Background: The aim of this study was to evaluate the safety and effectiveness of robot‐assisted thymectomy (RAT) in large anterior mediastinal tumors (AMTs) (size ≥6 cm) compared with video‐assisted thymectomy (VAT) and open surgery.Methods: A total of 132 patients with large AMTs who underwent surgical resection from January 2016 to June 2022 were included in this study. A total of 61 patients underwent RAT, 36 patients underwent VAT and 35 patients underwent open surgery. Perioperative outcomes were compared.Results: There were no significant differences in tumor size (p = 0.141), or pathological types (p = 0.903). Compared with the open group, the RAT and VAT groups were associated with a shorter operation time (115.00 vs. 160.00, p = 0.012; 122.50 vs. 160.00, p = 0.071), and less blood loss (50.00 vs. 200.00, p < 0.001; 50.00 vs. 200.00, p < 0.001), respectively. The rate of conversion in the RAT group was similar to that in the VAT group (6.56% vs. 13.89%, p = 0.229). Concomitant resection was less frequently performed in the VAT group than in the RAT and open groups (5.56% vs. 31.15%, p = 0.040; 5.56% vs. 31.43%, p = 0.006). VAT patients had a lower drainage volume (365.00 vs. 700.00 and 910.00 mL, p < 0.001), shorter duration of chest tube (2.00 vs. 3.00 and 4.00, p < 0.001), and shorter hospital stay (5.00 vs. 6.00 and 7.00, p < 0.001) than the RAT and open groups. There was no 30‐day mortality in any group. No difference was seen in R0 resection rates (p = 0.846). The postoperative complication rates were similar among the three groups (p = 0.309). Total in‐hospital costs (66493.90 vs. 33581.05 and 42876.40, p < 0.001) were significantly higher in the RAT group.Conclusions: RAT is safe and effective for the resection of large AMTs compared to VAT and open surgery. Vascular resection in RAT is technically feasible. A long‐term follow‐up is required.
Noise-induced phase transition between chaos and ordered states plays an important role in the self-organization of dissipative structures. Here, we numerically simulate the phase transitions between statistically stationary regimes in a laser cavity by using a stochastic complex cubic-quintic Ginzburg-Landau equation with a phase-diffusion model. Depending on the linear and nonlinear gains, the dissipative system converges from the initial noise floor to different regimes, including continuous wave, phase turbulence, amplitude turbulence (AT), and stable soliton structure. Around the regime boundaries, we found that only certain phase noise can cause the instability of phase turbulence. The defects formed due to the noise will continue to spread out and self-organize to form AT or stabilize the soliton structure in the defect area. Our results are helpful in understanding the dynamics of the buildup of mode locking and noiselike pulses in passively mode-locked lasers.
A residual convolutional neural network (RCNN) is introduced to retrieve the separation and relative phase of solitons in soliton molecules in passively mode-locked lasers (PMLs). It proved to be an effective method to explore complex soliton molecule dynamics.