Predicting Surgical Eligibility for Pediatric Osteosarcoma Patients Using Artificial Intelligence

Korea Institute of Radiological & Medical Sciences Develops AI Machine Learning Model to Accurately Predict Preoperative Chemotherapy Response and Metastasis

[Asia Economy Reporter Kim Bong-su] A technology has been developed that can more accurately predict the response to preoperative chemotherapy and metastasis in pediatric osteosarcoma patients using artificial intelligence (AI).

The Korea Institute of Radiological & Medical Sciences announced on the 8th that the research team led by Dr. Woo Sang-geun and Dr. Gong Chang-bae developed this AI machine learning model.

Osteosarcoma, which frequently occurs during the rapid bone growth period in children and adolescents, is treated by first administering adjuvant chemotherapy to reduce recurrence, followed by surgical removal of the tumor. Traditionally, the response to chemotherapy and metastasis have been predicted mainly by analyzing genetic information and diagnostic imaging (positron emission tomography/computed tomography, PET/CT). Genetic analysis methods take a long time and results can vary depending on the biopsy site, while PET/CT allows for rapid analysis but provides limited information about cancer cells, resulting in lower accuracy.

The research team compared and analyzed the accuracy of chemotherapy response and metastasis prediction using conventional methods and the machine learning model in 52 pediatric osteosarcoma patients before neoadjuvant chemotherapy (chemotherapy administered before surgery). When analyzing genetic information (Ki-67 and Ezrin), which is highly expressed in osteosarcoma, and PET/CT images separately, the prediction accuracy was 53% for genetic information analysis and 71% for PET/CT imaging.

On the other hand, the machine learning model trained by integrating genetic information (Ki-67 and Ezrin) and PET/CT images showed a prediction accuracy of 85% for chemotherapy response and metastasis.

The research team explained, "This study developed a machine learning model that integrates specific genetic information expressed in osteosarcoma and PET/CT images," adding, "It not only improved the prediction accuracy of chemotherapy response and metastasis but also opened the possibility of clinical application of AI programs to enhance cancer treatment effectiveness."

The results of this study were published in the online edition of the international oncology journal Cancers on the 28th of last month.

This content was produced with the assistance of AI translation services.