The mutational spectrum and molecular characteristics of acute myelomonocytic lineage leukemia, namely acute myeloid leukemia (AML) French–American–British (FAB) subtypes M4 and M5, are largely unknown. In order to explore the mutational spectrum and prognostic factors of FAB-M4 and -M5, next-generation sequencing (NGS) was performed to screen for mutated genes and fusion genes relevant to the pathogenesis of AML. Of the 63 patients enrolled in the study, 60% had more than three mutated genes. NPM1 had the highest mutation frequency, followed by DNMT3A, FLT3, NRAS, RUNX1,and TET2. Univariate analysis suggested These authors contributed equally: Zhiheng Cheng, Kai Hu, Lei Tian, that age ≥60 years was an independent factor for both poor Yifeng Dai. event-free survival (EFS) and overall survival (OS, P = These authors contributed equally: Jinlong Shi, Lin Fu. 0.009, 0.002, respectively), MYH11-CBFβ was associated with better EFS and OS (P = 0.029, 0.016, respectively). * Jinlong Shi However, multivariate analysis was not able to identify any email@example.com independent risk factor for survival in the cohort of FAB- * Lin Fu M4 and -M5 patients, including peripheral white blood cell firstname.lastname@example.org count, bone marrow blast percentage, MYH11-CBFβ, FLT3- ITD, mutations in NPM1 and DNMT3A, and allogeneic Translational Medicine Center, Huaihe Hospital of Henan hematopoietic stem cell transplantation (allo-HSCT). Our University, Kaifeng 475000, China study provided new insight into the mutational spectrum Department of Hematology and Lymphoma Research Center, and molecular characteristics of FAB-M4 and -M5. The Peking University Third Hospital, Beijing 100191, China clinical implications of the genetic signature of FAB-M4 Laboratory of Environmental Medicine and Developmental and -M5 need to be further elucidated by larger studies. Toxicology, Shantou University Medical College, Shantou 515041, China Department of Medicine, William Beaumont Hospital, Royal Oak, MI 48073, USA Introduction Department of Clinical Laboratory, Beijing Haidian Hospital, Beijing Haidian Section of Peking University Third Hospital, Beijing 100080, China The accumulation of somatic genetic changes in hemato- poietic progenitor cells, including mutations, copy number Department of Hematology, Huaihe Hospital of Henan University, Kaifeng 475000, China alterations, and chromosomal translocations, is the core of acute myeloid leukemia (AML) pathogenesis . Due to the Department of Respiratory, Huaihe Hospital of Henan University, Kaifeng 475000, China profound clinical and biological heterogeneity of AML, accurate diagnosis and prognostic evaluation is essential for Department of Hematology, The First Afﬁliated Hospital of Soochow University, Suzhou 215006, China individualized treatment . Next-generation sequencing (NGS) is characterized by massive parallel sequencing and Department of Biomedical Engineering, Chinese PLA General Hospital, Beijing 100853, China high-throughput multiplex assay. It has been employed to depict mutational proﬁles , characterize recurrent genetic Department of Medical Big Data, Chinese PLA General Hospital, Beijing 100853, China mutations  of various diseases. In recent years, NGS has 1234567890();,: 78 Z. Cheng et al. become an indispensable tool in AML research. With the merged into a single BAM ﬁle, and duplicate reads were help of NGS, a growing list of major mutations in AML removed using Picard 1.17, 1.22, or 1.25 (http://picard. leukaemogenesis has been identiﬁed, such as NPM1, FLT3, sourceforge.net). We used the Mutational Signiﬁcance in CEBPA, DNMT3A, IDH1/IDH2, EZH2, U2AF1, SMC1A, Cancer (MuSiC) package to categorize mutations, including and SMC3 . It also helped to demonstrate the complex AT transition, AT transversion, CG transition, CG trans- interplay of genetic alterations in AML . version, CpG transition, CpG transversion, and indel. Sta- Advances in identiﬁcation of prognostic genetic altera- tistical tests including convolution, Fisher’s test, and a tions have facilitated greater detailed risk stratiﬁcation [7, likelihood test were used to combine the category-speciﬁc 8]. However, few studies have addressed the mutational binomials to obtain an overall P-value. Signiﬁcantly muta- spectrum of the French–American–British (FAB) subtypes ted genes and pathways were identiﬁed with the above M4 and M5, two unique subgroups of AML that method. exhibit monocytic morphology and cytochemical features . In this study, we investigated the mutational Statistical analysis signatures of FAB-M4 and -M5 and their prognostic value. The clinical and molecular characteristics of patients were summarized using descriptive statistics. Data sets were Patients and methods described with median and/or range. EFS was deﬁned as the time from diagnosis to removal from the study due to Patients relapse, death, or failure to achieve complete remission. OS was deﬁned as the time from diagnosis to death from any Sixty-three AML patients derived from The Cancer Genome cause, or was censored at the last follow-up. Survival was Atlas (TCGA) database (https://cancergenome.nih.gov/) estimated using the Kaplan–Meier method and the log-rank were enrolled in this study , including 41 FAB-M4 and test. Univariate Cox proportional hazards models were used 22 FAB-M5 patients. Written informed consent was to determine clinical and molecular variables associated obtained from each patient, and the study protocol was with survival. Multivariate proportional hazards models approved by Washington University Institutional Review were constructed for EFS and OS, using a limited backward Board. NGS was performed to evaluate the mutational elimination procedure. P < 0.05 was considered statistically spectrum of all patients. Detailed descriptions of clinical and signiﬁcant for all analyses. All statistical tests were two molecular characteristics were publicly accessible from the sided and were performed by SPSS software 20.0 and TCGA website. Event-free survival (EFS) and overall sur- GraphPad Prism software 5.0. vival (OS) were the primary endpoints of this study. Next-generation sequencing and analysis Results The procedure described by Mardis et al.  was followed Clinical and biological characteristics of patients for library construction and whole-genome sequencing. Brieﬂy, Illumina DNA sequencing was used to generate Clinical and biological characteristics of patients are sum- between 58.5 and 155.9 billion base pairs of sequence data marized in Table 1. Median age was 58 (range 21–81) for each of the patients, with haploid coverages ranging years, with 29 cases older than 60 years. Thirty-four cases from 18.9 to 50.4 billion base pairs. Heterozygous single- were men. Forty-one was FAB-M4 and 22 cases were FAB- nucleotide polymorphism (SNP) detected in the whole- M5. The median white blood cell (WBC) count at diagnosis 9 9 genome sequencing data was compared with SNP array was 27.1 × 10 /L, and in 29 cases it was ≥30 × 10 /L. Forty genotypes, which conﬁrmed bi-allelic detection of between patients had a bone marrow blast percentage of more than 96.60 and 99.86% in all patients. Libraries for whole-exome 70% and 25 had peripheral blood blast more than 20%. sequencing were constructed and sequenced with either Twenty-eight patients had abnormal karyotypes. Thirty-nine Illumina HiSeq 2000 or Illumina GAIIX using 76 bp paired- (61.9%) patients had intermediate-risk AML. Chemother- end reads. Standard quality control metrics, including error apy was differed in two patients due to old age and poor rates, percentage passing ﬁlter reads, and total Gb produced, functional status. Twenty-four patients received hemato- were used to assess process performance before down- poietic stem cell transplantation (HSCT). Thirty-eight stream analysis. The Illumina pipeline generated data ﬁles patients had more than three recurrent genetic mutations. (BAM ﬁles) that contained the reads and the quality para- NPM1 had the highest mutation frequency (n = 24, 38%), meters. For each sample, reads were aligned using either followed by DNMT3A (n = 23, 37%), FLT3-ITD/TKD (n = Maq 0.6.8 or 0.7.1 or BWA 0.5.5 on a per-lane basis, Prognosis signiﬁcance of mutational spectrum in FAB M4 and M5 79 Table 1 Clinical and molecular characteristics of patients Table 1 (continued) Characteristics Median (range) or N/% Characteristics Median (range) or N/% Age (years) 58 (21–81) Mutation 6/9.5 <60 34/54.0 Wild type 57/90.5 ≥60 29/46.0 NRAS/KRAS Gender Mutation 10/15.9 Male 34/54.0 Wild type 53/84.1 Female 29/46.0 TET2 Race Mutation 5/7.9 Caucasian 55/87.3 Wild type 58/92.1 Others 8/12.7 WT1 FAB subtypes Mutation 3/4.8 M4 41/65.1 Wild type 60/95.2 M5 22/34.9 KIT WBC count/×10 /L 27.1 (1.6–298.4) Mutation 3/4.8 <30 34/54.0 Wild type 60/95.2 ≥30 29/46.0 U2AF1 BM blasts/% 75 (30–98) Mutation 3/4.8 <70 23/36.5 Wild type 60/95.2 ≥70 40/63.5 MYH11-CBFβ PB blasts/% 12 (0–90) Presence 8/12.7 <20 38/60.3 Absence 55/87.3 ≥20 25/39.7 MLL rearrangements Karyotype MLL-PTD 3/4.8 Normal 35/55.6 MLL-MLLT10 4/6.3 Abnormal 28/44.4 MLL-MLLT4 2/3.2 Risk MLL-MLLT3 2/3.2 Good 9/14.3 MLL-ELL 2/3.2 Intermediate 39/61.9 Absence 50/79.4 Poor 15/23.8 HSCT Recurrent gene mutations 6 (0–12) Matched unrelated donor 7/29.2 <3 25/39.7 Matched related donor 12/50.0 ≥3 38/60.3 Autologous 5/20.8 FLT3 Relapse FLT3-ITD 8/12.7 Yes 29/46.0 FLT3-TKD 12/19.0 No 34/54.0 Wild type 43/68.3 FAB French–American–British, WBC white blood cell, BM bone NPM1 marrow, PB peripheral blood, HSCT hematopoietic stem cell W288 24/38.1 transplantation Wild type 39/61.9 DNMT3A R882 17/27.0 20, 32%), NRAS (n = 7, 11%), RUNX1 (n = 6, 10%), and Non-R882 mutations 6/9.5 TET2 (n = 5, 8%) (Fig. 1). Wild type 40/63.5 IDH1/IDH2 Comparison of EFS and OS between different clinical R132 3/4.8 and molecular characteristic groups R140 3/4.8 Wild type 57/90.5 EFS and OS of different age (≥60 vs. <60 years), WBC RUNX1 counts (≥30 vs. <30 × 10 /L), bone marrow blasts (≥70% vs. <70%), peripheral blasts (≥20% vs. <20%), number of 80 Z. Cheng et al. Table 2 Comparison of EFS and OS between different clinical and molecular characteristic groups Variables EFS OS 2 2 χ P-value χ P-value Age (≥60 vs. <60 years) 7.306 0.007 10.089 0.001 WBC (≥30 vs. <30 × 10 /L) 0.525 0.469 0.962 0.327 BM blasts (≥70% vs. <70%) 2.535 0.111 3.604 0.058 PB blasts (≥20% vs. <20% 0.008 0.928 0.023 0.879 Mutated genes (≥3 vs. <3) 1.882 0.170 1.772 0.183 MYH11-CBFβ (positive vs. 5.486 0.019 6.916 0.009 negative) FLT3-ITD (positive vs. negative) 0.037 0.847 0.000 0.997 Fig. 1 The mutational spectrum of FAB-M4 and -M5 patients (n= 63). NPM1 (mutated vs. wild type) 0.014 0.905 0.007 0.936 NPM1 had the highest mutation frequency (n = 24, 38%), followed by DNMT3A (mutated vs. wild type) 0.728 0.394 0.978 0.323 DNMT3A (n = 23, 37%), FLT3 (n = 20, 32%), NRAS (n= 7, 11%), RUNX1 (n= 6, 10%), and TET2 (n= 5, 8%). In addition, IDH2, IDH2, IDH1/IDH2 (mutated vs. wild 0.007 0.934 0.019 0.892 KRAS, WT1, KIT, and U2AF1 also had ≥5% mutation frequency type) RUNX1 (mutated vs. wild type) 1.857 0.173 1.454 0.228 NRAS/KRAS (mutated vs. wild 0.011 0.915 0.000 0.985 recurrent genetic mutations (≥3 vs. <3), allo-HSCT (yes vs. type) no), FLT3-ITD (positive vs. negative), MYH11-CBFβ TET2 (mutated vs. wild type) 0.362 0.547 0.666 0.415 (positive vs. negative), and the mutation status of other WT1 (mutated vs. wild type) 0.804 0.370 0.012 0.912 common AML mutations were compared with the KIT (mutated vs. wild type) 0.046 0.831 0.354 0.552 Kaplan–Meier method and the log-rank test, as listed in U2AF1 (mutated vs. wild type) 0.982 0.332 2.048 0.152 Table 2. Older patients (age ≥60 years) had shorter EFS and Allo-HSCT (yes vs. no) 1.499 0.221 3.788 0.052 OS (P = 0.007, P = 0.001, respectively; Fig. 2a, b). Pre- sence of MYH11-CBFβ was associated with longer EFS and EFS event-free survival, OS overall survival, WBC white blood cell, BM bone marrow, PB peripheral blood, HSCT hematopoietic stem cell OS (P = 0.019, P = 0.009, respectively; Fig. 2c, d). How- transplantation ever, other variables had no effect on EFS and OS. Evaluation of possible prognostic factors where CEBPA, NPM1, DNMT3A, FLT3, NRAS, IDH2 and Univariate analysis showed that age ≥60 years was an WT1 were mutated in more than 10% of AML patients [13, unfavorable factor for both EFS and OS (P = 0.009, P = 14]. Our results indicated that FAB-M4 and -M5 might 0.002, respectively), whereas MYH11-CBFβ fusion was a have different mutational spectrum from other AML. favorable factor for EFS and OS (P = 0.029, P = 0.016, Studies in cytogenetically normal AML had demon- respectively) (Table 3). However, in multivariate analysis, strated that FLT3-ITD was associated with increased risk of WBC counts, bone marrow blasts, MYH11-CBFβ fusion, relapse  while NPM1 was a good prognostic factor . FLT3-ITD, mutations in NPM1 and DNMT3A, and allo- In our study, however, recurrent genetic mutations, geneic HSCT did not show prognostic signiﬁcance in our including FLT3-ITD, NPM1, DNMT3A, NRAS/KRAS, cohort of FAB-M4 and -M5 patients (Table 4). IDH1/2, RUNX1, TET2, WT1, KIT, and U2AF1 [17, 18], were not association with survival. The discrepancy could be related to the small sample size of our cohort, but could Discussion also be explained by possible interplay of mutated genes. Our results were consistent with previous ﬁndings that The genetic heterogeneity and the complex interaction MYH11-CBFβ fusion indicated better survival . among different oncogenic pathways of AML has been well AML in older patients generally had poorer prognosis demonstrated in previous studies . FAB-M4 and -M5 due to higher mutation burden, poorer baseline performance are unique subtypes of AML that belong to acute myelo- status, and co-morbidities . In our cohort of FAB-M4 monocytic lineage leukemia and their mutation spectrum and -M5, age ≥60 years had a negative impact on survival. has not been fully described. In our study, we found a high Of note, allogeneic HSCT (allo-HSCT), which was usually mutation frequency (≥10%) of NPM1, DNMT3A, FLT3 and considered the cure for AML, did not show survival beneﬁt NRAS. This was different from the results of other studies in our study. High mutation frequency in NPM1, a good Prognosis signiﬁcance of mutational spectrum in FAB M4 and M5 81 Fig. 2 Kaplan–Meier estimate of event-free survival (EFS) and overall had longer EFS and OS than those with negative MYH11-CBFβ. e, f survival (OS). a, b Patients older than 60 years had shorter EFS and Allogeneic hematopoietic stem cell transplantation did not confer OS than younger patients. c, d Patients with positive MYH11-CBFβ signiﬁcant survival beneﬁt prognostic factor as shown in other studies, could help is a retrospective study, which is not as good as prospective explain the relatively benign course of FAB-M4 and -M5 research. Third, the prognostic signiﬁcance of the related that transplant was not mandatory to prevent relapse and factors was the combination of chemotherapy and trans- prolong survival. To further clarify the beneﬁt and risk of plantation, it cannot reﬂect the effect of a single factor on allo-HSCT in FAB-M4 and -M5 patients, larger prospective the prognosis. studies would be warranted. In conclusion, our study provided new insight into Several limitations need to be acknowledge. First, due to mutational spectrum and molecular signatures of FAB-M4 the limitation of sample size, we did not stratify data more and -M5. These patients may not require allo-HSCT. More precisely based on factors that affect the prognosis. Hence, detailed mutational spectrum information and large pro- our results did not fully account for the effect of mutational spective studies are required in the future for better prog- spectrum and clinical data on prognosis. Second, our study nostication of FAB-M4 and -M5. 82 Z. Cheng et al. Table 3 Univariate analysis for EFS and OS Variables EFS OS HR (95%CI) P-value HR (95%CI) P-value Age (≥60 vs. <60 years) 2.201 (1.223–3.964) 0.009 2.538 (1.399–4.604) 0.002 WBC (≥30 vs. <30 × 10 /L) 0.806 (0.448–1.449) 0.471 0.747 (0.415–1.344) 0.330 BM blasts (≥70% vs. <70%) 1.642 (0.885–3.049) 0.116 1.806 (0.971–3.361) 0.062 PB blasts (≥20% vs. <20% 1.028 (0.568–1.859) 0.928 0.955 (0.528–1.728) 0.879 Mutated genes (≥3 vs. <3) 1.525 (0.829–2.806) 0.174 1.507 (0.819–2.775) 0.187 MYH11-CBFβ (positive vs. negative) 0.271 (0.084–0.877) 0.029 0.234 (0.072–0.761) 0.016 FLT3-ITD (positive vs. negative) 0.919 (0.389–2.170) 0.847 1.002 (0.423–2.370) 0.997 NPM1 (mutated vs. wild type) 0.964 (0.530–1.755) 0.905 1.025 (0.563–1.867) 0.936 DNMT3A (mutated vs. wild type) 1.294 (0.714–2.344) 0.396 1.348 (0.743–2.448) 0.326 IDH1/IDH2 (mutated vs. wild type) 1.040 (0.409–2.647) 0.934 0.938 (0.369–2.380) 0.892 RUNX1 (mutated vs. wild type) 1.814 (0.759–4.340) 0.181 1.696 (0.709–4.054) 0.236 NRAS/KRAS (mutated vs. wild type) 1.045 (0.467–2.339) 0.915 1.008 (0.451–2.254) 0.985 TET2 (mutated vs. wild type) 1.328 (0.524–3.364) 0.550 1.468 (0.579–3.723) 0.419 WT1 (mutated vs. wild type) 1.709 (0.521–5.605) 0.377 1.068 (0.331–3.446) 0.913 KIT (mutated vs. wild type) 0.857 (0.208–3.539) 0.831 0.653 (0.158–2.699) 0.556 U2AF1 (mutated vs. wild type) 1.797 (0.553–5.834) 0.330 2.311 (0.705–7.574) 0.166 Allo-HSCT (yes vs. no) 0.682 (0.368–1.265) 0.225 0.546 (0.293–1.015) 0.056 EFS event-free survival, OS overall survival, WBC white blood cell, BM bone marrow, PB peripheral blood, HSCT hematopoietic stem cell transplantation Table 4 Multivariate analysis for EFS and OS Variables EFS OS HR (95% CI) P-value HR (95% CI) P-value Age (≥60 vs. <60 years) 1.873 (0.985–3.562) 0.056 1.879 (0.982–3.596) 0.057 WBC (≥30 vs. <30 × 10 /L) 0.874 (0.406–1.880) 0.730 0.799 (0.387–1.651) 0.545 BM blasts (≥70% vs. <70%) 1.700 (0.792–3.650) 0.173 1.617 (0.769–3.400) 0.205 MYH11-CBFβ (positive vs. negative) 0.448 (0.105–1.900) 0.276 0.357 (0.086–1.474) 0.154 FLT3-ITD (positive vs. negative) 0.948 (0.375–2.394) 0.910 1.127 (0.448–2.839) 0.799 NPM1 (mutated vs. wild type) 0.704 (0.328–1.509) 0.367 0.704 (0.332–1.493) 0.360 DNMT3A (mutated vs. wild type) 1.398 (0.704–2.775) 0.339 1.358 (0.692–2.665) 0.373 Allo-HSCT (yes vs. no) 0.797 (0.409–1.552) 0.504 0.642 (0.325–1.269) 0.202 EFS event-free survival, OS overall survival, WBC white blood cell, BM bone marrow, HSCT hematopoietic stem cell transplantation Acknowledgements This work was supported by grants from the References National Natural Science Foundation of China (81500118, 61501519), the China Postdoctoral Science Foundation funded project (project 1. Schlenk RF, Döhner H. Genomic applications in the clinic: use in No.2016M600443), PLAGH project of Medical Big Data (Project treatment paradigm of acute myeloid leukemia. Hematol Am Soc No.2016MBD-025) and Jiangsu Province Postdoctoral Science Hematol Educ Program. 2013;2013:324–30. Foundation funded project (project No.1701184B). 2. Cazzola M, Della Porta MG, Travaglino E, Malcovati L. Classi- ﬁcation and prognostic evaluation of myelodysplastic syndromes. Semin Oncol. 2011;38:627–34. Compliance with ethical standards 3. Kohlmann A, Grossmann V, Nadarajah N, Haferlach T. Next- generation sequencing - feasibility and practicality in haematol- Conﬂict of interest The authors declare that they have no conﬂict of ogy. Br J Haematol. 2013;160:736–53. interest. Prognosis signiﬁcance of mutational spectrum in FAB M4 and M5 83 4. Garraway LA. Genomics-driven oncology: framework for an 13. Naoe T, Kiyoi H. Gene mutations of acute myeloid leukemia in emerging paradigm. J Clin Oncol. 2013;31:1806–14. the genome era. Int J Hematol. 2013;97:165–74. 5. Cancer Genome Atlas Research Network. Genomic and epige- 14. Grossmann V, Schnittger S, Kohlmann A, Eder C, Roller A, nomic landscapes of adult de novo acute myeloid leukemia. New Dicker F, et al. A novel hierarchical prognostic model of AML Engl J Med. 2013;368:2059–74. solely based on molecular mutations. Blood. 2012;120:2963–72. 6. Papaemmanuil E, Gerstung M, Bullinger L, Gaidzik VI, Paschka 15. Walter MJ, Ding L, Shen D, Shao J, Grillot M, McLellan M, et al. P, Roberts ND, et al. Genomic classiﬁcation and prognosis in Recurrent DNMT3A mutations in patients with myelodysplastic acute myeloid leukemia. New Engl J Med. 2016;374:2209–21. syndromes. Leukemia. 2011;25:1153–8. 7. Patel JP, Gönen M, Figueroa ME, Fernandez H, Sun Z, Racevskis 16. Ganguly BB, Kadam NN. Mutations of myelodysplastic syn- J, et al. Prognostic relevance of integrated genetic proﬁling in dromes (MDS): an update. Mutat Res/Rev Mutat Res. acute myeloid leukemia. New Engl J Med. 2012;366:1079–89. 2016;769:47–62. 8. Ofran Y, Rowe JM. Genetic proﬁling in acute myeloid leukae- 17. Papaemmanuil E, Gerstung M, Malcovati L, Tauro S, Gundem G, mia--where are we and what is its role in patient management. Br J Van Loo P, et al. Clinical and biological implications of driver Haematol. 2013;160:303–20. mutations in myelodysplastic syndromes. Blood. 9. Bennett JM, Catovsky D, Daniel MT, Flandrin G, Galton DA, 2013;122:3616–27. Gralnick HR, et al. Proposal for the recognition of minimally 18. Haferlach T, Nagata Y, Grossmann V, Okuno Y, Bacher U, Nagae differentiated acute myeloid leukaemia (AML-M0). Br J Haema- G, et al. Landscape of genetic lesions in 944 patients with mye- tol. 1991;78:325–9. lodysplastic syndromes. Leukemia. 2014;28:241–7. 10. Ley TJ, Miller C, Ding L, Raphael BJ, Mungall AJ, Robertson A, 19. Kurosawa S, Miyawaki S, Yamaguchi T, Kanamori H, Sakura T, et al. Genomic and epigenomic landscapes of adult de novo acute Moriuchi Y, et al. Prognosis of patients with core binding factor myeloid leukemia. New Engl J Med. 2013;368:2059–74. acute myeloid leukemia after ﬁrst relapse. Haematologica. 11. Mardis ER, Ding L, Dooling DJ, Larson DE, McLellan MD, Chen 2013;98:1525–31. K, et al. Recurring mutations found by sequencing an acute 20. Döhner H, Weisdorf DJ, Bloomﬁeld CD. Acute myeloid leuke- myeloid leukemia genome. New Engl J Med. 2009;361:1058–66. mia. New Engl J Med. 2015;373:1136–52. 12. Estey EH. Acute myeloid leukemia: 2013 update on risk- stratiﬁcation and management. Am J Hematol. 2013;88:318–27.
Cancer Gene Therapy – Springer Journals
Published: Feb 28, 2018
It’s your single place to instantly
discover and read the research
that matters to you.
Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.
All for just $49/month
Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly
Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.
Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.
Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.
All the latest content is available, no embargo periods.
“Hi guys, I cannot tell you how much I love this resource. Incredible. I really believe you've hit the nail on the head with this site in regards to solving the research-purchase issue.”Daniel C.
“Whoa! It’s like Spotify but for academic articles.”@Phil_Robichaud
“I must say, @deepdyve is a fabulous solution to the independent researcher's problem of #access to #information.”@deepthiw
“My last article couldn't be possible without the platform @deepdyve that makes journal papers cheaper.”@JoseServera