Abstract PD9-08: Accuracy of ultrasound in evaluating response to neoadjuvant therapy in invasive lobular carcinoma of the breast

Abstract

Abstract Background: Evaluating the size of invasive lobular carcinoma of the breast (ILC) poses a challenge due to higher rates of false negatives compared to invasive ductal carcinoma. Magnetic resonance imaging (MRI) is often considered the best imaging modality for ILC, but requires additional cost, time, and the use of intravenous contrast, unlike ultrasonography (U/S). We sought to determine whether U/S might provide sufficient information to evaluate response to therapy, thus enabling the use of U/S for surgical planning after NAT in ILC. Methods: We queried a database of 685 ILC cases at UCSF between 1981 and 2018, identifying all cases receiving NAT. We reviewed MRI and U/S reports and collected longest tumor diameter and imaging features. We used t-test, chi-squared test, and Pearson’s correlation to compare U/S and MRI longest diameter to true size on pathology. Data were evaluated in Stata 14.2. Results: 148 patients treated with NAT were identified. Of these, 84 (56.8%) received neoadjuvant chemotherapy and 64 (43.2%) received neoadjuvant endocrine therapy. Compared to the endocrine therapy group, those receiving chemotherapy were significantly younger (mean age 53.3 years vs. 62.8 years, p=0.0002), more likely to have tumors of the ER-PR-HER2- or HER2+ subtypes (p=0.016), and of higher pathologic stage (p=0.024). Among 43 cases with evaluable data, longest tumor diameter on post-treatment U/S strongly correlated with longest diameter on post-treatment MRI (r=0.93, p<0.0001). Post-treatment U/S longest diameter was significantly correlated with true size on pathology (r=0.56, p=0.0008). This correlation held true whether the patient had endocrine therapy (r=0.47, p=0.02) or chemotherapy (r=0.75, p=0.03). Finally, we sought to determine whether tumor phenotype on pre-treatment MRI was associated with accuracy of post-treatment U/S. For patients who had no non-mass enhancement (NME) on pre-treatment MRI, post-treatment U/S was significantly correlated with true tumor size on pathology (r=0.65, p=0.0013). However, if NME was present on pre-treatment MRI, there was no correlation between post-treatment U/S and true tumor size (p=0.48). In our cohort, 82 patients (55.8%) had no NME on pre-treatment MRI. Conclusions: Post-treatment MRI longest diameter and post-treatment U/S longest diameter were strongly correlated. The presence of NME on pre-treatment MRI meant that post-treatment U/S was no longer a useful predictor of true tumor size, but when NME was not present, post treatment U/S longest diameter was significantly correlated with true size. Therefore, for ILC patients with no NME on pre- treatment MRI, these findings suggest that treatment response evaluation of tumor size can be done with U/S only. Future work will explore the ability of U/S to determine nodal response compared to MRI. Additionally, MRI features such as functional tumor volume may provide additional response information beyond longest diameter, and ongoing research is needed. Table 1: Characteristics of study population.Neoadjuvant chemotherapy (n=84)Neoadjuvant endocrine therapy (n=64)P-valueMean age, range53.3 (29-81)62.8 (41-89)0.0002Biomarker and receptor status0.016ER+PR+HER2-41 (54.7%)35 (62.5%)ER+PR-HER2-19 (25.33%)20 (35.7%)ER-PR-HER2-4 (5.33%)0HER2+11 (14.7%)1 (1.8%)Grade0.11122 (27.9%)24 (37.5%)250 (63.3%)29 (60.9%)37 (8.9%)1 (1.56%)Pathologic stage0.024I30 (35.7%)34 (53.1%)II36 (42.9%)14 (21.9%)III18 (21.4%)16 (25.0%)Mean follow-up time (years, 95% confidence interval)5.9 (4.9-6.9)5.1 (4.2-6.1)0.15 Citation Format: Kelly Fahrner-Scott, Jasmine M Wong, Merisa Piper, Cheryl Ewing, Michael Alvarado, Laura J Esserman, Nola Hylton, Rita A Mukhtar MD. Accuracy of ultrasound in evaluating response to neoadjuvant therapy in invasive lobular carcinoma of the breast [abstract]. In: Proceedings of the 2019 San Antonio Breast Cancer Symposium; 2019 Dec 10-14; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2020;80(4 Suppl):Abstract nr PD9-08.