Abstract P1-15-15: Accuracy of MRI after neoadjuvant therapy for invasive lobular carcinoma of the breast

Abstract

Abstract Background: Invasive lobular carcinoma of the breast (ILC) has higher rates of false negative imaging than invasive ductal carcinoma, and lower rates of neoadjuvant therapy (NAT) use. We evaluated the accuracy of Breast Imaging Reporting And Data System (BIRADS) findings on magnetic resonance imaging (MRI) after either neoadjuvant chemotherapy or endocrine therapy, and determined whether imaging change correlates with disease free survival. Methods: We queried a database of 674 ILC cases treated at UCSF from 1981-2017 and identified all patients treated with NAT. We reviewed MRI reports and recorded BIRADS descriptors of findings, maximal tumor diameter for mass or non-mass enhancement (NME), and subjective radiologist comments on progression or improvement. We used the t-test, chi-squared test, Pearson's correlation, and Kaplan Meier survival estimates to evaluate the accuracy of MRI after NAT compared to true tumor size on pathology, and the relationship between imaging change and disease free interval in Stata 14.2. Results: Of 136 patients with ILC treated with NAT, we included 101 women who had a post-treatment breast MRI report available. Of these, 58.4% received neoadjuvant chemotherapy, and 41.6% neoadjuvant endocrine therapy. After NAT, MRI findings were mass only in 43%, both mass/NME in 33%, NME only in 18%, and neither in 5%. Maximal diameter of mass on post-treatment MRI underestimated true size by a mean of 3.3 cm (range -3.6 to 15.3 cm). NME size on post-treatment MRI underestimated true size by a mean of 1.87 cm (range -7.2 to 9.7 cm). Mass size on MRI underestimated true size by ≥1 cm in 61.5% of cases; this size discrepancy was associated with increased positive margins (46.4% versus 20%, p=0.011). NME size on MRI underestimated true size by at ≥1 cm in 65.6%. The correlation coefficient between mass size on MRI and true size was 0.34 (p=0.0041), which increased to 0.67 (p<0.0001) when excluding those with associated NME. The correlation coefficient between NME size on MRI and true size was 0.28 (p=0.1239). Subjective progression on post-treatment MRI was associated with increased recurrence rates (80% versus 18.3%, p=0.001). In those with subjective improvement on MRI, there was a trend towards longer disease free interval (89% versus 73% disease free at 4 years, p=0.13). Table 1.Patient and tumor characteristics. Neoadjuvant chemotherapy (n=59)Neoadjuvant endocrine therapy (n=42)P valueMean age (yrs, 95% CI)53.6 (50.9-56.3)61.3 (58.4-64.1)0.0002Subtype  0.105ER+ PR+ HER2-29 (53.7%)23 (62.16%) ER+ PR- HER2-14 (25.9%)13 (35.14%) ER- PR- HER2-1 (1.85%)0 HER2+10 (18.5%)1 (2.7%) Grade  0.076114 (25%)16 (38.1%) 237 (66.1%)26 (61.9%) 35 (8.93%)0 Surgical stage  <0.001I17 (28.81%)28 (66.67%) II26 (44.07%)6 (14.29%) III16 (27.12%)8 (19.1%) Mean follow-up time (yrs, 95% CI)5.6 (4.53-6.75)5.1 (3.84-6.27)0.48 Conclusions: Maximal tumor diameter on MRI after NAT in ILC vastly underestimates true tumor size. While these findings suggest using caution when using an MRI for surgical planning in patients with ILC, particularly if there is associated NME, the trend towards improved disease free survival in those with a subjective improvement is intriguing and suggests that MRI changes could become an early predictor of outcomes. Citation Format: Fahrner-Scott KE, Wong JM, Piper M, Ewing C, Alvarado M, Esserman LJ, Hylton N, Mukhtar RA. Accuracy of MRI after neoadjuvant therapy for invasive lobular carcinoma of the breast [abstract]. In: Proceedings of the 2018 San Antonio Breast Cancer Symposium; 2018 Dec 4-8; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2019;79(4 Suppl):Abstract nr P1-15-15.