Anna Marie Maier1, Jörg Heil1, Aba Harcos1, Hans-Peter Sinn2, Geraldine Rauch3, Lorenz Uhlmann4, Christina Gomez1, Anne Stieber5, Annika Funk1, Richard G Barr6, André Hennigs1, Fabian Riedel1, Benedikt Schäfgen1, Sarah Hug1, Frederik Marmé7, Christof Sohn1, Michael Golatta8. 1. Department of Gynecology, Breast Unit, Heidelberg University, Heidelberg, Germany. 2. Department of Pathology, Heidelberg University, Heidelberg, Germany. 3. Charité Universitätsmedizin Berlin, Institute of Biometry and Clinical Epidemiology, Corporate Member of Freie Universität Berlin, Humboldt-Universität Zu Berlin, and Berlin Institute of Health Berlin, Berlin, Germany. 4. Institute of Medical Biometry and Informatics, Heidelberg University, Heidelberg, Germany. 5. Department of Diagnostic and Interventional Radiology, Heidelberg University, Heidelberg, Germany. 6. Department of Radiology, Northeastern Ohio Medical University, Rootstown, Ohio and Radiology Consultants Inc., Youngstown, Ohio, USA. 7. Experimental & Translational Gynecological Oncology, University Hospital Mannheim, Heidelberg University, Mannheim, Germany. 8. Department of Gynecology, Breast Unit, Heidelberg University, Heidelberg, Germany. Electronic address: michael.golatta@med.uni-heidelberg.de.
Abstract
OBJECTIVE: To compare the validity of Shear Wave Elastography (SWE) for the preoperative assessment of pathological complete response (pCR) to standard clinical assessment in breast cancer patients undergoing neoadjuvant chemotherapy (NACT). MATERIALS AND METHODS: This prospective, consecutive clinical trial was conducted under routine clinical practice. Analysis included 134 patients. SWE served as index test, final pathology from surgical specimen as reference standard. PCR (ypT0) was defined as primary endpoint. Elasticity changes were compared for the pCR- vs. non-pCR group. To determine the validity of shear wave velocity (Vs), ROC analyses and diagnostic accuracy parameters were calculated and compared to the final standard clinical assessment by physical examination, mammography and B-mode ultrasound (ycT + vs. ycT0). RESULTS: Vs was significantly reduced in pCR and non-pCR groups during NACT (pCR: ΔVs(abs) = 3.90 m/s, p < 0.001; non-pCR: ΔVs(abs) = 3.10 m/s, p < 0.001). The pCR-group showed significant lower Vs for all control visits (t1,2,END: p < 0.001). ROC analysis of Vs yielded moderate AUCs for the total population (t0: 0.613, t1: 0.745, t2: 0.685, tEND: 0.718). Compared to standard clinical assessment, Vs(tEND) (cut-off: ≤3.35 m/s) was superior in sensitivity (79.6 % vs. 54.5 %), NPV (86.4 % vs. 77.5 %), FNR (20.4 % vs. 45.5 %), inferior in specificity (58.6 % vs. 77.5 %), PPV (46.3 % vs. 54.5 %), FPR (41.4 % vs. 22.5 %). CONCLUSION: SWE measures significant differences in tumour elasticity changes in pCR vs. non-pCR cases. SWE shows improved sensitivity compared to standard clinical assessment, high NPV and low FNR, but failed in specificity in order to predict pCR under routine conditions.
OBJECTIVE: To compare the validity of Shear Wave Elastography (SWE) for the preoperative assessment of pathological complete response (pCR) to standard clinical assessment in breast cancerpatients undergoing neoadjuvant chemotherapy (NACT). MATERIALS AND METHODS: This prospective, consecutive clinical trial was conducted under routine clinical practice. Analysis included 134 patients. SWE served as index test, final pathology from surgical specimen as reference standard. PCR (ypT0) was defined as primary endpoint. Elasticity changes were compared for the pCR- vs. non-pCR group. To determine the validity of shear wave velocity (Vs), ROC analyses and diagnostic accuracy parameters were calculated and compared to the final standard clinical assessment by physical examination, mammography and B-mode ultrasound (ycT + vs. ycT0). RESULTS: Vs was significantly reduced in pCR and non-pCR groups during NACT (pCR: ΔVs(abs) = 3.90 m/s, p < 0.001; non-pCR: ΔVs(abs) = 3.10 m/s, p < 0.001). The pCR-group showed significant lower Vs for all control visits (t1,2,END: p < 0.001). ROC analysis of Vs yielded moderate AUCs for the total population (t0: 0.613, t1: 0.745, t2: 0.685, tEND: 0.718). Compared to standard clinical assessment, Vs(tEND) (cut-off: ≤3.35 m/s) was superior in sensitivity (79.6 % vs. 54.5 %), NPV (86.4 % vs. 77.5 %), FNR (20.4 % vs. 45.5 %), inferior in specificity (58.6 % vs. 77.5 %), PPV (46.3 % vs. 54.5 %), FPR (41.4 % vs. 22.5 %). CONCLUSION: SWE measures significant differences in tumour elasticity changes in pCR vs. non-pCR cases. SWE shows improved sensitivity compared to standard clinical assessment, high NPV and low FNR, but failed in specificity in order to predict pCR under routine conditions.