Accelerated T2W Imaging with Deep Learning Reconstruction in Staging Rectal Cancer: A Preliminary Study

Area of Science:

• Radiology
• Medical Imaging
• Artificial Intelligence in Medicine

Background:

• Deep learning reconstruction (DLR) shows promise for accelerating MRI scan times.
• Limited research exists on DLR's efficacy in accelerated rectal cancer staging.
• Optimizing DLR for reduced scan times while maintaining diagnostic quality is crucial.

Purpose of the Study:

• To determine the optimal deep learning reconstruction (DLR) level for time savings in phantom experiments.
• To evaluate the feasibility of accelerated T2W imaging using DLR for rectal cancer staging in patients.
• To compare image quality and diagnostic performance between accelerated DLR and conventional MRI protocols.

Main Methods:

• Phantom experiments were conducted to assess resolution, signal-to-noise ratio (SNR), and image quality at various DLR levels.
• 52 rectal cancer patients underwent accelerated T2W MRI with highly-denoised DLR (DLR_H_40sec) and conventional reconstruction (ConR_2min).
• Image quality and diagnostic performance for TN-staging were evaluated by radiologists of varying experience levels.

Main Results:

• Phantom studies indicated DLR_H achieved superior SNR, detail conspicuity, and sharpness with minimal distortion in reduced scan times.
• DLR_H_40sec images demonstrated enhanced sharpness and SNR compared to ConR_2min.
• Agreement with pathological TN-stages improved with DLR_H_40sec across all observer experience levels.
• Area under the receiver operating characteristic curve (AUC) for identifying advanced tumors (T3-4) and nodal involvement (N1-2) was comparable between DLR_H_40sec and ConR_2min.

Conclusions:

• Accelerated T2W imaging with highly-denoised DLR (DLR_H_40sec) significantly reduces scan time by two-thirds.
• DLR_H_40sec provides improved image quality (sharpness, SNR) over conventional reconstruction.
• This accelerated DLR approach maintains diagnostic performance for rectal cancer TN-staging, offering a viable alternative to conventional MRI protocols.