Clinical study predicting prolonged postoperative ileus after laparoscopic colorectal cancer surgery with intelligent bowel sound auscultation system

Figure 1  Patient screening flowchart.

Figure 2 Schematic diagram of wireless bowel sound acquisition. After registering the intestinal sound patch, it is attached to the patient’s lower right abdomen for long-term wireless collection, and the data is transmitted to the computer for real-time analysis of intestinal sound characteristics.

Figure 3 Characteristics of intestinal sounds diagram display. A: Diagram showing changes in the number of bowel sounds at different times on post-operation 1 day, where blue represents the area of weakened intestinal sounds, green represents the area of normal intestinal sounds, and yellow represents the area of increased intestinal sounds; B: Frequency of chart of intestinal sounds, showing the relationship between detected intestinal sounds and time over a period; C: Display of frequency of bowel sounds analysis; D: Display of bowel sound vibration amplitude analysis. BR: Bowel rate.

Figure 4 Comparison chart of bowel sounds indicators on the first postoperative day among different groups of patients. A: Comparison of number of bowel sounds differences on the first postoperative day, with significant differences between the prolonged postoperative ileus (PPOI) group and the no-PPOI group as well as the blank control group; B: Comparison of recovery time of bowel sounds differences, with significant differences between the PPOI group and the no-PPOI group; C: Comparison of FBS differences on the first postoperative day, with no significant differences among the PPOI group, no-PPOI group, and blank control group; D: Comparison of bowel sound vibration amplitude differences on the first postoperative day, with no statistical significance between the PPOI group and the no-PPOI group, but significant differences between the PPOI group and no-PPOI group compared to the blank control group. ^aP < 0.01; ^bP < 0.0001. NBS: Number of bowel sounds; cpm: Counts per minute; RTBS: Recovery time of bowel sounds; FBS: Frequency of bowel sounds; BSVA: Bowel sound vibration amplitude; PPOI: Prolonged postoperative ileus.

Figure 5 Trend chart of intestinal sounds characteristics at different time points. A-C: Comparison of changes in numbers of bowel sounds (A) bowel sounds vibration amplitude (B) and frequency of bowel sounds (C) pre-operation, operation day, and post-operation 3 days in different groups of patients. NBS: Number of bowel sounds; cpm: Counts per minute; FBS: Frequency of bowel sounds; BSVA: Bowel sound vibration amplitude; PPOI: Prolonged postoperative ileus.

Figure 6 Analysis of the characteristics of bowel sounds and operation time on the cut-off values for predicting prolonged postoperative ileus. A: Number of bowel sounds predictive value for prolonged postoperative ileus (PPOI), cut-off value is 1.201 counts per minute; sensitivity 56.67%, specificity 80.58%; B: Recovery time of bowel sounds predictive value for PPOI, cut-off value is 16.9 hours; sensitivity 90.00%, specificity 43.75%; C: Bowel sound vibration amplitude predictive value for postoperative PPOI, no statistically significant difference; D: Operative time predictive value for PPOI, cut-off value is 205 minutes; sensitivity 90.00%, specificity 38.83%.

Figure 7 Variable screening schematic diagram. A: Boruta analysis indicates that four variables—hypoproteinemia, recovery time of bowel sounds, number of bowel sounds, and operation time—were selected; B: The LASSO regression identified two variables, number of bowel sounds and operation time. FBS: Frequency of bowel sounds; CEA: Carcinoembryonic antigen; CA19-9: Carbohydrate antigen 19-9; BSVA: Bowel sound vibration amplitude; BMI: Body mass index; RTBS: Recovery time of bowel sounds; NBS: Number of bowel sounds.

Figure 8 Performance evaluation graph of predictive model. A: The receiver operating characteristic curve assesses the predictive model for postoperative prolonged postoperative ileus in colorectal cancer, area under the curve = 0.816; B: Decision curve analysis is used to assess the predictive model, indicating it has a good clinical net benefit; C: Calibration curve reflects how closely the predicted risk matches the actual risk; D: The nomogram highlights how different factors influence disease diagnosis. AUC: Area under the curve; RTBS: Recovery time of bowel sounds; NBS: Number of bowel sounds.