Citations

Abstract Background: Biopsy diagnosis of benign breast disease (BBD) based on the most severe lesion in a sample predicts future breast cancer risk and has implications for screening and management. Lobules are the functional unit of the breast and the structures from which BBD arises. We developed and preliminarily validated an automated computational pathology algorithm to discriminate normal from BBD breast lobules as a step toward automated comprehensive characterization of benign biopsies. Methods: 152 BBD biopsies (27 training, 125 validation) from the Mayo Clinic were examined using scanned digital images of H&E stained sections. For each image, a pathologist annotated up to 10 representative lobules as normal or BBD using standard pathology criteria. A deep learning algorithm to quantify lobular features was developed using 129 lobules from 27 subjects. Nine features were identified that discriminate normal vs BBD lobules, expressing lobule size, acini size and number, acinar lumen size, proportion of lobular stroma, and capillaries. Here, we validate their performance to discriminate normal vs BBD lobules in a set of 1250 lobules from 125 subjects using area under the ROC curve (AUC) analysis. Random forest analysis was used for multivariable modeling; model performance was assessed with a tenfold cross-validation approach. Results: Median subject age was 52 years. Among the 125 validation subjects, BBD findings were nonproliferative in 39%, proliferative in 45%, and atypical hyperplasia in 16%. Sections included 552 (44%) normal lobules and 698 BBD (56%) lobules, with representation of both lobule types on each section. In univariate analyses, four individual features showed good discrimination between normal vs. BBD lobules, yielding the following AUCs: lobule size= 0.74, mean acini size= 0.75, epithelial area= 0.75, and number of acini with large lumens= 0.76. Lobule size and epithelial area were highly correlated (Spearman rank correlation r = 0.95), but both of these features showed lower correlation with mean acini size and number of acini with large lumens (each r < 0.5). With random forest modeling, number of acini with large lumens was the strongest discriminating feature, followed by mean acini size, lobule size, and epithelial area. With ten-fold cross validation of the multivariable random forest model, the overall AUC was 0.82 (95% CI: 0.79-0.85). Conclusion: Our validation showed that automated quantitative computational pathology assessment of breast lobules can discriminate normal versus BBD on a per lobule basis. This finding supports the feasibility of developing automated algorithms to classify every lesion in breast biopsies, expanding beyond visual assessment. Further studies using deep learning may reveal novel pathology features for classifying BBD biopsies with the potential to strengthen estimation of breast cancer risk. Citation Format: Amy C. Degnim, Thomas de Bel, Mark E. Sherman, Derek C. Radisky, Stacey J. Winham, Tanya L. Hoskin, Melody L. Stallings Mann, Marlene Frost, Robert A. Vierkant, Brendan T. Broderick, Ethan P. Heinzen, Rushin Brahmbhatt, Muhammad Arshad, Celine M. Vachon, Jodi M. Carter, Lori A. Denison, Daniel W. Visscher, Jeroen van der Laak. Discrimination of benign breast disease from normal lobules using an automated computational pathology algorithm [abstract]. In: Proceedings of the Annual Meeting of the American Association for Cancer Research 2020; 2020 Apr 27-28 and Jun 22-24. Philadelphia (PA): AACR; Cancer Res 2020;80(16 Suppl):Abstract nr 2113.

Abstract Background: The intrinsic immune response in triple negative breast cancer (TNBC) has both prognostic and predictive utility, and immune-related biomarkers are now actionable targets in TNBC. Current modalities for immune profiling are limited by lack of spatial context and precise quantitation. We used quantitative digital spatial profiling (DSP) to 1) characterize the immune architecture of tumoral and stromal regions in TNBC and 2) identify prognostic immune signatures. Methods: From the FinXX trial of adjuvant capecitabine therapy, tumor sections (FFPE) matched for patient characteristics, treatment arm and RFS-based outcome (N=44) were assessed with Nanostring GeoMxTM DSP. Briefly, with 3-plex immunofluorescence (pan-cytokeratin, CD45, CD68 with SYTO-13 dye for nuclei), spatially-defined CD45-enriched or CD68-enriched stromal segments and adjacent tumor segments (N=950) were selected for digital quantitation (NanoString nCounter) of 39 immune proteins (e.g. immune cell profiling proteins: CD3, CD4, CD8, CD11c, CD20, CD56, CD68, HLA-DR, PD-1, PD-L1, Granzyme B), immune drug targets (e.g. VISTA, STING, IDO1), and activation status-related proteins (e.g. ICOS, PD-L2, CD40, CD40L, CD44). Normalized, differentially-expressed (DE) proteins were evaluated, and weighted gene co-expression network analysis (WGCNA) was performed to identify modules of highly co-expressing immune proteins, and association with RFS as a categorical variable was determined. Results: 20 DE proteins in tumor segments were associated with RFS. Among them, those with the highest RFS-associated fold changes (FC) included CD20, CD40, CD56, HLA-DR, Granzyme B, and PD-L2 (FC): 2.1-4.4) (p < 0.05). In contrast, only 4 DE proteins were associated with RFS in stromal segments (Granzyme B, IDO-1, PD-L1, and PD-L2 (FC: 1.5-3.3) (p < 0.05). By WGCNA, 2 tumor-based immune modules [IM1 (lymphocyte infiltration) and IM2 (lymphocyte activation)] were identified based on their co-expression. The top three co-expressed proteins in IM1 were CD4, CD40, and CD45 (r >0.9). The three top co-expressed proteins in IM2 were PD-L1, PD-L2, and Granzyme B (r > 0.85). Together with CD56 expression, IM1 and IM2 comprised a set of 3 intratumoral immune features that associated with RFS. For recurring tumors, all 3 immune features clustered with low scores . Among non-recurrent tumors, two distinct, non-overlapping clusters with inverse characteristics were identified: high IM2, low IM1, low CD56 vs. low IM2, high IM1, and high CD56. Overall, across all 3 immune features, PDL2, CD40 and CD56 expression most strongly associated with RFS. Conclusion: With spatially-defined, precise quantification of immune proteins, we identified 3 immune features, including two distinct immune modules, IM1 and IM2 that correlate with durable RFS. These immune modules of highly co-expressing proteins were associated with RFS only when present in the tumor compartment but not tumor-adjacent, immune cell-rich stroma. Citation Format: Jodi M. Carter, Saranya Chumsri, Yaohua Ma, Xue Wang, Jennifer M. Kachergus, Douglas Hinerfeld, Heather Ann Brauer, Sarah Warren, Heikki Joensuu, Edith A. Perez, Aubrey Thompson. Quantitative digital spatial profiling reveals novel prognostic intratumoral immune signatures in triple-negative breast cancer [abstract]. In: Proceedings of the Annual Meeting of the American Association for Cancer Research 2020; 2020 Apr 27-28 and Jun 22-24. Philadelphia (PA): AACR; Cancer Res 2020;80(16 Suppl):Abstract nr 1004.

Abstract Over one million women in the United States receive biopsy diagnoses of benign breast disease (BBD) each year, which confer a 1.5–4.0-fold increase in breast cancer risk. Studies in the general population suggest that nonsteroidal anti-inflammatory agents (NSAID) lower breast cancer risk; however, associations among women with BBD are unknown. We assessed whether NSAID use among women diagnosed with BBD is associated with lower breast cancer risk. Participants included 3,080 women (mean age = 50.3 ± 13.5 years) in the Mayo BBD surgical biopsy cohort diagnosed between January 1, 1992 and December 31, 2001 who completed breast cancer risk factor questionnaires that assessed NSAID use, and whose biopsies underwent detailed pathology review, masked to outcome. Women were followed from date of BBD biopsy to breast cancer diagnosis (main outcome) or censoring (death, prophylactic mastectomy, reduction mammoplasty, lobular carcinoma in situ or last contact). Median follow-up time was 16.4 ± 6.0 years. Incident breast cancer was diagnosed among 312 women over a median follow-up of 9.9 years. Regular non-aspirin NSAID use was associated with lower breast cancer risk [HR = 0.63; 95% confidence interval (CI) = 0.46–0.85; P = 0.002] with trends of lower risk (highest tertiles of use vs. nonuse) for greater number of years used [HR = 0.55; 95% CI = 0.31–0.97; Ptrend = 0.003), days used per month (HR = 0.51; 95% CI = 0.33–0.80; Ptrend = 0.001) and lifetime number of doses taken (HR = 0.53; 95% CI = 0.31–0.89; Ptrend = 0.003). We conclude that nonaspirin NSAID use is associated with statistically significant lower breast cancer risk after a BBD biopsy, including a dose–response effect, suggesting a potential role for NSAIDs in breast cancer prevention among patients with BBD.

AbstractBackgroundLipomatosis of nerve (LN) is a complex peripheral nerve disorder characterized by fibrofatty nerve enlargement. MRI of this pathology is pathognomonic and obviates a diagnostic biopsy. Mutation in PIK3CA has been associated with LN cases with nerve‐territory overgrowth which may occur in some cases. We evaluate an association of LN of the sciatic nerve and early‐onset colonic diverticular disease and discuss the potential pathogenesis.MethodsOur institutional database was searched for LN cases. Available information of identified cases was reviewed, and cases with a confirmed diagnosis of LN affecting the lumbosacral plexus and/or sciatic nerve; available MRI of the affected nerve(s); and diverticular disease occurring in the area supplied by the nerve(s) affected by LN were further analyzed. PIK3CA mutation testing was performed on available tissue samples.ResultsWe identified 10 LN cases of lumbosacral plexus and/or sciatic nerve. Of these, three fulfilled our inclusion criteria. All three patients had concomitant colonic diverticular disease, diagnosed at a relatively young age. MRI studies of these cases showed LN involvement of the sacral nerves innervating the sigmoid colon. All three also had abnormal diagnostic workup including various GI tests and evidence of associated nerve‐territory overgrowth. Colonic tissue samples for PIK3CA mutation were negative.ConclusionWhile the pathogenesis of the colonic diverticular disease is increasingly recognized as being multifactorial, our observations are consistent with the potential role of autonomic nervous system dysfunction affecting either the pelvic floor musculature, or the colon itself (or both) in a subset of patients with early‐onset diverticular disease.

Aims Arthrofibrosis is a relatively common complication after joint injuries and surgery, particularly in the knee. The present study used a previously described and validated rabbit model to assess the biomechanical, histopathological, and molecular effects of the mast cell stabilizer ketotifen on surgically induced knee joint contractures in female rabbits. Methods A group of 12 skeletally mature rabbits were randomly divided into two groups. One group received subcutaneous (SQ) saline, and a second group received SQ ketotifen injections. Biomechanical data were collected at eight, ten, 16, and 24 weeks. At the time of necropsy, posterior capsule tissue was collected for histopathological and gene expression analyses (messenger RNA (mRNA) and protein). Results At the 24-week timepoint, there was a statistically significant increase in passive extension among rabbits treated with ketotifen compared to those treated with saline (p = 0.03). However, no difference in capsular stiffness was detected. Histopathological data failed to demonstrate a decrease in the density of fibrous tissue or a decrease in α-smooth muscle actin (α-SMA) staining with ketotifen treatment. In contrast, tryptase and α-SMA protein expression in the ketotifen group were decreased when compared to saline controls (p = 0.007 and p = 0.01, respectively). Furthermore, there was a significant decrease in α-SMA (ACTA2) gene expression in the ketotifen group compared to the control group (p < 0.001). Conclusion Collectively, these data suggest that ketotifen mitigates the severity of contracture formation in a rabbit model of arthrofibrosis. Cite this article: Bone Joint Res 2020;9(6):302–310.

e13106 Background: Chemotherapy resistance remains a significant barrier in the effective treatment of patients with triple negative breast cancer (TNBC). TGF-β signaling is a well characterized oncogenic pathway in TNBC that promotes chemoresistance by inducing epithelial to mesenchymal transition (EMT) and tumor stemness. Aurora-A kinase (AURKA) is a serine/threonine kinase responsible for chromosomal stability during mitosis. Significantly, aberrant expression of AURKA induces breast cancer progression. We hypothesized that aberrant activation of TGF-β and AURKA signaling pathways contributes to chemoresistance in TNBC by promoting EMT and tumor stemness and that dual-targeting of these oncoproteins will enhance chemosensitivity. Methods: RNA-Seq data were analyzed from patient derived xenografts (PDx) established from patients with Stage I-III TNBC who received pre-operative taxane and anthracycline based chemotherapy. Chemoresistance was defined as an RCB score of 1-3, and chemosensitivity was defined as an RCB score of 0. Unique TNBC cell lines developed from brain metastasis PDxs (TNBC M14, TNBC M25) were treated in vitro with 10nM docetaxel (DTX), 50nM LY2157299 (TGF-β inhibitor) and 50nM MLN8237 (AURKA inhibitor). EMT reprogramming was determined by measuring the expression of vimentin, ALDH activity and mammosphere growth. Apoptotic cells following drug treatment were labeled with Red Annexin-V marker and monitored in real time using the IncuCyte instrument. Results: RNA-Seq revealed that there was no baseline difference in expression of AURKA between chemoresistant and chemosensitive TNBC PDxs. However, there was a 2.7-fold increase in AURKA expression in the post-treatment PDx models compared to pre-treatment PDx models. In vitro treatment of the M14 and M25 cell lines with DTX demonstrated no significant increase in apoptotic cells compared to control, whereas treatment with the combination of DTX, LY2157299 and MLN8237 resulted in a two-fold increase in apoptosis compared to treatment with DTX alone (p = 0.0068 M14, p = 0.003 M25). Dual-targeting with LY2157299 and MLN8237 reduced the expression of vimentin and ALDH activity. Conclusions: TGF-β and AURKA play a central role inducing EMT and a stem-cell-like phenotype in TNBC that confers chemoresistance. AURKA is up-regulated after exposure to chemotherapy in chemoresistant TNBC PDx models. LY2157299 and MLN8237 reduce TNBC stemness in M14 and M25 cell lines and enhance DTX chemosensitivity. Dual-targeting of TGF-β and AURKA is a potentially promising approach in chemoresistant TNBC.

592 Background: Immune responses in the tumor microenvironment have prognostic and predictive value in BC. However, the potential of immune responses observed in peripheral blood as biomarkers in BC remains unclear. We have shown that a higher frequency of circulating monocytes and a lower frequency of antigen-experienced memory CD8+ T cells are associated with response to NAC in triple negative BC (Leon-Ferre et al SABCS 2019). Here, we used cytometry by time-of-flight (CyTOF) to evaluate associations between circulating immune cells, clinical features and response to T-based NAC in HER2+ BC. Methods: PBMC suspensions from 36 pts with stage I-III HER2+ BC were prospectively collected prior to initiation of T-based NAC, stained with 29 metal-tagged antibodies optimized to identify major human immune cell subsets, and acquired in the Helios CyTOF instrument. Differential abundance analysis of immune cells by clinical characteristics and by NAC response was evaluated using Wilcoxon rank sum test. % of immune cell subsets is presented as % of all PBMCs. Results: Most pts presented with ER- tumors (56%), measuring > 5cm (64%) and with nodal metastases (78%). After NAC, 16 pts (44%) achieved pathologic complete response (pCR). Analysis of preNAC PBMCs demonstrated a significantly higher number of B cells (8% vs 5%, p = 0.05) and effector memory CD8+ T cells (CD45RA-/CCR7-, 3 vs 1%, p = 0.02) in pts with pCR compared to those with residual disease. Of the B cell subsets, naïve B cells (CD24-/CD27-) were higher in pts who achieved pCR vs not (7% vs 4%, 0 = 0.04). Regarding clinical characteristics, cN+ pts at presentation exhibited a lower number of peripheral blood T cells compared to cN- pts (47% vs 63%, p = 0.03). Of the T cell subsets, overall CD4+ and naïve CD4+ T cells (CD45RA+/CCR7+) were lower in cN+ vs cN- pts (31% vs 45%, p = 0.05; and 11% vs 24%, p = 0.04). We also observed differences in CD56+/CD16- NK cells by ER status (ER- 1% vs ER+ 3%, p = 0.01), and a moderate negative correlation between age and % circulating CD8+ T cells (rho -0.4669, p = 0.004). Conclusions: Distinct peripheral blood immune cell profiles are observed in HER2+ BC at diagnosis, and are associated with response to T-based NAC and initial clinical characteristics. Notably, pts who later achieved pCR had a relative abundance of B cells and effector memory CD8+ T cells at diagnosis. These data suggest that immune cell phenotyping in peripheral blood may have potential as a biomarker to predict response to NAC in BC.

510 Background: Several studies have established the critical role of preexisting immune response in triple negative breast cancer (TNBC). Most studies evaluated the tumor infiltrating lymphocytes in stroma. However, limited data are available with regards to the importance of specific subtypes and spatial distribution of these immune infiltrates. Methods: NanoString IO360 gene expression analysis and Digital Spatial Profiling (DSP) were used. DSP was used to quantify 39 immune-related proteins in stromal and tumor-enriched segments from 44 TNBC samples from the FinXX trial (NCT00114816) and 335 samples from the Mayo Clinic (MC) cohort of centrally reviewed TNBC (Leon-Ferre BCRT 2018). In FinXX trial, 22 patients with recurrence and 22 patients without recurrence were included. In MC cohort, 217/335 patients received adjuvant chemotherapy while 118 patients had surgery only without adjuvant chemotherapy. Regions were segmented based on pancytokeratin staining. The general linear model was used for statistical analysis of differential expression with recurrence free survival (RFS) as a categorical variable (recur yes or no). Kaplan-Meier estimates and Cox regression models were also used for analysis. Results: In the FinXX trial, using global gene expression analysis with IO360, there was no signature significantly associated with RFS. However, using DSP, high protein expression of CD56 in the tumor-enriched segments was associated with significant improvement in RFS (HR 0.26, 95%CI 0.09-0.78, p 0.01). Nevertheless, CD56 expression in the stroma (HR 0.66, 95%CI 0.29-1.53, p 0.33) and all segments (HR 0.53, 95%CI 0.23-1.25, p 0.14) was not significantly associated with improved outcome. We further validated these findings in the MC TNBC cohort where intratumoral CD56 expression was associated with a significant improvement in RFS (HR 0.23, p 0.002) but not stromal CD56 (p 0.79). Interestingly, when evaluating the MC TNBC cohort according to receipt of chemotherapy, intratumoral CD56 was associated with improved outcome only in patients who received chemotherapy (p 0.02 vs. 0.07). In both cohorts, higher expressions of intratumoral PD-L1, HLA-DR, and CD8 were associated with improved outcome. Conclusions: Using an in-depth analysis with spatially defined context, we identify that intratumoral CD56-positive NK cells are associated with improved outcome in TNBC. Our study highlights the potential role of NK cells in TNBC and future implications for biomarkers and therapeutic targets.Support: W81XWH-15-1-0292, P50CA116201-9, P50CA015083. Clinical trial information: NCT00114816 .

Abstract Background: TNBC is the most aggressive sub-type of breast cancer and has limited targeted therapies. Atezolizumab was recently FDA-approved for patients with advanced PD-L1+ TNBC. In a well-characterized cohort of patients with early-stage TNBC, we used the FDA-approved PD-L1 SP142 companion assay for atezolizumab to score whole tumor sections for PD-L1 expression. We determined the frequency and scoring distribution of PD-L1 in these tumors and evaluated PD-L1 status for associations with clinicopathologic variables and outcomes. Methods: Whole tumor sections from 498 women with surgically-treated, non-(distant) metastatic TNBC (ER/PR &lt;1% & HER2-negative by central review) were semi-quantitatively scored for % of PD-L1-staining tumor cells (TC) or tumor-associated immune cells (IC) (PD-L1 SP142, Ventana, performed in CAP/CLIA-certified clinical laboratory). Per FDA assay guidelines, PD-L1+ tumors were defined as ≥1% of tumor-associated immune cells (IC) staining per tumor area. Baseline clinical characteristics, histopathologic features, Ki-67 proliferative indices (PI), and stromal tumor infiltrating lymphocytes (TILs) were correlated with PD-L1 status. The impact of tumoral PD-L1 status on invasive disease-free survival (IDFS) and overall survival (OS) was assessed using Cox proportional hazards models. Results: The majority of tumors were pT1-T2 (95%), pN0-1 (85%), and grade 3 (90%) with Ki-67 &gt;15% (80%). 228/498 (46%) tumors were PD-L1+, distributed as IC score 1-5%: 21%, 6-10%: 13%, 11-20%: 5% and &gt;20%: 7%. 270/498 (54%) were PD-L1-. 43 (8.6%) of tumors had PD-L1 staining in TC (all but 1 were also IC+). PD-L1 IC expression was associated with larger tumor size (p = 0.005), higher nodal stage (≥pN1: 43% vs. 31%, p &lt; 0.001), higher grade (grade 3: 95% vs. 86%, p&lt; 0.001), higher Ki-67 PI (&gt;15%: 88% vs. 69%, p &lt;0.001), and stromal TILs (median score: 40% vs. 10%; p &lt; 0.001). PD-L1 IC status was not associated with patient age. Among histologic subtypes, PD-L1+ tumors were more frequently medullary type (31% vs. 6%) and less frequently metaplastic type (4% vs 10%). At time of this analysis, median follow-up for IDFS and OS were 11 years and 7.2 years, respectively. By univariable analysis, absence of PD-L1 IC expression was associated with poorer IDFS (median: 15.5 vs. 20 years, p = 0.04) and OS (9.5 vs. 17.5 years, p &lt; 0.05).By multivariable analysis, tumor size, stromal TILs, nodal status, and adjuvant chemotherapy (CTX) were significantly associated with IDFS. Age, stromal TILs, nodal status, and receipt of adjuvant CTX were associated with OS. PD-L1+ tumors trended toward improved IDFS (statistical significance not reached), HR: 0.68 (95% CI: 0.45-1.03, p= 0.068) but not OS: 0.90 (95% CI: 0.52-1.57, p= 0.32). Conclusion: Using the PD-L1 SP142 companion assay, almost half of early-stage TNBC were PD-L1+ (IC). However, PD-L1 expression correlated with increasing tumor size and nodal burden, as well as histologic grade, Ki-67 PI, stromal TILs and lymphocyte-enriched subtypes. PD-L1 was associated with improved IDFS and OS in univariable analysis but not multivariable analysis. The majority of the PD-L1+ tumors (73%) had IC scores ≤10%, with heterogeneous tumoral staining. Given the low (1%) assay threshold for positivity, these data emphasize the utility of evaluating PD-L1 in whole tumor sections rather than tissue cores. Citation Format: Jodi M Carter, Mei-Yin C Polley, Jason P Sinnwell, Roberto A Leon-Ferre, Fergus J Couch, Krishna R Kalari, Judy C Boughey, Minetta C Liu, Daniel W Visscher, Matthew P Goetz. Frequency, characteristics and prognostic factors of PD-L1+ triple negative breast cancer using the PD-L1 SP142 companion assay [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 PD1-08.