Reduced field-of-view DWI outperforms conventional DWI in assessing tumor heterogeneity and HPV status in head and neck cancer
Reduced field-of-view diffusion-weighted imaging (reduced-FOV DWI) is a promising technique for assessing tumor heterogeneity and microstructure. In this study, we evaluated the ability of reduced-FOV DWI to identify specific patterns associated with human papillomavirus (HPV) exposure in head and neck squamous cell carcinoma (HNSCC). Whole-lesion analysis of the tumor was performed on reduced-FOV and conventional DWI using a multi b value sequence. The extracted parameters included the 10th-90th percentile of the apparent diffusion coefficient (ADC) and fractional blood volume (fBV), skewness, and kurtosis. The differences between HPV-negative and HPV-positive tumors were greater for reduced-FOV DWI. ADC was significantly lower in HPV-positive tumors (reduced-FOV DWI: p < 0.001, r<sub>90th</sub> = 0.50; conventional DWI: p < 0.001, r<sub>10th</sub> = 0.49). The fBV was significantly lower in HPV-positive tumors on reduced-FOV DWI (reduced-FOV DWI: p<sub>90th</sub> < 0.001, r<sub>90th</sub> = 0.47; conventional DWI: not statistically significant). Moreover, there were significant differences between parameters derived from reduced-FOV and conventional DWI (ADC: p<sub>10th-90th</sub> < 0.001, fBV: p<sub>10th-50th</sub> < 0.001; skewness<sub>ADC, fBV</sub>: p < 0.001, kurtosis<sub>ADC</sub>: p = 0.008; kurtosis<sub>fBV</sub>: p < 0.001). In conclusion, DWI parameters varied depending on HPV status and the differences between HPV-positive and HPV-negative tumors were greater for reduced-FOV DWI. The reduced-FOV DWI better reflected histopathology than did conventional DWI.
Authors
- Tord Hompland (ORCID: https://orcid.org/0000-0003-0529-6077)
- Knut Håkon Hole (ORCID: https://orcid.org/0000-0001-6885-8538)
- Marta D. Switlyk (ORCID: https://orcid.org/0000-0003-1246-6556)
- Einar Dale (ORCID: https://orcid.org/0000-0001-9483-2788)
- L G Willumsen
- Svetlana Tafjord
Institutions
- Oslo University Hospital (NO)
- University of Oslo (NO)
Publication Details
- Journal
- Scientific Reports
- Published
- 2025-08-29
- DOI
- https://doi.org/10.1038/s41598-025-17284-8
- Primary Topic
- Radiomics and Machine Learning in Medical Imaging
- Type
- article
- Field-Weighted Citation Impact
- 0.00