Iwase, Toshiaki

e15060 Background: In metastatic Breast Cancer (mBC) Circulating Tumor Cells (CTCs) are an established prognostic indicator of patients (pts) with highly aggressive disease, poor clinical outcomes, and poor response rates to therapy. However, CTCs are typically found in < 20% of mBC pts and many pts without CTCs also progress. Cancer associated macrophage-like cells (CAMLs) are inflammatory tumor macrophages also found in blood, in > 90% of mBC, and are indicators of poor clinical outcomes independent of CTCs. We conducted a prospective study to model CTC & CAML subtypes from 212 mBC pt blood samples prior to induction of new systemic therapies to optimize and validate risk models based on 2-year outcomes of Progression-free survival (PFS) & Overall survival (OS). Methods: An observational multi-institutional prospective study was conducted on 212 mBC pts prior to induction of a new line of therapy who were progressing on their current therapy. Anonymized (7.5ml) whole blood samples were taken prior to therapy induction and filtered by CellSieve filtration. Quantities and subtypes of CTCs & CAMLs were categorized by their expression of CK+/CD45+/CD14+ for CAMLs and CK+/CD45-/CD14- for CTCs. PFS (RECIST v1.1) and OS hazard ratios (HRs) were calculated by censored univariate & multivariate analysis at 2 years. Results: Pt median age was 51 years (range 25-92); TNBC (44%), ER/PR+ (30%), HER2+ (15%). Pts were treated with chemotherapy alone (18%), immune checkpoint inhibitor (46%), targeted (40%), hormone (8%) and unknown (5%). 38% (81/212) pts had ≥1 CTCs which correlated with significantly poorer PFS (HR = 1.9; 95%CI 1.3-2.7; p = 0.0016) and OS (HR = 2.0; 95%CI, 1.3-3.0; p = 00025), with the most significant risk at ≥8 CTCs (14% pts), PFS (HR = 3.8; 95%CI, 2.1-6.9; p < 0.0001) and OS (HR = 7.2; 95% CI, 3.4-14.9; p < 0.0001). Independently, 93% (n = 196/212) of pts had ≥1 CAML, with a threshold of >40um size (84% pts) had significantly poorer PFS (HR = 2.0; 95% CI, 1.2-3.1; p = 0.0032), but not OS (HR = 1.5; 95%CI, 0.9-2.7; p = 0.1169). However, a threshold of >125um size CAML (43% pts) also had significantly poorer PFS (HR = 1.8; 95% CI, 1.2-2.7; p = 0.0013) and OS (HR = 1.6; 95% CI, 1.0-2.5; p = 0.0233). By combining models, the best outcomes were seen in pts with 0 CTCs & 0 CAMLs (mPFS = 12.6 & mOS = 21.3), followed by 0 CTCs & ≥40um CAMLs (mPFS = 5.2 & mOS = 17.5) and 0 CTCs & ≥125um CAMLs (mPFS = 4.9 & mOS = 15.2). Poorer outcomes were seen with ≥1 CTCs (mPFS = 3.3 & mOS = 13.5) and the worst outcomes with ≥8 CTCs (mPFS = 2.5 & mOS = 4.8). Conclusions: CTCs were uncommon in mBC pts but correlated with very poor clinical outcomes. In parallel analysis, CAMLs were common with CAML size correlating with increasingly poorer outcomes. By combining CTC & CAML subtypes, mBC pts were more accurately stratified by risk of progression and death. Additional multivariate studies correlating treatment class and tumor response rates are ongoing.

e15045 Background: Increased cancer cell mitosis in tumor tissue is widely regarded as an indicator of aggressive disease, a predictor of shorter survival, and an indicator of response to specific anthracycline based chemotherapies. Circulating tumor cells (CTCs) are cancer cells that detach from primary tumors, enter the bloodstream, and seed metastatic sites, which in patients (pts) with metastatic breast cancer (mBC) are also predictors of poor prognosis and reduced therapy response. Furthermore, small preliminary studies have identified a specific subtype of mitotic CTCs that appears to correlate with poor survival rates compared to CTC detection alone. However, larger studies have yet to be conducted to validate the clinical utility of mitotic CTCs. In this large multi-institutional prospective study, we isolated CTCs from the blood of n = 155 mBC pts to subtype mitotic CTCs to assess progression-free survival (PFS) and overall survival (OS) over 2 years. Methods: Peripheral blood samples (7.5ml) were prospectively collected from n = 155 mBC pts progressing on systemic therapies and prior to new systemic therapy induction (i.e., chemotherapy n = 90, hormone therapy n = 13, immunotherapy n = 51, targeted therapy n = 62). Samples were processed using CellSieve microfilters to isolate CTCs via size exclusion and fluorescently stained for CD45, Cytokeratin, and DAPI. CTCs were imaged, enumerated, and subtyped based on the presence of ≥1 mitotic event, using established visual indicators. PFS and OS were assessed over 2 years by censored univariate and multivariate analyses. Results: CTCs were detected in ~45% (n = 70/155) of pts, while mitotic CTCs were identified in a subset of this group, ~23% (n = 35/155). The presence of any CTC was associated with poor outcomes (PFS HR = 2.16, 95%CI = 1.4-3.2, p = 0.0002 and OS: HR = 2.51, 95%CI = 1.5-4.0, p = 0.0002). The presence of mitotic CTCs correlated with significantly poorer outcomes (PFS: HR = 5.6, 95%CI = 3.1-10.2, p < 0.0001 and OS: HR = 7.3, 95%CI = 3.7-14.5, p < 0.0001). Overall, pts with no CTCs had a median PFS (mPFS) of 7 months and median OS (mOS) of 24 months, pts with non-mitotic CTCs had a mPFS of 6 months and mOS of 15 months; pts with both CTCs and mitotic CTCs had a mPFS of 3 months and mOS of 5 months. Notably, 97% of pts with a mitotic CTC progressed within 6 months, regardless of systemic therapy type. Interestingly at 2 years, a sole non-progressive pt with a mitotic CTC was receiving alpelisib (a specialized PI3KCA inhibitor). Conclusions: Overall, the detection of CTCs indicated poor outcomes in pts, with mitotic CTCs correlating with significantly more aggressive disease and further reduced survival rates regardless of systemic therapy type. This study highlights the need for larger, more comprehensive studies to expand upon and validate these findings.

3053 Background: Recently, early studies have identified circulating stromal cells (CStCs), i.e. Cancer Associated Macrophage-Like cells (CAMLs), in patients (pts) with newly diagnosed invasive cancers, which are phagocytic myeloid immune cells emanating from primary tumors. However, no study has evaluated CAMLs as a pan solid tumor screening tool nor their clinical impact on pt outcomes. We analyzed the blood of untreated newly diagnosed cancer pts (n=441), or in non-malignant conditions (i.e. pancreatic cysts, n=77), or healthy controls (n=100). We found that CAMLs are prevalent (82%) in cancer pt blood, absent in healthy controls (0%) and less common in non-malignant conditions (36%), with correlations to worse clinical outcomes. Methods: Anonymized peripheral blood was taken from n=441 pts before or after confirmation of invasive malignancy [stage I (n=105), stage II (n=125), stage III (n=114), stage IV (n=97) with pathologically confirmed lung (n=84), pancreas (n=61), breast (n=73), prostate (n=65), esophageal (n=34), Renal Cell (n=29), Sarcoma (n=27), or Other (n=68). In addition, anonymized blood was taken from pts with untreated non-malignant conditions including benign breast masses (n=19), Lupus (n=11), liver Cirrhosis (n=6), rising PSA (n=24), or other benign mass (n=17); and from healthy control volunteers (n=100). CAMLs were isolated by CellSieve microfiltration, defined as enlarged polyploid cells with cytokeratin+ and/or CD45/CD14+. Pts were monitored for 0-8 years (median=2.6) for progression free survival (PFS) and overall survival (OS). Results: CAMLs were found in 82% of all cancer pts averaging 5.4 CAMLs, specifically, in 66% of Stage I, 85% Stage II, 84% Stage III, and 92% Stage IV. No CAMLs were found in any healthy controls, but were in 36% of non-malignant pts. CAML sensitivity in cancer vs healthy was 82% (CI95% 78-85%), specificity=100% (CI95% 96-100%), PPV=100% (CI95% 99-100%), NPV=55% (CI95% 50-60%). CAML sensitivity in cancer vs benign was 82% (CI95% 78-85%), specificity=64% (CI95% 32-44%), PPV=93% (CI95% 90-95%), NPV=38% (CI95% 32-44%). CAML presence was significantly associated with worse PFS (HR=2.0, 95%CI 1.4-2.9, p<0.0001) & OS (HR=2.2, 95%CI 1.4-3.3, p=0.0004), with engorged phagocytic CAMLs (>50um), highly significant for worse PFS (HR=3.3, 95%CI 2.5-4.5, p<0.0001) & OS (HR=3.0, 95%CI 2.1-4.3, p<0.0001). Further, in benign pts positive for CAMLs, 2 pts were in situ, 16 had histologies with increased cancer risk (i.e. IPMN, atypical hyperplasia), and 5 pts died within 3 years. Conclusions: CAMLs appear to be a highly sensitive blood based biomarker that can identify invasive carcinoma in all stages of cancer regardless of subtype, but are not found in healthy controls & rare in non-malignant conditions. Further, enlarged phagocytic CAMLs appeared to correlate with shorter PFS and OS prior to treatment induction.

3054 Background: Tumor Macrophage Fusion Cells (TMFCs) are hybrids fusions of tumor cells and macrophage/immune cells that have recently been documented in advanced solid tumors, such as metastatic Breast Cancer (mBC). Further, TMFCs are found in the blood of patients (pts) as a CD45+/CD14+ binucleated subtype of standard circulating tumor cells (CTCs), which are Cytokeratin+ (CK+) & CD45 negative, but distinct from highly hyperploidy cells known as cancer associated macrophage-like cells. However, no study has explored the clinical meaning of TMFCs in blood. To better elucidate the clinical meaning of TMFCs, we prospectively pooled 137 mBC pt samples prior to induction of new therapy, ie baseline (BL), and after therapy induction (FU). Both TMFCs or standard CTCs were identified and enumerated to analyze their prognostic value for progression-free survival (PFS) and overall survival (OS). Methods: In this prospective study, we collected 7.5 mL blood samples from n=137 mBC pts enrolled in prospective trials from multiple institutions before starting new treatment lines for newly progressive mBC. If possible, an optional FU sample was collected (n=73) after BL (median=4.9 weeks). TMFCs & CTCs were isolated using a CellSieve microfilter and differentiated by staining for CK, CD45, CD14, and DAPI. Cox proportional regression hazard ratios (HRs) with 95% Confidence Intervals (CI) for PFS & OS by univariate & multivariate analyses were based on RECIST v1.1, determined by local institutional pathologist over 24 months (M). Results: TMFCs were detected in 26.3% (n=36/137) at BL and associated with significantly worse PFS (HR=3.3, CI 1.9-5.7, p<0.001) & OS (HR=2.7, CI 1.4-5.1, p=0.005). TMFCs were also detected in 27.4% (n=20/73) at FU (n=73) and prognostic for PFS (HR=5.2, CI 2.3-11.9, p<0.001) & OS (HR=5.1, 1.8-14.2, p=0.004). Further, pts with increases in TMFCs between BL and FU had worse PFS (HR=4.3, CI 1.9-10.1, p=0.001) & OS (HR=4.5, CI 1.6-12.7, p=0.011). In contrast, pts with CTCs were found in 39.4% (n=54/137) at BL, which were prognostic for worse PFS (HR=2.0, CI 1.3-3.2, p=0.005) and not prognostic for OS (HR=1.7, CI 0.9-3.0, p=0.103). At FU, pts with CTCs were found in 38.4% (n=28/73), which were prognostic for worse PFS (HR=3.6, CI 1.8-7.3, p=0.001) & OS (HR=3.2, CI 1.3-8.0, p=0.021). Overall at BL, pts without CTCs or TMFCs had a 5.7 M mPFS & >24M mOS, CTCs alone had a 5.0 M mPFS & 20.3 M mOS, TMFCs had 2.0 M mPFS & 5.8 M mOS. At FU, pts without CTCs or TMFCs had a 5.8 M mPFS & 14.5 M mOS, CTCs alone had a 3.3 M mPFS & 14.6 M mOS, while TMFCs pts had a 1.7 M mPFS & 4.8 M mOS. Conclusions: TMFCs appear to represent a CTC subgroup with additional prognostic value to standard CTCs, that may correlate with significantly faster PFS and OS versus pts with CTCs alone, or without CTCs. Subtyping treatment regimens is ongoing to determine if the TMFC populations are responsive to specific therapy types in these mBC pts.

e14548 Background: Micronuclei (MN) are small, excised DNA fragments resulting from biological DNA repair mechanisms in response to internal chromosomal instability, such as those caused by genotoxic events (i.e. chemotherapy and radiation), and whose formation in tumor stromal cells in blood circulation is prognostic for worse clinical outcomes (Fenech 2020). Further, mutagenic events occurring during the formation of MN are associated with the development of cancer neoantigens, which are a major element in the programmed cell death ligand (PD-L1) pathway that influences response to immunotherapies. We hypothesized that MN can be detected in metastatic breast cancer (mBC) patients (pts) within circulating stromal cells (CStCs) that prognosticates for clinical outcomes and may associate with PD-L1 cell expression. Methods: We initiated a prospective pilot study using a compiled set of n=124 mBC samples, from multiple institutions, enrolled prior to starting new lines of treatment for progressive mBC. We enumerated MN formation found in CStCs from whole peripheral blood (7.5mL), filtered for CStCs, then stained for PD-L1 and DAPI. DAPI was used to identify MN, defined by small (<3 µm) DAPI+ circular formations within the cytoplasm, separate from primary nuclei. In conjunction, all cells were stained with PD-L1 with expression normalized to background and evaluated against MN by T-test analysis. MN presence was compared to all available clinical variables, including induction of PD-L1 immunotherapy. Pts’ progression-free survival (PFS) and overall survival (OS) hazard ratios (HRs) were analyzed by censored univariate and multivariate analyses based on RECIST v1.1 over 24 months. Results: MN were identified in 56% (n=70/124) of CStCs and their presence was significantly prognostic for worse PFS (HR=2.4, CI= 1.53-3.76, p<0.001) & OS (HR= 1.89 CI=1.07-3.34, p=0.039) over 24 months. Overall, mPFS of pts with MN was 2.8 months and without MN was 12.1 months. The mOS of pts with MN was 12.7 months and without MN was 20.8 months. Additionally, MN positive CStCs had a significantly higher PD-L1 expression than MN negative CStCs (p<0.001), with 67% of MN positive pts (n=46/66) having high PD-L1 expression while 30% of MN positive pts (n=20/66) having low PD-L1. Further, evaluation of PD-L1 in CStCs showed a linear relationship (R=0.9806) between MN number and PD-L1 in cells, with 0 MN average=0.016, 1 MN=0.200, 2 MN=0.318, and ≥3 MN=0.400. Conclusions: These initial findings suggest that MN are detected in pts with advanced mBC and may identify pts with faster progression and worse survival outcomes. Further, the number of MN in CStCs appears to correlate with higher PD-L1 expression within cells. Follow up clinical information and subtyping studies are ongoing to also evaluate of the effect of MN positive CStCs on PD-L1 immunotherapies in these pts.