ALBERT

Description: Background The epitome of cancer treatment personalization is N=1 segmentation where a custom therapy is designed for every patient. Because most cancer aberrations are not actionable mutations and tumors can have more than one actionable mutation, this one biomarker/one drug approach to cancer personalization has inherent limitations due to its over simplification. Personalization 2.0 methodology creates a patient simulation avatar incorporating a patient’s genomic profile information holistically. Methods Bone marrow samples from two myeloma patients (P1 and P2) refractory to most recent treatment was collected, and P1’s sample was sorted into CD138+ and CD138- cells. The patient cells were analyzed for chromosomal alterations using Comparative Genomic Hybridization (aCGH) arrays by GenPath Diagnostics and cytogenetic chromosome analysis by Washington University School of Medicine and New York University (NYU), respectively. Using this information, a predictive simulation avatar model of each patient was created by Cellworks based on genomic profile of patients. A digital functional library of over 80 FDA-approved drugs and agents currently in clinical trials were simulated individually and in combination using the two patient avatars to create a personalized treatment for each patient. The findings were prospectively validated using patient cells ex vivo as assessed by MTT assay at New York University. Results P1 aberrations included trisomy of CCND1 and deletion of TP53 along with single copy losses in different arms of chromosomes 1, 6, 8, 12, 13, 14, 16, 17 and 22 and gains in different arms and regions of chromosomes X, 1, 4, 7, 9, 17, 3, 5, 11, 15 and 19, indicating the presence of hyperdiploid clones. Using this information, 897 gene perturbations were included to model this patient simulation avatar. Simulation predicted high beta-catenin (CTNNB1) activity with increased hedgehog and NOTCH pathways that were inherent causes of Bortezomib resistance. Significant activation of STAT3 and STAT5 due to amplification of IL6 pathway, JAK2 and JAK3 was noted. Amplifications of MET, IGFR and FGFR converged at ERK and AKT signaling loops. Along with deletion of TP53, this profile had amplification of many anti-apoptotic genes including survivin, MCL1 and XIAP. Modeling predicted sensitivity to the JAK inhibitor Tofacitinib, a drug approved for rheumatoid arthritis. This was prospectively validated ex vivo, and the experimental data correlated with the prediction showing a reduction in viability. P2 aberrations include losses in chromosomes X and 9 and a chromosome 11:14 translocation that is a common occurrence in MM. This translocation results in an amplification of CCND1 expression. The genomic aberrations reported include knockdown of tumor suppressors RXRA, TGFBR1, TJP2 and TSC1. TSC1 regulates the mTOR pathway, and its deletion causes an aberrant activation of mTOR and its downstream targets. Reduced expression of RXRA and TJP2 both in different manners leads to increase in AP1 activation. NFkB is also activated due to RXRA reduction. TGFBR1 reduction decreases the expression of cell cycle inhibitors via SMAD2/3 down-regulation. In this patient avatar, modeling predicted sensitivity to a combination of Sirolimus and Trametinib. Ex vivo validation confirmed this prediction of additive synergy of these two drug agents in the context of this patient. Conclusions This study demonstrates and validates the personalization of treatment through two patient cases based on creating predictive simulation avatar models using genomic profile information. This modeling holistically incorporates all genomic aberration information and is not limited to associating drugs to actionable mutations. Disclosures Doudican: Cellworks: Research Funding. Vali:Cellworks: Employment. Basu:Cellworks: Employment. Kumar:cellworks: Employment. Singh:Cellworks: Employment. Sultana:Cellworks: Employment. Abbasi:Cellworks: Employment, Equity Ownership.

Description: Introduction Development of resistance to single agent therapy is a significant clinical obstacle in the treatment of multiple myeloma (MM). Genetic mutations and the bone marrow micro-environment are major determinants of MM resistance mechanisms. Given the complexity of MM, the need for combinatorial therapeutic regimens targeting multiple biological mechanisms of action is pressing. Repurposing has the advantage of using drugs with known clinical history. Methodology We used a predictive simulation-based approach that models MM disease physiology in plasma cells by integrating and aggregating signaling and metabolic networks across all disease phenotypes. We tested the efficacy of over 50 repurposed molecularly targeted agents both individually and in combination across simulation avatars of the MM cell lines OPM2 and U266. OPM2 harbors mutations in KRAS, CDKN2A/2C, PTEN, RASSF1A and P53, whereas U266’s mutational components include BRAF, CDKN2A, P53, P73, RASSF1A and RB1. These cell lines were used as models because they possess mutations in genes classically known to be involved in myeloma. The predicted activity of novel combinations of existing drug agents was validated in vitro using standard molecular assays. MTT and flow cytometry were used to assess cellular proliferation. Western blotting was used to monitor the combinatorial effects on apoptotic and cellular signaling pathways. Synergy was analyzed using isobologram plots and the Bliss independence model. Results Through simulation modeling, we identified two novel therapeutic regimens for MM using repurposed drugs: (1) AT101 (Bcl2 antagonist) and tesaglitazar (PPAR α/γ agonist) and (2) Ursolic acid (UA, inhibitor of NFκβ) and SP600125 (pan-JNK inhibitor). Simulation predictions showed that combining the IC30 concentrations with respect to viability of AT101 and tesaglitazar reduced proliferation by 40% and viability by 50%. Similarly simulation predictions showed that the combination of the IC30 concentrations of UA and SP600125 reduced proliferation by 50% and viability by 40%. Corroborating our predictive simulation assays, 10 µM tesaglitazar and 2 µM AT101 caused minimal growth inhibition as single agents in OPM2 and U266 MM cell lines. Growth inhibition in these cell lines is synergistically enhanced when the drugs are used in combination, reducing cellular viability by 88% and 77% in OPM2 and U266 cells, respectively. Similarly, proliferation was reduced by 34% with 7.5 μM UA and 25% with 10 μM SP600125 in OPM2 cells. When used in combination, cellular proliferation was synergistically reduced by 64%. In addition, isobologram analysis predicted synergy of lowered doses of the drugs in combination. Both combinations synergistically inhibited proliferation and induced apoptosis as evidenced by an increase in the percentage sub-G1 phase cells and cleavage of caspase 3 and poly ADP ribose polymerase (PARP). Conclusions These results highlight and validate the use of our predictive simulation approach to design therapeutic regimens with novel biological mechanisms using drugs with known chemistries. This allows for design of personalized treatments for patients using their tumor genomic signature beyond the “one-gene, one-drug” paradigm. The reuse of existing drugs with clinical data facilitates a rapid translational path into clinic and avoids the uncertainties associated with new chemistry. The corroboration of these results with patient derived cell lines will be pursued and discussed. Disclosures: No relevant conflicts of interest to declare.

Description: Background: Hypomethylating agents (HMAs) (e.g., azacitidine (aza), decitabine (dec)) and lenalidomide (len) are approved agents and used in the treatment of patients with myelodysplastic syndromes (MDS) or acute myeloid leukemia (AML). Despite their widespread use, HMAs fail in the majority of MDS and AML patients, and len fails in 75% of non-del(5q) MDS. Unfortunately, no method exists to predict disease response, thus the management of MDS and AML patients is challenging. Predicting treatment response would improve treatment effectiveness, restrict treatment-related adverse events to those who would benefit, and reduce health care costs. Ideally, patient prediction would be based on disease biology. Aim: To determine the biological and clinical predictive values of a genomics-informed computational biology method in patients with AML and MDS who are treated with aza, dec or len. Methods: Patients with AML or MDS were recruited in a prospective clinical trial (NCT02435550) designed to assess predictive values by comparing computer predictions of treatment response to actual clinical response. Genomic profiling was conducted by conventional cytogenetics, whole exome sequencing (SureSelectXT Clinical Research Exome, Agilent), and array CGH (Agilent). These genomic results were inputted into computational biology software (Cellworks Group), which generates disease-specific protein network maps using PubMed and other online resources. Digital drug simulations were conducted by quantitatively measuring drug effect on a cell growth score, which is a composite of cell proliferation, viability and apoptosis. Each patient-specific protein network map was digitally screened for the extent by which aza, dec or len reduced simulated disease growth in a dose-respondent manner. Treatment was physician's choice based on SOC. Before initiating treatment, treating physicians were masked to the results of whole exome sequencing and computational predictions. Clinical outcomes were prospectively recorded. To be eligible for efficacy assessment, patients must have had at least 4 cycles of HMA treatment or 2 cycles of len treatment. For AML, CR+PR was used to define response (IWG 2003). For MDS, CR+PR+HI was used to define response (IWG 2006). To validate the predicted protein network perturbations, Western blot assays were performed on pertinent pathway proteins. Comparisons of computer-predicted versus actual responses were performed using 2x2 tables, from which prediction values were calculated. Fisher's exact test was used to compare prediction values of the genomics-informed computer method versus empiric drug administration. Results: Between June 2015 and June 2016, 80 patients were recruited. 40/80 (50%) had AML and 40/80 had MDS (50%). The median age was 66 (range 24-91). 44/80 (55%) were treatment-naïve and 36/80 (45%) were treatment-refractory. 99% completed all planned molecular tests and computational analyses. Laboratory validation study of computer-predicted, activated protein networks in 19 samples from 13 different patients showed correct prediction of 5 activated networks (Akt2, Akt3, PIK3CA, p38, Erk1/2) in 17 samples, exhibiting 89% accuracy. At the time of this report, 20/80 patients were eligible for efficacy evaluation. 6/20 patients showed clinical response to SOC therapy, while 14/20 did not achieve clinical response. 18 patients' outcome predictions were correctly matched to their actual clinical outcomes, and 2/20 were incorrectly matched, resulting in 90% prediction accuracy, 75% positive predictive value (PPV), 100% negative predictive value (NPV), 100% sensitivity, and 86% specificity. The accuracy of the genomics-informed computer method was significantly greater than empiric drug administration (p=1.664e-05). New genomic signature rules were discovered to correlate with clinical response after aza, dec or len. Conclusions: A computational method that models multiple genomic abnormalities simultaneously showed high predictive value of protein network perturbations and clinical outcomes after standard of care treatments. The network method uncovered molecular reasons for drug failure and highlighted resistance pathways that could be targeted to recover chemosensitivity. This technology could also be used to establish eligibility criteria for precision enrollment in drug development trials. Disclosures Vali: Cellworks Group: Employment. Abbasi:Cellworks: Employment. Kumar:Cellworks group: Employment. Kumar Singh:Cellworks group: Employment. Basu:Cellworks Group: Employment. Kumar:Cellworks Group: Employment. Husain:Cellworks Group: Employment. Wingard:Ansun: Consultancy; Merck: Consultancy; Fate Therapeutics: Consultancy; Astellas: Consultancy; Gilead: Consultancy.

Description: Background: Current tumor profiling analytics provide some insight into the various molecular abnormalities and their individual consequences on oncogenic signaling. However, these analyses are limited by their lack of integration where the combined effect of individual mutations, gene copy number variations and chromosomal aberrations are not consolidated to create the global molecular architecture that supports neoplastic growth, particularly in the context of drug resistance. Consequentially, identities of the preferential oncogenic pathway(s) tumor cells employ to oppose the effects of targeted therapies remain cryptic and unactionable. Here we present a simulation-based method, which not only replicates the molecular architecture of ibrutinib-resistant Waldenstroms Macroglobulinemia (WM, for which ibrutinib is the only FDA-approved agent) in silico, but also predicts cell sensitivity towards existing drugs, which we validated experimentally for potential clinical translation. Materials: We used the newly established human WM cell line, RPCI-WM1/IR, as a surrogate model of ibrutinib-refractory WM. Genomic data including whole exome sequencing (WES) and copy number analysis (CNA) was utilized for the creation of an avatar of RPCI-WM1/IR, which through simulation identified the salient and prominently dysregulated cellular pathways. Importantly, illustrating these pathways highlights common convergence points on increased proliferation and viability. These convergence points were then directly and indirectly targeted by simulated testing of a library of FDA approved drugs and those impacting these dysregulated pathways were nominated. Importantly, this simulation avatar approach not only looks for agents acting on the specific gene mutation, but also predicts the convergence points to be attacked. The personalized simulation avatar technology is a comprehensive functional proteomics representation of the WM physiology network. A standardized library of equations models all the biological reactions such as enzymatic reactions, allosteric binding and protein modulation by phosphorylation, de-phosphorylation, ubiquitination, acetylation, prenylation and others. Results: Several genomic aberrations were used to create the RPCI-WM1/IR simulation avatar. Functional activity (based on mutation or copy number alteration) of several ibrutinib targets or transcription factors associated with BTK activity such as FYN, SP1, BMX and FRK were predicted to be lost. Increased expression of CAV1, which also inhibits BTK mediated signaling, was increased. An increase in CSNK2B, which activates PU.1- a transcriptional target of BTK, was also observed. Of note, no CXCR4 mutations, which have been shown to impact ibrutinib response, were observed. Next, the cytotoxic potential of over 150 FDA approved drug (and some in experimental stages) were simulated individually and in combination on the RPCI-WM1/IR avatar. In silico modeling predicted aberrant activity of aurora kinase A (AURKA) and its associated signaling partners, which could be disrupted with the (AURKA) inhibitor, tozasertib. AURKA activation was predicted as upregulated due to alterations in several genes: RASA1 loss and SOS1 increase --〉 increased ERK --〉 increased ETS1 --〉 increased AURKA. High beta-catenin signaling (high CTNNB1 and FZD1/4 and low AXIN1 and GSK3B) was also shown to increase AURKA. The simulation predictions were experimentally validated in vitro where AURKA inhibition with tozasertib significantly inhibited proliferation of RPCI-WM1/IR cells (IC50~14nM) as well as inducing apoptosis (48hr, 20nM treatment) and cell-cycle arrest. Conclusions: Our data demonstrates the potential of in silico modeling in predicting novel drug targets, allowing guidance in 1.) Delineating operational oncogenic circuits in an ibrutinib-resistant state by reanimation of the molecular architecture in silico, 2.) Calculating the impact of individual genomic abnormalities and their collective influence on maintaining tumor survival and 3.) Performing a rapid in-silico drug-sensitivity screen directed by the pathway analyses, which can be validated experimentally using standard assays. This novel approach holds tremendous potential in creating highly personalized therapies for ibrutinib-refractory WM patients based on unique genetic signatures. Disclosures Vali: Cellworks Group, Inc.: Employment, Equity Ownership. Kumar:Cellworks Group, Inc.: Employment. Singh:Cellworks Group, Inc.: Employment. Abbasi:Cellworks Group, Inc.: Employment, Equity Ownership.

Description: Introduction: Immunotherapy is an exciting new option to stimulate a "host-vs.-cancer" effect. Such treatment regimens are dominated by Programmed Death-1 (PD-1) and Programmed Death Ligand-1 (PD-L1) immune checkpoint inhibitors. However, not all patients respond to PD-1/PD-L1 inhibition as confirmed by results from an "all-comers" clinical strategy. These inherent clinical gaps are addressed in studies by incorporating PD-L1 expression as inclusion criteria and combining PD-1/PD-L1 inhibitors with other drug agents. Unfortunately these outcomes have not been clinically beneficial and cause unnecessary toxicity to patients and unmet needs. Here we develop personalized methodologies to predict a response to PD-L1/PD-1 inhibitors based on two factors. First, PD-L1 is not the only immune checkpoint; other signaling between cancer and immune cells needs to be incorporated to predict non-responders to a PD-L1/PD-1 inhibitor. Second, precision medicine and personalization is driven by patient tumor genomics. A recent retrospective study correlated NRAS with PD-L1/PD-1 treatment responses. Beyond a single gene mutation or aberration, the holistic consideration of all gene mutations, copy number variations, and methylation status would impact immune signal activators and inhibitors from the cancer cell. We used a predictive simulation model of multiple myeloma and dendritic cells with corresponding in-vitro models. We developed an immunotherapy response phenotype, which is a function of bio-markers representing immune evasion, immune activation, metastasis, and dendritic cell infiltration in cancer cells. We then modeled two myeloma cell lines using available genomics information as proxy for myeloma patient cancer. Finally, we predicted these two representative patient classes would vary in response to PD-L1/PD-1 inhibitors. We validated our prediction that bio-markers contribute to the immunotherapy response. Methods: Predictive computational simulation models of myeloma cell lines, human myeloid dendritic cells, and myeloma cell + dendritic cell co-cultures were developed and used to predict extracellular (IL6, IL10, TGFB1, VEGFA) and cell-associated (CD47, FASLG, IDO1, PD-L1) biomarker readouts in an automated high-throughput system. Results were validated with myeloma cell lines MM.IS and U266B1 cultivated with and without dendritic cells in Transwell 12-well polystyrene plates. At 24 hours, IL6, IL10, TGFB1, and VEGFA concentrations in tissue culture media were determined using Millipore immunoassays. CD47, FASLG, IDO1, and PD-L1 concentrations in cell lysates were determined by ELISA and IHC. One-way fixed-effects ANOVA models were fit to log-transformed concentrations. Pairwise group comparisons were conducted using Tukey's Honest Significant Differences (JMP10, Version 10.0, SAS, Cary, NC) at a 0.05 level of significance. Results: Predictive simulation and experimental results were highly correlated. Predicted IL10, TGFβ1, IDO1, CD47, FASLG, IL6, VEGFA, and PD-L1 responses differed among MM.IS and U266B1 patient models and were similar to those observed in cell culture supernatants and cell lysates. Simulation models also predicted myeloma cell effects on dendritic cell biomarker responses. Twenty-three myeloma cell lines had high, moderate, and low expression effects on 19 dendritic cell biomarker readout responses. Experimental results using MM.IS and U266B1 cell lines + dendritic cell co-cultures confirmed these effects also occur in cell co-culture. Predicted IL10, TGFβ1, IDO1, CD47, FASLG, IL6, VEGFA, and PD-L1 responses were similar to that observed in cell culture supernatants and cell lysates. Conclusion: The ability to predict which patients respond to PD-L1/PD-1 inhibitors and adjuvant combinations of existing drugs will improve response rates and meet current needs. Here we show a novel predictive simulation based approach with a myeloma example to leverage patient genomics information for predicting non-responders. This approach predicts the immunotherapy index of a patient, which is correlated experimentally. We show that U266B1 has higher PD-L1 expression because of patient genomics. While correlation at the bio-marker level is important, future work will focus on predicting and confirming the response to PD-L1 inhibitors. Disclosures Brogden: Cellworks Group Inc.: Research Funding. Fischer:Cellworks Group Inc.: Research Funding. Bates:Cellworks Group Inc.: Research Funding. Lanzel:Cellworks Group Inc.: Research Funding. Gomez Hernandez:Cellworks Group Inc.: Research Funding. Treinen:Cellworks Group Inc.: Research Funding. Recker:Cellworks Group Inc.: Research Funding. Nair:Cellworks Group Inc.: Employment. Sundara:Cellworks Group Inc.: Employment. Joseph:Cellworks Group Inc.: Employment. Abbasi:Cellworks Group, Inc.: Employment, Equity Ownership. Vali:Cellworks Group, Inc.: Employment, Equity Ownership.

Description: Background: The myelodysplastic syndromes (MDS) are a heterogeneous group of malignancies characterized by multiple genomic abnormalities. New technologies are needed that bridge the complex molecular biology of MDS and available therapeutics. Methods: Therefore, we developed a predictive simulation software program to create MDS patient avatars based on disease-specific genomic profile. The avatar technology has three layers: the first layer is a proprietary mathematics solver engine architected to handle millions of differential equations representing biological cross-talk reactions needed for modeling MDS physiology; the middle layer is the comprehensive functional proteomics representation of disease physiology network; and the topmost layer is a semiconductor engineering-based automaton engine which allows high-throughput simulation of multi-million interventions through assembly code language. The MDS avatars were intended to map the complex interplay of dysregulated pathways and predict potential therapeutics. The human MDS-L cell line is one of few bona fide MDS cell lines and was analyzed by conventional G-banding karyotyping, FISH, array CGH and Sanger sequencing of MDS-relevant genes. Using this genomic information, a simulation patient avatar was created. Next, a library of over 80 FDA approved agents was simulated against the dysregulated pathways. A list of predicted drugs where then prospectively validated using in vitro culture of MDS-L cells. Results: The MDS-L cells harbored an activating NRAS mutation (c.35G〉C, p.G12A) and complex chromosome abnormalities. The simulation MDS-L patient avatar predicted activation of the RAF-ERK pathway due to NRAS mutation and various copy number variations (CNVs) including high CN of IGFR, AURKA, PAR5 that predicted high AKT activation, high CN of mTOR, high CN of IL6 and JAK3 with predicted activation of STAT3 and STAT5, high CN of MDM4 with predicted lower levels of TP53, high CN of RCE1, MAPK1 and MAP2K1, and low CN of RASA1 and DUSP1 (Figure 1). Given the strong dominance of AKT and ERK loops in this profile (Figure 1), we simulated potential drugs individually and in combinations to predict impact on these pathways. Iterative simulations identified four FDA-approved drugs: (A) nelfinavir and celecoxib acting on AKT, (B) sorafenib and trametinib that inhibit RAF-ERK targets, and (C) statins that inhibit the RAS-RAF-ERK pathway through inhibition of prenylation. In vitro validation experiments showed statistically significant reductions in MDS-L cell viability when treating with nelfinavir alone, celecoxib alone, and enhanced additive reduction in viability with the combination of nelfinavir and celecoxib (Table 1). Conclusions: This study shows how a novel simulation method can be employed to use patient-specific MDS genomic profiling to map dysregulated pathways and predict potentially therapeutic agents. The MDS case presented led to the discovery of a novel combination of nelfinavir and celecoxib. Results from this study serve as the basis for MDS clinical trials that assign treatment based on genetic mutations. Figure 1. Predicted Dysregulated Pathways in MDS-L Cells. MDS-L genomic data was used to generate a map of dysregulated pathways. NRAS mutation is highlighted in blue with mapping of consequent downstream effects. High Copy Number (CN) in red and Low (CN) mutations in purple are also mapped with predicted downstream pathway effects. The totality of dysregulated pathways is predicted to converge on increased cell proliferation and increased cell viability. FDA-approved drugs predicted to impact dysregulated pathways (i.e., celecoxib and nelfinavir) are also shown in the map at critical impact points. Figure 1. Predicted Dysregulated Pathways in MDS-L Cells. MDS-L genomic data was used to generate a map of dysregulated pathways. NRAS mutation is highlighted in blue with mapping of consequent downstream effects. High Copy Number (CN) in red and Low (CN) mutations in purple are also mapped with predicted downstream pathway effects. The totality of dysregulated pathways is predicted to converge on increased cell proliferation and increased cell viability. FDA-approved drugs predicted to impact dysregulated pathways (i.e., celecoxib and nelfinavir) are also shown in the map at critical impact points. Table 1. MDS-L Cell In Vitro Validation Experiments of Drug Treatments. Nelfinavir Drug Dose 0μM 5μM 10μM Celecoxib 0 μM 100.00% 81.45% ± 7.8% 53.49% ± 6.9% 30 μM 71.34% ± 4.89% 60.58% ± 3.25% 35.78% ± 3.59% MDS-L cells were treated with vehicle control, nelfinavir alone (incremental dose increases), celecoxib alone (incremental dose increases), and combination nelfinavir and celecoxib. MDS-L cell viability was quantified and compared among treatment groups. Both nelfinavir alone and celecoxib alone reduce MDS-L cell viability in a dose-dependent manner. The combination of nelfinavir and celecoxib showed enhanced additive reduction of MDS-L cell viability, again in a dose-dependent manner. Disclosures Cogle: PHAR: Consultancy. Vali:CellWorks: Employment. Kumar:CellWorks: Employment. Singh:CellWorks: Employment. Tyagi:CellWorks: Employment. Abbasi:CellWorks: Employment, Equity Ownership.

Description: Background: Multiple myeloma (MM) is characterized by the invasion of malignant plasma cells into the bone marrow. While first line treatment options result in significant clinical benefit to patients, spatiotemporal clonal evolution results in disease relapse and mortality. Advances in genomics have armed clinicians with unprecedented insight into the molecular architecture of MM cells, however, the clinical benefit derived by genomics-guided intervention has been limited. We present a novel computational biology modelling (CBM) tool, which takes into account the combined effect of individual mutations, gene copy number abnormalities and large scale chromosomal changes in order to predict the salient molecular pathways utilized by the MM cell for survival. By reverse-engineering MM cell architecture in silico, the CBM tool is able to predict drug response and resistance mechanisms. Thus, our aim was to determine the accuracy of the CBM tool in predicting treatment response of relapsed/refractory MM patients for future management of their disease, in a more individualized manner. Methods: Cytogenetics and somatic mutations (by targeted NGS) for 15 MM patients were input into the CBM model to predict responses to different therapeutic combinations. All patients were relapsed to prior treatment. CBM uses PubMed and other online resources to generate patient-specific protein network maps of activated and inactivated disease pathways. We simulated the specific combinations of the drugs per patient and measured the quantitative drug effect on a composite MM disease inhibition score (i.e., cell proliferation, viability, apoptosis and paraproteins). The actual clinical outcome of the treatments was compared with predicted outcomes. Results: Fifteen patients were analysed using CBM for prediction of treatment response after NGS was performed. 13/15 were clinically evaluable, of which 1 was a responder and 12 were non-responder. 6/13 patients were treated on clinical trial and 7/13 were on drug combinations per physician decision. CBM correctly predicted 1 responder and 11 non-responder with a PPV of 50%, NPV 100%, specificity 91.67%, sensitivity 100%. The accuracy of CBM prediction was 92.30%. CBM also predicted the response of prior drug therapies for its non-response at relapse. For prior drug treatment options, 14 patients were evaluable. All the 14 patients were clinically non-responders and CBM correctly predicted for 13 patients with NPV 100%, Specificity 92.85% and overall accuracy of 92.85%. The majority of patients did not respond to therapies recommended at relapse. As an example, the operative molecular pathways from 2 patients who did not respond to combination treatment, either pre-NGS or post-NGS profiling, are shown in Fig. 1 and Table 1. CBM identified amplification (AMP) of chromosome (chr) 1 (WNT3A, IL6R, CKS1B, MCL1, PIK3C2B, USF1), chr 3 (HES1, PIK3CA, CTNNB1, WNT7A, FANCD2), chr 5 (IL6ST, IRF1, GLRX, SKP2), chr 7 (CDK5, EZH2, IL6, CAV1, ABCB1), chr 9 (NOTCH1, HSPA5, FANCC, FANCG), chr 15 (DLL4, FANCI, ALDH1A2), chr 19 (ERCC1, ERCC2, USF2); deletion(DEL) of chr 13 (CUL4A) , chr 16 (AXIN1, CDH1) and TP53 mutation in different combinations, which confer resistance to therapies at relapse. Conclusions: The CBM technology represents a potential means to identify therapeutic options for MM patients based on the patients individual tumor-genome profile and which can also be deployed for uncovering drug resistance mechanisms. This tool may aid clinicians in decision making for recommending the most appropriate therapy based on standard of care agents or clinical trials; thus improving patient outcomes and reducing unnecessary costs or drug-related toxicities. Disclosures Singh: Cellworks Research India Private Limited: Employment. Sauban:Cellworks Research India Private Limited: Employment. Husain:Cellworks Research India Private Limited: Employment. Kumar:Cellworks Research India Private Limited: Employment. Kumari:Cellworks Research India Private Limited: Employment. Tyagi:Cellworks Research India Private Limited: Employment. Abbasi:Cell Works Group Inc.: Employment. Vali:Cell Works Group Inc.: Employment. Ailawadhi:Pharmacyclics: Research Funding; Takeda: Consultancy; Celgene: Consultancy; Amgen: Consultancy; Janssen: Consultancy.

Description: Background: Pediatric AML (pAML) treatment outcomes can vary due to genomic heterogeneity. Thus, selecting the right drugs for a given patient is challenging. There is a need for a priori means of predicting treatment responses based on tumor "omics". Computational biology modeling (CBM) is a precision medicine approach by which biological pathways of tumorigenesis are mapped using mathematical principles to yield a virtual, interactive tumor model. This model can be customized based on a patient's omics and analyzed virtually for response to therapies. Aim: To define prediction values of a CBM precision medicine approach in matching clinical response to ADE therapy in a cohort of pAML patients. Methods: Thirty pAML patients that were treated ADE chemotherapy were utilized with information on the clinical, genomic (cytogenetics, mutations) and protein expression data from this cohort of pAML patients used for the CBM. From cytogenetics results, gene copy number variations were coded as either knocked-down (KD) or over-expressed (OE). From NGS results (2 gene panel - CEBPA, NPM1), gene mutations were coded as either loss or gain of function (LOF or GOF). For protein expression data, proteins that were 〉2sigma from the mean were coded as KD if their value was 0. Proteins with values

Description: Background: Monosomy of chromosome 7/Del 7 (-7) or its long arm (del(7q)) is one of the most common cytogenetic abnormalities in pediatric and adult myeloid malignancies, particularly in adverse-risk acute myeloid leukemias (AMLs). Monosomy 7 with complex karyotype further worsens the prognosis. Therefore, predicting response of therapies in this segment of patients is urgently needed to improve disease management by customizing therapy to the profile genomics instead of the conventional method of trial and error or one-size-fits all treatments. Aim: Predictive analysis of disease characteristics of Monosomy 7 and Sub-clustering of this segment along with identification of therapy options for treatment of this cohort using computational biology modeling (CBM). Methods: We selected a cohort of 1168 AML patients whose genomics and clinical data was collated from public domain datasets and literature. Using CBM, patients with (-7) were identified from this larger cohort and were sub-clustered using a machine learning approach. The genomic data of the representative patients per sub-cluster of (-7) cohort were used to generate disease-specific protein network maps. Digital drug simulations were done by quantitatively measuring drug effect and calculating a disease inhibition score (DIS) that is a composite of proliferation, viability and apoptosis along with impact on disease specific biomarker score. Each patient-specific map was digitally screened for the extent by which standard of care (SOC) and combinations with Non-SOC inhibited the disease. Results: CBM identified 187 (-7) patients (16%) that grouped into 8 sub-clusters (Table 1). CBM analysis of genomics of each sub-cluster representative identified combination therapy options for this cohort with poor response to SOC, supported by genomics driven disease characteristics. Monosomy 7 aberrations would lead to decreased expression of EZH2, CARD11, EIF3, PMS2, HUS1, KMT2C (MLL3), CDK5 and IKZF1. Azacitidine (AZA) is predicted to be a non-responder in this cohort due to decreased expression of EZH2, that would impact DNA methylation via reduced recruitment of DNMTs. Lenalidomide (LEN) is also predicted to be a non-responder due to decreased expression of IKZF1, CARD11, and EIF3 despite presence of Del5q, a strong inclusion for LEN response. This segment has shown high microsatellite instability (MSI) characteristics and this could be linked to reduced mismatch repair (MMR) pathway due to reduced expression of PMS2, HUS1 and KMT2C. This characteristic could explain certain patient cases with (-7) who have responded to Cytarabine. (Figure 1) However, the response rate of chemotherapy in (-7) cohort is

Description: Background: In AML, leukemic transformation causes clonal expansion of immature cells through de-regulated cell division cycles. CDK4/CDK6 regulates neoplastic progression, which might represent an effective strategy for treating AML. But current clinical data shows either limited efficacy or elusive results. Bromodomain and extra-terminal (BET) inhibitors interferes with transcriptional complexes and disrupting gene transcription of key oncogenes such as MYC. Also, there is need to explore usage of other receptor tyrosine kinase inhibitors. Therefore, a drug combination strategy should be explored to overcome limited clinical efficacy. Aims: To create digital drug models for BET inhibition (BETi), CDK4/6 inhibition (CDKi), FLT3 inhibition (FLT3i) and validate the predicted responses in AML patient samples with ex vivo chemosensitivity testing. Furthermore, to validate the identified combination of BET inhibitor with CDK4/6 inhibitor or FLT3 inhibitor. Methods: The Beat AML project (supported by the Leukemia & Lymphoma Society) collects clinical data and bone marrow specimens from AML patients. Bone marrow samples are analyzed by conventional cytogenetics, whole-exome sequencing, RNAseq, and an ex vivo chemosensitivity assay. 33 patients were randomly chosen. Every available genomic abnormality was inputted into a computational biology model (CBM, Cell Works Group Inc.) that uses PubMed and other online resources to generate patient-specific protein network maps of activated and inactivated pathways. Digital drug simulations with BETi (JQ1), CDKi (palbociclib), FLT3i (sorafenib) were conducted by quantitatively measuring drug effect on a composite AML disease inhibition score (i.e., cell proliferation, viability, and apoptosis). Paired comparisons between the computational predictions and the sample's ex vivo chemosensitivity IC50 values were conducted. Results: None of the 33 AML patients showed ex vivo chemosensitivity to palbociclib alone, and the CBM method was highly accurate (94%) in predicting this lack of response (Table 1). Through CBM mapping the following genetic mutations were identified as potentially contributing to palbociclib resistance: loss of function (LOF) of RB1 (Retinoblastoma 1), LOF of PTCH1 (patched 1), LOF of FBXW7 (F-box and WD repeat domain containing 7), gain of function (GOF) of CCNE1/2 (Cyclin E1/2) or LOF in NPM1 (nucleophosmin 1). Additionally, the CBM method showed that NPM1 mutated AML cases that were resistant to palbociclib also showed a better response to the combination of palbociclib and JQ1. In 28/33 (85%) patients, this combination of palbociclib and BETi (JQ1) was toxic to AML cells in ex vivo chemosensitivity assay (Table 2). The CBM method predicted that 31/33 (94%) of AML patients would respond to palbociclib and JQ1 (Table 2), and the accuracy of matching between CBM and ex vivo chemosensitivity assay was 91% (Table 2). Another combination with high proportion of responding patients was FLT3i and BETi (Table 3), with accuracy of matching between CBM and ex vivo assay of 91% (Table 3). Furthermore, computational analysis of the combination of BETi and FLT3i revealed that patients with mutation of NPM1 along with FLT3 TKD/ITD mutation showed high degree of synergy at lower drug concentrations. Conclusion: Digital drug simulations of inhibitions of CDK4/6, BET, and FLT3 using AML patient genomic data accurately matched ex vivo chemosensitivity results. The integration of computational biology modeling and Beat AML data led to the identification of potential palbociclib resistance pathways in AML, which led to the rational design of new multi-drug regimens. This integrated system enabled novel inferences that informs future clinical trials for patients with AML. Disclosures Tyner: Janssen: Research Funding; Incyte: Research Funding; Takeda: Research Funding; Leap Oncology: Equity Ownership; Gilead: Research Funding; Syros: Research Funding; Seattle Genetics: Research Funding; Agios: Research Funding; Aptose: Research Funding; AstraZeneca: Research Funding; Genentech: Research Funding. Druker:Cepheid: Consultancy, Membership on an entity's Board of Directors or advisory committees; GRAIL: Consultancy, Membership on an entity's Board of Directors or advisory committees; Blueprint Medicines: Consultancy, Equity Ownership, Membership on an entity's Board of Directors or advisory committees; Vivid Biosciences: Membership on an entity's Board of Directors or advisory committees; Beta Cat: Membership on an entity's Board of Directors or advisory committees; Millipore: Patents & Royalties; MolecularMD: Consultancy, Equity Ownership, Membership on an entity's Board of Directors or advisory committees; ALLCRON: Consultancy, Membership on an entity's Board of Directors or advisory committees; Aptose Therapeutics: Consultancy, Equity Ownership, Membership on an entity's Board of Directors or advisory committees; Gilead Sciences: Consultancy, Membership on an entity's Board of Directors or advisory committees; Aileron Therapeutics: Consultancy; Patient True Talk: Consultancy; Celgene: Consultancy; Amgen: Membership on an entity's Board of Directors or advisory committees; McGraw Hill: Patents & Royalties; Bristol-Meyers Squibb: Research Funding; Third Coast Therapeutics: Membership on an entity's Board of Directors or advisory committees; Henry Stewart Talks: Patents & Royalties; Oregon Health & Science University: Patents & Royalties; Leukemia & Lymphoma Society: Membership on an entity's Board of Directors or advisory committees, Research Funding; Novartis Pharmaceuticals: Research Funding; Monojul: Consultancy; ARIAD: Research Funding; Fred Hutchinson Cancer Research Center: Research Funding. Vidva:Cellworks Research India Private Limited: Employment. Narvekar:Cellworks Research India Private Limited: Employment. Agrawal:Cellworks Research India Private Limited: Employment. Gera:Cellworks Research India Private Limited: Employment. Prasad:Cellworks Research India Private Limited: Employment. Shyamasundar:Cellworks Research India Private Limited: Employment. Tunwer:Cellworks Research India Private Limited: Employment. Abbasi:Cell Works Group Inc.: Employment. Vali:Cell Works Group Inc.: Employment. Cogle:Celgene: Other: Steering Committee Member of Connect MDS/AML Registry.