Search

e13056 Background: Invasive lobular carcinoma (ILC) accounts for only 10-15% of all invasive breast cancers but has distinct clinicopathologic characteristics and genomic profiles. In particular, metastatic lobular cancers (mILC) have unique sites of metastasis and it is unclear if the response to initial endocrine therapy differs from metastatic ductal cancers (mIDC). Therefore we have undertaken a single-institution, retrospective analysis to compare overall outcomes and response to initial endocrine therapy (ET) in patients (pts) with metastatic ILC and IDC. Methods: An IRB approved retrospective review of medical records was conducted evaluating pts treated for metastatic IDC and ILC at The Ohio State University Comprehensive Cancer Center from January 1, 2004 to December 31, 2014. Pts diagnosed with mIDC were matched on age, year of diagnosis, estrogen receptor/progesterone receptor and HER2 status and site of metastasis 2:1 to patients diagnosed with mILC. Overall survival (OS) was defined as the time from metastasis to death or last known follow-up. Progression-free survival (PFS) was defined as time from metastasis to progression on first-line ET. Time to chemotherapy (TTC) was defined as time from starting ET for metastasis to initiation of chemotherapy. Kaplan Meier (KM) methods were used to calculate median OS, PFS and TTC. Results: A total of one hundred sixty one pts with metastatic breast cancer were included in this analysis. The demographic features between the two groups were well balanced and included in the table below. The median OS was 2.6 yrs (95% CI: 1.55, 3.22) for mILC and 2.2 yrs (95% CI: 1.95, 2.62) for mIDC. A subset of 111 patients who started on endocrine therapy were used in the PFS and TTC analyses. The median PFS following first-line ET was 2.2 yrs (95% CI: 0.1.0, 2.7) for mILC and 1.4 yrs (95% CI: 0.91, 1.90) for mIDC. Median TTC was 2.1 yrs (95% CI: 1.71, 4.92) for mILC and 2.4 yrs (95% CI: 1.90, 4.77) for mIDC. There was no statistically significant difference in outcomes between the two groups. Conclusions: Outcomes in patients with ILC and IDC have been varied, with several studies reporting that patients with ILC have worse outcomes and response to chemotherapy. Our retrospective study examining outcomes in mILC in comparison with mIDC showed no difference in OS. Given the concern of resistance to conventional therapies in patients with lobular cancers, it is reassuring to see that the median PFS on first line ET and TTC was similar to metastatic ductal cancers.[Table: see text]

1025 Background: Analysis of circulating tumor DNA (ctDNA) over time allows non-invasive evaluation of tumor genomic evolution. We characterize changes in tumor fraction (TFx), somatic copy number alterations (SCNAs), and somatic mutations (muts) over time in patients (pts) with BRCA1/2 muts and metastatic breast cancer (mBC) who received a PARP inhibitor (PARPi) or platinum chemotherapy. Specifically, we seek to identify the frequency of BRCA1/2 reversion muts. Methods: Pts with mBC and germline or somatic BRCA1/2 muts were identified on a banking protocol of prospectively-collected serial samples of blood and plasma. Control pts without a BRCA1/2 mut were matched 2:1 by age and hormone receptor (HR) status. Ultra-low-pass whole genome sequencing (ULPWGS) with 0.1x depth was performed on all plasma samples (n = 103) and the ichorCNA algorithm was used to determine TFx and SCNAs. Targeted panel sequencing (TPS) of 402 cancer-related genes was performed at 10,000x depth on plasma samples, and one blood sample per pt. The panel includes BRCA1/2 and 38 other DNA damage repair (DDR) genes. Somatic muts were identified by joint calling with Mutect2 across plasma timepoints with paired pt normal blood. Germline variant calling from TPS on blood with HaplotypeCaller was used to confirm germline muts in BRCA1/2.Results: We identified 10 pts with mBC with a germline (n = 7) or somatic (n = 3) BRCA1 (n = 2) or BRCA2 (n = 8) mut and banked blood and plasma samples at 3-9 timepoints at a median of 8 weeks apart (range 1-43). The control cohort of 20 pts with mBC and wildtype BRCA1/2 was well matched by age and HR status. All pts with BRCA1/2 muts received a PARPi and/or platinum chemotherapy at some point during sample collection. Half of control pts received platinum chemotherapy. Germline BRCA1/2 muts were confirmed in all 7 pts with known germline muts. Among the BRCA1/2 mut cohort, median TFx was 0.04 (range 0-0.57) with 20% of samples having TFx > 0.10. A median of 1.5 (range 0-39) somatic muts per pt were found in DDR genes. Four pts (40%) had secondary non-reversion muts in BRCA1/2. A reversion mut of a germline BRCA2 mut, restoring the open reading frame of BRCA2, was discovered at the last timepoint from 1 pt while receiving carboplatin. A germline BRCA1/2 reversion mut in this cohort occurred in 2.3% of samples, 14.3% of pts. There was no significant difference in the percent of genome with a SCNA between the first and last time point, nor before and after PARPi/platinum. The somatic mut landscape and clonal evolution of TPS using PyClone will be presented. Conclusions: Evaluation of serial ctDNA samples for TFx, SCNAs, and somatic muts from banked plasma and blood from pts with mBC is feasible. The frequency of reversion muts in BRCA1/2 was low, suggesting that either their incidence is low or ctDNA TPS is not sensitive enough to detect them. Secondary non-reversion muts in BRCA1/2 and other somatic DDR muts were more common. SCNAs were stable over time.

e18739 Background: Hair loss is a well-known side effect of chemotherapy. The Paxman Hair Loss Prevention System, a scalp cooling device, has been shown to be effective in reducing chemotherapy induced alopecia in patients receiving chemotherapy (Nangia, JAMA, 2017). The National Comprehensive Cancer Network and the European Society for Medical Oncology guidelines have recommended scalp cooling as category 2A and 2B options, respectively. Methods: The real world use of scalp cooling using the Paxman device, as documented by orders through the Paxman Hub during the years of 2017-2020 was summarized. Descriptive statistics were used to summarize demographics and utilization. Results: Data from 6649 patients who used scalp cooling were reviewed. Patients with breast cancer were the most common users of scalp cooling (78%, n=5197) followed by gynecology (12%, n=775), gastrointestinal (3%, n=201), lung (1%, n=81) and genitourinary (1%, n=52). The majority of patients were between the ages of 45-65 (55%), followed by 65-74 (18%), older than 75 (5%), and 25-44 (2%). Average number (#) of cycles of cooling completed was 6.53 (range of average # of cycles 4.50-12). Scalp cooling with this device was commonly used in 39 out of 50 states. Conclusions: This is the largest report of scalp cooling usage in the real world setting in the USA, including scalp cooling usage in older adults. Uptake of scalp cooling across various cancers has not been uniform and this deserves further study.

e12596 Background: Triple negative breast cancers (TNBC), characterized by the lack of expression of estrogen receptors and progesterone receptors as well as human epidermal growth factor receptor 2, are associated with high distant recurrence rate and death. As a result, the majority of patients with TNBC are treated with perioperative chemotherapy with the goal of eradicating micrometastases and preventing distrant relapse. The preoperative systemic therapy offers the advantages of permitting an assessment of chemo-sensitivity, increased rates of breast conserving surgery and the ability to adapt postoperative therapies depending on the response. Recently, neoadjuvant chemotherapy has been used at an increasing frequency. The response to neoadjuvant chemotherapy, as measured by the residual cancer burden index, for example, is correlated with the long-term prognosis of TNBC and HER2 expressing breast cancers. Previous studies suggest that tumor-infiltrating lymphocytes (TILs) may correlate with pathological complete response (pCR) rates in TNBC patients treated with neoadjuvant chemotherapy. The pathologic evaluation of TILs in TNBC has been recommended by the International TILs working group since 2014. In this study we sought to analyze the association of TILs with pCR in a cohort of TNBC patients treated with neoadjuvant chemotherapy. Methods: An IRB-approved single-institution retrospective analysis was performed on 127 patients diagnosed with TNBC who received neoadjuvant anthracycline and taxane based chemotherapy at the Ohio State University Comprehensive Cancer Center between January 1st, 2012 and November 31st, 2018. We analyzed TILs as a continuous variable as a part of a secondary analysis of this data. Whole tissue sections from archived H&E stained glass slides were scanned using Philips UltraFast Scanner at ×40 magnification with a single-focus layer. TIL scoring was performed according to guideline recommendations from the International TILs Working Group (2014). Results: A total of 127 female patients with TNBC were identified. The median age at diagnosis was 52.0 years (range 32.0, 74.0) and patients were predominately white (103, 81%), post-menopausal (68, 53.5%) and presented with invasive ductal cancer (113, 89%), stage II (88, 69%), and high grade (108, 85%). Of those patients, 56 had TILs measurement available. pCR was associated with statistically higher level of TILs in core biopsies taken prior to chemotherapy had (Wilcoxon rank-sum test, p = 0.04). Conclusions: The long-term prognosis of patients with TNBC is predicted by the response to neoadjuvant chemotherapy. Consistent with other studies, our study revealed that TILs are associated with a higher probability of pCR. Our future goals are to identify which TIL subsets correlate best with pCR and to identify the mechanism for the increased chemotherapy responsiveness of lymphocyte-infiltrated tumors.

e12528 Background: Chemotherapy promotes oxidative stress and neuroinflammation through microglial activation. In preclinical models, minocycline is a potent mitigator of microglial activation. We evaluated the efficacy of co-administration of minocycline with chemotherapy and its effect on inflammatory markers, PET/CT imaging, and cognitive changes in patients with BC. Methods: Women initiating first line adjuvant or neoadjuvant chemotherapy for BC were randomized to receive 100mg twice daily minocycline or placebo during chemotherapy treatment. Depression, anxiety, and cognitive function symptoms were measured with CES-D, STAI, BRIEF-A, and MFMQ testing. Inflammatory changes were evaluated by cytokine panels and PET/CT brain imaging. PET/CT imaging was performed on a Philips Vereos with a single bed dynamic acquisition over the brain acquired continuously for 75 minutes after injection of 3mCi 18F-fluorodeoxyglucose (FDG). Linear mixed effects models were used for statistical analysis. Results: 60 BC patients were randomized (stage distribution: 18.3% IA, 30% IIA, 10% IDC, 8.3% IB, 18.3% IIB, 13.3% IIIA). Median age was 52 years (range 26 - 71). 27 patients (45%) received an anthracycline (AC)-containing regimen, while 33 (55%) received a non-AC-containing regimen. Minocycline group had 42.9% ER+/PR+, HER2- and 32.1% HER2+ disease, while placebo had 51.9% ER+/PR+, HER2- and 22.2% HER2+. Inflammatory cytokine levels increased over time with chemotherapy regardless of group (Log10 of IL-8, TNF-alpha, and TNF-RII +0.5, +0.21, +1.7 with p < 0.001). There were no statistically significant differences between minocycline and placebo groups in changes in anxiety (STAI, -7.9 vs -7.6, p = 0.93), depression (CES-D, -0.3 vs -2.1, p = 0.39), cognitive function (BRIEF-A: +5.8 vs -0.4, p = 0.20; MFMQ: -1.8 vs -5.1, p = 0.49), or cytokine levels. 20 patients underwent PET/CT at treatment onset, 15 repeated imaging at cycle 4, and 13 at 6 months post-treatment. Although not significantly related to treatment group, distinct changes in brain metabolic characteristics in key regions such as the hippocampus were identifiable and quantifiable on an individual basis, and a significantly higher rate of FDG uptake was found over 75 minutes in the left hippocampus in patients who received AC-based chemotherapy compared to non-AC chemotherapy (p = 0.0271). Conclusions: Inflammation increased during chemotherapy. Minocycline co-administration did not mitigate changes in anxiety, depression, cognitive function, cytokine levels, or FDG-uptake in PET/CT imaging. The hippocampal differences in patients receiving AC verses non-AC chemotherapy suggest that AC therapy was more pro-inflammatory. This highlights the potential of PET/CT to monitor microglial activation, providing an objective assessment of neurocognitive function in future studies. Clinical trial information: NCT02203552.

e14004 Background: Depression and anxiety are frequent causes of excess morbidity for cancer patients, but access to mental health care, including outpatient psychiatry, is limited across academic and community settings. The Collaborative Care Model (CoCM) is a team-based model of care that is proven to leverage limited mental health resources across a wider population in primary care (Raney L, Am J Psychiatry, 2015). Evidence suggests that CoCM can be effectively applied to cancer populations too (Walker J, Gen Hosp Psychiatry, 2009). Methods: The investigators adapted CoCM to treat depression and anxiety in an academic BC clinic. A quality protocol was approved by the institutional board. All English-speaking BC patients were screened for depression using the PHQ-2 and PHQ-9. Patients with PHQ-9 ≥15 or clinical suspicion of severe anxiety or depression were referred to CoCM. The CoCM intervention included social work care management and regular case reviews by a psychiatrist. Medical oncologists prescribed all psychiatric medications. Patients were referred to mental health services in the community as needed. Results: From November 2018 through January 2020, a total of 74 patients were enrolled in CoCM. Median age was 50 years (range: 28-74 years) with BC stages I-III (n = 49, 66.2%) and stage IV (n = 25, 33.8%). Treatments within the cohort include: endocrine therapy (n = 39, 52.7%), chemotherapy (n = 18, 24.3%), observation (n = 9, 12.1%), single agent HER2 inhibitor targeted therapy (n = 4, 5.4%), immunotherapy (n = 2, 2.7%), single agent CDK4/6 inhibitor (n = 1, 1.4%), and radiation (n = 1, 1.4%). Of the 74 patients, 28 had PHQ-9 scores ≥15 at enrollment. On average, ending PHQ-9 scores decreased 39% from the initial score (average beginning score of 19.3 and ending score of 11.3 [n = 19]). 50 patients had GAD-7 ≥10 at enrollment. On average, ending GAD-7 scores decreased 36% from the initial score (average beginning score of 15.4 and ending score of 9.9 [n = 32]). On a 5-point scale, the average patient satisfaction score was 4.3 [range: 4.1-4.5] and the average medical oncology satisfaction score was 4.6 [range: 4.5-4.7]. Financial viability is promising based on projections that 93.4% of psychiatry costs (10% salary + benefits) are covered by reimbursements for care and 2 existing social workers serving as care managers. Conclusions: The collaborative care model is an effective and financially sustainable approach to promptly address depression and anxiety symptoms in BC. Further studies are needed to assess its applicability to patients with other forms of cancers.

e13554 Background: Leptomeningeal carcinomatosis (LMC) is a complication of advanced malignancies wherein primary tumors metastasize to the leptomeninges surrounding brain and spinal cord. LMC complicates 4-15% of malignant solid tumors with incidence increasing as survival of patients with advanced cancer improves. Diagnostic methods include magnetic resonance imaging (MRI) and cerebrospinal fluid (CSF) cytology. MRI findings may be nonspecific, and the gold standard of diagnosis is malignant cytology on CSF analysis. We assessed detection methods, incidence, and outcomes of LMC at The Ohio State University Comprehensive Cancer Center from 2005-2015. Methods: This was an IRB-approved single-institution retrospective study of 160 patients with confirmed diagnosis of LMC who were treated at the OSUCCC-James between Jan 1, 2005 and Dec 31, 2015. Patients with hematologic and central nervous system malignancies were excluded. Descriptive statistics were used to summarize demographic and clinical characteristics. Overall survival (OS) was defined as time from LMC diagnosis to death or last known follow-up, and was generated using Kaplan-Meier methods. Results: Median age of LMC diagnosis was 55.8 years (range: 48, 62.5). 69 (43%) patients had primary breast cancer, 41 (26%) had lung cancer, and 17 (11%) had melanoma. 73 patients (46%) presented with stage IV disease at initial diagnosis of the primary cancer, 41 (26%) with stage III disease, and 26 (16%) with stage II disease. Median time from diagnosis of primary cancer to diagnosis of LMC was 2 years (range: 0, 31.2). 158 (99%) patients had metastases at the time of LMC diagnosis, predominantly in bone (36%) or brain (36%). Median OS was 1.9 months (CI: 1.3, 2.5). 160 (100%) patients had an MRI of the brain or spine and 155 (97%) had MRI findings consistent with LMC. 75 (47%) patients underwent lumbar puncture, and 39 (52%) had CSF cytology positive for malignancy. Conclusions: Patients with LMC commonly presented with stage IV breast cancer, lung cancer, or melanoma with metastases to the brain or bone. Despite treatment, prognosis remains poor and confirmation of diagnosis can be challenging. Clinicians should have a low threshold for investigating LMC in high risk patients presenting with neurologic signs or symptoms.

202 Background: Fractional CO2 laser therapy is an emerging treatment for GSM. The objective of this study was to determine the feasibility, tolerability and preliminary efficacy of fractional CO2 laser therapy in BC survivors. Methods: This was a single arm feasibility study of BC survivors with dyspareunia and/or vaginal dryness. Participants received three treatments with office-based fractional CO2 laser on at least 30 day intervals and returned for a one-month follow-up. Feasibility was defined as treatment completion without serious adverse events (SAE) in a minimum of 80% of patients. Primary efficacy was evaluated using the mean change (∆) in the score on the Vaginal Assessment Scale (VAS). Secondary efficacy endpoints included mean ∆ in scores on the Female Sexual Function Index (FSFI) and Urogenital Distress Inventory (UDI). Descriptive statistics (means and 95% confidence intervals (CI) for continuous variables and proportions for categorical variables) were used. Results: The study is ongoing with 65 patients enrolled. To date, 37 patients have completed all study treatments and follow-up. Median age for those who have completed treatment was 57 years (range 34-72). Most were ER/PR positive (78%) and Her 2 negative (81%) with stage I (43%) or II (41%) BC. Ninety-five percent were receiving endocrine therapy, most commonly aromatase inhibitors (73%). No SAEs were reported in the 37 patients who have completed study treatments and their outcomes are as follows. Based on the VAS, 78% reported moderate-severe dyspareunia and 89% reported moderate-severe vaginal dryness at baseline. At follow up, 28% reported moderate-severe dyspareunia and 28% reported moderate-severe vaginal dryness. The VAS score improved from baseline to follow up (mean ∆ 4.1; 95% CI [3.1, 5.1]). Similarly, the FSFI score improved (mean ∆ -10.0; 95% CI [-13.2, -6.9]) and the UDI score improved (mean ∆ -5.7; 95% CI [-10.1, -1.3]). Final efficacy analysis will be reported once all patients have completed all time points. Conclusions: Fractional CO2 laser treatment is feasible and tolerable in BC survivors and may reduce symptom burden from GSM. A randomized controlled trial with sham laser is currently in development. Clinical trial information: NCT03307044.

TPS1127 Background: Heat shock protein 90 (HSP90) is a molecular chaperone which is necessary for proper folding and stabilization of proteins. Client proteins of HSP90 include many oncogenic proteins known to be over-activated in triple negative breast cancer such as AKT, EGFR, members of RAS/MAPK signaling pathway and androgen receptor. High expression of HSP90 in breast cancer has been associated with poor outcome. In addition, over-expression of HSP90 client proteins such as AKT and c-RAF has been implicated in paclitaxel resistance. Onalespib (AT13387) is a synthetic non-ansamycin small molecule that acts as an inhibitor of HSP90 by binding to the amino terminal of the protein and has dissociation constant (Kd) of 0.71 nM. Methods: Patients with inoperable or metastatic, triple negative or < 10% hormone receptor positive breast cancer are treated with onalespib and paclitaxel on days 1, 8, 15 every 28 days. Paclitaxel is given at a standard dose of 80 mg/m2 while the dose of onalespib is gradually increased using standard 3+3 design (see table). In order to assess the effect of each drug on pharmacokinetics of the other drug, onalespib is given on day -7 prior to cycle 1 and skipped on day 1 of cycle 1 during which paclitaxel is administered alone. The primary objective of the study is to determine recommended phase II dose and assess the toxicity profile of the combination. The secondary objectives include pharmacokinetic of each agent. Overall response rate, response duration and progression-free survival will also be assessed. The study is currently enrolling patients to dose level 2. Clinical trial information: NCT02474173. [Table: see text]

e12126 Background: Even with neo-adjuvant chemotherapy (NAC) many breast cancer (BC) patients (pts) relapse, especially triple negative pts. The incorporation of checkpoint inhibitors into NAC for BC is being tested in clinical trials. How NAC affects checkpoint receptor expression is not known. Such information could aid in the rational selection of checkpoints to target during NAC. We sought to characterize changes in the frequency of circulating CD4 and CD8 T cells expressing PD1, CTLA4, LAG3, TIM3, and OX40 over the course of NAC. Methods: In this prospective trial, expression of PD-1, CTLA-4, Lag3, Tim3 and Ox40 on circulating CD4 and CD8 T cells were measured by FACS analysis in pts with operable breast cancer (BC) prior and at the end of NAC. The primary objective was to explore the association between NAC and expression levels of the immune checkpoints. Results: 1, 20 and 3 pts had clinical stage I, II, IIIA, respectively. Median age was 48. 11, 6 and 7 pts were triple negative (TN), HER2+ and hormone receptor (HR)+, respectively. Complete pathologic response rate was 45.8%. Globally CD4 T cells expressing CTLA4, Lag3, Ox40 and PD1 decreased following NAC (all p < 0.01). Conversely, CD8 T cells expressing CTLA4, Lag3 and Ox40 significantly increased (all p < 0.01). More CD8 T cells from HER2+ pts expressed Lag3 prior to therapy compared to HR+ pts (p < 0.05) with a similar trend compared to TN pts. Prior to therapy more CD8 T cells from HER2+ and TN pts expressed Tim3 compared to HR+ pts (p < 0.05 for each). Post therapy more CD4 T cells from HER2+ pts expressed PD1 compared to HR+ and TN pts (p = 0.027 and 0.018 respectively). Clinical response did not predict change in checkpoint expression. An interaction analysis revealed that HER2+ disease predicted a drop in CTLA4 CD4 T cells and a drop in Lag3 CD4 and CD8 T cells over NAC (p < 0.05). Conclusions: This analysis identified changes in checkpoint receptor expression by CD4 and CD8 T cells in BC pts after NAC. Differences in checkpoint receptor expression were found between BC subgroups. This data provides a starting point for understanding checkpoint receptor expression changes with NAC, and could help guide the selection and timing of incorporating checkpoint inhibitors in BC.

e20508 Background: The completion of advance directives (AD) and the discussion of a patient’s end of life goals result in a more patient-centered approach to care. The Adena Cancer Center develope...

339 Background: Biomarkers used to identify patients at high risk for recurrence after surgical resection of colorectal cancer (CRC) lack predictive capacity and are applicable only to localized disease. As the use of adjuvant therapy in this setting can be controversial, the development of a biomarker that is prognostic, predictive, and applicable after resection of primary or metastatic disease may dramatically improve patient outcomes. Methods: We compiled microarray data from 850 patients with predominantly primary CRC from four publically available datasets to predict activity of 19 oncogenic pathways to generate patterns of pathway deregulation for each patient tumor. Mixture modeling was applied to the samples using affinity propagation to subgroup tumors bearing unique patterns of pathway deregulation. The model was then applied to a dataset of 133 predominantly metastatic CRC samples from patients undergoing curative surgical resection at the Duke University Medical Center. Results: Mixture modeling resulted in the identification of 6 distinct molecular subgroups of CRC (MSCC) based on pathway deregulation for both primary and metastatic CRC. Kaplan-Meier analysis of the primary data set revealed differences in recurrence free survival among the 6 MSCC (p=0.0004) with patients assigned to MSCC3 having the poorest prognosis and patients assigned to MSCC4 having the best prognosis. A similar result was obtained when the model was applied to the metastatic data set (p

447 Background: Current biomarkers for colorectal cancer sub-classify tumors based on single mutations, such as KRAS; however, studies of single mutations belie the molecular complexity of colorectal cancer in which an average of 14 key genes per tumor is dysregulated. We hypothesize that colorectal cancer may be molecularly sub-classified based on an oncogenic pathway prediction model in which tumors are grouped based on patterns of oncogenic pathway dysregulation/expression. Methods: Affymetrix microarray data from 850 patients with primary colorectal cancer from publically available datasets were combined using Bayesian factor regression modeling normalization. The activity of 19 separate oncogenic pathways was predicted among tumors to generate patterns of pathway activity for each sample. Mixture modeling was applied to these samples to identify subgroups of tumors with unique patterns of pathway dysregulation. Validation of subclasses was performed on a dataset of 133 primary and metastatic colorectal cancer samples of patients undergoing curative surgical resection at our institution. Tumors were subgrouped according to our previous model and recurrence free survival was calculated. In vivo validation was performed by treating NOD/SCID mice bearing patient derived tumors with everolimus with changes in tumor size calculated between day 0 and day 21. Results: Mixture modeling resulted in 6 individual subgroups of colorectal cancer based on pathway dysregulation. Kaplan Meier curves revealed that patients in subclass 4 had the poorest prognosis while patients in subclass 6 had the best prognosis (p=0.05). Further, tumors in subclass 4 were generally enriched for high mTOR pathway activation and patient derived explants from subclass 4 with high predicted mTOR activity were found to be sensitive to the MTOR pathway inhibitor everolimus. Conclusions: Prediction of oncogenic signaling pathway activity is a powerful tool that may be used to molecularly sub-classify colorectal cancer into biologically relevant subgroups. These subgroups have prognostic and predictive implications for recurrence following surgical resection and responsiveness to targeted therapy.

e14064 Background: Current chemotherapy regimens for colorectal cancer exhibit only 50% response rates; we hypothesize that responses to chemotherapy agents can be predicted by gene expression prof...