Artificial Intelligence Imaging Diagnosis Using Super-Resolution and Three-Dimensional Shape for Lymph Node...

Area of Science:

• Oncology
• Medical Imaging
• Artificial Intelligence

Background:

• Accurate preoperative diagnosis of lymph node metastasis is crucial for optimizing treatment strategies in low rectal cancer.
• Current diagnostic methods for lymph node metastasis in low rectal cancer have limitations in accuracy.

Purpose of the Study:

• To develop a high-precision diagnostic method for lymph node metastasis in low rectal cancer using artificial intelligence (AI).

Main Methods:

• A retrospective observational study was conducted at a single cancer center and engineering college in Japan.
• Patients with low rectal adenocarcinoma underwent contrast-enhanced multidetector row CT (≤1 mm slice).
• Pelvic lymph nodes were extracted and pathologically diagnosed for AI model training and validation.

Main Results:

• Four AI diagnostic methods were compared, incorporating super-resolution images and 3D shape data.
• The combination of super-resolution and 3D shape data yielded the highest diagnostic ability for sensitivity (0.964), negative predictive value (0.966), and accuracy (0.968).
• Super-resolution imaging alone demonstrated superior specificity (0.994) and positive predictive value (0.993).

Conclusions:

• AI, particularly using super-resolution images and 3D shape data, shows significant potential for improving the diagnosis of lymph node metastasis in low rectal cancer.
• The developed AI method could be a game-changer in the diagnosis and treatment of low rectal cancer.
• Limitations include a small patient cohort from a single center and the absence of external validation.