Support vector machines improve the accuracy of evaluation for the performance of laparoscopic training tasks.

BACKGROUND: Despite technological advances in the tracking of surgical motions, automatic evaluation of laparoscopic skills remains remote. A new method is proposed that combines multiple discrete motion analysis metrics. This new method is compared with previously proposed metric combination methods and shown to provide greater ability for classifying novice and expert surgeons. METHODS: For this study, 30 participants (four experts and 26 novices) performed 696 trials of three training tasks: peg transfer, pass rope, and cap needle. Instrument motions were recorded and reduced to four metrics. Three methods of combining metrics into a prediction of surgical competency (summed-ratios, z-score normalization, and support vector machine [SVM]) were compared. The comparison was based on the area under the receiver operating characteristic curve (AUC) and the predictive accuracy with a previously unseen validation data set. RESULTS: For all three tasks, the SVM method was superior in terms of both AUC and predictive accuracy with the validation set. The SVM method resulted in AUCs of 0.968, 0.952, and 0.970 for the three tasks compared respectively with 0.958, 0.899, and 0.884 for the next best method (weighted z-normalization). The SVM method correctly predicted 93.7, 91.3, and 90.0% of the subjects' competencies, whereas the weighted z-normalization respectively predicted 86.6, 79.3, and 75.7% accurately (p < 0.002). CONCLUSIONS: The findings show that an SVM-based analysis provides more accurate predictions of competency at laparoscopic training tasks than previous analysis techniques. An SVM approach to competency evaluation should be considered for computerized laparoscopic performance evaluation systems.

Connecting multiple low-flow intravenous infusions in the newborn: problems and possible solutions.

OBJECTIVE: To compare the efficiency of a stopcock system and a newly designed device to titrate low-flow infusions to critically ill infants. DESIGN: In vitro study comparing fluid delivery through readily available and custom-built equipment. SETTING: Engineering laboratory of a university-affiliated hospital. PATIENTS/SUBJECTS: None. INTERVENTIONS: Two infusion pumps were used to deliver five 0.1-mL/hr infusions and an 8-mL/hr carrier through a conventional stopcock system. A diluted dye was used to quantize the amount of any given infusion reaching the patient. Samples were collected over a 90-min period and analyzed by spectophotometry. Thirty minutes into each trial, the red dye infusion rate was doubled. A multiple infusion connector was manufactured and tested in place of the stopcock system. Without a stopcock system or multiple infusion connector, both an infusion and a syringe pump were tested for accuracy of delivery of 0.1 mL/hr and 0.2 mL/hr of dye infusions. MEASUREMENTS AND MAIN RESULTS: The infusion pump was more accurate than a 60-mL syringe pump in generating infusion rates of both 0.1 mL/hr and 0.2 mL/hr. When delivering a 0.1-mL/hr dye infusion through the distal port of an array of six stopcocks, the actual delivery of a diluted red dye infusion took approximately 32 mins to double after being increased from 0.1 mL/hr to 0.2 mL/hr. When using the multiple infusion connector to connect the low-volume drips together, the same change in flow rate caused the actual delivery of dye to double in <8 mins (a result comparable to the data from the proximal port of a stopcock). This result was independent of which port was used for dye delivery. Streaming of the red dye within the stopcock system was observed. CONCLUSIONS: When using the conventional stopcock array to titrate low-rate infusions, significant delays were observed. A device designed specifically for infusions in infants may substantially improve this system.

Thin-film nitinol (NiTi): a feasibility study for a novel aortic stent graft material.

OBJECTIVE: Although technological improvements continue to advance the designs of aortic stent grafts, miniaturization of the required delivery systems would allow their application to be available to a wider range of patients and potentially decrease the access difficulties that are encountered. We performed this feasibility study to determine if thin-film NiTi (Nitinol) could be used as a covering for stent grafts ranging from 16 mm to 40 mm in diameter. Specifically, we wished to determine the profile reduction attainable and improve the flexibility of our design. METHODS: Using a novel hot-sputter deposition technique, we created sheets of thin-film NiTi (TFN) with a tensile strength of >500 Megapascal (MPa) and thickness of 5-10 microns. TFN was used to cover stents, which were then deployed in vitro. Patterned thin film was fabricated via a lift-off technique; grafts were constructed with stents ranging from 16-40 mm and deployed in a pulsatile flow system from the smallest diameter polymer tubing into which the stent and TFN would fit. The bending/stiffness ratio vs similar sized expanded polytetrafluoroethylene (ePTFE)-covered stents was also determined. RESULTS: TFN was created in both non-patterned and patterned forms, with a tensile strength of >100 MPa for the latter. We created devices that were successfully deployed via delivery systems half the size of fabric-covered stent grafts (ie, the 16 mm stent graft that originally was delivered via a 16French (F) system was reduced to 8F, and the 40 mm stent graft delivered via a 24F system was reduced to 12F). No migration of the devices was observed with deployment in both straight and curved tubing, which was sized so that the stent grafts were oversized by 20%. Both forms of the thin-film were noted to be more flexible than the same sized ePTFE stent graft, and the patterned graft had an additional 15-30% flexibility vs the non-patterned film. CONCLUSION: These in vitro results demonstrate the feasibility of TFN for covering stent grafts designed for placement in the aorta. The delivery profile can be significantly reduced across a wide range of sizes, while the material remained more flexible than ePTFE.

Creation of transcatheter aortopulmonary and cavopulmonary shunts using magnetic catheters: feasibility study in swine.

Surgical shunts are the basic form of palliation for many types of congenital heart disease. The Glenn shunt (superior cavopulmonary connection) and central shunt (aortopulmonary connection) represent surgical interventions that could potentially be accomplished by transcatheter techniques. We sought to investigate the efficacy of using neodymium iron boron (NdFeB) magnetic catheters to create transcatheter cavopulmonary and aortopulmonary shunts. NdFeB magnets were machined and integrated into catheters. "Target" catheters were placed in the pulmonary arteries (PAs), and radiofrequency "perforation" catheters were placed in either the descending aorta (DAo) for central shunts or the superior vena cava (SVC) for Glenn shunts. The magnet technique or "balloon target" method was used to pass wires from the DAo or the SVC into the PA. Aortopulmonary and cavopulmonary connections were then created using Atrium iCAST covered stents. Magnet catheters were used to perforate the left pulmonary artery from the DAo, thereby establishing a transcatheter central shunt. Given the orientation of the vasculature, magnetic catheters could not be used for SVC-to-PA connections; however, perforation from the SVC to the right pulmonary artery was accomplished with a trans-septal needle and balloon target. Transcatheter Glenn or central shunts were successfully created in four swine.

Voltage-induced strain clocking of nanomagnets with perpendicular magnetic anisotropies.

Nanomagnetic logic (NML) has attracted attention during the last two decades due to its promise of high energy efficiency combined with non-volatility. Data transmission in NML relies on Bennett clocking through dipole interaction between neighboring nanomagnetic bits. This paper uses a fully coupled finite element model to simulate Bennett clocking based on strain-mediated multiferroic system for Ni, CoFeB and Terfenol-D with perpendicular magnetic anisotropies. Simulation results demonstrate that Terfenol-D system has the highest energy efficiency, which is 2 orders of magnitude more efficient than Ni and CoFeB. However, the high efficiency is associated with switching incoherency due to its large magnetostriction coefficient. It is also suggested that the CoFeB clocking system has lower bit-density than in Ni or Terfenol-D systems due to its large dipole coupling. Moreover, we demonstrate that the precessional perpendicular switching and the Bennett clocking can be achieved using the same strain-mediated multiferroic architecture with different voltage pulsing. This study opens new possibilities to an all-spin in-memory computing system.

Use of a covered stent modification to produce a transcatheter valve: laboratory and animal testing.

Stent-based transcatheter valves continue to require large sheaths inappropriate for deployment in children. Low-profile covered stent valves (CSVs) were constructed by removing triangular sections from two sides of partially expanded Palmaz P308 stents before covering the stents with 0.1 mm polytetrafluoroethylene. Valves were carefully crimped onto balloon catheters and deployed in a pulsatile flow loop. With fixed afterload and pump output, flow, degree of stenosis, effect on pulse pressure, and ease of deployment were determined for each valve. In vivo transcatheter feasibility studies were then performed by disabling the aortic valve of two 25-kg pigs, and deploying transcatheter CSVs into their descending aorta. All transcatheter valves deployed consistently via sheaths three French sizes larger than the recommended sheath for their balloon and none created significant obstruction. With the bicuspid and supravalvar CSVs, the flow was 64% and 79% (respectively) of a commercially available valve. Angiograms revealed excellent acute CSV function after deployment with only mild regurgitation and without significant obstruction to systolic flow. Although long-term testing is required, a modified CSV design may have utility in low-profile pediatric transcatheter valve replacement.

Thin film nitinol covered stents: design and animal testing.

Interventionalists in many specialties have the need for improved, low profile covered stents. Thin films of nitinol (<5-10 microns) could be used to improve current covered stent technology. A "hot target" sputter deposition technique was used to create thin films of nitinol for this study. Covered stents were created from commercially available balloon-inflatable and self-expanding stents. Stents were deployed in a laboratory flow loop and in four swine. Uncovered stent portions served as controls. Postmortem examinations were performed 2-6 weeks after implantation. In short-term testing, thin film nitinol covered stents deployed in the arterial circulation showed no intimal proliferation and were easily removed from the arterial wall postmortem. Scanning electron microscopy showed a thin layer of endothelial cells on the thin film, which covered the entire film by 3 weeks. By contrast, significant neointimal hyperplasia occurred on the luminal side of stents deployed in the venous circulation. Extremely low-profile covered stents can be manufactured using thin films of nitinol. Although long-term studies are needed, thin film nitinol may allow for the development of low-profile, nonthrombogenic covered stents.