AI Algorithm Analysis for CT Three-Dimensional Diagnosis and Accurate Assessment of Posterior Cruciate Ligament Tibial Avulsion Fractures

Abstract

BACKGROUND: Surgical decision-making for posterior cruciate ligament avulsion fractures is highly dependent on imaging evaluation. Traditional methods rely on subjective interpretation of CT images, which suffer from limitations such as difficulties in quantifying three-dimensional spatial displacement parameters and insufficient precision in assessing rotational angles. Given the advancements in artificial intelligence (AI) technology, there is a need to develop automated, intelligent image recognition software based on AI algorithms. OBJECTIVE: To investigate the intelligent diagnostic capabilities of AI algorithms for posterior cruciate ligament tibial avulsion fractures in 3D CT images and their effectiveness in accurately assessing 3D parameters of fracture fragments. METHODS: Knee CT data from 24 patients with posterior cruciate ligament tibial avulsion fractures who were treated at the Wangjing Hospital of the China Academy of Chinese Medical Sciences between December 1, 2022, and August 30, 2024, were retrospectively collected. Three-dimensional reconstruction, intelligent fracture point recognition, and simulated reduction were performed using self-developed AI image recognition software. Translational and rotational parameters of the fracture fragments along the X, Y, and Z axes were obtained. These measurements were compared with those from traditional radiology reading software (PACS system) using rank-sum tests, Bland-Altman analysis, and linear regression models to assess consistency, and coefficients of variation were calculated to verify software stability. RESULTS AND CONCLUSION: ①There were no significant differences between AI software and traditional methods in measuring fracture fragment displacement (X/Y/Z axis translation and rotation) (P > 0.05). ②Bland-Altman analysis showed good consistency between the two methods, with no significant differences (P > 0.05). ③Linear regression models for X, Y, Z axis displacement and angles showed R² values > 0.99. ④The coefficients of variation for three repeated fracture point identifications by the AI software showed that for total fracture identification, 21 cases had coefficients of variation < 20%, and for articular surface fracture points, 18 cases had coefficients of variation < 20%. ⑤These findings indicate that the AI image recognition software can accurately quantify three-dimensional parameters of posterior cruciate ligament avulsion fracture fragments, with measurement results consistent with traditional methods and good stability. It can assist doctors in judging the degree of displacement and provide precise data support for preoperative planning. The software has good application prospects in avulsion fractures, and future studies should expand the sample size and further verify its impact on surgical outcomes.

1. Introduction

Posterior cruciate ligament avulsion fractures are common acute knee injuries. Due to frequent soft tissue interposition between the fracture fragment and the tibia, conservative treatment often fails to achieve anatomical reduction; therefore, surgical treatment is preferred for displaced fractures [1-2]. Surgical decision-making relies heavily on imaging evaluation. Traditional methods depend on two-dimensional CT images combined with radiology reading software (e.g., PACS systems) for subjective interpretation. When measuring rotational parameters, it is often necessary to calculate different rotational angles based on the extension lines of the fracture fragment edges and the bone defect area. However, when multiple fragments are present in the same region, accurate measurement becomes difficult, leading to limitations such as difficulties in quantifying three-dimensional spatial displacement parameters and insufficient precision in assessing rotational angles [3-4], directly affecting surgical strategy formulation.

3D printing technology allows surgeons to visually assess displacement and offers advantages in implant design and plate pre-bending [5-8], but due to the high time and economic costs of production, it is often applied to more complex intra-articular fractures and is less used in simple avulsion fractures.

In recent years, AI tools based on machine learning models have been widely applied in clinical scenarios such as early disease screening, auxiliary diagnosis, prognosis prediction, and medical image processing [9-11]. In medical image processing, the main applications include automatic lesion localization, automatic image segmentation, and image feature extraction, greatly improving clinical efficiency and outcomes [12-14]. The author team applied automatic lesion localization to fracture recognition scenarios, proposing a digital algorithm-based intelligent CT three-dimensional image diagnosis and evaluation system. Unlike 3D printing, this software does not require model fabrication; instead, it generates three-dimensional models via computer and performs preoperative planning. The system automatically identifies fracture regions through digital algorithms, combines three-dimensional reconstruction technology to generate three-dimensional models of fracture fragments, and intelligently performs simulated reduction based on anatomical landmarks.