GIÁ TRỊ TÍN HIỆU T2W CHUẨN HÓA VÀ THÔNG SỐ HISTOGRAM TRÊN CỘNG HƯỞNG TỪ 3.0 TESLA TRONG CHẨN ĐOÁN UNG THƯ TUYẾN TIỀN LIỆT

Abstract

Objectives: To identify the value of normalized T2-weighted (nT2W) signal intensity and histogram-based texture parameters on 3.0T MRI in the diagnosis of prostate cancer. Methods: A retrospective, cross-sectional descriptive study comparing 3.0T MRI of 139 patients with suspected prostate cancer. T2W signal intensity and histogram parameters, including mean, median, standard deviation, maximum, minimum, skewness, kurtosis, entropy, and variance, were measured. The nT2W signal was calculated as the ratio of the mean T2W signal intensity to that of the internal obturator muscle. Data were compared between the prostate cancer group (Gleason ≥ 6) and the non-cancer group (Gleason < 6). Results: Peripheral zone: Prostate cancer lesions showed significantly lower values than non-cancer lesions for nT2W (3.49 ± 0.55 vs. 4.99 ± 1.43) and several histogram parameters (mean, median, minimum, entropy). Transition zone: Prostate cancer lesions had lower values than non-cancer for nT2W (3.47 ± 0.50 vs. 4.97 ± 0.89) and several histogram features (mean, median, minimum, maximum, skewness, kurtosis), with p ≤ 0.001. ROC curve analysis showed that normalized T2W was the most reliable parameter for diagnosing prostate cancer in the peripheral zone (AUC: 0.947; cut-off: 4.0; sensitivity: 89.1%; specificity: 96.2%) and in the transition zone (AUC: 0.945; cut-off: 4.09; sensitivity: 94.4%; specificity: 82.0%). Conclusion: The nT2W signal on 3.0T MRI is the most reliable marker for diagnosing prostate cancer. Histogram parameters such as minimum, mean, median, and entropy in the peripheral zone, as well as minimum, mean, median, skewness, and kurtosis in the transition zone, provide valuable additional diagnostic information.