De Santi B, Salvi M, Giannini V, Meiburger K M, Marzola F, Russo F, Bosco M, Molinari F
Annu Int Conf IEEE Eng Med Biol Soc. 2020 Jul;2020:1671-1674. doi: 10.1109/EMBC44109.2020.9176307.
In the last decade, multiparametric magnetic resonance imaging (mpMRI) has been expanding its role in prostate cancer detection and characterization. In this work, 19 patients with clinically significant peripheral zone (PZ) tumours were studied. Tumour masks annotated on the whole-mount histology sections were mapped on T2-weighted (T2w) and diffusion-weighted (DW) sequences. Gray-level histograms of tumoral and normal tissue were compared using six first-order texture features. Multivariate analysis of variance (MANOVA) was used to compare group means. Mean intensity signal of ADC showed the highest showed the highest area under the receiver operator characteristics curve (AUC) equal to 0.85. MANOVA analysis revealed that ADC features allows a better separation between normal and cancerous tissue with respect to T2w features (ADC: P = 0.0003, AUC = 0.86; T2w: P = 0.03, AUC = 0.74). MANOVA proved that the combination of T2-weighted and apparent diffusion coefficient (ADC) map features increased the AUC to 0.88. Histogram-based features extracted from invivo mpMRI can help discriminating significant PZ PCa.
在过去十年中，多参数磁共振成像（mpMRI）在前列腺癌检测和特征描述方面的作用不断扩大。在这项研究中，对19例患有具有临床意义的外周带（PZ）肿瘤的患者进行了研究。将全层组织学切片上标注的肿瘤掩码映射到T2加权（T2w）和扩散加权（DW）序列上。使用六个一阶纹理特征比较肿瘤组织和正常组织的灰度直方图。采用多变量方差分析（MANOVA）比较组均值。表观扩散系数（ADC）的平均强度信号在受试者操作特征曲线（AUC）下显示出最高面积，等于0.85。MANOVA分析表明，与T2w特征相比，ADC特征能更好地区分正常组织和癌组织（ADC：P = 0.0003，AUC = 0.86；T2w：P = 0.03，AUC = 0.74）。MANOVA证明，T2加权和表观扩散系数（ADC）图特征的组合将AUC提高到了0.88。从体内mpMRI中提取的基于直方图的特征有助于鉴别有意义的PZ前列腺癌。
