Low b-values in apparent diffusion coefficient calculations overestimate diffusion in rectal cancer

Area of Science:

• Radiology
• Oncology
• Medical Imaging

Background:

• Apparent diffusion coefficient (ADC) quantification in rectal cancer typically uses a monoexponential model.
• Optimal b-value selection for ADC in rectal cancer remains undefined.
• Diffusion-weighted MRI (DWI) is essential for assessing rectal cancer characteristics.

Purpose of the Study:

• To evaluate the influence of various b-value combinations on ADC values in rectal cancer.
• To determine optimal b-value selection for accurate ADC quantification in rectal cancer using a monoexponential model.
• To compare ADC values derived from different b-value combinations against a biexponential reference.

Main Methods:

• Prospective observational study of 23 rectal cancer patients (OxyTarget study, NCT01816607).
• Diffusion-weighted MRI (DWI) acquired with seven b-values (0-1300 s/mm²) on a 1.5T scanner.
• Whole-tumour volumes of interest delineated by two radiologists; ADC calculated using 18 b-value combinations.

Main Results:

• Tumour ADC values varied significantly based on b-value combinations.
• Excluding low b-values (≤ 100 s/mm²) resulted in reduced ADC values.
• Using b-values ≥ 500 s/mm² (two or three values, including 1000 s/mm²) showed the smallest deviation from the biexponential reference, while including b=0 s/mm² caused overestimation.

Conclusions:

• Rectal cancer ADC quantification using the monoexponential model is highly sensitive to b-value selection.
• Eliminating low b-values (≤ 100 s/mm²) reduces calculation uncertainty by minimizing perfusion effects.
• Recommended use of b-values > 100 s/mm², including a high b-value (≥ 1000 s/mm²), for reliable ADC assessment in rectal cancer.