Scholarlink

Purpose <p>To develop and validate a biparametric MRI (bpMRI)-based nomogram for distinguishing malignant or potentially malignant uterine smooth muscle tumors from atypical benign leiomyomas.</p> Materials and methods <p>In this retrospective study, patients who underwent MRI for atypical uterine masses from January 2012 to January 2023 and subsequently underwent surgery were identified. MRI was performed using 1.5 Tesla or 3 Tesla scanners, including T2-weighted, diffusion-weighted sequences, and ADC maps. Two experienced radiologists evaluated lesions for qualitative features (i.e., T1/ T2 signal intensity, margins, presence of haemorrhage, DWI restriction) and quantitative mean ADC values. Histopathology served as the reference standard. Logistic regression identified significant predictors of malignancy, which were incorporated into a point-based nomogram. Receiver Operating Characteristic (ROC) analysis assessed model performance; interobserver agreement was evaluated using Cohen’s kappa.</p> Results <p>Eighty-nine lesions from 77 patients (median age, 45.6 years) were included. Malignant and potentially malignant lesions were 26 (30%), and 63 (70%) were benign. On multivariate analysis, irregular margins, hyperintense T2 signal, hyperintense DWI signal, and low ADC values (≤ 0.998 × 10⁻<sup>3</sup> mm<sup>2</sup>/s) were independent predictors of malignancy. These were incorporated into the point-based nomogram. Patients scoring ≥ 132 on the nomogram were classified as likely malignant. The derived nomogram model demonstrated high accuracy (AUC = 0.975; 95% Confidence Interval (CI): 0.91–0.99), with 96.1% sensitivity and 93.6% specificity. Interobserver agreement ranged from substantial to excellent (κ = 0.72–0.89).</p> Conclusion <p>A biparametric MRI-based nomogram using reproducible, contrast-free imaging features can help accurately differentiate malignant from benign uterine smooth muscle tumors. This tool may guide preoperative management, particularly when CE-MRI is unavailable or contraindicated.</p> Graphical abstract <p></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

Biparametric MRI-based nomogram for differentiating malignant from atypical benign uterine smooth muscle tumors

  • 期刊论文

  • 发表日期：

    2025年11月24日

查看全文
  • Aisha Lakhani,
  • Harshaveena Reddy,
  • Antony Augustine,
  • Betty Simon,
  • Anu Eapen,
  • Vinotha Thomas,
  • Sherin Daniel,
  • K. Reka,
  • Anuradha Chandramohan

摘要

Purpose

To develop and validate a biparametric MRI (bpMRI)-based nomogram for distinguishing malignant or potentially malignant uterine smooth muscle tumors from atypical benign leiomyomas.

Materials and methods

In this retrospective study, patients who underwent MRI for atypical uterine masses from January 2012 to January 2023 and subsequently underwent surgery were identified. MRI was performed using 1.5 Tesla or 3 Tesla scanners, including T2-weighted, diffusion-weighted sequences, and ADC maps. Two experienced radiologists evaluated lesions for qualitative features (i.e., T1/ T2 signal intensity, margins, presence of haemorrhage, DWI restriction) and quantitative mean ADC values. Histopathology served as the reference standard. Logistic regression identified significant predictors of malignancy, which were incorporated into a point-based nomogram. Receiver Operating Characteristic (ROC) analysis assessed model performance; interobserver agreement was evaluated using Cohen’s kappa.

Results

Eighty-nine lesions from 77 patients (median age, 45.6 years) were included. Malignant and potentially malignant lesions were 26 (30%), and 63 (70%) were benign. On multivariate analysis, irregular margins, hyperintense T2 signal, hyperintense DWI signal, and low ADC values (≤ 0.998 × 10⁻3 mm2/s) were independent predictors of malignancy. These were incorporated into the point-based nomogram. Patients scoring ≥ 132 on the nomogram were classified as likely malignant. The derived nomogram model demonstrated high accuracy (AUC = 0.975; 95% Confidence Interval (CI): 0.91–0.99), with 96.1% sensitivity and 93.6% specificity. Interobserver agreement ranged from substantial to excellent (κ = 0.72–0.89).

Conclusion

A biparametric MRI-based nomogram using reproducible, contrast-free imaging features can help accurately differentiate malignant from benign uterine smooth muscle tumors. This tool may guide preoperative management, particularly when CE-MRI is unavailable or contraindicated.

Graphical abstract