Scholarlink

Purpose <p>To establish a dual-task assessment framework using time-dependent diffusion MRI (td-dMRI) for: (1) differentiating benign from malignant breast lesions classified as BI-RADS category 4, and (2) preoperatively predicting the proliferation marker Ki-67 in invasive ductal carcinoma (IDC).</p> Methods <p>This prospective study enrolled 152 consecutive patients between June 2024 and June 2025. Patients with BI-RADS 4 lesions detected on ultrasound/mammography, no prior biopsy or treatment, and scheduled for surgical resection. Modalities included MRI using the IMPULSED protocol with PGSE and OGSE sequences. A two-compartment biophysical model quantified microstructural parameters including intracellular volume fraction (fin), mean cell diameter (d-mean), cellularity, extracellular diffusion coefficient (Dex), and ADC values. The final cohort comprised 42 benign and 102 malignant lesions (82 IDC cases). Multivariable logistic regression identified independent predictors, with diagnostic performance assessed through ROC analysis.</p> Results <p>A total of 144 patients (mean age, 52.2 ± 9.5&#xa0;years) were evaluated. All td-dMRI parameters demonstrated significant differences between benign and malignant lesions (p &lt; 0.05). Malignant lesions exhibited significantly lower ADC (PGSE), ADC (OGSE), d-mean, and Dex, but higher cellularity and fin compared with benign lesions. The combined diagnostic model (cellularity + d-mean + ADC (OGSE33Hz)) achieved the highest performance with an AUC of 0.892 (95% CI 0.834–0.950), sensitivity of 94%, and specificity of 92%, significantly outperforming conventional ADC (PGSE). Among 82 IDC patients, fin showed a strong correlation with the Ki-67 index (p &lt; 0.001). The Ki-67 prediction model (fin + d-mean + ADC (OGSE33Hz)) yielded an AUC of 0.825 (95% CI 0.736–0.914), significantly outperforming individual parameters (p &lt; 0.05).</p> Conclusion <p>The td-dMRI-based technique effectively differentiated benign from malignant breast lesions and provided noninvasive prediction of Ki-67 status in IDC, offering valuable tools for clinical decision-making.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Time-dependent diffusion MRI for non-invasive differentiation of benign and malignant breast lesions and evaluation of proliferation index biomarkers

  • 期刊论文

  • 发表日期：

    2026年3月11日

查看全文
  • Luqing Ru,
  • Yanlin Tang,
  • Jigeng Zhang,
  • Ran Shi,
  • Dawei Yin,
  • Yong Zhang,
  • Yuan Kong,
  • Jun Du,
  • Peng Wang

摘要

Purpose

To establish a dual-task assessment framework using time-dependent diffusion MRI (td-dMRI) for: (1) differentiating benign from malignant breast lesions classified as BI-RADS category 4, and (2) preoperatively predicting the proliferation marker Ki-67 in invasive ductal carcinoma (IDC).

Methods

This prospective study enrolled 152 consecutive patients between June 2024 and June 2025. Patients with BI-RADS 4 lesions detected on ultrasound/mammography, no prior biopsy or treatment, and scheduled for surgical resection. Modalities included MRI using the IMPULSED protocol with PGSE and OGSE sequences. A two-compartment biophysical model quantified microstructural parameters including intracellular volume fraction (fin), mean cell diameter (d-mean), cellularity, extracellular diffusion coefficient (Dex), and ADC values. The final cohort comprised 42 benign and 102 malignant lesions (82 IDC cases). Multivariable logistic regression identified independent predictors, with diagnostic performance assessed through ROC analysis.

Results

A total of 144 patients (mean age, 52.2 ± 9.5 years) were evaluated. All td-dMRI parameters demonstrated significant differences between benign and malignant lesions (p < 0.05). Malignant lesions exhibited significantly lower ADC (PGSE), ADC (OGSE), d-mean, and Dex, but higher cellularity and fin compared with benign lesions. The combined diagnostic model (cellularity + d-mean + ADC (OGSE33Hz)) achieved the highest performance with an AUC of 0.892 (95% CI 0.834–0.950), sensitivity of 94%, and specificity of 92%, significantly outperforming conventional ADC (PGSE). Among 82 IDC patients, fin showed a strong correlation with the Ki-67 index (p < 0.001). The Ki-67 prediction model (fin + d-mean + ADC (OGSE33Hz)) yielded an AUC of 0.825 (95% CI 0.736–0.914), significantly outperforming individual parameters (p < 0.05).

Conclusion

The td-dMRI-based technique effectively differentiated benign from malignant breast lesions and provided noninvasive prediction of Ki-67 status in IDC, offering valuable tools for clinical decision-making.