Whole-Lesion Apparent Diffusion Coefficient-Based Entropy-Related Parameters for Characterizing Cervical Cancers:...

Area of Science:

• Radiology and Imaging
• Oncology
• Biomedical Engineering

Background:

• Intratumoral heterogeneity in cervical cancer impacts treatment response and prognosis.
• Accurate assessment of heterogeneity is crucial for effective cancer management.
• Current methods for assessing heterogeneity may have limitations.

Purpose of the Study:

• To develop and evaluate apparent diffusion coefficient (ADC)-based entropy parameters for assessing intratumoral heterogeneity in cervical cancer.
• To compare these parameters between cervical tumors and adjacent normal cervical tissues.
• To determine the potential of these parameters in differentiating cancerous from normal tissues.

Main Methods:

• Prospective study of 51 women with biopsy-confirmed cervical cancer using 3-T pelvic diffusion-weighted MRI.
• Calculation of first-order and second-order ADC-based entropy parameters from whole tumor volumes and adjacent normal tissues.
• Statistical analysis included intraclass correlation coefficient, Wilcoxon test, Kruskal-Wallis test, and ROC curve analysis.

Main Results:

• All developed entropy parameters demonstrated excellent interobserver agreement (>0.900).
• Cervical cancers exhibited significantly higher values for entropy, entropy(H)0, entropy(H)45, entropy(H)90, entropy(H)135, and entropy(H)mean compared to normal tissues (P < .0001).
• Second-order entropies showed superior performance in differentiating cervical cancers from normal tissues, with entropy(H)45, entropy(H)90, entropy(H)135, and entropy(H)mean achieving the highest area under the ROC curve (0.867).

Conclusions:

• Whole-lesion ADC-based entropy parameters were successfully developed for cervical cancer.
• These parameters show initial potential for characterizing intratumoral heterogeneity.
• The findings suggest these novel parameters can aid in distinguishing cervical tumors from adjacent normal tissues.