Researchers at the University of Tsukuba have made significant strides in the field of pancreatic neuroendocrine tumors (PanNETs), specifically focusing on non-functional PanNETs (NF-PanNETs). To address the challenges associated with a preoperative diagnosis of lymph node metastasis in NF-PanNETs, Gu et al., developed an innovative imaging model. This model combines the power of radiomics, which involves the analysis of data extracted from radiological images, with deep learning techniques. By leveraging contrast-enhanced computed tomography (CT) images, the researchers created a predictive model capable of noninvasively determining preoperative lymph node metastasis.
The study has been published in the journal eClinicalMedicine and showcases the integration of radiomics features extracted from both CT and magnetic resonance imaging (MRI) images using artificial intelligence deep-learning techniques. The researchers found that this approach yielded promising results, with the model demonstrating an impressive 89% success rate in predicting lymph node metastasis. Furthermore, when the model was validated using data from an external medical setting, the success rate increased to 91%.
A significant aspect of the model’s performance is its consistency, irrespective of the size of the tumor. Current clinical guidelines lack a clear consensus on the necessity of surgery for tumors smaller than 2 cm, making this an area of considerable controversy. The developed model, however, exhibited consistent and reliable results regardless of tumor size, addressing a critical gap in existing diagnostic methods.
The researchers have stated that the implications of their research go beyond just accurately diagnosing before surgery. The model they developed could be a valuable tool for surgeons in selecting the most suitable surgical procedures and treatment strategies. This is particularly important in the difficult medical landscape of NF-PanNETs, where treatment decisions are significantly influenced by the presence or absence of lymph node metastasis.
Gu and colleagues have found that the integrated imaging model is a groundbreaking development in the treatment of non-functional pancreatic neuroendocrine tumors (NF-PanNETs). This model has shown consistently high success rates and has the potential to significantly improve patient outcomes by providing surgeons with more precise information to make informed decisions during the complex treatment of NF-PanNETs. The model can predict LNM in NF-PanNETs before surgery, which can help address the current insufficiency of clinical guidelines concerning the 2-cm threshold for tumor lymph node dissection and provide more accurate therapeutic strategies.
Reference
Gu W, Chen Y, Zhu H, Chen H, Yang Z, Mo S, et al. Development and validation of CT-based radiomics deep learning signatures to predict lymph node metastasis in non-functional pancreatic neuroendocrine tumors: a multicohort study. eClinicalMedicine [Internet]. 2023 Nov 1 [cited 2024 Jan 22];65.