The Intelligent Medical Computing Laboratory (IMCL, also known as GEM) at Macao Polytechnic University has made new progress in multimodal medical image analysis. A research paper jointly conducted by Macao Polytechnic University and the Netherlands Cancer Institute, titled "Leveraging Modality-Guided Pre-training for Dual-Prompt-Driven Multi-cancer PET-CT Segmentation," has been accepted by Medical Image Analysis, a leading international journal in the field of medical image analysis.
The study was conducted by Xinglong Liang, Professor Ritse Mann, and researchers from Macao Polytechnic University. It focuses on multi-cancer PET-CT tumor segmentation, a challenging task that requires the effective integration of complementary information from different imaging modalities. To address the difficulties of accurately segmenting complex lesion regions, the team proposed an intelligent segmentation framework based on modality-guided pre-training and a dual-prompt-driven strategy.
PET-CT plays an important role in cancer diagnosis, staging, and treatment evaluation. PET images provide valuable information about tissue metabolic activity, while CT images offer clear anatomical structural details. Effectively leveraging the complementary strengths of these two modalities is essential for achieving more accurate and robust tumor segmentation.
The acceptance of this paper demonstrates the team's continued research efforts and growing expertise in multimodal medical image analysis, AI-assisted tumor segmentation, and advanced medical artificial intelligence methods. This achievement further highlights IMCL's commitment to developing intelligent computational tools that support precision medicine and clinical decision-making.
Congratulations to the research team on this important accomplishment.