Background Not all non-small cell lung cancer (NSCLC) patients benefit from chemotherapy when added to immune checkpoint inhibitors (ICI). We used comprehensive genomic profiling coupled with a computational biology model (Cellworks) to create an algorithm to identify patients who may benefit from adding chemotherapy to ICI (ICI+C). Patients and Methods The algorithm (Therapy Response Index, or TRI) was trained in a retrospective cohort of 553 NSCLC patients from the U.S. Veteran’s Health Administration (VHA) system who received comprehensive genomic profiling. The TRI, computational model and clinical threshold were locked and validated in 710 advanced NSCLC front-line patients receiving either ICI or ICI+C, obtained from the Flatiron Health-Foundation Medicine NSCLC clinico-genomic database. Results The classifier was significantly associated with OS in a multivariate analysis. (LR P = .0355). Patients with a low TRI (TRI ≤ 32) received an estimated incremental benefit in median OS of ∼ 3 months with ICI+C (LR P = .04). In contrast, patients with a high TRI (TRI > 32) showed no improvement in median OS when receiving ICI+C. A statistical test of interaction between TRI and chemotherapy met prespecified criteria of P < .25 (LR P = .1361), suggesting that TRI may be predictive of chemotherapy benefit. Conclusions TRI was predictive of OS and incremental chemotherapy benefit in patients with NSCLC receiving ICI or ICI+C. These results support the use of the classifier to identify patients who may benefit from ICI+C and those unlikely to respond to ICI alone, independent of PD-L1 levels.

Immune checkpoint inhibitors (ICI) have become a standard treatment option in patients with high microsatellite instability (MSI-H). Although immune checkpoint inhibitors are largely effective in these patients, MSI is an imperfect biomarker and other therapies may be warranted. A mechanistic Computational Biology Model (CBM) was developed by Cellworks that can be personalized based on a patient’s tumor-based genomic profile, revealing signaling pathway dysregulation and patient-specific drug response. Model output was used to identify MSI-H patients who may have a poorer response to ICIs. Computational biosimulation was performed using real-world retrospective cohorts of 423 STAD patients and 534 CRC patients (TCGA). MSI measurements were provided by TCGA. Efficacy scores (ES) based on biosimulated composite cell growth in response to disease and therapy were generated on all patients for pembrolizumab. Molecular rationales for ICI resistance were identified for MSI-H patients with low pembrolizumab ES (ES-L). Model efficacy scores for pembrolizumab were significantly higher in MSI-H patients for both STAD (average ES 20.5 vs 3.2 respectively, p-value < 0.001) as well as CRC (average ES 13.4 vs 2.4 respectively, p-value < 0.001). Of the MSI-H patients, 59% and 81% were ES-L (STAD and CRC respectively) MSI-H/ES-L STAD patients showed a significantly higher percentage of NOTCH2, EGFR, and EZH2 amplifications as well as a higher percentage of TP53-SOF mutations, whereas MSI-H/ES-L CRC patients showed a significantly higher percentage of MYC amplifications (p-value < 0.05) In this study, a therapy efficacy score produced through biosimulation was significantly associated with microsatellite instability levels, and markers of ICI resistance were identified in MSI-H/ES-L patients. Future studies will be performed to confirm the hypothesis that biosimulation has utility in deciding which MSI-H patients might benefit from therapies other than ICI. Adity Ghosh, Mamatha Patil, Anuj Tyagi, Ansu Kumar, Chandan Kumar, Muthiyah MR, Rahul KR, Swati Khandelwal, Jyoti Chauhan, Michael Castro, Shweta Kapoor, James Wingrove. Use of biosimulation to predict immune checkpoint inhibitor resistance in patients with high microsatellite instability [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2025; Part 1 (Regular Abstracts); 2025 Apr 25-30; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2025;85(8_Suppl_1):Abstract nr 3308.

Background: Differentiation arrest resulting from aberrant methylation represents a key pathogenic and phenotypic feature of MDS. Hypomethylating agents (HMA) can reverse epigenetic dysregulation, but as single agents fail to control the disease in approximately 50% of patients, many of whom progress to AML and carry a dismal prognosis. We hypothesized that Differentiation Scoring (DS) derived from Cellworks Computational Omics Biology Model (CBM) of each patient's genomic aberrations could serve as a biomarker to predict response to HMA. Methods: We studied 169 MDS patients treated with azacytidine (n = 86) or decitabine (n = 83) whose genomic aberrations and clinical outcomes were available from public databases. Biosimulation of the disease state was computationally derived from pathways that were up- or down-regulated by individual genomic aberrations. DS was defined as the ratio of the quantified impact of key differentiation biomarkers in myeloid malignancies (CEBPA, GATA1, SPI1 and ITGAM) divided by the impact of HOXA9 transcription, a known repressor of differentiation, and normalized on a log2 scale. We also used CBM to determine the change in DS attributable to HMA to define an efficacy score (ES) as a parameter of drug response. A Pearson correlation coefficient was obtained using the two variables. DS in clinical responders (R) and non-responders (NR) was evaluated with the Student t-test. We defined DS > 1.5 as DS-high and < 1.5 as DS-low. Survival data were available for 101 patients and correlated with DS-high (n=49) and DS-low (n=52) groups using a KM-curve. Results: Median age of this cohort was 69 years (range, 20-87 years), including 56 females, 111 males, and 2 of unknown sex. Regarding HMA response, 62 patients were R while 107 were NR. Prior to therapy, 86 patients had DS-high and 83 had DS-low scores. Post-treatment DS were correlated with HMA ES (R = 0.58, p < 2.2e-16), and showed a favorable HMA response for DS-high patients. Biosimulation showed the incremental benefit on DS of HMA was greater for people with DS-high than DS-low scores (44% vs 16%, p = 7.51 e-17). Median DS was higher in R than NR (p=0.00232). Patients with DS-high scores had superior survival compared to those with DS-low scores (log rank p = 0.0016). (Figure 1a) Different genomic aberrations contributed to the DS scores. DS-high patients had more TP53 mutations, 17p del, 8q amp, and 5q del. On the other hand, DS-low patients more often had mutations in ASXL1, SRSF2, RUNX1, DNMT3A, EZH2, NF1, NRAS, CBL, or 7q deletion. (Figure 1b) Loss of TP53 due to mutation or deletion upregulates DNMT1 and silences HOXA9 transcription, leading to DS-high scores. KAT6A (8q) amp also upregulates EZH2 and CpG methylation, while 5q del causes GATA1 transcription, increasing CpG methylation via SALL4 to produce high ß-catenin signaling that led to DS-high scores. By contrast, 7q loss, ASXL1, SRSF2, and EZH2 mutations downregulated EZH2, diminished CpG methylation and unsilenced HOXA9 transcription. As a result, HOXA9 arrested differentiation results in DS-low scores. Additionally, NF1 or NRAS mutations upregulate ERK signaling which inhibits CEBPA, and also caused DS-low scores. Conclusions: Computational biosimulation based on patients’ individual genomic aberrations reveals a spectrum of DS among patients with MDS. DS strongly predicted HMA response. In general, NR had lower DS and survival than DS-high patients. While patients with DS-low scores might have fared even worse without HMA therapy, we propose that DS-low patients should be considered for novel combination therapy at the time of diagnosis given their unsatisfactory outcomes with HMA therapy as a single agent. Figure 1View largeDownload PPTFigure 1View largeDownload PPT Close modal

Abstract BACKGROUND TMZ-induced G:T mismatches trigger MMR to perform futile repair of O6-methylguanine leading to apoptosis. Without MMR, the G:T mutation is genomically incorporated to produce DNA mutation signature #11. Hypermutation ensues and may compromise survival without the redemptive benefit of immunotherapy. MMR genes are infrequently mutated or deleted. More commonly, MMRD results from epigenetic silencing, transcription failure, or micro-RNA compromising translation. METHODS Comprehensive genomic profiling and Cellworks biosimulation was utilized to diagnose MMRD and correlated with survival in 38 TCGA patients with newly diagnosed, IDH wildtype, m-MGMT GBM treated with adjuvant TMZ. The signaling pathway consequences for MutSα and MutLα were assessed. Kaplan-Meier curves were constructed for PFS and OS. RESULTS Patients were characterized: MMR proficient (Grp 1) and deficient (Grp 2). Half (19/38) had compromise of 1-10 pathways impacting MMR: 3 had deletions of MLH1 or PMS2. Others included deletions of EP300, CREBBP, KMT2A-D, ARID1A, HUS, or EXO1 and amplifications of KDM4A/C or MIR21/155. Grp 1 had significantly higher MMR biosimulation scores than Grp 2. (p=0.00082). The median PFS was 10.51 and 3.58 months (p=0.0072) and median OS was 16.96 and 9.40 (p=0.0059) months in Group 1&2, respectively. CONCLUSIONS Up to half of GBM patients have MMRD caused by pathway dysregulation. Biosimulation of MMRD predicts early progression on TMZ, echoing the long-held observation that TMZ does not trigger apoptosis in MMRD cancers. The study also reports inferior OS for MMRD compared to the historical experience of unmethylated-MGMT patients, suggesting that TMZ-induced hypermutation may compromise survival. As lomustine does not rely on intact MMR, 2nd-line lomustine may have blunted the impact of MMR on OS. Alternatively, upfront lomustine might have produced superior disease control in MMRD patients. Computational biosimulation offers the opportunity to diagnose patients who should not receive TMZ despite m-MGMT and who could benefit from alternative adjuvant strategies.

Background:Relapsed glioblastoma (GBM) is often an imminently fatal condition with limited therapeutic options. Computation biological modeling, i.e., biosimulation, of comprehensive genomic information affords the opportunity to create a disease avatar that can be interrogated in silico with various drug combinations to identify the most effective therapies.Case Description:We report the outcome of a GBM patient with chromosome 12q amplification who achieved substantial disease remission from a novel therapy using this approach. Following next generation sequencing (NGS) was performed on the tumor specimen. Mutation and copy number changes were input into a computational biologic model to create an avatar of disease behavior and the malignant phenotype. In silico responses to various drug combinations were biosimulated in the disease network. Efficacy scores representing the computational effect of treatment for each strategy were generated and compared to each other to ascertain the differential benefit in drug response from various regimens. Biosimulation identified CDK4/6 inhibitors, nelfinavir and leflunomide to be effective agents singly and in combination. Upon receiving this treatment, the patient achieved a prompt and clinically meaningful remission lasting 6 months.Conclusions:Biosimulation has utility to identify active treatment combinations, stratify treatment options and identify investigational agents relevant to patients' comprehensive genomic abnormalities. Additionally, the combination of abemaciclib and nelfinavir appear promising for GBM and potentially other cancers harboring chromosome 12q amplification.

Background: DNA methyltransferase inhibition (DNMTi) with the hypomethylating agents (HMA) azacitidine (AZA) or decitabine, remains the mainstay of therapy for the majority of high-risk Myelodysplastic Syndromes (MDS) patients. Nevertheless, only 40-50% of MDS patients achieve clinical improvement with DNMTi. There is a need for a predictive clinical approach that can stratify MDS patients according to their chance of benefit from current therapies and that can identify and predict responses to new treatment options. Ideally, patients predicted to be non-responders (NR) could be offered alternative strategies while being spared protracted treatment with HMA alone that has a low likelihood of efficacy. Recently, an intriguing discovery of immune modulation by HMA has emerged. In addition to the benefits of unsilencing differentiation genes and tumor suppressor genes, HMA's reactivate human endogenous retroviral (HERV) genes leading to viral mimicry and upregulation of the immune response as a major mechanism of HMA efficacy. Although the PD-L1/PD1 blockade plus HMA has been recognized as a beneficial combination, there are no established markers to guide decision-making. We report here the utility of immunomic profiling of chromosome 9 copy number status as a significant mechanism of immune evasion and HMA resistance.

Background: DNA methyltransferase inhibition (DNMTi) with hypomethylating agents (HMA), azacitidine (AZA) or decitabine (DAC), remains the mainstay of therapy for most high-risk Myelodysplastic syndrome (MDS) patients. However, only 40-50% of MDS patients achieve clinical improvement with DNMTi. Previously, combinations of HMA and histone deacetylase (HDAC) inhibitors have been explored in MDS with varying clinical outcomes. However, the heterogeneity of genomic aberrations in MDS portend widely divergent responses from HDAC inhibition, implying that a predictive clinical decision support tool could select patients most likely to benefit from this combination. We explored the molecular basis of observed clinical response in a group of patients treated with DAC and Valproic-Acid (VPA).

Background: Mantle Cell Lymphoma (MCL) accounts for 3-10% of all non-Hodgkin lymphomas with a median overall survival of 3-4 years. Hyper-CVAD (CVAD) with or without Rituximab constitutes first line therapy for treatment of MCL, yet the use of this combination is associated with high toxicity and only modest efficacy. On the other hand, impressive clinical efficacy has been reported in relapsed MCL patients treated with rituximab and cladribine (RC). Prediction of response based on cancer genomics heterogeneity creates an opportunity to personalize treatment and avoid toxic therapy which has little chance of response. We conducted a study using the Cellworks Biosimulation Platform to identify novel genomic biomarkers associated with response to CVAD and RC among MCL patients.

Further characterization of thymic epithelial tumors (TETs) is needed. Genomic information from 102 evaluable TETs from The Cancer Genome Atlas (TCGA) dataset and from the IU-TAB-1 cell line (type AB thymoma) underwent clustering analysis to identify molecular subtypes of TETs. Six novel molecular subtypes (TH1-TH6) of TETs from the TCGA were identified, and there was no association with WHO histologic subtype. The IU-TAB-1 cell line clustered into the TH4 molecular subtype and in vitro testing of candidate therapeutics was performed. The IU-TAB-1 cell line was noted to be resistant to everolimus (mTORC1 inhibitor) and sensitive to nelfinavir (AKT1 inhibitor) across the endpoints measured. Sensitivity to nelfinavir was due to the IU-TAB-1 cell line’s gain-of function (GOF) mutation in PIK3CA and amplification of genes observed from array comparative genomic hybridization (aCGH), including AURKA, ERBB2, KIT, PDGFRA and PDGFB, that are known upregulate AKT, while resistance to everolimus was primarily driven by upregulation of downstream signaling of KIT, PDGFRA and PDGFB in the presence of mTORC1 inhibition. We present a novel molecular classification of TETs independent of WHO histologic subtype, which may be used for preclinical validation studies of potential candidate therapeutics of interest for this rare disease.

7027 Background: ATRA combined with arsenic trioxide revolutionized the treatment of APL. Based on promising in vitro data, several clinical trials evaluated ATRA combinations in non-APL AML, in which some patients seemed to benefit from the addition. Thus, predicting response a priori is imperative to determine the optimal treatment for each patient. The CBM was used to evaluate the impact of initial therapy with ATRA combined with cytarabine, etoposide, idarubicin (ATRA-CEI) to assess the biomarkers responsible for response in adults with AML. Methods: AML patients participating in clinical trial NCT00151242 had their leukemia sequenced as part of the trial, and genomic profiles were used for computational modeling by the CBM, which uses curated data about genomic aberrations from PubMed as input to generate disease-specific protein network maps and predict drug responses. Disease biomarkers unique to each patient were identified using biosimulation. Digital drug simulations were conducted by measuring the effect of ATRA-CEI on a composite cell growth score of cell proliferation, apoptosis and other hallmarks of cancer. ATRA-CEI was mapped to the patient genome along with a mechanism of action and validated based on the genomic profile and its biological consequences. Results: Of 171 patients treated with ATRA-CEI, 107 (63%) responded (R) and 64 did not (NR). A subset of 18 patients with favorable genomic features were found to be NR and their non-response was correctly predicted by CBM in all 18 cases. Mutations of DNMT3A, EZH2, ASXL, FLT-3, and GART amplification emerged as novel biomarkers of ATRA-CEI failure (only 37 of 107 responders (35%) with these findings, compared to 70 of 107 responders (65%) without these findings (p = 0.0027)). DNMT3A, EZH2, ASXL1 loss of function mutations activate FABP5, a key mechanism of ATRA resistance, and also activate ABCC1 (PgP), which reduces the efficacy of etoposide and idarubicin by upregulating MDR1. In general, monosomy 7 is expected to confer ATRA resistance due to the presence of EZH2 and KMT2E gene deletions. Indeed, 18 of 32 patients with monosomy 7 did not respond. However, the 14 who responded had co-occurrence of deletions involving IGFBP3, PMS2, HUS1, CDK5, XRCC2/4, AKR1B10, and others that overcame ATRA resistance associated with monosomy 7 and were identified by CBM. Use of CBM helps avoid unnecessary use of ATRA in patients unlikely to respond (19% of cases) thus reducing toxicity and cost without changing efficacy, and also identifies those likely to respond, even when they have monosomy 7, where non-response is the norm. Conclusions: ATRA benefits a subset of patients with non-APL AML. CBM predicted response using computational modeling of all genetic alternations, which explains its success versus traditional one-gene-one-drug approaches.

PURPOSE KRAS-mutated ( KRASMUT) non–small-cell lung cancer (NSCLC) is emerging as a heterogeneous disease defined by comutations, which may confer differential benefit to PD-(L)1 immunotherapy. In this study, we leveraged computational biological modeling (CBM) of tumor genomic data to identify PD-(L)1 immunotherapy sensitivity among KRASMUT NSCLC molecular subgroups. MATERIALS AND METHODS In this multicohort retrospective analysis, the genotype clustering frequency ranked method was used for molecular clustering of tumor genomic data from 776 patients with KRASMUT NSCLC. These genomic data were input into the CBM, in which customized protein networks were characterized for each tumor. The CBM evaluated sensitivity to PD-(L)1 immunotherapy using three metrics: programmed death-ligand 1 expression, dendritic cell infiltration index (nine chemokine markers), and immunosuppressive biomarker expression index (14 markers). RESULTS Genotype clustering identified eight molecular subgroups and the CBM characterized their shared cancer pathway characteristics: KRAS MUT/ TP53 MUT, KRAS MUT/ CDKN2A/ B/ C MUT, KRAS MUT/ STK11 MUT, KRAS MUT/ KEAP1 MUT, KRAS MUT/ STK11 MUT/ KEAP1 MUT, KRAS MUT/ PIK3CA MUT, KRAS MUT/ ATM MUT, and KRAS MUT without comutation. CBM identified PD-(L)1 immunotherapy sensitivity in the KRAS MUT/ TP53 MUT, KRAS MUT/ PIK3CA MUT, and KRAS MUT alone subgroups and resistance in the KEAP1 MUT containing subgroups. There was insufficient genomic information to elucidate PD-(L)1 immunotherapy sensitivity by the CBM in the KRAS MUT/ CDKN2A/ B/ C MUT, KRAS MUT/ STK11 MUT, and KRAS MUT/ ATM MUT subgroups. In an exploratory clinical cohort of 34 patients with advanced KRASMUT NSCLC treated with PD-(L)1 immunotherapy, the CBM-assessed overall survival correlated well with actual overall survival ( r = 0.80, P < .001). CONCLUSION CBM identified distinct PD-(L)1 immunotherapy sensitivity among molecular subgroups of KRASMUT NSCLC, in line with previous literature. These data provide proof-of-concept that computational modeling of tumor genomics could be used to expand on hypotheses from clinical observations of patients receiving PD-(L)1 immunotherapy and suggest mechanisms that underlie PD-(L)1 immunotherapy sensitivity.

e16091 Background: Using a 5FU backbone, a variety of combination chemotherapy regimens have been adopted for gastric cancer. However, the genetic heterogeneity of the disease suggests certain drugs would lead to better outcomes for specific patients. In principle, the incorporation of molecular profiling information into treatment selection can generate superior survival for patients. Therefore, we conducted a pilot study using the CBM to identify novel genomic biomarkers of response and resistance to several 5FU-based regimens. Methods: 12 gastric cancer patients treated with 5FU-cisplatin (N = 4), 5FU-VP16 (N = 5) and 5FU-cisplatin-docetaxel (N = 3) were selected from the TCGA database. Mutation and copy number aberrations from each case served as input for CBM to generate patient-specific protein network maps generated from PubMed and other resources. Disease-biomarkers unique to each patient were identified within protein network maps. Drug impact on the disease network was digitally simulated to determine treatment efficacy by measuring effect of chemotherapy on the cell growth score, i.e., a composite of cell proliferation, viability, apoptosis, metastasis, DNA damage and other cancer hallmarks. Effectively, the mechanism of action of each drug was mapped to each patient’s genome and the biological consequences of genomic abnormalities were correlated with response. Results: Of the 12 patients in the study, 10 patients responded to 5FU-based regimens and 2 were non-responders. CBM correctly predicted the therapy response in each therapy segment. Key response criteria were: 5FU-cisplatin - ARID1A (LOF), RAD51 (LOF), MBD1 (LOF), DUT (LOF). 5FU-VP16 - FBXW7 (DEL), RAD51D (DEL), RAD51C (DEL), XRCC2 (DEL), XRCC3 (DEL). 5FU-cisplatin-docetaxel - ARID1A (DEL), CHEK1 (DEL), REV3L (DEL), XRCC3(DEL) and TUBB (AMP). Resistance biomarkers were also identified for these drugs. FBXW7 loss caused cisplatin resistance, but also predicted sensitivity to VP16. Interestingly, all five 5FU-VP16 responders harbored FBXW7 loss. Similarly, in two 5FU-cisplatin-docetaxel non-responders, KIFC1 (GOF), YBX1(AMP), CDC20 (AMP) were key biomarkers for docetaxel resistance. In one of these patients, ATM (DEL), RBBP8 (DEL) and GSTP1 (DEL) predicted response to VP16. Conclusions: The efficacy of individual cytotoxic agents varies in the population. Mutations and copy number changes in DNA checkpoint, DNA repair, and tumor suppressor genes impact responsive to chemotherapy and should be routinely assessed. Notably, 5FU-VP16 combination could have delivered better antitumor efficacy for several non-responders or 25% of the patients in this small cohort. CBM is able to identify both the most active as well inefficacious components of the treatment upfront. We suggest that this approach should be prospectively evaluated in a larger cohort.

e21208 Background: Paclitaxel and carboplatin (PC) is used to treat a wide variety of malignancies including gynecologic, breast, lung, and occult primary cancers. In NSCLC, PC led to a substantial improvement in 1-yr survival from 10% (P alone) to approximately 50% seen with the combination. Nevertheless, a large proportion of patients do not respond. An optimal cytotoxic strategy for managing NSCLC and the discovery of chemotherapy biomarkers to guide treatment selection remain unmet needs in the clinic. Cellworks CBM platform identified a unique chromosomal signature which permits a stratification of which patients are most likely to respond to PC treatment. Methods: 22 patients treated with PC were published in TCGA dataset and selected for analysis. The mutation and copy number aberrations from individual cases served as input into the CBM (generated from PubMed and other online resources) to create a patient-specific protein network map. Disease-biomarkers unique to each patient were identified within protein network maps. Digital drug simulations were conducted by measuring effect of PC on a cell growth score comprised of a composite of cell proliferation, apoptosis, and other cancer hallmarks. Drug simulations were systematically conducted to identify and evaluate therapeutic efficacy. The drug combination was mapped to the patient genome along with a rational mechanism of action and validated based on the genomic profile and its biological consequences. Results: Of the 22 patients treated with PC, 13 had clinical responses and 9 were non-responders. The computer simulation correctly predicted response in 16/22 with 72.73% accuracy, 55.56% specificity and 84.62% sensitivity. CBM identified novel amplified segments of Chromosome 11p and 1p were responsible for non-responsiveness to PC. Key genes on these chromosomes were identified belonging to the autophagy, reactive oxygen species (ROS) scavenging, DNA repair, and microtubule polymerization pathways. Amplification of AMBRA1, ATG13 and TRAF6 (11p) led to autophagy upregulation resulting in low ROS level, a well-documented resistance loop for chemotherapy. SIRT3 and CAT (11p), ROS scavenging genes, were also upregulated due to increase in copy number. CTH (1p) is another key enzyme involved in GSH-mediated ROS scavenging and was also upregulated. Biosimulation indicated a low ROS level was the key reason of resistance to PC. Heightened DNA damage repair due FANCF and ZNF143 (11p amp) and USP1 (1p amp), was another cause of PC resistance. These discoveries suggest that a combination of an autophagy inhibitor / BCL2 mimetic might prove useful to reverse PC resistance associated with 11p and 1p amp. Conclusions: This study highlights how CBM simulation platform can help to identify novel patient segments for therapy response prediction and use drug re-purposing to overcome chemotherapy resistance.

Background: Genomic heterogeneity in leukemic blasts characterizes Acute Myeloid Leukemia (AML) patients and is associated to variable drug response. However, use of genomics to guide therapy has generally been restricted to a single-gene approach, which rarely has sufficient predictive power to be clinically useful. Comprehensive DNA sequencing and biosimulation of the Computational Omics Biology Model (CBM) provide the opportunity and means of predicting treatment outcome in advance of treatment.

Background: The optimal treatment strategy for managing Acute Myeloid Leukemia (AML) and the use of reliable and predictive biomarkers to guide selection of cytotoxic chemotherapy regimens among patients with diverse genomic profiles remain unmet needs in the clinic. The combination of MEC [mitoxantrone (MIT), etoposide (VP16), and cytarabine (ARA-C)] is a commonly used regimen for relapsed or refractory AML patients. Unfortunately, many patients do not respond to MEC, and which of the three drug agents matters most for each individual patient is not known. Predictors of response are needed urgently.

e21211 Background: Cancer management using cytotoxic drugs is hampered by limited efficacy. Hence, a personalized treatment approach matching chemotherapy with appropriate patients remains a persistent and unmet need in the clinic. Genomic heterogeneity among patients creates an opportunity to discern key genomic aberrations and pathways that confer resistance and response to standard treatment options. Pemetrexed, an antifolate, primarily inhibits thymidylate synthase (TYMS) while platinum compounds (carboplatin/cisplatin) trigger DNA damage similar to alkylating agents. When the DNA damage exceeds repair, apoptosis results. We conducted a pilot study using CBM to novel genomic biomarkers of response and resistance to pemetrexed–platinum treatment. Methods: 25 patients who received pemetrexed–platinum therapy were selected from TCGA data. Mutation and copy number aberrations from each case served as input into the CBM (generated from PubMed and other online resources) to create a patient-specific protein network map. Disease-biomarkers unique to each patient were identified within protein network maps. Drug impact on the disease network was digitally simulated to determine treatment efficacy by measuring effect of chemotherapy on the cell growth score, i.e., a composite of cell proliferation, viability, apoptosis, metastasis, DNA damage and other cancer hallmarks. Effectively, the mechanism of action of each drug was mapped to each patient’s genome and biological consequences due to genomic abnormalities were correlated with response. Results: Among the 25 patients, 23 responders (R) and 2 non-responders (NR). The computer simulation correctly predicted response in 19/25 with 76% accuracy, 100% specificity and 73.91% sensitivity. CBM identified co-occurrence of deleted segments of chromosome 6q, 13q, 14q and 17q were responsible for pemetrexed-platinum response. The key genes governing response on these chromosomes included drug transporters and DNA repair pathways. ATP7B (13q) which exports platinum-based drugs was deleted in responders. The loss of REV3L (6q ), MBD1 (6q ), ERCC1 (6q), BRCA2 (13q), BRCA1 (17q), FANCM (14q), RAD51B (14q), XRCC3 (14q ) led to failure of DNA repair. RB1 (13q) del also led to failure of BRG1 mediated DNA repair . Combinations of these repair genes and transporters formed the major response criteria among the 23 responders while these characteristics were absent in non-responders. Conclusions: This pilot study highlights how CBM simulation platform can identify patients for therapy response prediction. Copy number changes impact responsiveness to chemotherapy and should be routinely assessed. We suggest that this approach should be validated prospectively in a larger patient cohort.

Background:Early T-cell Precursor Acute Lymphoblastic Leukemia (ETP-ALL), an orphan disease, is a sub-type of T-Cell Acute Lymphoblastic Leukemia (T-ALL) with very poor prognosis and limited therapy options. ETP-ALL is a heterogeneous disease with many distinct genomic profiles, often with more myeloid than lymphoid characteristics. However, standard of care (SOC) drugs for acute myeloid leukemia (AML) have shown limited efficacy for ETP-ALL (PMID: 32733662, 25435716). The genomic profiles of ETP-ALL patients have more complex cytogenetics and larger numbers of genomic aberrations when compared to non-ETP-ALL (T-ALL) profiles (PMID: 22237106, 30641417). We present an alternative multi-gene analysis approach using the Cellworks Omics Biology Model (CBM) workflow to identify unique, intersecting protein pathways in patient-specific disease profiles. The CBM predictive workflow was used to design novel personalized therapy options for an ETP-ALL representative PEER human lymphoid cell line in comparison to a T-ALL JURKAT cell line. The predicted combination therapies were then validated in a lab model. Methods:A PEER cell line was selected to represent ETP-ALL and a JURKAT cell line was selected as a representative for non-ETP T-ALL. Next Generation Sequencing (NGS) was performed for the PEER cell line. For the JURKAT cell line, publicly available NGS whole exome sequencing from cBioPortal and Sanger, along with array CGH from Agilent, were used. The genomic data for the PEER and JURKAT cell lines were used as inputs to the CBM to generate dynamic patient-specific disease protein network maps. Biomarkers and pathway characteristics unique to the PEER and JURKAT cell lines were identified. A digital drug library of targeted FDA-approved agents was simulated on the disease models using both single drug agents and drug combinations at varying doses. The treatment impact was assessed by quantitatively measuring drug effect on a cell growth score, which is a composite of the quantified values of cell proliferation, survival and apoptosis along with impact on the patient-specific disease biomarker score. Comparative dose response studies were run to assess IC50 differences for both cell lines. Cellworks VenturaTM predicted novel therapy combinations for the ETP-ALL representative PEER cell line, which were then prospectively validated by in vitro experiments. The same therapy options were predicted to be less effective in the T-ALL representative JURKAT cell line, which was also confirmed by in vitro studies. Results:The CBM predicted three novel combination therapies for the ETP-ALL representative PEER cell line: nilotinib + cytarabine, bortezomib + cytarabine and bortezomib + idarubicin. All three therapies were predicted to be less effective in JURKAT cells. In vitro, PEER cells were sensitive to all 3 combinations, as predicted by the CBM; whereas, JURKAT cell lines were not sensitive to the first 2 combinations (as predicted), but were sensitive to bortezomib + idarubicin. The CBM analysis is supported by scientific rationales for these combinations based on the genomics-driven disease characteristics of the cell-line. The reasons for drug sensitivity and resistance were determined. These combinations were then prospectively validated in vitro on both cell lines and the experimental responses matched the predicted outcomes. Conclusion:The Cellworks Omics Biology Model integrates the multiple genomic abnormalities in a patient to identify disease network characteristics unlike other NGS analytic tools that attempt to interpret the impact of each genomic alteration in isolation. CBM identified 3 novel therapy options for ETP-ALL that were validated in vitro, similar to anecdotal experience in vivo. This predictive technology can improve clinical decision-making and identify novel treatment options. Disclosures Howard: Cellworks:Consultancy;Servier:Consultancy, Other: Speaker;EUSA Pharma:Consultancy;Sanofi:Consultancy, Other: Speaker;Boston Scientific:Consultancy.Kumar:Cellworks Research India Private Limited:Current Employment.Pampana:Cellworks Research India Private Limited:Current Employment.Ullal:Cellworks Research India Private Limited:Current Employment.Tyagi:Cellworks Research India Private Limited:Current Employment.Lala:Cellworks Research India Private Limited:Current Employment.Kumari:Cellworks Research India Private Limited:Current Employment.Joseph:Cellworks Research India Private Limited:Current Employment.Raju:Cellworks Research India Private Limited:Current Employment.Balakrishnan:Cellworks Research India Private Limited:Current Employment.Mundkur:Cellworks Group Inc.:Current Employment.Macpherson:Cellworks Group Inc.:Current Employment.Nair:Cellworks Research India Private Limited:Current Employment.Kapoor:Cellworks Research India Private Limited:Current Employment.

Background: Monosomy 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). In general, (-7) is associated with poor response to induction chemotherapy (PMID 12393746). At the same time, not all patients fare poorly so the ability to identify responders and non-responders remains a high priority. Aim: To predict the response for induction chemotherapy in AML patients with (-7) and identify novel genomic signatures of response and resistance. Methods: Genomic data from 13 consecutive patients with (-7) were analyzed using the Cellworks Omics Biology Model (CBM) to generate patient-specific protein network models. All data was anonymized, de-identified and exempt from IRB review. For each model, disease simulations were performed and patients were segregated into HOXA-upregulated and HOXA-downregulated cohorts based on the simulation levels of HOXA5 and HOXA9. Digital drug simulations for induction chemotherapy were accomplished by measuring the impact of drug effect on a cell growth score, a composite of cell proliferation, viability and apoptosis indices. Each patient-specific model was analyzed to identify mechanisms underlying treatment outcomes. Results: 7/13 (54%) of (-7) patients failed to achieve remission after induction chemotherapy (Table 1) which highlighted that (-7) alone does not confer resistance to chemotherapy. CBM identified other genomic alterations that determine chemotherapy response, including DNA repair deficiency genes, mismatch repair (MMR), and homologous recombination repair (HRR) genes. DNA methylation and histone methylation (H3K27me) impacting HOXA gene expression, mainly HOXA5 and HOXA9, were identified as upstream regulators of DNA repair genes. Loss or reduced levels of EZH2 is associated with lower H3K27 methylation and thereby higher expression of HOXA5 and HOXA9 gene targets. High active levels of HOXA correlated with low rates of successful remission induction (22%, n=7) and lower activity levels of HOXA correlated with successful induction chemotherapy (100%, n=4) (Table 1). CBM identified that (-7) results in a decreased expression of EZH2, CARD11, EIF3, PMS2, HUS1, KMT2C (MLL3), CDK5 and IKZF1 genes. Since EZH2 abnormalities alone are not implicated in poor prognosis for AML patients, we sought other aberrations that prevent HOXA upregulation. Using CBM, we identified multiple accompanying aberrations which regulate HOXA genes. Deletions of KAT6A, ASXL1, DNMT3B, DNMT3L genes and high KMT2A-partial tandem duplication correspond to HOXA-upregulation whereas EED amplification, gain of function mutations in DNMT3A, mutations in IDH1/2, MYC and KAT6A amplification and KDM4A deletion result in HOXA-downregulation. Conclusion: Alterations of chromatin regulation have consequences for transcription factors that regulate expression of DNA repair genes. Under conditions where DNA repair is enhanced, induction chemotherapy was 78% less likely to effect remission in (-7) AML patients undergoing induction chemotherapy. Loss of H3K27 methylation associated with loss of PRC2 function by any means resulted in HOXA-upregulation and upregulation of DNA repair genes induced resistance to induction therapy. On the other hand, CBM analysis identified genetic signatures associated with a 100% remission rate from AML induction therapy despite the presence of (-7). Generation of H3K27me caused by PRC2 activation resulting from numerous mechanisms led to HOXA-downregulation and 100% response to induction therapy. Stratification of patients harboring (-7) by HOXA biomarker analysis could inform treatment planning, avoid drug-related adverse events and reduce treatment costs after validation with a larger prospective dataset. Disclosures Castro: Cellworks Group Inc: Consultancy. Watson:Cellworks Group Inc.: Consultancy, Membership on an entity's Board of Directors or advisory committees; Cellmax Life Inc.: Consultancy, Membership on an entity's Board of Directors or advisory committees; Mercy Bioanalytics, Inc.: Consultancy, Membership on an entity's Board of Directors or advisory committees; SEER Biosciences, Inc.: Consultancy, Membership on an entity's Board of Directors or advisory committees; BioAI Health Inc.: Consultancy, Membership on an entity's Board of Directors or advisory committees. Kumar:Cellworks Research India Private Limited: Current Employment. Nair:Cellworks Research India Private Limited: Current Employment. Grover:Cellworks Research India Private Limited: Current Employment. Sahu:Cellworks Research India Private Limited: Current Employment. Mohapatra:Cellworks Research India Private Limited: Current Employment. G:Cellworks Research India Private Limited: Current Employment. Agarwal:Cellworks Research India Private Limited: Current Employment. Suseela:Cellworks Research India Private Limited: Current Employment. Ganesh:Cellworks Research India Private Limited: Current Employment. Sauban:Cellworks Research India Private Limited: Current Employment. Kumar:Cellworks Research India Private Limited: Current Employment. Raman:Cellworks Research India Private Limited: Current Employment. Singh:Cellworks Research India Private Limited: Current Employment. Basu:Cellworks Research India Private Limited: Current Employment. Lunkad:Cellworks Research India Private Limited: Current Employment. Mundkur:Cellworks Group Inc.: Current Employment. Macpherson:Cellworks Group Inc.: Current Employment. Kapoor:Cellworks Research India Private Limited: Current Employment. Howard:Servier: Consultancy, Other: Speaker; Boston Scientific: Consultancy; Sanofi: Consultancy, Other: Speaker; EUSA Pharma: Consultancy; Cellworks: Consultancy.