Session Item

Functional MRI provides quantitative imaging biomarkers that can be used to evaluate tumour response to radiotherapy. This work uses an intra-voxel incoherent motion (IVIM) model to quantify perfusion and diffusion characteristics of cervical tumours from diffusion-weighted (DW) MRI, evaluating the model’s suitability and assessing model parameter changes during therapy.

Imaging was performed at 1.5 T in 8 patients with locally advanced (stages IIB-IVA) squamous cell carcinoma of the cervix, following research ethics committee approval (REC: 20/NW/0377). All patients underwent standard of care treatment: weekly cisplatin chemotherapy prescribed at 40 mg/m²; combined chemoradiation/brachytherapy prescribed to reach a final dose of 85-90 Gy EQD2. DW-MRI was performed (b = 0, 20, 40, 60, 80, 100, 150, 300, 500, 800 s/mm², TR = 2800 ms, TE = 61 ms, voxel = 2.9 x 2.9 x 6.0 mm³) at up to 3 timepoints: pre-treatment, week 3, and week 5 of treatment.

IVIM and apparent diffusion coefficient (ADC) models were fitted voxel-wise to data at all b-values, and were statistically compared using the corrected Akaike information criterion (AICc). Along with IVIM parameter maps (diffusion coefficient, D, and perfusion signal fraction, f ), model preference maps were generated showing which model was favoured in each voxel. ADC-favoured voxels are expected to reflect regions with a single tissue diffusion component, while IVIM-favoured voxels are expected to reflect well-perfused regions. Tumour ROIs were defined on b = 800 s/mm² images, and median IVIM parameter values were obtained, along with the fraction of voxels in which the IVIM model was favoured, termed p-IVIM. In addition, a conventional ADC fit was performed using b = 150,300,500,800 s/mm².

Model preference maps (Figure 1) reflect expected trends based on tissue characteristics, with ADC tending to be favoured in the bladder and uterine cavity fluid, while IVIM tends to be preferred in the fibroid, which is expected to be highly vascular. Across all tumours at the pre-treatment timepoint, p-IVIM = 0.47 ± 0.19 (mean±SD), indicating that the ADC model is favoured in a subset of tumour voxels for each patient.

Figure 2 plots median D and f values for all tumours and timepoints. D significantly increased from baseline to week 3 (p = 0.013, paired t-test for patients scanned at both timepoints), while f showed no significant change (p = 0.13). Median ADC from the conventional fit did not change significantly between baseline and week 3 (p = 0.11). The fraction of tumour voxels favouring IVIM showed large inter-patient variability, which was maintained throughout treatment. Reliable contouring was hampered for some tumours at week 5 due to their small size.

The diffusion coefficient from IVIM is sensitive to early therapy-induced changes in cervical tumours. However, the IVIM model is not favoured in all tumour voxels, demonstrating that model suitability should be evaluated as part of imaging biomarker validation.