Search

Background: The advance of next generation sequencing (NGS) leads to give the abundant genetic information of breast cancer (BC). We developed the prediction model of neoadjuvant chemotherapy (NAC) response and long term survival using genetic characteristics in cooperation with clinicpathologic features. Methods: Early and locally advanced BCs which would be planned to receive standard NAC (four cycles of anthracycline plus cyclophosphamide and four cycles of docetaxel or docetaxel plus trastuzumab for human epidermal growth factor receptor 2[HER2+] disease) followed by curative surgery. We prospectively collected tumor tissue and matched blood three times for each patients: at BC diagnosis (T1), three weeks after the first cycle of NAC (T2), and curative surgery (T3). Whole exome sequencing (WES) and RNASeq were performed to detect somatic mutation, mutational signature, tumor mutational burden (TMB) and PAM50 prediction was to classify intrinsic subtype. In terms of clinical variables, clinical stage and IHC subtype at diagnosis, pathologic complete response (pCR), residual cancer burden (RCB) class and distant recurrence free survival (DRFS) were used for prediction model development. Logistic regression was used for predicting RCB class and pCR with clinical and genomic variables at T1. Univariate and multivariate Cox regression were performed to identify prognostic factors for DRFS. Results: In total, 210 patients who were treated with scheduled NAC were enrolled. Finally, WES in 252 BC tissues (T1:123, T2:106 and T3:23) RNASeq in 198 BC tissues (T1:100, T2:77 and T3:21) were conducted from 123 patients. In NAC response, 52 patients achieved pCR(42.3%) and 14 patients were in RCB class 3, 39 in class 2, 14 in class 1 and 51 in class 0. Median follow up duration was 44months and distant recurrence was observed in 13 patients. TP53 mutation (60%) was the most commonly detected genetic alteration followed by TTN(31%), ERBB2(25%) and PIK3CA(18%). Eight BCL11A (3%) and 3 MPL mutations (1%) were also observed respectively. Serial tumor sequencing suggested that mutation profile have not changed during NAC. In terms of mutational signature, signature 3 was most frequently observed and no difference was observed in serial sequencing data. In addition, APOBEC (sig.2) and HRD (sig.3) were observed differently according to intrinsic subtype. In terms of intrinsic subtype, basal subtype decreased during NAC and luminal A type increased. Median TMB was 87 (range, 14-570). In prediction model, clinical N3(cN3) stage (Odds ratio [OR] of cN3 vs. not : 4.536, 95% confidence interval [CI]: 0.925, 44.016, p=0.042) and intrinsic subtype at BC diagnosis (OR of non-luminal vs. luminal types: 0.208, 95% CI: 0.062, 0.604, p=0.001) were associated with pCR. Clinical stage (OR of clinical stage III vs. II: 4.115, 95% CI: 1.094, 19.217, p=0.021), TMB (OR of TMB high vs. low: 0.255, 95% CI: 0.043,1.042, p.value=0.045), MPL genetic alteration (OR of mutation vs. wild type: 16.347, 95% CI: 0.797,1013.88, p.value=0.037), BCL11A (OR of mutation vs. wild type: 16.347, 95% CI: 0.797,1013.88, p.value=0.037) were associated with RCB class (3 vs. others). In terms of DRFS, prediction model consisted of clinical stage (Hazard ratio [HR] of stage III vs. II: 3.496, 95% confidence interval [CI]:1.337, 5.654), PIK3CA (HR of mutation vs. wild type: 1.572, 95% CI:-0.008, 3.152) TMB (HR of high vs. low: -1.999. 95% CI:-3.669, -0.329) had c-index of 0.861(95% CI:0.805, 0.916). Conclusions: During NAC, somatic mutation and mutation signature were consistently maintained while intrinsic subtype dynamically changed. In prediction model, TMB, PIK3CA mutation and clinical stage showed predictive roles on DRFS of BC in NAC setting. Citation Format: Ji-Yeon Kim, Kyunghee Park, Woong-Yang Park, Jeong Eon Lee, Seok Won Kim, Seok Jin Nam, Se Kyung Lee, Zhengyan Kan, Yeon Hee Park. Prediction model of the response of neoadjuvant chemotherapy and long term survival according to multi-omic profiling in cooperation with clinicopathologic features in patients with breast cancer [abstract]. In: Proceedings of the 2021 San Antonio Breast Cancer Symposium; 2021 Dec 7-10; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2022;82(4 Suppl):Abstract nr P1-08-07.

Background: Neutrotrophin receptor tyrosine kinase (NTRK) gene fusions (NTRK1, NTRK2, or NTRK3) are oncogenic drivers of various tumor types. The NTRK fusion was detected in less than 5% of breast, colorectal, lung or any other types of cancers. However, large scaled next generation sequencing data for NTRK fusion in breast cancer have not existed. In this study, we performed RNASeq and fusion analysis including NTRK genes. Methods: We prospectively collected BC tumor tissues from the translational research conducted in Samsung Medical Center. Fusion was predicted from RNAseq using the following seven softwares(SWs): ChimeraScan, DeFuse, MapSplice, TophatFusion, STAR.Arriba, STAR.fusion, and STAR.SEQR. To remove false-positive fusion calls, calls less than 3 left/right spanning reads and 10 total supporting reads were removed. After filtering out the blacklist fusion calls which were recurrently detected, calls commonly predicted in more than two SWs were analyzed. Results: In total 629 BC samples, 613 samples were finally analyzed after quality control (QC) of RNASeq. According to immunohistochemistry (IHC) profile, 356(58.7%) was hormone receptor (HR) positive human epidermal growth factor receptor 2 (HER2) negative BC, 42 (6.9%) of HR positive HER2 positive BC, 53 (8.7%) of HR negative, HER2 positive and 155(25.6%) of triple negative BC (TNBC). PAM50 prediction informed that 174(28.9%) of luminal A, 151(24.6%) of luminal B, 170 (27.7%) of basal-like, 85 (13.9%) of HER2-enriched and 33 (5.4%) of normal. In median number of fusion events, 12 was called by ChimeraScan (Interquartile range [IQR]: 5, 33), 57 by DeFuse (IQR: 33, 82), eight by Mapsplice (IQR: 5, 12), two by TophatFusion (IQR: 0, 4), five by STAR.Arriba (IQR: 2, 12), two by STAR.fusion (IQR:0, 5) and three by STAR.SEQR fusion caller (IQR: 1, 7) after call filtering. After initial fusion call, we excluded the results from ChimeraScan and DeFuse fusion callers because of discrepancy of number of called fusion events. In five fusion callers, median number of fusion events was eight (IQR :2,20) per BC sample. In terms of NTRK fusion, we detected NTRK2-BANCR fusion event in one TNBC patients (1/425, 0.2%). This fusion was detected in four of five SWs for fusion detection with significant number of supporting reads in RNASeq. NTRK2-BANCR fusion was the out-of-frame fusion, which C-terminal truncated protein kinase domain of NTRK2 and its partner long non-coding RNA BANCR was combined and RNA expression of this fusion was increased. Other fusion events of BCs were NCOR2-PARP4 (3.7%), BRD4-NWD1 (3.7%), ESR1-RGS17 (1.8%), FGFR1-TACC1 (0.2%) and MKRN1-BRAF (0.2%). In BC subtype according to IHC, fusion events were more frequently observed in TNBC compared with other subtypes regardless of the type of fusion filters. In terms of intrinsic subtype, fusion events were most frequently observed in basal like type and least in normal like intrinsic subtype (all ps Citation Format: Ji-Yeon Kim, Kyunghee Park, Woong-Yang Park, Jeong Eon Lee, Seok Won Kim, Seok Jin Nam, Jonghan Yu, Young-Hyuck Im, Jin Seok Ahn, Yeon Hee Park. Fusion analysis including NTRK fusion in breast cancers (BC): From RNASeq data analysis from 629 BC tissue samples [abstract]. In: Proceedings of the 2021 San Antonio Breast Cancer Symposium; 2021 Dec 7-10; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2022;82(4 Suppl):Abstract nr P3-13-08.

CDK4/6 inhibitors such as palbociclib in combination with endocrine therapy (ET) have remarkablyimproved the outcome of patients with ER+/HER2- metastatic breast cancer (MBC). However, manypatients are intrinsically resistant to CDK4/6i therapy, and those who respond eventually acquireresistance. Although high baseline CCNE1 expression and rare alterations in RB1 and FAT1 geneshave been shown to be associated with CDK4/6i resistance, the molecular mechanisms of CDK4/6iresistance are complex and remain poorly understood. To better understand and overcome CDK4/6iresistance, we performed multi-omics profiling of paired tumor biopsies from ER+/HER2- MBCpatients treated with palbociclib combined with ET. Tumor biopsies taken at pre-treatment, on-treatment, and progressive disease (PD) from 71 patients were profiled using whole-exomesequencing (WES), whole-transcriptome sequencing (RNA-Seq) and IHC analysis. Ourcomprehensive analysis identified several tumor intrinsic molecular markers associated with worsePFS, including the Luminal B subtype (p=0.012, HR=2.593), BRCA1/2 pathogenic mutation (p=0.012,HR=2.67) and mutation signatures linked to APOBEC enzymatic activity (p=0.002, HR=3.19).Conversely, the estrogen response signature (p=0.006, HR=0.43) was associated with favorableprognosis. Unsupervised analysis revealed a cluster of tumors enriched in homologousrecombination deficiency (HRD) linked genomic scars that was associated with poor prognosis(p=0.005, HR=2.49). Of note, these HRD-high tumors responded even more poorly to treatment whenco-occurring with TP53 somatic mutations. Integrative analysis further identified three poorprognosis clusters (IC2-4) enriched in Luminal B, proliferative and HRD features when compared tothe favorable prognosis cluster (IC1).Comparing baseline vs. PD samples, we observed a pattern of post-treatment enrichment for the poorprognosis markers. In addition, breast cancer-associated genes such as BRCA1/2, TP53 and PTENharbored a higher prevalence of genomic alterations including somatic mutation, amplification,. deletion and gene fusion at PD. Cell cycle gene expression and signatures also markedly increased atPD compared to baseline whereas estrogen response signatures decreased. Upon diseaseprogression, tumors had frequently switched to molecular subtypes with aggressive and estrogenindependent characteristics, demonstrating high plasticity in response to CDK4/6i and ET treatment.These patterns of acquired resistance were validated by IHC analysis of cyclins E1 and E2, Ki67 andpRb. To investigate the genomic alterations responsible for acquired resistance, we compared 21paired baseline and PD samples. We observed that PD-specific RB1 loss-of-function events occurredwith higher prevalence than previously reported, underscoring a major role of cell cycle de-regulation in conferring resistance to CDK4/6 inhibition. In this prospective longitudinal multi-omicsstudy, we identified novel candidate biomarkers that can be used to improve prediction of responseto CDK4/6i. In addition, we derived new insights into the molecular mechanisms of drug resistanceto palbociclib plus ET that will help guide therapeutic strategies and drug development inHR+/HER2− MBC. Citation Format: Zhengyan Kan, Seock-Ah Im, Kyunghee Park, Ji Wen, Kyung-Hun Lee, Yoon-La Choi, Won-Chul Lee, Ahrum Min, Vinicius Bonato, Seri Park, Sripad Ram, Dae-Won Lee, Ji-Yeon Kim, Su Kyeong Lee, Won-Woo Lee, Jisook Lee, Miso Kim, Scott L. Weinrich, Han Suk Ryu, Tae Yong Kim, Stephen Dann, Diane Fernandez, Jiwon Koh, Song Yi Park, Shibing Deng, Eric Powell, Rupesh Kanchi Ravi, Jadwiga Bienkowska, Paul A. Rejto, Woong-Yang Park, Yeon Hee Park. Serial genomic profiling reveals molecular mechanisms of breast cancer resistance to palbociclib [abstract]. In: Proceedings of the 2021 San Antonio Breast Cancer Symposium; 2021 Dec 7-10; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2022;82(4 Suppl):Abstract nr PD2-08.

Background: Young-PEARL study showed median progression-free survival (PFS) of 20·1 months in the palbociclib plus endocrine therapy group versus 14·4 months in the capecitabine group (hazard ratio 0·659 [95% CI 0·437-0·994], log-rank p=0·0235). We conducted exploratory biomarker analysis to predict the efficacy of the trial. Methods: This was a phase II trial that randomized 184 patients with HR+ MBC in premenopausal women to palbociclib plus exemestane with GNRH agonist (Arm A, n=79) versus capecitabine (Arm B, n=62). We performed targeted sequencing (CancerSCANTM) containing 375 cancer-related genes (141 patients) and whole transcriptome sequencing (165 patients) using baseline tumor samples to examine genomic alteration in relation to drug response in terms of PFS. Result: By research-based PAM50 subtyping, 73% of patients classified as Luminal (Luminal A and Luminal B), and showed better prognosis in all patients and Arm A (p Conclusions: The alteration of a few genes including Rb1 loss may be associated with resistance of palbociclib in HR-positive premenopausal population with MBC. Luminal type showed better prognosis, and BRCA2 pathogenic mutation showed worse prognosis regardless luminal/non-luminal type. ESR1 mutation was found in low population frequency because all patients didn’t received AI therapy. Further exploration of molecular variables is warranted to determine and validate biomarkers of efficacy. Clinical trial information: NCT02592746. Citation Format: Kyunghee Park, Gun Min Kim, Kyung Hae Jung, Seok Yun Kang, In Hae Park, Jee Hyun Kim, Hee Kyung Ahn, Woong-Yang Park, Seock-Ah Im, Yeon Hee Park. Exploratory biomarker analysis of Young-PEARL [palbociclib plus exemestane with GnRH agonist versus capecitabine in premenopausal women with HR (hormone receptor)-positive, HER2-negative metastatic breast cancer (MBC)] study [abstract]. In: Proceedings of the 2020 San Antonio Breast Cancer Virtual Symposium; 2020 Dec 8-11; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2021;81(4 Suppl):Abstract nr PS5-19.

Introduction Recent transcriptome analysis developed a holistic tumor microenvironment (TME) classification platform based on immune and fibrotic markers. This TME classification had four categories, immune enriched, fibrotic (IE/F); immune enriched, non-fibrotic (IE); Fibrotic (F); and Immune Desert (D). And these four TME subtypes are a predictive biomarker to immunotherapy in multiple cancer and four subtypes have been changed during treatment. Previously our study suggested baseline immune features and dynamic immune response on-treatment were predictive of treatment outcome in BC with neoadjuvant chemotherapy (NAC). In this study, we evaluated the impact of TME classification and dynamic change of TME classification on treatment outcome in BC with NAC. Methods Early and locally advanced BCs which would be planned to receive standard NAC (four cycles of anthracycline plus cyclophosphamide and four cycles of docetaxel or docetaxel plus trastuzumab for human epidermal growth factor receptor 2[HER2+] disease or six cycles of docetaxel, carboplatin, trastuzumab and pertuzumab for HER2+ disease) followed by curative surgery. We prospectively collected tumor tissue and matched blood three times for each patients: at BC diagnosis (T1), three weeks after the first cycle of NAC (T2), and curative surgery (T3). RNASeq was performed to classify TME subtype. In terms of clinical variables, clinical stage and IHC subtype at diagnosis, pathologic complete response (pCR), distant recurrence free survival (DRFS) and overall survival (OS) were used. Generalized logistic regression was used for predicting RCB class and pCR with clinical and genomic characteristics at T1. Kaplan-Meier analysis were performed to analysis DRFS and OS. Results In total, 210 patients who were treated with scheduled NAC were enrolled. Finally, RNASeq in 240 BC tissues (T1:119, T2:91 and T3:30) were conducted from 142 patients. In 119 BCs which was performed RNASeq at T1, hormone receptor(HR)+, HER2- BC was 32 (26.9%), 29 of HR+HER2+ BC (24.4%), 18 of HR-HER2+ BC (15.1%) and 44 of triple negative BC (TNBC) (37.0%). In TME classification, immune desert (D) was most frequently observed (45.3%), followed by IE (35.3%), F (10.1%) and IE/F (9.2%). The association between BC subtype and TME subtype suggested that HR+HER2- BC was frequently categorized into D (22 of 32, 68.8%) whereas TNBC was into IE (24 of 44, 54.5%) (p< 0.001). TME subtype has been dynamically changed during NAC. At T2, IE subtype was most frequently observed (27.5%) followed by D (25.3%), IE/F (24.2%) and F (23.1%). The inclination of TME change were different according to NAC response. In BC achieved pCR, IE/F subtype had increased (4 at T1 and 10 at T2) and decrease of D subtype (15 at T1 and 3 at T2). In BC with non-pCR, IE/F subtype had slightly increased at T2 (7 at T1 and 12 at T2) but there was no IE/F subtype at T3 point. Contrarily, D subtype had decreased at T2 but increased at T3 (39 at T1, 20 at T2 and 24 at T3). The impact of TME subtype was different according to pCR status. In BC with pCR, F subtype had poor prognosis in DRFS and OS compared to other subtype ([5year DRFS rate for F vs. others: 66.7% vs. 93.2%, p=0.028], [5year OS rate for F vs. others: 70.7% vs. 100%, p< 0.001]). In BC without pCR, there was no different DRFS and OS according to TME subtypes. Conclusion Our data suggested that TME subtype has been changed during NAC and the subtype switching was affected by the NAC response. Moreover, TME subtype may have prognostic role in DRFS and OS according to pCR status. Citation Format: ji-Yeon Kim, Kyunghee Park, woong-Yang Park, Jeong Eon Lee, Seok Won Kim, Seok Jin Nam, Se Kyung Lee, Yeon Hee Park. Tumor microenvironment subtype influence the response of neoadjuvant chemotherapy in breast cancer [abstract]. In: Proceedings of the 2022 San Antonio Breast Cancer Symposium; 2022 Dec 6-10; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2023;83(5 Suppl):Abstract nr P6-01-37.

Background/Purpose: Next-generation sequencing (NGS) is being incorporated rapidly into clinical practice. Fine-needle aspiration biopsy (FNAB) specimens have been used feasibly in molecular analysis including direct sequencing and microarrays. They are readily available and enriched in malignant cells, thus providing opportunities for genomic analysis for more clinical samples. In this study, we assessed the feasibility and sensitivity of FNAB for the detection of somatic mutations by NGS compared to bulk tissue. Methods: Bulk tissue and FNAB was sampled via skin superficial to the palpable tumor from surgically resected breast cancer specimen. DNA was extracted from the bulk tissues and FNAB samples obtained from twelve patients. Somatic mutations detected from whole exome sequencing (WES) by next-generation sequencing (NGS) (HiSeq 2500, Illumina) were analyzed for corresponding pairs of bulk tissue and FNAB. Verification of somatic mutations detected exclusively from FNAB and known to be clinically relevant to breast cancer was carried out by Sanger sequencing. Invasive tumor percentages of bulk tissues were evaluated using hematoxylin and eosin (H&E)-stained sections. Results: Average depth of coverage were 158.8x and 158.3x for bulk tissue and FNAB, respectively. Number of detected somatic mutations ranged from 2 to 153 (median 18.5) and 19 to 210 (median 39.5) for bulk tissue and FNAB, respectively. Ten specimens had more mutations detected exclusively from FNAB than from bulk tissue. Allele fractions plotting of corresponding pairs of bulk tissue and FNAB showed good, intermediate, and poor correlation in five, two, and five specimens, respectively. H&E-stained sections of bulk tissue from the five specimens with good correlation contained an invasive tumor percentage of 45 to 98%, whereas those from five specimens with poor correlation contained 0 to 25%. Three of the poorly correlated bulk tissues were judged to have 0% of invasive tumor. Among mutations detected exclusively from FNAB, eighteen different genes of interest in 22 foci were evaluated for both FNAB and corresponding bulk tissue by Sanger sequencing. In the results, three mutations (PIK3CA, TP53 x2) were verified in FNAB samples but not in the bulk tissue. Conclusion: WES was successfully carried out in all pairs of bulk tissue and FNAB from twelve breast cancer patients. In samples with high tumor content somatic mutation profiles showed high correlation between the two samples whereas samples with low tumor content failed to show correlation. The failure was mostly due to the scarcity of tumor portions in the bulk tissues, indicating that FNAB more reliably retained malignant tumor portion. This study suggests that FNAB is an easy and feasible method, and furthermore, provides a more reliable specimen for NGS analysis where somatic mutations could be identified for potential prognostic or therapeutic benefits. Citation Format: Han-Byoel Lee, Jisun Kim, Kyung-Min Lee, Je-Gun Joung, Hae-ock Lee, Min Kyoon Kim, Eunshin Lee, Jongjin Kim, Tae-Kyung Yoo, Yun-Gyoung Kim, Young Joon Kang, Han Suk Ryu, In-Ae Park, Hyeong-Gon Moon, Dong-Young Noh, Woong-Yang Park, Wonshik Han. Feasibility and sensitivity of fine-needle aspiration biopsies for the detection of somatic mutations using next-generation sequencing in breast cancer [abstract]. In: Proceedings of the Thirty-Seventh Annual CTRC-AACR San Antonio Breast Cancer Symposium: 2014 Dec 9-13; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2015;75(9 Suppl):Abstract nr P2-05-07.

Tumor microenvironment interacts with cancer cells, causing tumor progression. Tumor microenvironment consists of cellular and acelluar components. We investigated cellular components of tumor microenvironment and heterogeneity of ovarian cancer. We performed single cell RNA sequencing analysis of cellular components of malignant ascites from a patient with advanced ovarian cancer. Among the various cell types including cancer cells, B cells, macrophages and fibroblasts, macrophage was the most dominant in malignant ascites. To determine the heterogeneity of ovarian cancer, we obtained primary ovarian cancer tissue from ovarian cancer patient. We performed single cell RNA sequencing analysis of cellular components derived from ovarian cancer tissue. We determined intra-heterogeneity of ovarian cancer and cellular components of ovarian cancer tissue. Taken together, we determined cellular components of ovarian cancer tissue and ascites and intra-heterogeneity of ovarian cancer. This study can provide insight of tumor microvironment and help understanding of ovarian cancer heterogeneity. Citation Format: Untack Cho, Yong Sang Song, Woong-Yang Park, Hae-Ock Lee, Youngjin Han. Evaluating heterogeneity of tumor microenvironment using single cell RNA sequencing [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2019; 2019 Mar 29-Apr 3; Atlanta, GA. Philadelphia (PA): AACR; Cancer Res 2019;79(13 Suppl):Abstract nr 124.

Although targeted deep sequencing of cell-free DNA (cfDNA) was recently used to investigate tumor somatic mutations in particular subtypes of non-Hodgkin lymphomas (NHLs), the immense genetic heterogeneity across subtypes poses a hurdle to design a universal gene panel applicable for diverse subtypes of NHLs. We designed a panel targeting 66 genes associated with NHLs and performed targeted deep sequencing to analyze plasma cfDNA from patients with various subtypes of NHLs. Genetic profiling in plasma cfDNA using the method resulted in 88.0% sensitivity and >99% specificity in detecting mutations present at a frequency greater than 20% in the tumor biopsies. Furthermore, the level of ctDNA significantly decreased and increased depending on designated clinical responses to therapy and disease progression. These results demonstrated that ctDNA sensitively indicated the presence of cancer and reliably correlated with tumor burden, suggesting potential utility of the method for patients with various subtypes of NHLs. Citation Format: Yeon Jeong Kim, Seung-Ho Shin, Danbi Lee, Woong-Yang Park, Seok Jin Kim, Won Seog Kim, Donghyun Park. Analysis of circulating tumor DNA by targeted ultra-deep sequencing across various non-Hodgkin lymphoma subtypes [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2019; 2019 Mar 29-Apr 3; Atlanta, GA. Philadelphia (PA): AACR; Cancer Res 2019;79(13 Suppl):Abstract nr 2511.

Background: Impairment in DNA damage response and repair (DDR) pathway is known as a predictive biomarker of platinum sensitivity in many cancer types. In the era of cancer immunotherapy, DDR alteration is re-emphasized as a predictive biomarker of immune-checkpoint inhibitor due to its positive correlation to tumor mutation burden (TMB). Patients and methods: Target gene sequencing (381 genes) was conducted from 100 extensive disease (ED) and 66 limited disease (LD) small cell lung cancer (SCLC) patients. DDR-related genes were pre-defined by literature review, and mutations were classified as double-strand breaks (DSB, n=82): homologous recombination (n=54), non-homologous end joining (NHEJ, n=19), and Fanconi anemia (FA, n=32); or single-strand breaks (SSB, n=31): mismatch repair (MMR, n=19), base excision repair (BER, n=7), and nucleotide excision repair (NER, n=6). Based on the mutation profile, correlation of TMB and survival outcomes were analyzed. Results: Compared to patients with an intact DDR pathway (n=70), a higher TMB was observed in patients with homologous recombination (P Conclusions: DDR pathway alterations, both DSB and SSB, in SCLC have a positive correlation with high TMB. However, it has demonstrated limited value in prediction of platinum efficacy. Tumor mutation burden and its correlation with DNA damage related gene setsFAHR(-) n=131(+) n=32P(-) n=112(+) n=54PLow TMB69 (52.7%)14 (43.8%)0.47965 (58.0%)20 (37.0%)0.018High TMB62 (47.3%)18 (56.2%)47 (42.0%)34 (63.0%)NHEJBER(-) n=147(+) n=19P(-) n=156(+) n=7P*Low TMB78 (53.1%)7 (36.8%)0.27780 (51.3%)3 (42.9%)0.716High TMB69 (46.9%)12 (63.2%)76 (48.7%)4 (57.1%)MMRNER(-) n=147(+) n=19P*(-) n=157(+) n=6P*Low TMB81 (55.1%)4 (21.1%)0.00781 (51.6%)2 (33.3%)0.437High TMB66 (44.9%)15 (78.9%)76 (48.4%)4 (66.7%)DSBSSB(-) n=84(+) n=82P(-) n=135(+) n=31PLow TMB50 (59.5%)35 (42.7%)0.04476 (56.3%)9 (29.0%)0.006High TMB34 (40.5%)47 (57.3%)59 (43.7%)22 (71.0%) Citation Format: Sehhoon Park, Hayoon Lee, Boram Lee, Jong-Mu Sun, Woong-Yang Park, Jin Seok Ahn, Myung-Ju Ahn, Se-Hoon Lee, Keunchil Park. DNA damage response and repair pathway alteration and its association with tumor mutation burden and platinum-based chemotherapy in small cell lung cancer [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2019; 2019 Mar 29-Apr 3; Atlanta, GA. Philadelphia (PA): AACR; Cancer Res 2019;79(13 Suppl):Abstract nr 3137.

The molecular bases underlying neoadjuvant chemotherapy (NAC) response are poorly understood. To elucidate the effects of NAC on breast tumor biology and its association with clinical outcome, we have conducted WES and RNA-Seq profiling of a longitudinal breast cancer (BC) cohort consisting of 146 cases (281 tumors, 109 pairs), including 55 (38%) that achieved pathologic complete responses (pCR) and 91 (62%) that harbored residual diseases at time of surgery. Tumor biopsies were collected for each patient at three time points - pre-treatment, three weeks after the first cycle of anthracycline and cyclophosphamide (AC) and at the time of surgery, after 3 more cycles of AC followed by 4 cycles of taxane. In addition to somatic mutations and copy number alterations, we also derived a comprehensive set of genomic and molecular features for each tumor including chromosomal instability, loss-of-heterozygosity, mutation burden, mutation signatures and expression signatures for oncogenic signaling pathways and immune cell subsets. Virtual microdissection analysis inferred 14 factors that represent distinct tissue compartment including a tumor infiltrating lymphocyte (TIL) factor and revealed that initial NAC treatment increased stromal and adjacent normal tissue fractions while reducing tumor cellularity. NAC also induced dynamic changes in immune gene expressions over time, a pattern that was validated through detecting and quantifying the density of TILs from H&E images. To investigate NAC induced changes in tumor intrinsic biology we classified tumors into five oncogenic cellular states on a reference Onco-GPS map defined by transcriptional signatures from breast cancer cell lines. We observed that, as a result of NAC treatment, tumors often change from one oncogenic state to another, transiently upregulating an EMT program that appears to mediate drug resistance and increase the likelihood of residual disease. Multiple regression and multivariate analyses were performed to identify predictive biomarkers of pCR status while adjusting for BC subtypes and tumor purity. We found that ER+ subtype and estrogen response signature but not Ki-67 were independently associated with NAC response. Within TNBC, the immunomodulatory subtype was enriched in responders while the basal-like subtype had the poorest response. Pretreatment TIL and changes in TIL level over time were independently associated with NAC response, implicating anti-tumor immunity in mediating the efficacy of chemotherapies. Through multi-omics characterization of longitudinally paired tumor biopsies, we have revealed dynamic changes in the tumor molecular states in BC patients undergoing NAC treatment, identified molecular markers of treatment outcome and derived insights into the mechanism of action as well as resistance to an important class of therapy. Citation Format: Samir Lal, Ying Ding, Jeong Eon Lee, Soo-Hyeon Lee, Se Kyung Lee, Jae-Yong Nam, Jong Han Yu, Yoon-la Choi, Seok Won Kim, Seok Jin Nam, Ji-Yeon Kim, Sripad Ram, Eric Powell, Keith A. Ching, Pablo Tamayo, William Kim, Huwate Yeerna, Soo Youn Cho, Vinicius Bonato, Shibing Deng, Jinho Kim, Hyuntae Shin, Woong-Yang Park, Paul A. Rejto, Jadwiga Bienkowska, Yeon-Hee Park, Zhengyan Kan. Multi-omics profiling of breast cancers during neoadjuvant chemotherapy identified distinct molecular changes and biomarkers associated with clinical response [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2019; 2019 Mar 29-Apr 3; Atlanta, GA. Philadelphia (PA): AACR; Cancer Res 2019;79(13 Suppl):Abstract nr 1684.

Background and Purpose In several breast cancer analyses of targeted gene panels, there was considerable concordance of mutations observed between primary tumors and matched metastases. We are to investigate to gain an understanding of the underlying genetics leading to BC metastasis to compare genomics data using paired biopsies between primary breast and several metastatic sites. Methods We performed CancerSCANTM of forty-eight patients with paired primary and metastatic tumors, which is 380-gene targeted panel sequencing using HiSeq 2500 sequencing platform (Illumina). We made variant calls using Mutect for SNV, Pindel for INDEL, Contra for CNV, and in-house software for Fusion with default parameters. Tumor mutational burden (TMB) was measured by number of mutations per megabase (mb), and TMB was divided into two groups: high TMB (>=20mutations/mb) and low TMB ( Results Seven among 48 patients (14.6%) showed discordant subtypes between primary and metastatic tumors, discordant patterns were diverse. Lung and pleural was the most frequent organ in metastatic tumors, and liver and lymph node were followed. There were 15 shared, 6 primary-specific, 4 metastatic-specific mutations per patient on average, and no significant difference in overall survival (OS) with respect to mutation-sharing patterns. TMB and number of mutations related to mismatch-repair (MMR) genes such as NF1 (P=0.0398) and PMS2 (P=0.1020) tended to be reduced in metastatic tumors compared to paired primary tumors. Mutational signature 3 (BRCA1/2) was increased in metastatic tumors, and signature 5 (General Cancer) was decreased (P>0.05). defect in MMR genes (P=0.0445) and germline BRCA mutation (P=0.0186) were significantly associated to OS, but TMB status, and mutational signature related to MMR respectively were not significant variables for OS according to the molecular profiles in primary tumors. Patients with PMS2 (P=0.156) and NF1 (P=0.0734) multi-hit mutations were better prognosis. Combining molecular profiles in primary tumors such as TMB and mutational signature 6 (MMR) and signature 13 (APOBEC) showed significant differences in OS: low TMB with high signature 6 plus 13 as worse, high TMB as intermediate, and low TMB with low signature 6 plus 13 as better prognosis (P=0.0228). Conclusions With the combination of TMB and MMR-related molecular profiles, we could stratify patients into three groups according to OS. Citation Format: Kyung Hee Park, Eunjin Lee, Seri Park, Se Kyung Lee, Seok Won Kim, Jeong Eon Lee, Seok Jin Nam, Ji-Yeon Kim, Hee Kyung Ahn, Woong-Yang Park, Yeon Hee Park. Analysis of the targeted sequencing with RNA Seq. results between primary breast cancer and metastatic sites: Matched paired analysis [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2019; 2019 Mar 29-Apr 3; Atlanta, GA. Philadelphia (PA): AACR; Cancer Res 2019;79(13 Suppl):Abstract nr 2520.

Purpose: Although both gBRCA1/2 mutations are known to increases incidences of breast and ovarian cancer by abruption in homologous recombination pathway, gBRCA1 mutant mainly present with a molecular subtype of triple negative breast cancer (TNBC) and gBRCA2 with hormone receptor (HR) positive which we have little understanding regarding the underlying pathophysiology. In this study, we have comprehensively analyzed clinical characteristics and somatic mutation in gBRCA1/2 mutant patients. Methods: Patients with gBRCA mutation tested (n=2720) in Samsung Medical Center between Jan 2007 to Oct 2018 were retrospectively reviewed. 386 patients were identified as gBRCA mutant and follow-up period less than 2 years (n=105), insufficient clinical data (n=22) and no surgical treatment (n=6) were excluded. Total of 259 patients, gBRCA1 (n=128), gBRCA2 (n=126) and both gBRCA1/2 mutation (n=5), were analyzed. Among the study population, 46 patients had deep target sequencing data with 40 patients also available for matched transcriptome outcome. Results: Median age was 40 years old (range 23-68). Patients were mostly under pre-menopausal status (81.2%) with invasive ductal carcinoma pathology (86.2%). Initial stage at diagnosis were stage 0 (4.7%), 1 (30.3%), 2 (40.6%) and 3 (24.4%). A novel mutation was observed in 16.9% of the study population and frameshift (44.9%) and nonsense (31.5%) mutations were predominant. Molecular subtypes were as follows: HR-positive (n=125, 49.2%), HR/HER2 positive (n=10, 3.9%), HER2 positive (n=3, 1.2%) and TNBC (n=116, 46.7%). TNBC was observed mostly in gBRCA1 mutant patients (71.9% vs. 19.1%) and HR-positive in gBRCA2 mutant patients (76.2% vs. 22.7%). Sixty-seven patients (26.4%) experienced disease recurrence and median time to recurrence were 145.0 months (95% confidential interval 107.3-170.2). Looking into co-altered somatic mutation using target sequencing (n=45), notably, alteration in other DNA-damage response pathway-related genes, such as TP53 (60%), ATR (27%), CHEK2 (20%), MSH6 (20%), were identified. Evaluated for mutual exclusivity, only gBRCA2 was significantly (P Conclusions: From our dataset, we have witnessed similar clinical characteristics but the difference in molecular subtype between gBRCA1/2 mutant. Despite the fact that both gBRCA1/2 mutations disrupt the homologous recombination pathways, this study provides early evidence regarding the hypothesis that difference in co-occurred somatic alteration between gBRCA1/2 lead to different carcinogenesis mechanism which consequences different molecular subtype. Detail analyses using paired transcriptome result will be presented in the poster. Citation Format: Sehhoon Park, Eunjin Lee, Seri Park, Sohee Lee, Joon Young Hur, Sang Eun Yoon, Kangkook Lee, Jang Ho Cho, Ji-Yeon Kim, Jin Seok Ahn, Young-Hyuck Im, Woong-Yang Park, Yeon Hee Park. Clinical characteristics and combined somatic mutation using target sequencing among breast cancer patients with germline BRCA mutation [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2019; 2019 Mar 29-Apr 3; Atlanta, GA. Philadelphia (PA): AACR; Cancer Res 2019;79(13 Suppl):Abstract nr 3502.

Three different histopathological growth patterns (HGP) have been identified in colorectal cancer liver metastases (CRCLM) resected from patients: the desmoplastic HGP, the pushing HGP, and the replacement HGP. Evidence suggests that the predominant growth pattern with which a CRCLM presents has clinically relevant prognostic implications. We have shown that patients with replacement HGP lesions who received bevacizumab plus chemotherapy had a worse pathological response and five-year overall survival than those with desmoplastic HGP lesions receiving the same treatment. We have also demonstrated that CRCLM with the replacement HGP promote vessel co-option for vascularization, rather than sprouting angiogenesis as seen in desmoplastic HGP metastases. These findings point to the growth patterns in CRCLM possibly serving as predictive biomarkers of response to anti-angiogenic therapy, when no such biomarker has been validated to date. However, HGP scoring is performed on resected liver metastatic tissue, implying that preoperative treatment precedes growth pattern assessment. Therefore, surrogate markers for the HGPs that can be appraised prior to surgery would be helpful to inform clinical decision-making about whether a patient with CRCLM may benefit from anti-angiogenic treatment. This project aims to identify genes that are differentially expressed between CRCLM presenting with either the replacement or desmoplastic HGP. RNA sequencing (RNA-Seq) was used to compare the transcriptional profiles of these distinct CRCLM. A gene expression signature was generated from the RNA-Seq data to include genes that were upregulated in the replacement HGP liver metastases. Immunostaining of select genes from this signature was performed to validate these findings on human tumor tissue. Initial analysis of our RNA-Seq data identified 525 genes that are differentially expressed between the replacement and desmoplastic HGP CRCLM, of which 53 genes met the criteria for the gene expression signature. Pathway analysis of these genes identified pathways involved in the immune system and extracellular matrix (ECM) to be upregulated in the replacement HGP lesions. Subsequent immunostaining revealed the expression pattern of lysyl oxidase like-4 (LOXL4) protein to be distinct between the HGPs - significantly greater quantities of LOXL4-expressing neutrophils were detected in the replacement HGP tumor microenvironment. Characterizing differences in gene expression between replacement and desmoplastic HGP CRCLM is expected to provide some insight into the mechanisms responsible for generating these distinct growth patterns. Our findings suggest that further investigations into the role of tumor-associated leukocytes and ECM remodelling may ultimately lead to the development of biomarkers and new therapeutic targets for replacement HGP CRCLM. Citation Format: Vincent Palmieri, Anthoula Lazaris, Stephanie Petrillo, Hussam Alamri, Abdellatif Amri, Woong-Yang Park, Zu-hua Gao, Peter Metrakos. LOXL4-expressing neutrophils are found in colorectal cancer liver metastases resistant to anti-angiogenic therapy [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2019; 2019 Mar 29-Apr 3; Atlanta, GA. Philadelphia (PA): AACR; Cancer Res 2019;79(13 Suppl):Abstract nr 2777.

The RUNX1-RUNX1T1 fusion is a frequent chromosomal alteration in acute myeloid leukemias (AMLs). Although RUNX1-RUNX1T1 fusion protein has pivotal roles in the development of AMLs with the fusion, RUNX1-RUNX1T1 fusion protein is difficult to target since it lacks kinase activities. Here, we used a sophisticated bioinformatic tool to elucidate targetable signaling pathways in AMLs with RUNX1-RUNX1T1 fusion. After analysis of 93 AMLs from the TCGA database, the expressions of 293 genes were significantly correlated with the expression of the RUNX1-RUNX1T1 fusion gene. Based on the 293 genes, the COX, VEGFR, PDGFR, and FGFR pathways are predicted to be specifically activated in AMLs with RUNX1-RUNX1T1 fusion. Compared with the AML cells without the fusion, the specific activations of the AML cell lines with RUNX1-RUNX1T1 fusion were confirmed by the increased phosphorylation of VEGFRs, PDGFRs, and FGFRs. Moreover, the in vitro proliferation of AML cells with RUNX1-RUNX1T1 fusion decreased significantly more than that of AML cells without the fusion when the pathways were inhibited pharmacologically. The results indicate that novel targetable signaling pathways could be identified by the analysis of the gene expression features of AMLs with non-targetable genetic alterations. The elucidation of specific molecular targets for AMLs that have a specific genetic alteration would promote personalized treatment of AMLs and improve treatment outcomes for AML patients in clinic. Note: This abstract was not presented at the meeting. Citation Format: Jae Won Yun, Yoon Kyung Bae, So Yeong Cho, Hee-Jin Kim, Do-Hyun Nam, Sun-Hee Kim, Sejong Chun, Kyeung Min Joo, Hye Won Lee, Woong-Yang Park. Elucidation of novel therapeutic targets for acute myeloid leukemias with RUNX1-RUNX1T1 fusion [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2019; 2019 Mar 29-Apr 3; Atlanta, GA. Philadelphia (PA): AACR; Cancer Res 2019;79(13 Suppl):Abstract nr 1723.

Tumor heterogeneity influences the clinical outcome of patients with cancer, and the diagnostic method to measure the tumor heterogeneity needs to be developed. We analyzed genomic features on pairs of primary and multiple metastatic lymph nodes from six patients with lung cancer using whole-exome sequencing and RNA sequencing. Although somatic single-nucleotide variants were shared in primary lung cancer and metastases, tumor evolution predicted by the pattern of genomic alterations was matched to anatomic location of the tumors. Four of six cases exhibited a branched clonal evolution pattern. Lymph nodes with acquired somatic variants demonstrated resistance to the cancer treatment. In this study, we demonstrated that multiple biopsies and sequencing strategies for different tumor regions are required for a comprehensive understanding of the landscape of genetic alteration and for guiding targeted therapy in advanced primary lung cancer. Cancer Res; 76(22); 6568–76. ©2016 AACR.

Background and Purpose Studies have suggested that the antigenicity of tumor cell is highly correlated with response to immune checkpoint inhibitors. Checkpoint blockade has been found to be particularly effective in tumors with high tumor mutational burden (TMB) in certain indications such as melanoma and NSCLC, while cancers with lower TMB have been shown to be less responsive to checkpoint inhibitors. In HER2-positive metastatic breast cancer (MBC), TMB has potential to be a candidate prognostic or predictive marker for conventional treatments including HER2 targeting agents. This research was conducted to explore the clinical implication of somatic TMB in a well-defined HER2-positive MBC population who were previously treated but progressed. Methods Forty-six patients with HER2-positive MBC who have progressed from more than two HER2-directed therapies were enrolled from 2007 to 2016. Whole exome sequencing was performed on FFPE tumor samples and matched normal tissue. Sequencing reads were aligned to reference human genome using BWA-MEM followed by standard pre-processing steps and GATK to generate analysis ready BAM files. MuTect suite was used to generate somatic single nucleotide variants (SNV) calls using default parameters by comparing BAM files from tumor and matched normal samples. TMB for a subject was defined as the sum of somatic non-synonymous SNVs that passed all the filters described. Results Among 46 patients, 13 (28.3%) patients were determined to estrogen receptor (ER)-positive and 9 (19.6%) patients were determined to progesterone receptor (PR)-positive in immunohistochemistry (IHC) analysis. The median age 48 years (range: 29-68) and all (100%) patients were female. 20 patients (43.5%) were recurrent MBC compared to de novo (n = 26, 56.5%). 16 (34.6%) patients demonstrated more than 100 mutations (and defined as high TMB in this population). The median follow up of duration was 57.5 months (95% CI: 55.4 - 60.6). The median overall survival (OS) according to low or high TMB status (44.9 months vs. 85.8 months) was significantly different (p = 0.016). In a multivariate Cox regression analysis, TMB was the only independent prognostic factor for good OS when adjusted for age and recurrence (Hazard ratio = 0.32, 95% CI, 0.103 - 0.998, p = 0.049). Conclusions Data implicated that high TMB can be a prognostic marker that predicts good OS outcomes with conventional HER2-directed treatments and chemotherapy. Furthermore, future trials using TMB as a stratification marker could identify the right population with potential to respond to immune checkpoint inhibitor after recurrence following HER2-directed treatments. To our knowledge this is the first report that tumor mutation burden (TMB) associate with response to standard of care therapy in HER2+ breast tumors. Citation Format: Yeon Hee Park, Kyunghee Park, Song Ee Park, Eunjin Lee, Ji-Yeon Kim, Jin Seok Ahn, Young-Hyuck Im, Hun Jung, Chooghoon Lee, Woong-Yang Park, Razvan Cristescu. Clinical implication of mutation load in patients with HER2-positive refractory metastatic breast cancer [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2018; 2018 Apr 14-18; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2018;78(13 Suppl):Abstract nr 2602.

Breast cancers (BC) in younger, premenopausal patients (YBC) tend to be more aggressive with worse prognosis, higher chance of relapse and poorer response to endocrine therapies compared to breast cancers in older patients. The proportion of YBC (age ≤ 40) among BC in East Asia is estimated to be 16-32%, significantly higher than the 7% reported in Western countries. In addition, approximately half of the Asian BC patients were premenopausal compared to 15-30% in the West. To characterize the molecular bases of Asian YBC, we have performed whole-exome sequencing (WES) and whole-transcriptome sequencing (WTS) on tumor and matched normal samples from 168 Korean BC patients consisting of 106 YBC cases (age ≤ 40) and 62 OBC cases (age > 40). We then performed comparison analyses with the TCGA BC cohort consisting of 1,116 tumors from primarily Caucasian patients, also grouped by age into YBC (age ≤ 40), IBC (40 < age ≤ 60) and OBC (age > 60). We performed logistic regression analyses to identify differentially expressed (DE) genes and pathways among age-based cohorts while controlling for the confounding effects of molecular subtype, tumor purity and stage. Within the Asian cohort, we found that estrogen response, endocrine therapy resistance, and various metabolism pathways are up-regulated in YBCs while cell cycle, proliferation and inflammatory pathways are up-regulated in OBCs. To separately examine molecular signatures from tumor, stroma and normal compartments, we used non-negative matrix factorization (NMF) analyses to virtually dissect bulk tumor expression data and identified 14 factors including 3 factors associated with normal tissues, 1 factor associated with stroma and 1 factor associated with tumor infiltrating leukocytes (TILs). By examining the correlation between pathway gene expression and NMF factors, we inferred that DE pathways such as fatty acid metabolism, bile acid biosynthesis, and epithelial-to-mesenchymal transition (EMT) were mainly active in stromal and normal tissue compartments. The TIL factor was significantly enriched in Asian BCs relative to Caucasian BCs with the highest TIL factor weight observed in Asian OBCs. Using gene expression signatures representing distinct types of TILs, we classified the combined cohort into three subtypes of varying TIL activities. Consistent with results from the NMF analysis, the TIL-high subtype is also significantly enriched in Asian BCs relative to Caucasian BCs. To our knowledge, this is the first large-scale multi-omics study of Asian breast cancer. Comparative analyses of multi-omics profiles from Asian and primarily Caucasian BC cohorts identified distinguishing molecular signatures associated with Asian BCs. Further, many signatures appeared to be specific to non-tumor compartments within bulk tumor, indicating that young Asian BCs may harbor distinctive tumor microenvironment. Citation Format: Yeon Hee Park, Ying Ding, Soo-Hyeon Lee, Hae Hyun Jung, Woosung Chung, Soonweng Cho, Jin-Ho Kim, Shibing Deng, Yoon-la Choi, Julio Fernandez, Se Kyung Lee, Seok Won Kim, Jeong Eon Lee, Ji-Yeon Kim, Jin Seok Ahn, Young-Hyuck Im, Seok Jin Nam, Woong-Yang Park, Zhengyan Kan. Comparative analyses of multi-omics profiles reveal distinctive molecular signatures of young Asian breast cancers [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2017; 2017 Apr 1-5; Washington, DC. Philadelphia (PA): AACR; Cancer Res 2017;77(13 Suppl):Abstract nr 3370. doi:10.1158/1538-7445.AM2017-3370

Background/Objectives: Sorafenib is the only approved targeted agent as first-line treatment for advanced hepatocellular carcinoma (HCC). Unfortunately, many HCC patients are initially not responsive to sorafenib. Due to anatomic limitations and underlying liver dysfunction, genomic studies of HCC patients have not been actively pursued yet compared to other solid tumors; especially, in sorafenib-treated subset. In this study, we have retrospectively identified advanced HCC patients who had archival tumor tissues available for targeted genomic sequencing and who had been treated with sorafenib. Methods: We performed targeted genomic profiling for 381 cancer-related genes from 42 HCC (32 hepatitis B virus-related, 1 alcohol-related, 7 unknown) patients treated with sorafenib at Samsung Medical Center, Seoul, Korea from July 2008 to October 2013. We excluded 2 patients who were lost to follow up for evaluation of treatment response to sorafenib. Results: Of the 40 patients, there were 5 confirmed PRs, 7 SDs, and 28 PDs. In 40 patients, there were 6 CDKN2B amplifications, 4 NTRK1 amplifications, 2 MET amplifications, 2 FGF19 amplifications, 2 CCND1 amplifications, 1 EGFR amplification, 1 FGF23 amplification and 1 CCNE1 amplification. In cell cycle-related genes, those patients with aberrations were less likely to respond to sorafenib. Of note, two patients whose tumor had CCND1 amplifications had de novo resistance to sorafenib. Conclusions: Our study demonstrates inter-patient heterogeneity in advanced HCC patients and the role of cell cycle-associated genes in resistance mechanism to sorafenib is being investigated. Citation Format: Wonseok Kang, Kyung Kim, Joon Hyeok Lee, Ho Yeong Lim, Woong-Yang Park, Jeeyun Lee, Yong Han Paik. Targeted genome profiling in patients with advanced hepatocellular carcinoma treated with sorafenib [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2017; 2017 Apr 1-5; Washington, DC. Philadelphia (PA): AACR; Cancer Res 2017;77(13 Suppl):Abstract nr 426. doi:10.1158/1538-7445.AM2017-426

Background & Aims: Next-generation sequencing (NGS) that enables the analyses of massively parallel sequences of DNA can advance the understanding of the underlying molecular pathophysiologies of cancer. Such recent genomic analyses have revealed a complex mutational landscape for PDACs. The aims of this study were to investigate the genomic profile to predict the chemotherapy response in unresectable PDACs. Methods: The total of 80 pathologically confirmed PDACs were enrolled and genomic DNA was extracted and quality control metrics of DNA analytes were measured. The specimens that passed a quality control test underwent targeted deep sequencing using a customized cancer panel (CancerSCAN) enriched in the exons of 83 genes. Results: Clinical prognostic factors associated with survival in PDAC were gender, tumor mass size, stage and chemotherapy response rate. (P=0.078, 0.009, 0.052 and Conclusion: Targeted sequencing using EUS-FNA specimens in PDAC showed excellent compatibilities to analyze genomic profiles of PDACs. Furthermore, novel genes associated with survival, metastasis and chemotherapy response in PDAC were identified. Citation Format: Joo Kyung Park, Kwang Hyuck Lee, Donghyun Park, Woong-Yang Park, Dae-Soon Son, Jong Kyun Lee, Kyu Taek Lee. Detection of genetic alterations to predict the chemotherapeutic responses in unresectable pancreatic ductal adenocarcinomas [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2017; 2017 Apr 1-5; Washington, DC. Philadelphia (PA): AACR; Cancer Res 2017;77(13 Suppl):Abstract nr 423. doi:10.1158/1538-7445.AM2017-423

Background: Patient-derived cell (PDC) model is one of prospective technologies to investigate drug response for precision medicine. Previously drug sensitivity derived PDC model has been studied in colorectal cancer organoid culture or lung cancer model. They revealed that PDC model recapitulates genomic characteristic and drug high-throughput screening is amenable. We established PDC models from metastatic cancer cohort of multiple cancer types, and revealed genomic characteristics and matched drug response. Method: Target sequencing was performed using OncoPanel. We called SNV, and CNV from target sequencing using GATK, then clinically unassociated SNPs were eliminated. Somatic submodules were identified from integrative variant profiles using HotNet referring HPRD. Drug high-throughput screening was performed SEMCO platform. In order to identify sensitive drug and variant gene pairs, drug sensitivity was tested by ANOVA using IC50 values. Result: Target sequencing of exome were generated from our 165 PDC models, and drug high-throughput screening was performed from 30 PDCs and ant-cancer 15 drugs. We tried integrative analysis approaches from genomic profile including mutation and copy number alteration, and identified genomic modules related with Fanconi, Hedgehog, FGFR receptor, Cytokine-cytokine receptor pathway. Variants of known oncogenes were successfully identified in MET, BRAF, FGFR and PIK4CA. Especially novel and recurrent mutations were identified in FGFR2, and corresponding samples indicated AZD4547 sensitivity as well as known drug response case such as BRAF V600E targeting Vemurafenib. Conclusion: Our study explains that PDC cohort sustains cancer related pathways, and its drug response follows genomic features. It is expected that our research would contribute to facilitate precision medicine from PDC model to utilize genomics and drug response. Citation Format: Seung T. Kim, Charny Park, Sun Kim, Won Lee, Joon Park, Kyoung-Mee Kim, Woongyang Park, Jeeyun Lee, young park, Jiryeon Jang. High-throughput screening using patient derived tumor cell to conduct personalized medicine. [abstract]. In: Proceedings of the 107th Annual Meeting of the American Association for Cancer Research; 2016 Apr 16-20; New Orleans, LA. Philadelphia (PA): AACR; Cancer Res 2016;76(14 Suppl):Abstract nr 66.

Next-generation sequencing (NGS) that enables the analyses of massively parallel sequences of DNA can advance the understanding of the underlying molecular pathophysiologies of cancer. Such recent genomic analyses have revealed a complex mutational landscape in surgical specimens of PDAC. Endoscopic ultrasound-guided fine-needle aspiration (EUS-FNA) is used in the diagnosis of pancreatic cancer. Several genomic analyses using EUS-FNA in PDAC have been undertaken in a small number of patients. This study evaluated the feasibility of targeted NGS of specimens obtained using EUS-FNA to clarify genetic markers associated with the prognosis of pancreatic cancer. One hundred sixty six patients who underwent EUS-FNA with pathologically confirmed PDAC were included. Genomic DNA was extracted and quality control metrics of DNA analytes were measured. The specimens that passed a quality control test (N = 101, 61%) underwent NGS using a customized cancer panel (CancerSCAN) enriched in the exons of 83 genes. Clinical prognostic factors associated with survival in PDAC were age, performance status, CA 19-9 level, curative resection, chemotherapy, tumor mass size, tumor location, stage and metastasis site in univariate analysis (P = 0.020, Citation Format: Joo Kyung Park, Kwang Hyuk Lee, Jong Kyun Lee, Kyu Taek Lee, Woong-Yang Park, Dae-Soon Son. Targeted sequencing from endoscopic ultrasound-guided fine needle aspirates of pancreatic ductal adenocarcinoma. [abstract]. In: Proceedings of the 107th Annual Meeting of the American Association for Cancer Research; 2016 Apr 16-20; New Orleans, LA. Philadelphia (PA): AACR; Cancer Res 2016;76(14 Suppl):Abstract nr 96.

Metastatic renal cell carcinoma (mRCC) evolves from primary RCC (pRCC) and harbors multiple subpopulations with distinct molecular and phenotypic features. Such underlying intratumoral heterogeneity (ITH) imposes difficulties in designing marker-based clinical trials because targeted mono-therapy eliminates a specific subpopulation of tumor cells while leaving others unharmed. Accordingly, a rational combination strategy that minimizes the survival of the drug-resistant subpopulation in a given heterogeneous tumor is essential for long-term therapeutic efficacy. Here, we examined the ITH of a paired mRCC and pRCC using single-cell mRNA sequencing (scRNA-seq) to identify specific tumor cell populations with drug target pathway activation. From the single cell expression profiles, we found the highly activated status of the EGFR and Src signaling pathways in the mRCC compared to the pRCC, with supporting in vitro high-throughput drug screening results. Distinct features of intratumoral expression variability across mRCC single cells that were masked in the bulk measurement prompted us to test the co-targeting strategy for the EGFR and Src pathways with increased likelihood for complete response. This combinatorial strategy showed significantly better treatment effects on mRCC-derived xenograft platforms in vitro and in vivo than monotherapies. Taken together, our findings show clinical implications of scRNA-seq in designing effective treatment regimens for overcoming treatment failure to conventional monotherapies, and also provide novel insights to the unmet clinical needs in effective personalized treatments. Citation Format: Kyu-Tae Kim, Hye Won Lee, Hae-Ock Lee, Hye Jin Song, Da Eun Jeong, Sang Shin, Hyunho Kim, Yoojin Shin, Do-Hyun Nam, Byong Chang Jeong, David G. Kirsch, Kyeung Min Joo, Woong-Yang Park. Single-cell transcriptome analysis guides tailored combinatorial therapeutics in refractory kidney cancer. [abstract]. In: Proceedings of the 107th Annual Meeting of the American Association for Cancer Research; 2016 Apr 16-20; New Orleans, LA. Philadelphia (PA): AACR; Cancer Res 2016;76(14 Suppl):Abstract nr 165.

Intratumoral heterogeneity has emerged as a problem to be resolved for precise diagnosis and effective treatment of cancer. Breast cancer is one of the most heterogeneous cancer in its molecular subtypes and its intratumoral genomic diversity has been demonstrated as well. Single-cell sequencing approach would reveal heterogeneity in gene expression signatures and the common gene expression characteristics across tumor cells within a tumor. In this study, we isolated live individual cells and amplified cDNAs using C1™Single-cell Auto Prep System (Fluidigm) then performed RNA sequencing using HiSeq2500 System (Illumina). Single-cell transcriptomes were acquired from four primary tumor (ER+/HER2-, ER-/HER2+, ER+/HER2+ and triple negative breast cancer) and two matched lymph node metastases (ER+/HER2+ and triple negative breast cancer). Because the tumor tissue isolates may include stromal and immune cells in the tumor microenvironment, we defined ‘epithelial cell adhesion molecule signature’ to filter out non-tumor populations (Poster# presented by Woosung Chung et. al) After the removal of non-tumor cells, collected tumor single-cell transcriptomes largely recapitulated bulk tumor-expression profiles. The ER+/HER2- tumor cells showed highly activated estrogen receptor-associated pathways. Although both ER-/HER2+ and ER+/HER2+ tumors had the HER2 gene amplification and overexpression, HER2/HER3 pathway is activated only in ER-/HER2+ tumor cells. The association of gene expression for cancer stem cells and chemoresistance, such as IL8 and CXCL1, with HER2/HER3 signaling pathway activation was prominent. By contrast, the ER+/HER2+ tumor cells had highly activated estrogen receptor signaling, presenting them as the luminal type. The triple negative breast cancer cells highly expressed adhesion molecules and proteases suggesting their invasive nature, yet demonstrated extensive gene expression heterogeneity. Collectively these data revealed the power of single cell gene expression profiling in the precise characterization of tumor cell properties. The explicit gene expression signatures for patient specific tumors may be defined as subtype-specific signatures and helpful to unmask targets for molecular directed therapies. Citation Format: Hye Hyeon Eum, Hae-Ock Lee, Woosung Chung, Kyu-Tae Kim, Kyung-Min Lee, Arum Jo, WonShik Han, Woong-Yang Park. Single-cell RNA sequencing presents explicit expression signatures in primary breast cancer. [abstract]. In: Proceedings of the 107th Annual Meeting of the American Association for Cancer Research; 2016 Apr 16-20; New Orleans, LA. Philadelphia (PA): AACR; Cancer Res 2016;76(14 Suppl):Abstract nr 161.

Therapies targeting the tyrosine kinase activity of Epidermal Growth Factor Receptor (EGFR) have been proven to be effective in treating a subset of non-small cell lung cancer (NSCLC) patients harboring activating EGFR mutations. Inevitably these patients develop resistance to the EGFR-targeted tyrosine kinase inhibitors (TKIs). Here, we performed integrated genomic analyses using an in vitro system to uncover alternative genomic alterations responsible for acquired resistance to EGFR-TKIs. Specifically, we identified 80 genes whose expression is significantly increased in the resistant clones and RNAi-based systematic synthetic lethal screening revealed that suppression of one upregulated transcript, SCRN1, a secernin family member, restores sensitivity to erlotinib by enhancing inhibition of PI3K/AKT signaling pathway. Furthermore, we detected increased levels of SCRN1 in 5 of 11 lung tumor specimens from EGFR-TKIs refractory patients by immunohistochemistry. Taken together, we propose that upregulation of SCRN1 is an additional mechanism associated with acquired resistance to EGFR-TKIs in a subset of lung cancer patients and that its suppression serves as a novel therapeutic strategy to overcome drug resistance in these patients. Citation Format: Nayoung Kim, Ahye Cho, Hideo Watanabe, Yoon-La Choi, Meraj Aziz, Michelle Kassner, Je-Gun Joung, Angela KJ Park, Joshua Francis, Joon Seol Bae, Soo-min Ahn, Kyoung-Mee Kim, Joon-Oh Park, Woong-Yang Park, Myung-Ju Ahn, Keunchil Park, Hongwei Holly Yin, Jeonghee Cho. Integrated genomic approaches identify upregulation of SCRN1 as a novel mechanism associated with acquired resistance to erlotinib in non small cell lung cancer cells with oncogenic EGFR mutation. [abstract]. In: Proceedings of the 107th Annual Meeting of the American Association for Cancer Research; 2016 Apr 16-20; New Orleans, LA. Philadelphia (PA): AACR; Cancer Res 2016;76(14 Suppl):Abstract nr 2931.

Breast cancers (BC) in younger, premenopausal patients (YBC) tend to be more aggressive with worse prognosis, higher chance of relapse and poorer response to endocrine therapies compared to breast cancers in older patients (OBC). The proportion of YBC (age ? 40) among BC in East Asia is estimated to be 16-32%, significantly higher than the 7% reported in Western countries. Genomic and molecular characterizations have deepened our understanding of breast cancer biology in areas ranging from intrinsic subtypes to treatment responses, however, the molecular bases of Asian YBC remains poorly characterized. We have performed whole-exome sequencing (WES), whole-transcriptome sequencing (WTS) and high coverage targeted sequencing on tumor and matched normal samples from 133 Korean BC patients consisting of 74 YBC cases (age ? 40). We further performed immunohistochemistry (IHC) analyses to characterize tumor-infiltrating lymphocytes (TILs) in 46 tumors using four markers (CD45, CD4, CD8 and CD163). We found that BRCA1/2 germline deleterious mutations are enriched in YBC and the ER+/HER2- subtype, indicating that Asian ER+ YBC has a significant germline contribution. MutSig analysis4 identified ARID1A as a significantly mutated gene, implicating chromatin modeling as a cancer driver in Asian BC. Differential expression analyses suggested that Asian YBC differ in energy metabolism and are more active in protein synthesis than OBC tumors, whereas OBC is more proliferative than YBC. Using gene expression signatures representing distinct immune cell types and immunohistochemistry, we classified our cohort into four subtypes of varying TIL activities: high, medium, low and quiet. The majority of immunogenic cases with high TIL levels lie in ER+ or HER2+ subtypes although higher proportion is seen in TNBC. Moreover, YBC tumors appear to harbor lower levels of TIL activities than OBC, suggesting that younger patients may be less likely to benefit from immunomodulatory therapies than older patients. To our knowledge, this is the first large-scale multi-omics study of Asian breast cancer and would significantly contribute to the compendium of molecular data available for young, premenopausal breast cancer. While the major landmarks in the molecular and immune landscape of Asian BC look similar to that of the predominantly Caucasian BC cohorts, we have identified a number of distinguishing characteristics pointing to distinctive oncogenic mechanisms underlying Asian BC. Citation Format: Yeon Hee Park, Ying Ding, Soonweng Cho, Soo-Hyeon Lee, Hae Hyun Jung, Woosung Chung, Jinho Kim, Woong-Yang Park, Eric Powell, Pamela Vizcarra, Shibing Deng, Se Kyung Lee, Seok Won Kim, Jeong Eon Lee, Ji-Yeon Kim, Jin Seok Ahn, Young-Hyuck Im, Seok Jin Nam, Zhengyan Kan. Multi-omics and immuno-oncology profiling of an Asian breast cancer cohort enriched in young and premenopausal patients. [abstract]. In: Proceedings of the 107th Annual Meeting of the American Association for Cancer Research; 2016 Apr 16-20; New Orleans, LA. Philadelphia (PA): AACR; Cancer Res 2016;76(14 Suppl):Abstract nr LB-325.

Summary: Breast cancer is a heterogeneous type of cancer accompanying extensive immune cell infiltration. Although its mortality rate has been decreasing with early diagnosis and introduction of molecular targeted therapies, it is still one of the world leading cause of cancer death in women. In breast cancer, molecular subtyping is used for the patient stratification and treatment decisions, emphasizing the importance of inter-patient tumoral heterogeneity in the tumor cell behavior. Tumor cells within a patient may also manifest intra-tumoral heterogeneity, which may cause treatment resistance. Single cell gene expression profiling would reveal both inter- and intra-tumoral heterogeneity in breast cancer encompassing tumor cells and associated stromal and immune cells. Experimental Process: From 4 different subtype primary breast cancer and 2 lymph node metastases, single cells were captured and amplified through C1™Single-cell Auto Prep System (Fluidigm) with the SMARTer Ultra Low RNA Kit. Sequencing libraries were constructed with the Nextera XT DNA sample Prep Kit and sequenced through a Hiseq2500 (Illumina) as 100-bp paired end mode of the TruSeq Rapid PE Cluster kit and Truseq Rapid SBS kit. Computational Process: RNA sequencing reads were aligned to the modified human genome reference using 2-pass mode of STAR_2.4.0b, and transcript per million (TPM) values were obtained as the relative gene expression levels of single cells through RSEM v1.2.17. Results: As most breast cancers originate from the mammary epithelium, we distinguished carcinoma cells from associated non-tumor cells by epithelial cell adhesion molecule (EPCAM) signature genes. We validated the distinction with chromosomal gene expression patterns and gene set enrichment analyses using tumor, stromal, or immune genesets. With or without removal of non-tumor cells, TNBC tumor cells showed the highest gene expression heterogeneity, recapitulating the known inter-patient heterogeneity. TNBC tumor cells from a single patient were categorized into 6 different TNBC-subtypes, further demonstrating intra-tumoral heterogeneity. In addition, immune cells identified in the TNBC tumor were naïve B cells and regulatory T cells, suggesting non-activated status of the immune system. Conclusions: The high resolution gene expression profiling revealed features of breast cancer transcriptome which may provide clues for molecular directed therapies targeting the tumor or the immune compartment. Citation Format: Woosung Chung, Hye Hyeon Eum, Kyu-Tae Kim, Kyung-Min Lee, Arum Jo, WonShik Han, Hae-Ock Lee, Woong-Yang Park. Identification of the carcinoma and immune cells in the breast cancer by single-cell RNA sequencing. [abstract]. In: Proceedings of the 107th Annual Meeting of the American Association for Cancer Research; 2016 Apr 16-20; New Orleans, LA. Philadelphia (PA): AACR; Cancer Res 2016;76(14 Suppl):Abstract nr 160.

One of the most detrimental hallmarks of glioblastoma multiforme (GBM) is cellular invasiveness, which is considered a potential cause of tumor recurrence. Infiltrated GBM cells are difficult to completely eradicate surgically and with local therapeutic modalities. Although much effort has focused on understanding the various mechanisms controlling GBM invasiveness, its nature remains poorly understood. In this study, we established highly serial intracranial transplantation. U87R4 cells were highly invasive and displayed stem cell-like properties, as compared to noninvasive but proliferative U87L4 cells. Microarray analysis during serial transplantation revealed that apoptosis-inducing genes (caspase3 and PDCD4) were downregulated whereas several cancer stem cell–relevant genes [Frizzled 4 (FZD4) and CD44] were upregulated in more invasive cells. U87R4 cells were resistant to anticancer drug–induced cell death, partly due to downregulation of caspase3 and PDCD4, and they retained activated Wnt/β-catenin signaling due to upregulation of Frizzled 4, which was sufficient to control neurosphere formation. We also found that FZD4 promoted expression of the epithelial to mesenchymal transition regulator SNAI1, along with acquisition of a mesenchymal phenotype. Taken together, our results argue that Frizzled 4 is a member of the Wnt signaling family that governs both stemness and invasiveness of glioma stem cells, and that it may be a major cause of GBM recurrence and poor prognosis. Cancer Res; 71(8); 3066–75. ©2011 AACR.

Background: Cancers are composed of cells with distinct genetic and phenotypic characteristics. Tumor heterogeneity in each patient may influence on personalized medicine strategies that depend on results from single tumor-biopsy samples. However, tumor heterogeneity in advanced primary lung cancer has not been well studied. The purpose of this study was to investigate the tumor heterogeneity in advanced primary lung cancer. Methods: We sequenced fourteen regions from four patients who were suspicious for primary lung cancer with multiple lymph node metastasis. Tumor samples were obtained from primary tumor and metastatic lymph node(s) or distant metastasis using bronchoscopic forcep biopsy, endobronchial ultrasound-guided transbronchial needle aspiration, and percutaneous needle biopsy. The profiles of somatic mutations, copy number alterations and gene expression were compared among primary tumor and metastatic sites in each patient. Clonal evolution was evaluated using the concept of cancer cell fraction by SciClone analysis. Results: We performed multiregion whole exome sequencing and RNA sequencing on three non-small cell lung cancers (2 adenocarcinomas [TH1 and TH4] and 1 squamous cell carcinoma [TH2]) and one small cell lung cancer (TH3). EML4-ALK fusion was noticed in TH1 and TH4. There was profound tumor heterogeneity in terms of somatic mutation, copy number alteration, or gene expression in each patient. The frequency of private somatic mutation from spatially distinct regions were 59% in TH1, 44% in TH2, 13% in TH3, and 30% in TH4. Known driver mutation was also heterogeneous in TH1 and TH2. Analysis of the clonal architecture confirmed the presence of clonal heterogeneity in each patient. Phylogenetic reconstruction revealed the evidence of branched clonal evolution. Conclusions: Tumor heterogeneity was noticed from spatially distinct regions from advanced primary lung cancers. The single biopsy specimens may not represent the whole landscape of genetic alteration in advanced primary lung cancers. Citation Format: Sang-Won Um, Je-Gun Joung, Su Yeon Lee, Joon Seol Bae, Jinha Park, Woong-Yang Park. Tumor heterogeneity of advanced primary lung cancer evaluated by multiregion sequencing. [abstract]. In: Proceedings of the 106th Annual Meeting of the American Association for Cancer Research; 2015 Apr 18-22; Philadelphia, PA. Philadelphia (PA): AACR; Cancer Res 2015;75(15 Suppl):Abstract nr 4815. doi:10.1158/1538-7445.AM2015-4815

The anaplastic lymphoma kinase (ALK) gene is frequently rearranged in various types of cancer, and is highly responsive to the target therapeutics. We developed a pipeline to detect ALK rearrangement in a large group of Korean thyroid cancer patients. We screened 474 malignant or benign thyroid tumor cases to find ALK fusion. The expression and translocation of the ALK gene were analysed by immunohistochemistry (IHC), fluorescence in situ hybridisation (FISH), and digital multiplexed gene expression (DMGE) using formalin-fixed paraffin embedded tissues. Four cases of ALK rearrangement were screened out of 474 patients by IHC, which were validated by performing FISH for 189 samples. On the other hand, DMGE analysis using Nanostring could detect three out of four IHC positive cases. Two ALK rearrangements were Striatin (STRN)-ALK fusion, identified by 5′ RACE analysis. ALK rearrangements were found exclusively in the BRAF wild type papillary carcinoma. Given the wide availability and accuracy for detecting ectopic ALK expression in the thyroid, we suggest that IHC based screening can be used as a practical method for identifying patients with ALK rearranged differentiated thyroid cancer. Citation Format: Gahee Park, Tae Hyuk Kim, Hae-Ock Lee, Jung Ah Lim, Jae-Kyung Won, Hye Sook Min, Kyu Eun Lee, Do Joon Park, Young Joo Park, Woong-Yang Park. Standard immunohistochemistry efficiently screens for ALK rearrangements in differentiated thyroid cancer. [abstract]. In: Proceedings of the 106th Annual Meeting of the American Association for Cancer Research; 2015 Apr 18-22; Philadelphia, PA. Philadelphia (PA): AACR; Cancer Res 2015;75(15 Suppl):Abstract nr 2134. doi:10.1158/1538-7445.AM2015-2134

BACKGROUND: Tumor progression within a brain is the process of acquiring malignant genetic mutations from glial progenitor cells, dodging their original developmental fate and deteriorating functional connectivity of CNS. In order to devise effective targeted therapeutic strategies, understanding distinct molecular and phenotypic evolutionary features of cancer clones is eminent. METHODS: To examine molecular signatures associated with malignant glioma progression, we analyzed transcriptome dynamics of 311 single-cells derived from 3 patients showing distinct disease patterns. Combined with single-cell level signatures, clonal architecture of tumors was reconstructed based on Bayesian-model clustering with integrated information of SSNVs and SCNAs identified from bulk population-level using aCGH and WES. RESULTS: We found evidence of branched evolutionary patterns within and across two histopathological separated tumor regions. Both tumors have high clonal probable somatic mutations, and evolutionary inference showed clonal lineage between two tumor regions was divided at early time during tumor progression. For all patients, the gene expression signatures within a tumor were heterogeneous but converged towards mesenchymal phenotype in the aggressive region. In the first case, the left tumor was self-limiting and showed mixed proneural and classical gene expression signatures whereas the right side tumor has expanded to the corpus callosum with EGFR overexpression and mesenchymal gene expression. In a case of a monocentric tumor with PDGFRA amplification, cancer cells disseminated from the core mass with massive immune cell infiltration underwent transcriptome switching from proneural to mesenchymal gene expression, with a concomitant FGFR1 copy number gain and mutation. Lastly, in a secondary glioblastoma with IDH1 mutation, GLIOLAN positive glioblastoma was surrounded by grade III oligodendroglioma. The two tumor areas showed intermingled heterogeneous gene expression signatures, with upregulation of astrocytic/mesenchymal genes in the GLIOLAN positive tumor cells. CONCLUSIONS: Our results show distinct genetic patterns of clonal expansion across patients undergoing malignant glioma progression. We also provide the transcriptome-wide view of gene expression signatures directed by genomic alterations. Overall our study reveals heterogeneity of individual tumor but elucidates common genomic features of glioblastoma progression. Citation Format: Kyu-Tae Kim, Jin-Ku Lee, Dong-Hyun Park, Do-Hyun Nam, Hae-Ock Lee, Woong-Yang Park. Transcriptome landscape of brain tumor progression at single-cell resolution. [abstract]. In: Proceedings of the 106th Annual Meeting of the American Association for Cancer Research; 2015 Apr 18-22; Philadelphia, PA. Philadelphia (PA): AACR; Cancer Res 2015;75(15 Suppl):Abstract nr 3748. doi:10.1158/1538-7445.AM2015-3748

Distinct genomic signature of a patient's cancer could provide a clue to design and predict accurate therapeutic responses to targeted medicine. Current approaches regarding heterogeneous cancer cells en masse as a pooled population, however, hardly reflect complete genomic landscape of tumor diversity, missing potentially important minor clones' implications regarding metastasis and drug resistance. Here we report a comprehensive analysis to discover different patterns of subclonal behaviors in functional modalities and signaling pathways out of heterogeneous population derived from a patient with lung squamous cell carcinoma, using single-cell mRNA sequencing. As a result of nonnegative matrix factorization as a clustering, we identified discrete 4 subgroups with a major group in accordance with overall expression profile of a pooled population of cells, and the other 3 minor groups that were masked in a pooled sample. Genomic variants were also reliably identified and compared across single cells. We integrated these genomic signatures from single cells to address the different responses in drug screening. Together, our approach provides insights into the more accurate strategy to identify minor but potentially important subclones that are relevant to drug resistance. Citation Format: Kyu-Tae Kim, Hye-Won Lee, Kyeung Min Joo, Sang-Chul Kim, Jin-Ku Lee, Nakho Chang, Jason K. Sa, Yu Jin Cho, Do-Hyun Nam, Hae-Ock Lee, Woong-Yang Park. Single-cell transcriptome analysis on lung squamous cell carcinoma derived from a patient. [abstract]. In: Proceedings of the 105th Annual Meeting of the American Association for Cancer Research; 2014 Apr 5-9; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2014;74(19 Suppl):Abstract nr 4664. doi:10.1158/1538-7445.AM2014-4664