters, and safety and efficacy outcomes
Procedural metrics were comparable between groups. Setup time (in room to procedure start) was similar between groups (40.1±13.8 vs. 38.9±16.6 minutes, P=0.820). Mean fluoroscopy time was 10.8±4.5 minutes (range 4.1–18.8) in the foot pedal group versus 13.7±7.6 minutes (range 2.2–26.6) in controls (P=0.171). Mean total radiation dose was 1,071.7±403.5 mGy·cm (range 510.1–2,180.4) versus 1,157.6±839.1 mGy·cm (range 125.0–3,281.7) (P=0.690). Vessel selection was 4.2±1.4 (range 1–6) in the foot pedal group versus 3.4±1.5 (range 1–6) in controls (P=0.123).
Procedure duration was 48.3±21.7 minutes in the foot pedal group versus 56.7±21.6 minutes in controls (P=0.253). When normalized to number of vessels, procedure time per vessel was significantly shorter in the foot pedal group (11.5±4.4 vs. 18.9±10.5 min/vessel, P=0.010), representing a 39% improvement in procedural efficiency. Three-dimensional rotational angiography was performed in 15 foot pedal cases (78.9%) and 10 control cases (52.6%, P=0.170). The foot pedal group underwent significantly more 3D runs per case (1.3±1.0 vs. 0.6±0.7, P=0.030) (Table 1).
The closed-system contrast delivery protocol demonstrated perfect adherence across all cases. There was 100% adherence to the intended 90% contrast dilution ratio, with no deviations or manual adjustments required. The standardized puff injection protocol delivered 2–3 puff injections per vessel for diagnostic runs. Total puff injection analysis (including both navigation and diagnostic puffs) displayed a mean of 8.3±2.8 puff injections per case (range 2–12), directly correlating with the number of vessels selected for evaluation. The automated injector system displayed a mean digital volume output of 93.4±27.6 mL per case (range 46–138 mL) (Table 2).
Contrast utilization was comparable between groups. In the foot pedal group, total contrast mixture used was 98.5±34.4 mL (range 42–150), injector-recorded contrast was 93.4±27.6 mL (range 46–138), and actual contrast delivered accounting dilution was 88.7±30.9 mL (range 38–135). Actual contrast delivered in controls was 88.2±42.5 mL (range 25–160) (P=0.966). Contrast efficiency showed a trend favoring the foot pedal group: 23.4±9.4 mL per vessel versus 27.4±10.4 mL per vessel in controls (P=0.226), representing a 15% reduction that did not reach statistical significance (Table 2).
Safety and efficacy outcomes were excellent in both groups. All 38 cases (100%) achieved diagnostic adequacy. Neither group experienced procedural complications or technical failures. In the foot pedal group, no device malfunctions or air bubble introductions occurred. Two patients in the control group required access site modifications (10.5%) compared to none in the foot pedal group (P=0.486). Operator satisfaction was rated 3/3 in all cases in both groups (Table 2). Operator adaptation to the foot pedal technique occurred rapidly, within 1 to 2 cases based on anecdotal feedback, demonstrating the intuitive nature of the system.
DISCUSSION
This study demonstrates that closed-system contrast delivery using foot pedal-controlled puff injections is non-inferior to traditional manual injection for diagnostic cerebral angiography. These findings are strengthened by the observation that the foot pedal group underwent significantly more 3D rotational angiography runs yet maintained comparable safety and radiation metrics, suggesting the true benefit may be underestimated. The significantly shorter procedure time per vessel (39% reduction, P=0.010) likely reflects streamlined hands-free workflow, though this must be interpreted cautiously given the imbalance in clinical indications between groups. The trend toward lower contrast per vessel (15% reduction, P=0.226), if confirmed in larger studies, could benefit patients with renal insufficiency or contrast allergies [16,17].
Technical Performance and Workflow
The technique offers several practical advantages [18]. Accurate contrast accounting revealed that actual dilution-corrected contrast (88.7±30.9 mL) was approximately 5 mL less than the injector-recorded digital volume (93.4±27.6 mL), consistent with the 90% dilution ratio applied to the total contrast mixture (98.5×0.9≈88.7 mL). Notably, the 5 mL difference approximates the access site contrast injection volume, suggesting that the closed-system digital measurement closely accounts for total contrast delivered. This finding aligns with recent studies emphasizing the importance of contrast media optimization in interventional procedures. It is important to note, though, that technician contrast estimates are subjective and may vary based on individual experience.
The closed system eliminates air bubble introduction from syringe filling and reduces blood loss from repeated aspiration cycles [9,10,19]. By maintaining a sealed pathway for contrast delivery from the injector directly to the patient, the closed system approach entirely removes this risk. Traditional techniques, particularly when using double-flush methods without continuous heparin lines, require repeated 5–10 mL blood aspirations to confirm catheter position. Over multi-vessel diagnostic angiograms, this cumulative loss can be clinically significant in smaller or anemic patients. The dual-mode manifold capability, which allows toggling between heparin flush state for catheter patency and injector/puff state for contrast delivery, eliminates this blood loss entirely.
Ergonomically, the foot pedal leaves the operator’s hands free for catheter manipulation and table positioning [20]. Unlike hand-operated injection controls, the foot pedal does not require sterility since it remains outside the sterile field. This approach addresses critical occupational radiation exposure concerns, as interventional radiologists face significant radiation exposure risks during prolonged procedures [20-22].
Because fluoroscopy is already foot-controlled, dual-foot operation with a second pedal for contrast injection may be ergonomically challenging and could limit operator independence during complex navigation. A finger-based hand switch alternative could offer more intuitive control but would require sterile preparation, negate hands-free advantages, and occupy a hand needed for catheter manipulation. In practice, the foot pedal and fluoroscopy pedal were not typically activated simultaneously; puff injections occurred during brief pauses in navigation, while fluoroscopy was coupled to diagnostic runs. Future iterations could explore wireless hand switches with sterile sheaths, voice-activated commands, or dual-function pedals with lateral rocker activation.
Closed-Circuit Dual-Port Injector System and Navigation Puff Delivery
The central innovation is the closed-circuit system enabled by the dual-port automatic injector (Nemoto Press Duo Elite), which eliminates all manual tableside contrast handling by integrating navigation puff delivery and diagnostic run injection into a single platform. A key component is the extension of power injector puff delivery to vessel navigation, not just diagnostic runs. Traditional angiographic practice reserves power injectors for subtracted 2D angiographic injections and/or 3D angiographic injection while relying on manual hand syringes for catheterization. Our navigation puff protocol (2 mL over 2 seconds) provided sufficient visualization for safe vessel selection across all cases (Fig. 2), with the 6/6 modified puff mode for common carotid roadmaps demonstrating system versatility. This establishes that manual tableside contrast syringes are unnecessary for diagnostic angiography when using programmable puff parameters.
Comparison of right common carotid artery roadmap images from two representative cases. (A) Conventional hand-injected roadmap. (B) Foot pedal-controlled puff roadmap (2 mL at 1 mL/sec). Both techniques demonstrate comparable vessel opacification adequate for catheter navigation and vessel selection.
The standardized vessel-specific injection protocols (CCA 7/11, ICA 6/9, ECA 4/4, VA 6/6, SA 10/15) ensured consistent, reproducible contrast delivery that eliminates operator-dependent variability in injection force, volume, and timing [12]. The mean of 8.3±2.8 total puff injections per case (including both navigation and diagnostic puffs) directly correlated with the 4.2±1.4 vessels selected, demonstrating protocol adherence and predictable contrast usage patterns.
The 300 mL total system capacity (150 mL per port) proved adequate for all diagnostic cases in this series.
Implications for Remote Operation
This work addresses one component of remote robotic angiography by demonstrating operator-controlled injection timing from a distance. Current robotic platforms provide remote table positioning and X-ray control, but contrast injection has required tableside presence [13]. The foot pedal puff capability designed into this platform (though currently implemented as a wired connection with sufficient cable length to reach tableside) could enable transmission of control signals over fiber optic networks, allowing operators to control all angiographic aspects from radiation-shielded locations.
Beyond radiation safety for high-volume operators, remote capabilities could enable remote neuroangiography. Expert neurointerventionalists could guide procedures in rural or underserved hospitals by remotely controlling imaging parameters and contrast injection timing while local operators perform catheter navigation. This preserves bedside expertise for immediate complication response while enabling specialized decision-making from remote locations, potentially expanding access to subspecialty neurointerventional care.
However, reliable robotic catheter manipulation, network latency management, fail-safe mechanisms, and regulatory approval are parallel challenges that are being investigated [1]. Complete remote operation remains a future goal rather than an immediate capability [23].
Limitations and Future Directions
This study has several constraints. The sample size of 19 patients per group from a single institution limits statistical power and generalizability, and the study may be underpowered to detect clinically meaningful differences in contrast utilization. All cases were diagnostic procedures performed by an experienced operator; therapeutic angiography (coiling, thrombectomy, embolization) involves larger contrast volumes, longer procedures, and more complex navigation that may present different challenges.
The use of historical controls rather than a randomized design introduces potential temporal confounding, though both groups were treated by the same operator using similar protocols. The imbalance in clinical indication distribution and the higher rate of 3D rotational angiography in the foot pedal group are important confounders for the efficiency, contrast, and radiation comparisons. A prospective randomized trial with stratification by indication, or propensity score matching, would provide stronger evidence and clarify whether the observed trends reflect true benefits of the closed-system approach.
Several practical considerations also warrant further study. The system requires capital investment in the Nemoto Press Duo Elite platform, and cost-effectiveness has not been evaluated; savings from reduced contrast wastage and workflow efficiency must be weighed against equipment, maintenance, and training costs. Specific technical challenges include the need for control room personnel to adjust injection parameters when switching between navigation and diagnostic run settings—a tableside control interface would address this. The wired pedal connection, though long enough to reach tableside, occasionally proved obtrusive during operator movement. Setup time was not formally quantified, but anecdotally, programming vessel-specific parameters took similar time to preparing multiple contrast syringes in the traditional workflow. Long-term hardware durability over thousands of pedal activations remains to be established, and the platform’s wireless transmission capability has not been tested for the latency, reliability, or fail-safe behavior required for remote operation.
Despite these limitations, this study demonstrates the feasibility and safety of a novel closed-system approach to diagnostic cerebral angiography. Its immediate clinical value lies in accurate digital contrast accounting, reduced air embolism risk, improved single-operator ergonomics, and streamlined workflow, with additional implications for future remote and robotic procedures.
CONCLUSION
The closed-circuit dual-port injector system presented here achieves non-inferiority to traditional manual injection for diagnostic cerebral angiography, with significantly shorter procedure time per vessel (39% reduction, P=0.010). By integrating automated navigation puff delivery and diagnostic run injection into a single platform—with foot pedal–triggered puff control eliminating all manual tableside contrast handling—this system enables precise digital contrast accounting and reduces air embolism risk. This closed-circuit approach establishes proof-of-concept for remote injection control as a component of future robotic endovascular applications.
Notes
Acknowledgments
The authors thank the clinical and research staff for their assistance with data collection and patient care. The research presented in this manuscript was conducted during Dr. Choudhri's faculty appointment at the University of Pennsylvania.
Fund
None.
Ethics Statement
This study was approved by the University of Pennsylvania Institutional Review Board (IRB) (protocol 844179). The requirement for informed consent was waived by the IRB given the retrospective nature of the study and minimal risk to participants. We anonymized patient information that could identify an individual.
Conflicts of Interest
Dr. Omar Choudhri serves as a consultant for Nemoto Kyorindo Co., Ltd. The other authors have no conflicts to disclose.
Author Contributions
Concept and design: MR and OC. Analysis and interpretation: OG. Data collection: OG, SD, NY, and JK. Writing the article: MR, OG, and SA. Critical revision of the article: MR, OG, SA, SD, NY, WA, JK, GL, LB, and OC. Final approval of the article: OC. Statistical analysis: OG. Obtained funding: none. Overall responsibility: OC.