The value of MRI for downgrading of breast suspicious lesions detected on ultrasound

Zongyu Xie, Wenjie Xu, Hongxia Zhang, Li Li, Yongyu An, Guoqun MaoBMC Medical Imaging 2023 · DOI

TL;DR

This study investigates the effectiveness of Magnetic Resonance Imaging (MRI) in reducing unnecessary breast biopsies for lesions initially classified as BI-RADS 4A and 4B on ultrasound (US). The findings suggest that MRI can accurately downgrade these suspicious lesions, significantly improving the positive predictive value of biopsies and reducing false positives without missing malignancies.

What Problem Does This Paper Address?

Breast cancer remains a significant health concern, and early detection is crucial for improving outcomes. Ultrasound (US) is a widely used, convenient, and cost-effective imaging modality for breast cancer screening. However, the BI-RADS (Breast Imaging Reporting and Data System) classification system, particularly categories 4A and 4B, often leads to a high rate of benign biopsies. This means many patients undergo invasive procedures that are ultimately unnecessary, incurring costs and potential patient anxiety. While MRI is known for its high sensitivity and negative predictive value in detecting breast cancer, its specific utility in downgrading BI-RADS 4A and 4B lesions identified by US, thereby avoiding unnecessary biopsies, has not been extensively studied. This paper aims to fill that gap by evaluating MRI's performance in this specific clinical scenario.

Methods

This retrospective study included 167 patients with 186 suspicious lesions initially classified as BI-RADS 4A or 4B on ultrasound. Two radiologists evaluated the MRI scans, including dynamic contrast-enhanced MRI (DCE-MRI) and diffusion-weighted imaging (DWI), aware of the US findings but blinded to pathological results. They assigned MRI BI-RADS categories based on consensus. The reference standard for diagnosis was histopathology (biopsy or surgical resection) or imaging follow-up for at least 12 months. The study calculated sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) for MRI findings, along with their 95% confidence intervals.

Datasets

The study utilized data from 167 patients who underwent breast MRI examinations between March 2014 and November 2019 at a single center. The lesions were initially classified using ultrasound and then evaluated with MRI. The final diagnosis was established through histopathology or imaging follow-up. No new datasets were created or introduced by this study.

Key Insights and Ideas

  • MRI as a Biopsy Gatekeeper: The study strongly supports MRI's role in acting as a "gatekeeper" for biopsies in BI-RADS 4A/4B US lesions. By providing a more refined assessment, MRI can help filter out benign cases that would otherwise lead to unnecessary invasive procedures.
  • DWI's Contribution: The integration of DWI alongside DCE-MRI appears to be crucial for improving diagnostic accuracy. This suggests that combining functional (DWI) and morphological/perfusion (DCE-MRI) information offers a more comprehensive picture of lesion characteristics.
  • The Challenge of "Additional Findings": While MRI is effective at downgrading, it can also detect additional suspicious lesions not seen on US. This highlights the need for careful follow-up and potentially further investigation of these new findings, as a significant proportion can be malignant.
  • Addressing the "Benign Biopsy Epidemic": The paper directly tackles the problem of high rates of benign biopsies in BI-RADS 4 categories, a known issue in breast imaging that leads to increased healthcare costs and patient burden.

Key Results

  • High Diagnostic Performance: MRI demonstrated excellent diagnostic performance, with a sensitivity of 100% and an NPV of 100% for detecting malignancy in the studied lesions. The specificity was 92.6% and the PPV was 87.8%.
  • Reduced False Positives: MRI significantly reduced false-positive biopsies by 92.6% (112 out of 121 benign lesions were correctly downgraded).
  • Increased PPV: The PPV of biopsies for suspicious lesions increased from 34.9% (based on US alone) to 87.8% when MRI was used to select lesions for biopsy.
  • Additional Lesions: 2.2% (4/186) of lesions were additionally detected on MRI, and 75% (3/4) of these were malignant, emphasizing the need for further work-up of these findings.

Limitations and Future Work

  • Single-Center Study: The study was conducted at a single institution, which may limit the generalizability of the findings.
  • Small Sample Size: While 167 patients and 186 lesions were included, a larger, multi-center study could provide more robust statistical power.
  • Uncertainty in Negative MRI Cases: 23 lesions with US BI-RADS 4A showed negative MRI findings and were not biopsied, relying solely on imaging follow-up. While these were considered benign, this lack of definitive pathology could slightly influence sensitivity and NPV calculations.
  • Inter-Observer Variability: The study did not assess inter-observer variability in MRI interpretation, which is a common aspect of diagnostic imaging.

Future work could involve multi-center validation of these findings, exploring the role of artificial intelligence in MRI interpretation for these lesions, and developing standardized protocols for managing additional suspicious lesions detected by MRI.

Summary updated

BibTeX
@article{xie2023value,
  author = {Zongyu Xie and Wenjie Xu and Hongxia Zhang and Li Li and Yongyu An and Guoqun Mao},
  title = {{The value of MRI for downgrading of breast suspicious lesions detected on ultrasound}},
  journal = {BMC Medical Imaging},
  year = {2023},
  doi = {10.1186/s12880-023-01021-6},
}
Figure