Objective:To explore the feasibility and consistency of a new digital classification system of pelvic fractures named as JST classification based on close reduction techniques.Methods:A retrospective collection was conducted of the data from the 63 patients with pelvic fracture who had undergone surgical treatment after JST classification at Department of Orthopaedics and Traumatology, Beijing Jishuitan Hospital from March 2021 to March 2023. Digital classification of the pelvic fractures was performed based on their locations and displacements. The classification first divides the pelvis into 4 parts: left half pelvis and right half pelvis; sacral Denis Ⅲ area and pubic symphysis. The symmetrical left and right sacral Denis Ⅰ and Denis Ⅱ areas are also included in the left/right half pelvis. Subsequently, the left half pelvis and right half pelvis are divided into 4 regions and marked by capitalized English letters: Sacrum Area (including Denis Ⅰ and Denis Ⅱ, denoted as S), Sacroiliac Joint Area (denoted as J), Iliac Area (denoted as I), and Pubic Area (denoted as P); to distinguish right/left, R and L are used as prefixes. The 2 asymmetric parts are also marked with English letters: Denis Ⅲ area of the sacrum (denoted as Sac), and pubic symphysis (denoted as C). Afterwards, the fracture line morphology and displacement in each region are marked digitally to form a complete JST classification system. The inter- and intra-observer reliabilities (Fleiss' and Cohen's Kappa) of the JST classification system were tested by 3 observers with more than 10 years of experience in pelvic fracture treatment.Results:Consistency analysis of the JST classification results showed that the mean κ value of the intra-observer reliability was 0.818 (from 0.658 to 0.946, P<0.001) and the inter-observer reliability 0.873 (from 0.674 to 1.000, P<0.001), both indicating excellent agreement. Of the 63 patients, 59 obtained successful closed reduction with the assistance of the Rossum Robot R-Universal intelligent orthopedic surgical robot system after fracture classification by the JST system, yielding a success rate of 93.7% (59/63). Conclusions:The new JST classification system for pelvic fractures demonstrates strong intra and inter-observer reliabilities compared with traditional classification systems. As JST classification system labels each fracture site and key bones, it is of great significance for the deep learning and intraoperative operations of intelligent fracture robots.

Objective:To elucidate the recent therapeutic efficacy of the intelligent fracture reduction robotic system in managing pelvic fractures.Methods:A retrospective evaluation of 49 pelvic fracture patients treated using the intelligent fracture reduction robotic system at Beijing Jishuitan Hospital's trauma orthopedics department between March 2021 and December 2022 was conducted. The cohort included 30 males and 19 females, with a mean age of 51.51±18.71 years (20-92 years range). Fractures were classified according to the Tile system: B1 type in 2 cases, B2 in 7, B3 in 3, C1 in 30, and C2 in 3. The median interval between injury and surgery was 6 days, with a range of 2-22 days. The robotic system assisted in pelvic fracture reduction and stabilization surgeries. Preoperative and postoperative evaluations involved pelvic CT scans, anteroposterior, inlet, and outlet radiographic images. Fracture displacement and reduction outcomes were assessed via X-ray imagery. Data captured included intraoperative blood loss, duration of surgery, fracture stabilization techniques, and postoperative monitoring period. The Majeed scoring system gauged functional outcomes.Results:Of the patients, 48 underwent minimally invasive interventions with robotic assistance, while one case necessitated open reduction and internal fixation due to an unsuccessful reduction. The duration between injury and operation ranged from 2 to 22 days. Average surgical time stood at 206.5±7.1 minutes (105-440 min range), and median intraoperative blood loss was 100ml (10-600 ml range). Using the Matta reduction criteria, 30 postoperative cases exhibited excellent and 9 good outcomes for posterior pelvic ring displacement, translating to a 93% (38/41) positive rate. For anterior pelvic ring shifts, 45 showed excellent and 3 good outcomes, culminating in a 100% (48/48) success rate. Follow-up for the 48 cases lasted 11.0 months (3-23 months range), with the Majeed functional score averaging 81.9±17.0 points (42-100 point range). 27 cases scored excellent, and 11 good, yielding a combined positive outcome rate of 79.2% (38/48).Conclusion:Employing the intelligent fracture reduction robotic system in pelvic fracture treatments facilitates minimally invasive interventions and yields favorable short-term clinical results.

Objective:To evaluate the emergency iliosacral screw fixation assisted by TiRobot for unstable posterior pelvic ring fracture.Methods:The 26 patients with unstable pelvic fracture were analyzed retrospectively who had undergone emergency iliosacral screw fixation at Department of Orthopedics & Traumatology, Beijing Jishuitan Hospital from June 2018 to December 2020. They were divided into 2 groups depending on whether orthopaedic TiRobot was used to assist screw insertion. In the observation group of 14 cases subjected to TiRobot-assisted insertion of iliosacral screws, there were 10 males and 4 females with an age of (45.9 ± 10.1) years; in the control group of 12 cases subjected to conventional manual insertion of iliosacral screws, there were 9 males and 3 females with an age of (49.2 ± 11.3) years. All the surgeries were conducted within 24 hours after injury. The 2 groups were compared in terms of screw insertion time, pin insertion, intraoperative blood loss, fluoroscopy time, postoperative screw position, fracture reduction and Harris hip score at the final follow-up.Results:The 2 groups were comparable because there was no significant difference between them in their preoperative general clinical data or follow-up time ( P>0.05). The screw insertion time [(16.1 ± 3.4) min] and fluoroscopy time [(8.1 ± 3.3) s] in the observation group were significantly shorter than those in the control group [(26.4 ± 5.4) min and (25.2 ± 7.4) s], and the pin insertions [1 (1, 2) times] and intraoperative blood loss [(10.5 ± 6.4) mL] in the former were significantly less than those in the latter [6 (3, 8) times and (24.8 ± 6.7) mL] (all P<0.05). Postoperatively, the sacroiliac screw position was excellent in 18 cases and good in 2 in the observation group while excellent in 14 cases, good in 2 and poor in 2 in the control group; the fracture reduction was excellent in 12 cases, good in one and fair in one in the observation group while excellent in 10 cases, good in one and fair in one in the control group, showing insignificant differences in the above comparisons ( P>0.05). There was no significant difference either in the Harris hip score at the final follow-up between the 2 groups ( P>0.05). Conclusion:Compared with conventional manual insertion of iliosacral screws, emergency iliosacral screw fixation assisted by TiRobot can effectively decrease surgical time and reduce operative invasion due to a higher accuracy rate of screw insertion.

Objective:This study aims to compare the application value of Autoregressive Integrated Moving Average model (ARIMA ) and deep learning model inprediction and analysis of liposuction operation data.Methods:The patients who met inclusion criteria and underwent liposuction surgery in the Plastic Surgery Hospital of Chinese Academy of Medical Sciences from January 2019 to September 2019 were enrolled in this study. For each patient, 250~400 s operation data including kinematics and mechanical data were collected by a senior plastic surgeon, usingthe liposuction operation recording system which consists of optical tracking and force sensing equipment. After pretreatment, the collected data were divided into one liposuction reciprocating cycle as one set of data. ARIMA model and deep learning model were used to analyze the collected datafor establishing prediction models of liposuction operation. Using Matlab2017, 30 couplesofliposuction data setwereextractedby simple random sampling, andthe DTW valueofeachcoupleofdatasetswascalculated as test standard. Then, the DTW values between 30 sets of predicted data and actual data based on the ARIMA model and the deep learning model were calculated respectively and compared with the test standard to verify the prediction result of the two models. Matlab2017 was used for statistical analysis. Independent sample t-test was used to compare the two groups, and P<0. 05 indicates that the difference is statistically significant.Results:18 patients were enrolled. All patients were females at 23-49 years old, with the mean age of 36. 6 years old. Liposuction was performed in the abdomen, thighs, and waist. A total of 16 800 sets of liposuction cycle data were obtained. The mean DTW value of test standard was 0. 048±0. 028. The meanDTW value between the ARIMA model predicted data and the actual data was 0. 660±0. 577, which was statistically significant compared with the test standard ( P< 0. 05) . The meanDTW value between the deep learning model predicted data and the actual data was 0. 052±0. 030, which was no significant difference compared with the test standard ( P> 0. 05 ). Conclusions:Compared with ARIMA model, deep learning model can predict liposuction operation data more accurately, and has better adaptability and real-time performance.

Objective:This study aims to compare the applicability value of autoregressive integrated moving average model(ARIMA) and deep learning model inprediction and analysis of liposuction operation data.Methods:The patients who met inclusion criteria and underwent liposuction surgery in the Plastic Surgery Hospital of Chinese Academy of Medical Sciences from January 2019 to September 2019 were enrolled in this study. For each patient, 250-400 s operation data including kinematics and mechanical data were collected by a senior plastic surgeon, using the liposuction operation recording system which consists of optical tracking and force sensing equipment. After pretreatment, the collected data were divided into one liposuction reciprocating cycle as one set of data. ARIMA model and deep learning model were used to analyze the collected data for establishing prediction models of liposuction operation. Using Matlab 2017, 30 couples of liposuction data set were extracted by simple random sampling, and the dynamic time warping (DTW) value of each couple of data sets was calculated as test standard. Then, the DTW values between 30 sets of predicted data and actual data based on the ARIMA model and the deep learning model were calculated respectively and compared with the test standard to verify the prediction results of the two models. Matlab 2017 was used for statistical analysis. Independent sample t-test was used to compare the two groups, and P<0.05 indicated statistically significant difference. Results:Eighteen patients were enrolled. All patients were females at 23-49 years old, with the mean age of 36.6 years old. Liposuction was performed in the abdomen, thighs, and waist. A total of 16 800 sets of liposuction cycle data were obtained. The mean DTW value of test standard was 0.048±0.028. The mean DTW value between the ARIMA model predicted data and the actual data was 0.660±0.577, which was statistically significant compared with the test standard ( P<0.05). The mean DTW value between the deep learning model predicted data and the actual data was 0.052±0.030, which was not significantly different compared to the test standard ( P>0.05). Conclusions:Compared with ARIMA model, deep learning model can predict liposuction operation data more accurately, and has better adaptability and real-time performance.

Objective:To introduce an innovative simulation training system of liposuction and compare the effect of the traditional training method with this system in the liposuction training for medical students.Methods:Thirty medical postgraduates (18 males and 12 females, aged 22 to 30 years) at Peking Union Medical College without liposuction experiences were selected. All the participants were randomly divided into two groups. In the traditional training group, the trainees were trained on the phantom, while the teachers gave explanations and demonstrations. In the simulation training system group, the trainees were trained by themselves on the simulation training system. Before and after the training, the two groups were required to perform a liposuction simulation test on the simulated training system. The resistance of liposuction cannula, the acceleration of liposuction cannula and the uniformity degree of operation of the two groups were recorded, and the differences in the training effects between the two groups were compared. R 3.5 and Python 3.7 were used for analysis. Application of the t test for measurement data was in accordance with normal distribution, and the results were expressed as Mean±SD deviation. Application of Wilcoxon signed-rank test or Wilcoxon rank sum test for the measurement data did not conform to the normal distribution. The results were expressed as M( P25, P75). P< 0.05 indicated statistical differences. Results:After the training, the area of liposuction in the traditional training group was more moderate than that before the training [skewness: -0.22(-0.38, -0.14) vs. -0.07(-0.24, 0.02)( V=20, P=0.022); kurtosis: 2.32(2.09, 2.58) vs. 1.96(1.90, 2.00)( V=112, P=0.002)]. After training, the number of times of lateral resistance[7.0(3.5, 13.5) vs. 0(0, 0)( V=111.5, P=0.004)] and acceleration [7.0(5.0, 17.5) vs. 3.0(2.0, 12.5)( V=102, P=0.002)] over-threshold were significantly reduced, the angle of liposuction coverage [131.18°(117.71°, 137.88°) vs. 169.89°(162.96°, 180.00°)( V=0, P<0.001)] was significantly improved, and the area of liposuction [skewness: -0.17(-0.33, 0.03) vs. -0.01(-0.13, 0.06)( V=21, P=0.026); kurtosis: 2.35(2.08, 2.50) vs. 1.94(1.83, 2.00)( V=118, P<0.001)] was more evenly distributed. The differences before and after training were analyzed between the simulation training system group and the traditional training group. The simulation training system group was superior to the traditional training group in the number of times of lateral resistance[-7.5±7.4 vs.-1.4±9.0 ( t=111.5, P=0.026)], the number of times of acceleration [-3.0(-6.5, -2.0) vs. -1.0(-4.0, 2.0)( W=156.5, P=0.035)] and the angle of coverage[(-40.24±18.88)° vs. (-11.10±25.54)° ( t=3.553, P<0.001)]. Conclusions:Simulation training system is an effective method in liposuction training to enhance the skills of trainees.

Objective:This study aims to compare the application value of Autoregressive Integrated Moving Average model (ARIMA ) and deep learning model inprediction and analysis of liposuction operation data.Methods:The patients who met inclusion criteria and underwent liposuction surgery in the Plastic Surgery Hospital of Chinese Academy of Medical Sciences from January 2019 to September 2019 were enrolled in this study. For each patient, 250~400 s operation data including kinematics and mechanical data were collected by a senior plastic surgeon, usingthe liposuction operation recording system which consists of optical tracking and force sensing equipment. After pretreatment, the collected data were divided into one liposuction reciprocating cycle as one set of data. ARIMA model and deep learning model were used to analyze the collected datafor establishing prediction models of liposuction operation. Using Matlab2017, 30 couplesofliposuction data setwereextractedby simple random sampling, andthe DTW valueofeachcoupleofdatasetswascalculated as test standard. Then, the DTW values between 30 sets of predicted data and actual data based on the ARIMA model and the deep learning model were calculated respectively and compared with the test standard to verify the prediction result of the two models. Matlab2017 was used for statistical analysis. Independent sample t-test was used to compare the two groups, and P<0. 05 indicates that the difference is statistically significant.Results:18 patients were enrolled. All patients were females at 23-49 years old, with the mean age of 36. 6 years old. Liposuction was performed in the abdomen, thighs, and waist. A total of 16 800 sets of liposuction cycle data were obtained. The mean DTW value of test standard was 0. 048±0. 028. The meanDTW value between the ARIMA model predicted data and the actual data was 0. 660±0. 577, which was statistically significant compared with the test standard ( P< 0. 05) . The meanDTW value between the deep learning model predicted data and the actual data was 0. 052±0. 030, which was no significant difference compared with the test standard ( P> 0. 05 ). Conclusions:Compared with ARIMA model, deep learning model can predict liposuction operation data more accurately, and has better adaptability and real-time performance.

Objective:This study aims to compare the applicability value of autoregressive integrated moving average model(ARIMA) and deep learning model inprediction and analysis of liposuction operation data.Methods:The patients who met inclusion criteria and underwent liposuction surgery in the Plastic Surgery Hospital of Chinese Academy of Medical Sciences from January 2019 to September 2019 were enrolled in this study. For each patient, 250-400 s operation data including kinematics and mechanical data were collected by a senior plastic surgeon, using the liposuction operation recording system which consists of optical tracking and force sensing equipment. After pretreatment, the collected data were divided into one liposuction reciprocating cycle as one set of data. ARIMA model and deep learning model were used to analyze the collected data for establishing prediction models of liposuction operation. Using Matlab 2017, 30 couples of liposuction data set were extracted by simple random sampling, and the dynamic time warping (DTW) value of each couple of data sets was calculated as test standard. Then, the DTW values between 30 sets of predicted data and actual data based on the ARIMA model and the deep learning model were calculated respectively and compared with the test standard to verify the prediction results of the two models. Matlab 2017 was used for statistical analysis. Independent sample t-test was used to compare the two groups, and P<0.05 indicated statistically significant difference. Results:Eighteen patients were enrolled. All patients were females at 23-49 years old, with the mean age of 36.6 years old. Liposuction was performed in the abdomen, thighs, and waist. A total of 16 800 sets of liposuction cycle data were obtained. The mean DTW value of test standard was 0.048±0.028. The mean DTW value between the ARIMA model predicted data and the actual data was 0.660±0.577, which was statistically significant compared with the test standard ( P<0.05). The mean DTW value between the deep learning model predicted data and the actual data was 0.052±0.030, which was not significantly different compared to the test standard ( P>0.05). Conclusions:Compared with ARIMA model, deep learning model can predict liposuction operation data more accurately, and has better adaptability and real-time performance.

Objective:To introduce an innovative simulation training system of liposuction and compare the effect of the traditional training method with this system in the liposuction training for medical students.Methods:Thirty medical postgraduates (18 males and 12 females, aged 22 to 30 years) at Peking Union Medical College without liposuction experiences were selected. All the participants were randomly divided into two groups. In the traditional training group, the trainees were trained on the phantom, while the teachers gave explanations and demonstrations. In the simulation training system group, the trainees were trained by themselves on the simulation training system. Before and after the training, the two groups were required to perform a liposuction simulation test on the simulated training system. The resistance of liposuction cannula, the acceleration of liposuction cannula and the uniformity degree of operation of the two groups were recorded, and the differences in the training effects between the two groups were compared. R 3.5 and Python 3.7 were used for analysis. Application of the t test for measurement data was in accordance with normal distribution, and the results were expressed as Mean±SD deviation. Application of Wilcoxon signed-rank test or Wilcoxon rank sum test for the measurement data did not conform to the normal distribution. The results were expressed as M( P25, P75). P< 0.05 indicated statistical differences. Results:After the training, the area of liposuction in the traditional training group was more moderate than that before the training [skewness: -0.22(-0.38, -0.14) vs. -0.07(-0.24, 0.02)( V=20, P=0.022); kurtosis: 2.32(2.09, 2.58) vs. 1.96(1.90, 2.00)( V=112, P=0.002)]. After training, the number of times of lateral resistance[7.0(3.5, 13.5) vs. 0(0, 0)( V=111.5, P=0.004)] and acceleration [7.0(5.0, 17.5) vs. 3.0(2.0, 12.5)( V=102, P=0.002)] over-threshold were significantly reduced, the angle of liposuction coverage [131.18°(117.71°, 137.88°) vs. 169.89°(162.96°, 180.00°)( V=0, P<0.001)] was significantly improved, and the area of liposuction [skewness: -0.17(-0.33, 0.03) vs. -0.01(-0.13, 0.06)( V=21, P=0.026); kurtosis: 2.35(2.08, 2.50) vs. 1.94(1.83, 2.00)( V=118, P<0.001)] was more evenly distributed. The differences before and after training were analyzed between the simulation training system group and the traditional training group. The simulation training system group was superior to the traditional training group in the number of times of lateral resistance[-7.5±7.4 vs.-1.4±9.0 ( t=111.5, P=0.026)], the number of times of acceleration [-3.0(-6.5, -2.0) vs. -1.0(-4.0, 2.0)( W=156.5, P=0.035)] and the angle of coverage[(-40.24±18.88)° vs. (-11.10±25.54)° ( t=3.553, P<0.001)]. Conclusions:Simulation training system is an effective method in liposuction training to enhance the skills of trainees.

Objective:This study aims to discuss the detection effect of a novel liposuction cannula movement recording system based on optical tracking technique and force sensor during liposuction and analyze the relationship between recorded data and liposuction efficiency.Methods:The patients who met inclusion criteria underwent liposuction from January 2019 to September 2019 were enrolled in this study and operated on by two surgeons with extensive liposuction experience. During a given procedure, one surgeon performed liposuction on one side of the targeted liposuction area, another surgeon performed liposuction on the contralateral side. The trajectory and force of liposuction cannula were recorded for approximately 300 seconds, and the volume of upper layer fat in liposuction aspirate was measured during experiment. The average amplitude, average frequency, and average forward resistance of the liposuction needle movement of the two surgeons were compared, and the liposuction efficiency was calculated and compared. The data were analyzed by paired t-test and signed rank sum test. Results:Eighteen patients were enrolled. All patients were females at 23-49 years old, with the mean age of 37 years old. Liposuction was performed in the abdomen, thighs, and waist. The movement amplitude and forward resistance of the liposuction cannula of surgeon A were (11.43±1.23) cm and (9.35±2.24) N, which were higher than those of surgeon B (10.00±2.33) cm and (8.20±3.05) N, and the differences were significant ( t=2.780, P=0.013; t=2.328, P=0.033). The frequency of liposuction cannula movement of surgeon A was (2.14±0.19) Hz, and that of surgeon B was (2.19±0.55) Hz. There was no significant difference between the two surgeons ( t=-0.366, P=0.719). The liposuction efficiency of operator A was (19.20±9.36) ml/min, and the efficiency of operator B was (15.27±8.05) ml/min. A was 3.93 ml/min higher than B. The difference was statistically significant ( t=3.736, P=0.002). Conclusions:The liposuction cannula movement recording system based on optical tracking and force sensing technology established in this study can record and analyze the operation trajectory and resistance of the liposuction cannula. When the movement amplitude and forward resistance of the cannula are larger, the liposuction efficiency is higher.