Fat saturation on T2-weighted imaging in the assessment of renal masses and its impact on clear cell likelihood score categorization

Objectives To evaluate the impact of fat saturation on T2-weighted (T2W) imaging in the assessment of small renal masses (SRMs; cT1a ≤ 4 cm) with the clear cell likelihood score (ccLS). Methods This retrospective study included 147 histologically confirmed SRMs evaluated with MRI including non-fat saturated (NFS) and fat saturated (FS) T2W fast spin echo acquisitions. Two readers (R1, R2) assigned qualitative T2 signal categories (hypo-, iso-, hyperintense), with sequence-specific ccLS scores recalculated while all non-T2 ccLS components were held constant. Agreement, sequence-related reclassification, and inter-reader agreement were assessed for T2 signal intensity and ccLS scores. Cumulative link mixed models (CLMMs) were used to assess sequence-related shifts in ordinal category. Quantitative mass-to-cortex signal-intensity ratios were compared between sequences using Bland-Altman analysis. Results Qualitative agreement between sequences was 74% and 65% for R1 and R2, respectively. FS-T2W was associated with over four-fold greater cumulative odds of a higher qualitative T2 signal category than NFS-T2W (OR, 4.68; 95% CI, 2.91–7.52; p < 0.001), with upgrading in 20% (R1) and 30% (R2) of lesions versus downgrading in 5% for both readers. Qualitative inter-reader agreement was substantial (κ = 0.77 for NFS-T2W; κ = 0.73 for FS-T2W). Mass-to-cortex ratios were higher on FS-T2W (median 1.30 vs. 1.15; p < 0.001), corresponding to a mean proportional increase of 17.9% (95% CI, 12.4–23.4%). ccLS concordance remained high between NFS-T2W and FS-T2W (R1: 94% and R2: 90%; weighted κ = 0.944 and 0.918). FS-T2W was associated with greater cumulative odds of a higher ccLS category than NFS-T2W (OR, 4.27; 95% CI, 1.52–12.0; p = 0.006). Inter-reader agreement for ccLS was almost perfect (weighted κ = 0.98 for NFS-T2W; 0.95 for FS-T2W). Conclusion Fat saturation introduces a systematic upward bias in qualitative T2 signal assessment of small renal masses. Although ccLS reclassification was less frequent and overall agreement remained high, this bias may alter ccLS assignment, particularly in lesions with inherently low T2 signal intensity, supporting standardized NFS-T2W acquisition when applying the ccLS.