Researchers at the Postgraduate Institute of Medical Education and Research (PGIMER) in Chandigarh, in collaboration with the Indian Institute of Technology (IIT), New Delhi, have developed and validated a deep learning (DL) model for detecting gall bladder cancer (GBC) using abdominal ultrasound. Their study, published in The Lancet Regional Health – Southeast Asia journal, demonstrated that the DL model’s diagnostic performance was comparable to experienced radiologists.
Gall bladder cancer is known for its poor detection and high mortality rate, making early diagnosis crucial. However, its similarity in imaging features to benign gallbladder lesions poses a challenge for radiologists.
Key findings of the study include:
•The DL model achieved a sensitivity of 92.3%, specificity of 74.4%, and an area under the receiver operating characteristic curve (AUC) of 0.887 for detecting GBC in the test set.
•The DL model’s performance was on par with that of two independent radiologists.
~It exhibited high sensitivity for detecting various types of GBC, including cases with stones, contracted gallbladders, small lesion sizes, and neck lesions.
•The DL model showed higher sensitivity for detecting the mural thickening type of GBC compared to one of the radiologists.
The study’s authors emphasized the need for further multicenter studies to validate the DL-based approach’s potential for GBC diagnosis. While acknowledging the study’s limitations, including its reliance on a single-center dataset and a knowledge cutoff date in 2021, they highlight the promising role of AI in enhancing GBC detection.
•Researchers at PGIMER and IIT, New Delhi, developed a DL model for detecting gall bladder cancer (GBC) using abdominal ultrasound.
•The DL model’s performance, with a sensitivity of 92.3% and specificity of 74.4%, was comparable to experienced radiologists.
•The DL-based approach exhibited high sensitivity for detecting various GBC types and lesion characteristics.
•Multicenter studies are recommended to explore the potential of DL-based GBC diagnosis further.
The study underscores the promise of AI in improving GBC detection.