Support Vector Machine-Based Analysis of Holmium Laser Lithotripsy in Treatment of Urinary Calculus from Ultrasound Images

Abstract

This study was to investigate the application value of ultrasound images optimized by support vector machine (SVM) algorithm in the efficacy analysis of holmium laser lithotripsy in the treatment of urinary calculi. 92 patients for treatment of UC were selected as research subjects, with 46 cases in each group. The control group received pneumatic lithotripsy for diagnosis and treatment. The observation group received holmium laser lithotripsy for calculus treatment. The perimeter and area of the defect and the length and width of the external distance of the most effective area of the defect were used as classification features, and a classifier based on SVM was constructed and applied to it. After treatment, the success rate, operation duration, stone clearance time, and hospital stay of the two groups were comprehensively evaluated. The results showed that the success rate of the observation group adopting holmium laser lithotripsy was 100%. The duration of operation in the observation group was (29.7 ± 7.65) min, and the time to clear calculus was (6.99 ± 5.29) days. The length of hospital stay was (3.67 ± 2.9) days. The probability of complications in the observation group was 3.11%. The observation group was superior to the control group in all surgical indicators (95%, 40.7 ± 8.36 minutes, 14.1 ± 7.21 days, and 5.12 ± 3.72 days), with considerable differences between groups ( $P<0.05$ ). The strong support structure similarity information improved the detection and diagnostic analysis ability of ultrasonic images. In conclusion, after ultrasound image evaluation based on SVM algorithm, the adoption of holmium laser lithotripsy can effectively improve the success rate of patients with urinary system disease, which was worthy of clinical adoption and promotion.