際際滷

際際滷Share a Scribd company logo

Summary of Dr. Berg Radiology 2012

?
?
Nasos Papapostolou
Nasos Papapostolou

Shear-wave Elastography Improves the Specificity of Breast US: The BE1 Multinational Study of 939 Masses,￣ published in Radiology 2012

1 of 10
Downloaded 39 times
ShearWave? elastography improves the specificity of breast ultrasound Summary of the article ＾Shear-wave Elastography Improves the Specificity of Breast US: The BE1 Multinational Study of 939 Masses,￣ published in Radiology 2012; 262:435-449 by Berg WA, Cosgrove DO, Dor└ CJ, Skyba DM, Henry JP, Gay J, Cohen-Bacrie C; the BE1 Study Group. 1 Summary of Berg et al. Radiology 2012; 262:435-449 SSID02661-01
Purpose of the study ? To determine if and which SWE? parameters added to conventional BI-RADS? criteria improve the characterization of breast masses ? Strong focus on the reclassification of the masses within categories 3 and 4a ? among BI-RADS? 3: better characterization of the 2% malignant masses in order to avoid delays for treatment ? among BI-RADS? 4a: better characterization of the benign lesions in order in order to replace biopsy by follow-up 2 Summary of Berg et al. Radiology 2012; 262:435-449 SSID02661-01
Take home messages (1/2) 1. Any of the SWE? features increases the AUC* of BI-RADS? assessment alone 2. Emax is the best SWE parameter to add to conventional BI-RADS? criteria to improve characterization of breast masses 3. BI-RADS? 2, supposedly benign masses: remain BI-RADS 2 regardless of SWE 4. BI-RADS? 5, 4c, 4b, moderate-to-highly suspicious masses: remain BI-RADS 5, 4c, 4b regardless of SWE *AUC: Area Under the ROC Curve. Indicative for both sensitivity and specificity. Values:1 perfect test; 0.5 useless test. 3 Summary of Berg et al. Radiology 2012; 262:435-449 SSID02661-01
Take home messages (2/2) 5. BI-RADS? 3 and 4a Reclassification rules with Emax added to conventional BI-RADS? criteria ? BI-RADS 3 + Emax− 160 kPa : upgrade the lesion to 4 Biopsy recommended Delay avoided in the early management of malignant masses ? BI-RADS 4a + Emax＋ 80 kPa : downgrade the lesion to 3 Follow-up recommended Biopsy of benign masses avoided ? BI-RADS 3 and 4a, subset of oval and circumscribed masses: Reclassification of low risk lesions with Ecol: Sensitivity increases from 0% to 100% (p= 0.046) Specificity increases from 79.1% to 90.4% (p<0.001) This excellent sensitivity allows to better identify the few malignant masses and to adapt the management without delay 4 Summary of Berg et al. Radiology 2012; 262:435-449 SSID02661-01
Materials and methods (1/2) ? Participants, investigators and reviewer These points are detailed in the document ＾Shear wave elastography is highly reproducible＾ SSID02658-01 ? Statistics C.J.Dor└, BSC, independent statistical expert from Medical Research Council of the United Kingdom Validation of the results with different statistical tools: Fisher exact test, weighted k values, McNemar test, Mann-Whitney U test, Spearman rank correlation 5 Summary of Berg et al. Radiology 2012; 262:435-449 SSID02661-01
Materials and methods (2/2) For each participant ? Imaging, information collected by the investigator and information collected during the blind review: see the document SSID02658-01 ? The most relevant SWE? features studied ? Qualitative SWE? features of the lesion and the surrounding parenchyma Esha: lesion shape Ehomo: homogeneity of the elasticity (or stiffness) Ecol: maximum elasticity on a six-point color scale ? Quantitative SWE? features of the lesion and surrounding parenchyma Emax: maximum value of elasticity 6 Summary of Berg et al. Radiology 2012; 262:435-449 SSID02661-01
Results (1/4) Global performances improved BI-RADS? category 3 C 5 Addition of SWE? features to conventional BI-RADS? assessments ? Any of the SWE features increases the AUC of BI-RADS assessment alone ? Ecol: the best performing qualitative feature, AUC increased from 0.940 to 0.971 ? Emax the best performing quantitative feature, AUC increased from 0.940 to 0.962 ? Excellent correlation between qualitative Ecol and quantitative Emax ? No improvement by adding more than one SWE feature 7 Summary of Berg et al. Radiology 2012; 262:435-449 SSID02661-01
Results (2/4) Improved specificity, constant sensitivity Reclassification of BI-RADS? 3 C 4a: Addition of Ecol to BI-RADS assessments alone 135 black or blue not suspicious downgraded 132 benign, 3 malignant Using B Mode Using B Mode criteria criteria + E color of which 18 cancers of which 18 cancers BIRADS 4a 193 78 20 red 303 418 suspicious BIRADS 3 upgraded of which 8 cancers of which 8 cancers 17 benign, 3 malignant ? Specificity increases from 61.1% to 78.5% (397/650 to 512/650) p<0.0001. i.e. + 17% ? PPV of biopsy for BI-RADS 4a: from 9.3% to 23% (18/193 to 18/78) i.e. + 14% ? Rate of malignancy for BI-RADS 3: never worse than the 2.6% initially observed Ehom: same improvement in specificity with an insignificant decrease of sensitivity from 97.2% to 96.2% Esha: significant improvement in specificity with an insignificant increase in sensitivity to 97.9% 8 Summary of Berg et al. Radiology 2012; 262:435-449 SSID02661-01
Results (3/4) Improved specificity, constant sensitivity Reclassification of BI-RADS? 3 C 4a: Addition of Emax to BI-RADS assessments alone 139 ＋ 80 kPa not suspicious downgraded 134 benign, 5 malignant Using B Mode Using B Mode criteria criteria + E color of which 18 cancers of which 18 cancers BIRADS 4a 193 87 33 − 160 kPa 303 409 suspicious BIRADS 3 upgraded of which 8 cancers of which 8 cancers 28 benign, 5 malignant ? Specificity increases from 61.1% to 77.4% (397/650 to 503/650) p<0.001. i.e. + 16% ? PPV of biopsy for BI-RADS 4a: from 9.3% to 20.7% (18/193 to 18/87) i.e. + 11% ? Rate of malignancy for BI-RADS 3: 1.9%, better than the 2.6% initially observed 9 Summary of Berg et al. Radiology 2012; 262:435-449 SSID02661-01
Results (4/4) Special benefits for low risk masses Reclassification of BI-RADS? 3 C 4a: Addition of Ecol to BI-RADS assessments alone for oval and circumscribed masses Using B Mode Using B Mode criteria criteria + E color 28 black or blue not suspicious of which 0 cancer of which 4 cancers downgraded BIRADS 4a all benign 37 17 8 red 144 164 suspicious BIRADS 3 upgraded of which 4 cancers of which 0 cancer 4benign, 4 malignant ? Sensitivity increases from 0% to 100% (0/4 to 4/4) (p= 0.046) ? Specificity increases from 79.1% to 90.4% (140/177 to 160 /177) (p<0.001) ) i.e. + 21% ? Accuracy increases from 77.3% to 90.6% (140/181 to 164/181) (p<0.001) i.e. + 23% The end 10 Summary of Berg et al. Radiology 2012; 262:435-449 SSID02661-01

More Related Content

Summary of Dr. Berg Radiology 2012

  • 1. ShearWave? elastography improves the specificity of breast ultrasound Summary of the article ＾Shear-wave Elastography Improves the Specificity of Breast US: The BE1 Multinational Study of 939 Masses,￣ published in Radiology 2012; 262:435-449 by Berg WA, Cosgrove DO, Dor└ CJ, Skyba DM, Henry JP, Gay J, Cohen-Bacrie C; the BE1 Study Group. 1 Summary of Berg et al. Radiology 2012; 262:435-449 SSID02661-01
  • 2. Purpose of the study ? To determine if and which SWE? parameters added to conventional BI-RADS? criteria improve the characterization of breast masses ? Strong focus on the reclassification of the masses within categories 3 and 4a ? among BI-RADS? 3: better characterization of the 2% malignant masses in order to avoid delays for treatment ? among BI-RADS? 4a: better characterization of the benign lesions in order in order to replace biopsy by follow-up 2 Summary of Berg et al. Radiology 2012; 262:435-449 SSID02661-01
  • 3. Take home messages (1/2) 1. Any of the SWE? features increases the AUC* of BI-RADS? assessment alone 2. Emax is the best SWE parameter to add to conventional BI-RADS? criteria to improve characterization of breast masses 3. BI-RADS? 2, supposedly benign masses: remain BI-RADS 2 regardless of SWE 4. BI-RADS? 5, 4c, 4b, moderate-to-highly suspicious masses: remain BI-RADS 5, 4c, 4b regardless of SWE *AUC: Area Under the ROC Curve. Indicative for both sensitivity and specificity. Values:1 perfect test; 0.5 useless test. 3 Summary of Berg et al. Radiology 2012; 262:435-449 SSID02661-01
  • 4. Take home messages (2/2) 5. BI-RADS? 3 and 4a Reclassification rules with Emax added to conventional BI-RADS? criteria ? BI-RADS 3 + Emax− 160 kPa : upgrade the lesion to 4 Biopsy recommended Delay avoided in the early management of malignant masses ? BI-RADS 4a + Emax＋ 80 kPa : downgrade the lesion to 3 Follow-up recommended Biopsy of benign masses avoided ? BI-RADS 3 and 4a, subset of oval and circumscribed masses: Reclassification of low risk lesions with Ecol: Sensitivity increases from 0% to 100% (p= 0.046) Specificity increases from 79.1% to 90.4% (p<0.001) This excellent sensitivity allows to better identify the few malignant masses and to adapt the management without delay 4 Summary of Berg et al. Radiology 2012; 262:435-449 SSID02661-01
  • 5. Materials and methods (1/2) ? Participants, investigators and reviewer These points are detailed in the document ＾Shear wave elastography is highly reproducible＾ SSID02658-01 ? Statistics C.J.Dor└, BSC, independent statistical expert from Medical Research Council of the United Kingdom Validation of the results with different statistical tools: Fisher exact test, weighted k values, McNemar test, Mann-Whitney U test, Spearman rank correlation 5 Summary of Berg et al. Radiology 2012; 262:435-449 SSID02661-01
  • 6. Materials and methods (2/2) For each participant ? Imaging, information collected by the investigator and information collected during the blind review: see the document SSID02658-01 ? The most relevant SWE? features studied ? Qualitative SWE? features of the lesion and the surrounding parenchyma Esha: lesion shape Ehomo: homogeneity of the elasticity (or stiffness) Ecol: maximum elasticity on a six-point color scale ? Quantitative SWE? features of the lesion and surrounding parenchyma Emax: maximum value of elasticity 6 Summary of Berg et al. Radiology 2012; 262:435-449 SSID02661-01
  • 7. Results (1/4) Global performances improved BI-RADS? category 3 C 5 Addition of SWE? features to conventional BI-RADS? assessments ? Any of the SWE features increases the AUC of BI-RADS assessment alone ? Ecol: the best performing qualitative feature, AUC increased from 0.940 to 0.971 ? Emax the best performing quantitative feature, AUC increased from 0.940 to 0.962 ? Excellent correlation between qualitative Ecol and quantitative Emax ? No improvement by adding more than one SWE feature 7 Summary of Berg et al. Radiology 2012; 262:435-449 SSID02661-01
  • 8. Results (2/4) Improved specificity, constant sensitivity Reclassification of BI-RADS? 3 C 4a: Addition of Ecol to BI-RADS assessments alone 135 black or blue not suspicious downgraded 132 benign, 3 malignant Using B Mode Using B Mode criteria criteria + E color of which 18 cancers of which 18 cancers BIRADS 4a 193 78 20 red 303 418 suspicious BIRADS 3 upgraded of which 8 cancers of which 8 cancers 17 benign, 3 malignant ? Specificity increases from 61.1% to 78.5% (397/650 to 512/650) p<0.0001. i.e. + 17% ? PPV of biopsy for BI-RADS 4a: from 9.3% to 23% (18/193 to 18/78) i.e. + 14% ? Rate of malignancy for BI-RADS 3: never worse than the 2.6% initially observed Ehom: same improvement in specificity with an insignificant decrease of sensitivity from 97.2% to 96.2% Esha: significant improvement in specificity with an insignificant increase in sensitivity to 97.9% 8 Summary of Berg et al. Radiology 2012; 262:435-449 SSID02661-01
  • 9. Results (3/4) Improved specificity, constant sensitivity Reclassification of BI-RADS? 3 C 4a: Addition of Emax to BI-RADS assessments alone 139 ＋ 80 kPa not suspicious downgraded 134 benign, 5 malignant Using B Mode Using B Mode criteria criteria + E color of which 18 cancers of which 18 cancers BIRADS 4a 193 87 33 − 160 kPa 303 409 suspicious BIRADS 3 upgraded of which 8 cancers of which 8 cancers 28 benign, 5 malignant ? Specificity increases from 61.1% to 77.4% (397/650 to 503/650) p<0.001. i.e. + 16% ? PPV of biopsy for BI-RADS 4a: from 9.3% to 20.7% (18/193 to 18/87) i.e. + 11% ? Rate of malignancy for BI-RADS 3: 1.9%, better than the 2.6% initially observed 9 Summary of Berg et al. Radiology 2012; 262:435-449 SSID02661-01
  • 10. Results (4/4) Special benefits for low risk masses Reclassification of BI-RADS? 3 C 4a: Addition of Ecol to BI-RADS assessments alone for oval and circumscribed masses Using B Mode Using B Mode criteria criteria + E color 28 black or blue not suspicious of which 0 cancer of which 4 cancers downgraded BIRADS 4a all benign 37 17 8 red 144 164 suspicious BIRADS 3 upgraded of which 4 cancers of which 0 cancer 4benign, 4 malignant ? Sensitivity increases from 0% to 100% (0/4 to 4/4) (p= 0.046) ? Specificity increases from 79.1% to 90.4% (140/177 to 160 /177) (p<0.001) ) i.e. + 21% ? Accuracy increases from 77.3% to 90.6% (140/181 to 164/181) (p<0.001) i.e. + 23% The end 10 Summary of Berg et al. Radiology 2012; 262:435-449 SSID02661-01