Enhancing parametrial invasion assessment in cervical squamous cell carcinoma: the collaborative impact of diffusion kurtosis imaging and T2-weighted imaging, exploring tumour core and 5-mm peritumoural tissue

Objective To evaluate the efficacy of magnetic resonance diffusion kurtosis imaging (DKI) combined with MRI T2-weighted imaging (T2WI) in assessing parametrial invasion (PI) in cervical squamous cell carcinoma.Methods 30 patients with cervical cancer underwent routine MRI and DKI scans. DKI parameters (mean diffusivity [MD], mean kurtosis [MK], fractional anisotropy [FA], and kurtosis anisotropy [KA]) were measured in the tumour parenchyma and surrounding 5 mm tissue. The integrity of the low-signal ring around the cervix on T2WI was recorded. LASSO regression identified optimal DKI parameters and ROC curves compared the diagnostic performance of each parameter and T2WI.Results Compared to the non-parametrial infiltration group (NPI), the parametrial infiltration group (PI) had higher values of MKT, KAT, and KAP (P = .018, .008, .042), while MDT was higher in NPI (P = .038). LASSO regression showed strong correlations between MKT, KAT, and KAP with PI. ROC analysis revealed the AUC, sensitivity, and specificity for MKT, KAT, and KAP were 0.765, 0.706, 0.846; 0.778, 0.882, 0.615; and 0.719, 0.529, 0.923, respectively. Combining T2WI with DKI (MKT + T2WI, KAT + T2WI, KAP + T2WI) improved AUCs to 0.846, 0.828, and 0.774. MKT + KAP and KAT + KAP yielded AUCs of 0.792 and 0.787, with sensitivity of 0.706 and specificity of 0.846.Conclusion DKI parameters (tumour MK, KA, and peritumoural KA) are valuable for assessing PI. Combining tumour and peritumoural parameters, along with T2WI, enhances diagnostic accuracy.Advances in knowledge This study presented an approach that combined DKI parameters with T2WI, integrating tumour and peritumoural parameter analysis to enhance the accuracy of assessing PI.