Abbreviated Breast MRI for Screening

Kuhl CK, Schrading S, et. al.

Introduction

  • MRI associated with high direct and indirect costs
    • conventional protocols time consuming to acquire and read
  • limited access to screening MRI

Introduction

  • Propose abbreviated MRI protocol
    • early postcontrast period only
    • MIP for fast overview of imaging volume
  • Reduce time and cost associated with screening breast MRI
  • Increase access

Methods - design

  • Prospective, 2009 - 2010
  • Hypothesized abbreviated protocol (AP) associated with reduced diagnostic accuracy
  • May be worth it for increased speed/decreased cost

Methods - protocol

  • T1 pre
  • T1 post (1st acquisition)
  • T1 subtracted
  • MIP
  • T1 post (2nd acquisition)
  • T1 post (3rd acquisition)
  • T1 post (4th acquisition)
  • T1 post (5th acquisition)
  • Coronal T1
  • Axial T2

Methods - inclusion

  • Women referred for screening MRI on clinical grounds
    • 24% Dense breast tissue
    • 26% Family history
    • 50% Personal history (screening contralateral breast)

Methods - interpretation

  • Two readers, 18 and 6 years experience
  • Images read immediately
    • 1st with AP
      • MIPs read first, recorded presence of abnormal enhancement
      • then rest of AP, given BI-RADS category
    • then with FDP (full diagnostic protocol)
  • Time to establish diagnosis recorded

Results - participants

  • 443 women (163 underwent 2 annual screening rounds)
  • 606 MRIs
  • 82% of exams in women with mildly or moderately increased risk
  • All were asymptomatic and had negative or benign mammogram at time of MRI
  • 427 of 606 had negative or benign US

Results - cancer yield

  • 11 cancers diagnosed (incidence = 2.5%)
    • 4 DCIS
    • 7 invasive
    • All had (false) negative or benign mammograms
    • 10 of 11 had (false) negative or benign US, 11th had negative targeted US at site of MRI detected cancer
  • 8 high-risk lesions
    • 3 papillomas
    • 2 LCIS
    • 2 ADH
    • 1 radial scar

Results - timing

  • Acquisition time
    • AP: 3 m, 4 s
    • FDP: 17 m, 4 s
  • Reading time
    • MIP: 2.8 s
    • AP: 28 s

Results - diagnostic accuracy

p=0.317

Discussion

  • Cancer yield and diagnostic accuracy equivalent between abbreviated and full protocols
  • Substantially reduced time of acquisition and reading
  • NPV of MIPs alone 99.8%
  • Sens/spec of AP identical to FDP

Discussion

  • FDP mainly needed for lesion characterization
  • 38% of BI-RADS 3 lesions diagnosed on FAST images downgraded to BI-RADS 2 after reading full protocol
    • Will decreased costs of AP justify increase in BI-RADS 3 diagnoses?

Discussion - limitations

  • High volume MRI readers, perhaps not transferrable to community practice

Editorial - Elizabeth Morris

  • MRI superior to mammography and US
    • Likelihood of finding cancer at biopsy < 10% if performed on basis of US findings, 30% if on basis of MRI findings
    • MRI geared to detect more biologically relevant cancers (neovascularity, inflammation)

Editorial

  • Currently, MRI limited to high risk populations
    • Limited access
    • expensive
    • time-consuming

Editorial

  • FAST MRI could reduce barriers to MRI screening
  • MRI screening should be offered to all eligible patients
  • MRI should be next additional screening modality, not US
Made with Slides.com