Background: We analyzed clinical associations of MET alterations in the plasma of patients with diverse malignancies. Methods: Digital sequencing of circulating tumor DNA (ctDNA) (54–70 genes) was performed in 438 patients; 263 patients also had tissue sequencing (182–315 genes). The most represented tumor types were gastrointestinal (28.1%), brain (24.9%), and lung (23.2%). Most patients (71.2%) had recurrent/metastatic disease. Results: MET alterations were observed in 31 patients (7.1%) and correlated with bone metastasis (P =0.007), with TP53 (P =0.001) and PTEN (P = 0.003) abnormalities, and with an increased number of alterations (median, 4 vs 1, P =0.001) (all multivariable analyses). Patients with MET alterations demonstrated a significantly shorter median time to metastasis/recurrence (1.0 vs 10.4 months, P = 0.044, multivariable) and a poorer survival (30.6 vs 58.4 months, P = 0.013, univariate only). Of the 31 patients with MET alterations, 18 also had tissue testing; only two also had tissue MET alterations (11.1%); MET alterations were detected at a lower frequency in tissue (1.14%) compared to ctDNA (7.1%), with P = 0.0002. Conclusions: In conclusion, the detection of MET alterations by liquid biopsy is feasible. MET ctDNA alterations were associated with a poorer prognosis, higher numbers of genomic abnormalities, and bone metastases. The correlation with bone metastases may explain the higher frequency of MET alterations in blood ctDNA than in tissue (since bones are rarely biopsied) and the previous observations of bone-predominant responses to MET inhibitors. The high number of co-altered genes suggests that MET inhibitors may need to be combined with other agents to induce/ optimize responses. Keywords: MET, ctDNA, cfDNA, Liquid biopsies, Bone metastasis Background MAPK, PI3K-Akt/PKB, Crk-Rap, and Rac-Pak pathways MET, also called c-MET or hepatocyte growth factor . These pathways form distinct branches that interact (HGF) receptor, is a receptor tyrosine kinase discovered to regulate cell proliferation, invasion, migration, angio- as an oncogene in the 1980s [1, 2]. Independent research genesis, development, organ regeneration, and tumori- found that HGF or scatter factor (SF) was a ligand for genesis . MET . Upon binding of HGF to MET, the kinase Activating MET mutations are found in diverse human domain phosphorylates growth factor receptor-bound cancer . For instance, activating mutations in the protein 2 (GRB2) and GRB2-associated binding protein kinase domain are a feature of both hereditary and some 1 (GAB1) and activates diverse downstream signaling non-hereditary forms of papillary renal cell carcinoma pathways important in cancer, including the ERK/ . MET amplification is seen in 5 to 20% of non-small cell lung cancer (NSCLC) and gastric cancer . MET overexpression correlates with poor treatment outcome * Correspondence: email@example.com; firstname.lastname@example.org; email@example.com in some malignancies . Department of Medicine, Center for Personalized Cancer Therapy, Division Drugs that target MET include inhibitors of the HGF/ of Hematology/Oncology, University of California, San Diego, Moores Cancer -MET pathway, MET antibodies, and MET kinase Center, 3855 Health Sciences Drive, #0658, La Jolla, CA 92093-0987, USA Full list of author information is available at the end of the article © The Author(s). 2018 Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated. Ikeda et al. Journal of Hematology & Oncology (2018) 11:76 Page 2 of 10 inhibitors . Cabozantinib and crizotinib are both Food each patient) and sequencing libraries were prepared and Drug Administration (FDA)-approved multikinase with custom in-line barcode molecular tagging and inhibitors that are also potent suppressors of MET [9, 10]. complete sequencing at 15,000× read depth (~ 4000 Crizotinib induced responses in some patients with highly unique double-stranded cfDNA fragments, each repre- MET-amplified lung cancer . Studies with the sented by 3–5 sequencing reads). The panels utilize hy- multi-target MET inhibitor cabozantinib have shown sig- brid capture followed by NGS of the critical exons in a nificant activity against a variety of solid tumors including panel of 54–70 genes (Additional file 1: Table S1 to S3) melanoma, as well as renal and non-small cell lung, liver, and report all four major types of genomic alterations medullary thyroid, breast, and ovarian cancer, but this ac- (point mutations, indels, fusions, and copy number am- tivity is likely due to other targets of cabozantinib, such as plifications). Post-sequencing bioinformatics matches the VEGFR or RET . In the phase I setting, patients with complementary strands of each barcoded DNA fragment MET amplification did not respond to MET inhibitors to remove false positive results . The variant allele (but the number of treated patients was small) . fraction (VAF) is computed as the number of mutated Because of their non-invasive nature, liquid biopsies DNA molecules divided by the total number (mutated are increasingly used in the clinical setting. Indeed, nu- plus wild type) of DNA fragments at that allele; VAF is merous studies showed a relatively good correlation with reported as a percentage. The majority of cell-free DNA tissue sequencing and the potential to detect actionable is wild type (germline); thus, the median VAF of somatic alterations [12–16]. In this study, we analyzed MET al- alterations is < 0.5%. The analytic sensitivity reaches de- terations in the plasma-derived circulating tumor DNA tection of one to two single-mutant fragments from a (ctDNA) of 438 patients with diverse malignancies and 10-ml blood sample (0.1% limit of detection), and the explored the relationship between MET alterations, analytic specificity is greater than 99.9999% . demographics, as well as other molecular alterations and For 144 patients, a 54-gene panel was used, which clinical outcomes. identified potential tumor-related alterations in 54 cancer-related genes (Additional file 1: Table S1) includ- Methods ing copy number amplifications in ERBB2, EGFR, and Patients MET (indels and fusions were not detected as part of We reviewed the clinicopathology and clinical outcomes this panel). For 272 patients, a 68-gene version of the of 438 consecutive patients with cancer for whom ctDNA original panel (expanded to all four major alteration testing had been performed and who were seen at the UC types) was used, and for 22 patients, the most recent San Diego Moores Cancer Center from June 2014 to July 70-gene panel version (further expanded to amplifica- 2016. Pathology was reviewed at UCSD. Data was ab- tions in 18 genes and fusions in 6 genes) was applied stracted from the electronic medical record. This study (Additional file 1: Table S2 and S3). Only non-synonymous was performed and consents were obtained in accordance alterations were included in our analysis. with the UCSD Institutional Review Board guidelines In addition, 263 patients (~ 60%) of the 438 patients [PREDICT-UCSD (Profile Related Evidence Determining with ctDNA test also had CLIA/CAP-accredited Individualized Cancer Therapy); NCT02478931]. next-generation sequencing (NGS) performed on tumor tissue (FoundationOne™, Cambridge, Massachusetts, http:// Next-generation sequencing www.foundationone.com.) (N = 182 to 315 gene panels). Digital sequencing of ctDNA (DNA) in all patients was performed by Guardant Health, Inc. (Guardant360, Statistical and outcome analysis Redwood City, California, http://www.guardanthealth.- Patient characteristics were summarized using descrip- com/guardant360/), a Clinical Laboratory Improvement tive statistics. Medians and respective 95% confidence Amendment (CLIA)-certified and College of American intervals and range were calculated, whenever possible. Pathologists (CAP)-accredited clinical laboratory. The Associations between categorical variables were tested analytical and clinical validation of Guardant360 was using a binary logistic regression model. Linear variables conducted in accordance with evidentiary standards were tested using the Mann-Whitney U test for univari- established by the Standards for Reporting of Diagnostic able analyses and a multiple linear regression model for Accuracy (STARD), REporting of tumor MARKer Stud- multivariable analyses. ies (REMARK), Evaluation of Genomic Applications in Time to metastasis/recurrence was defined as the time Practice and Prevention (EGAPP), and the recent interval between diagnosis and first metastasis/recur- Next-generation Sequencing: Standardization of Clinical rence (whichever came first) or last follow-up date Testing (Nex-StoCT) biomarker guidelines . As de- (patients who had not recurred/developed metastases at scribed in Lanman et al. , 5–30 ng of ctDNA was last follow-up were censored on that date). Overall sur- isolated from plasma (two 10 ml Streck tubes drawn for vival (OS) was defined as the time from diagnosis to Ikeda et al. Journal of Hematology & Oncology (2018) 11:76 Page 3 of 10 death or last follow-up date for patients who were alive lung (23.2%, N = 102), and breast (11.6%, N = 51) cancers. (patients still alive at the last follow-up were censored The majority of the patients had recurrent or metastatic on that date). Estimations for the time to first metasta- disease at the time of blood draw used for testing (71.2%, sis/recurrence and OS were done using Kaplan-Meier N =312) (Table 1). analyses and were compared among subgroups by the log-rank test for univariable analysis or Cox regression MET alterations and associations with patient characteristics models for multivariable analysis. All statistical analyses Overall, MET alterations were observed in 31 of the 438 were performed by author MS with SPSS version 24.0. patients whose ctDNA was tested (7.1%). Sixteen pa- tients had a MET amplification only; 13 had a somatic Results mutation only; and two had both an amplification and a Patient demographic characteristics somatic MET mutation (Fig. 1a). The median age of patients at diagnosis was 57.5 years (CI 95%, 54.5–59.1). Women comprised 52.1% (N = 228) Univariate analyses of the population. The majority of patients were Caucasian MET alterations were significantly (positively) associated (69.2%, N = 303). The most represented tumor types were with lung cancers (11.8%) and the presence of metasta- gastrointestinal (28.1%, N = 123), brain (24.9%, N = 109), sis/recurrence at the time of blood draw and negatively Table 1 Demographics comparison of 438 patients with or without MET alterations Characteristics Total patients, MET wild type, MET alteration(s), P values* N = 438 (100%) N = 407 (92.9%) N = 31 (7.1%) Age at diagnosis (years) (median, CI 95%) 57.5 (54.5–59.1) 57.7 (55.3–59.2) 53.8 (49.0–62.2) 0.791 Gender 0.288 Women 228 (52.1%) 209 (91.7%) 19 (8.3%) Men 210 (47.9%) 198 (94.3%) 12 (5.7%) Ethnicity Caucasian 303 (69.2%) 284 (93.7%) 19 (6.3%) 0.326 Asian 52 (11.9%) 46 (88.5%) 6 (11.5%) 0.188 Hispanic 30 (6.8%) 28 (93.3%) 2 (6.7%) 0.928 African American 9 (2.1%) 8 (88.9%) 1 (11.1%) 0.637 Middle Eastern 5 (1.1%) 5 (100%) 0 (0%) 0.998 Unknown 39 (8.9%) 36 (92.3%) 3 (7.7%) 0.875 Type of cancer Gastrointestinal 123 (28.1%) 118 (95.9%) 5 (4.1%) 0.132 Brain 109 (24.9%) 107 (98.2%) 2 (1.8%) 0.026** Lung 102 (23.2%) 90 (88.2%) 12 (11.8%) 0.039** Breast 51 (11.6%) 45 (88.2%) 6 (11.8%) 0.172 Genitourinary 18 (4.1%) 15 (83.3%) 3 (16.7%) 0.120 Head and neck 10 (2.3%) 9 (90.0%) 1 (10.0%) 0.717 Gynecologic 7 (1.6%) 6 (85.7%) 1 (14.3%) 0.465 Melanoma 5 (1.1%) 4 (80.0%) 1 (20.0%) 0.285 Hematologic 3 (0.7%) 3 (100%) 0 (0%) 0.998 Other 10 (2.3%) 10 (100%) 0 (0%) 0.998 Presence of metastasis or recurrence at the time of blood draw Yes 312 (71.2%) 284 (91.0%) 28 (9.0%) 0.024** No 126 (28.8%) 123 (97.6%) 3 (2.4%) The percentages on the first column are expressed over the total number of patients (N = 438); for the second and third columns, percentages are expressed over the total number of patients for each variable Lymphoma (n =2), sarcoma (n=2), thymoma (n = 2), desmoid tumor, neurofibromatosis, peripheral nerve sheath tumor, and carcinoma of unknown primary (each n =1) *P values were computed using the independent sample Mann-Whitney U test for linear variables (age at diagnostic) and the logistic binary regression analysis for categorical variables, as appropriate **MET alterations were negatively associated with brain tumors but positively associated with lung tumors and metastasis or recurrence at the time of blood draw Ikeda et al. Journal of Hematology & Oncology (2018) 11:76 Page 4 of 10 Fig. 1 a MET alterations representation. Pie chart representing the different types of MET alterations identified in our cohort comprising 438 patients. b Overall survival and time to first metastasis/recurrence. Kaplan-Meier curves depicting the overall survival (left panel) and the time to first metastasis/ recurrence (both from diagnosis time). P values are from univariable analysis. For more details, refer to Tables 4 and 5 associated with brain tumors 1.8% (all P < 0.05) (Table 1). The univariate association between MET alterations and We then examined the associations with other genomic lung cancer, the negative association with brain tumors, alterations and found that MET alterations correlated in and the correlation with metastases at the time of blood univariable analysis with alterations in TP53, EGFR, draw did not remain significant in multivariate analysis. PIK3CA, BRAF, ARID1A, ALK, and PTEN genes (all P < Multivariable analysis also demonstrated that MET alter- 0.05, Additional file 1: Table S4). We also investigated ations correlated with an increased number of alterations the potential associations between MET alterations and (median of 4 alterations vs 1, P = 0.001) (Table 3). These the location of metastatic sites and showed a significant conclusions remained valid if the patients with brain tu- correlation with lymph nodes and bone metastasis (both mors, lymphoma/leukemia, thymoma, and NF were ex- P < 0.02, Additional file 1: Table S4). Of note, patients cluded, with P values of < 0.01 for association between with MET alterations had a median of two metastatic MET alterations and TP53 or PTEN alterations, bone me- sites versus a single site for patients without MET alter- tastases, and increased number of metastases. ations (P = 0.001). The latter conclusions remained un- The other variables that were significantly associated changed if the patients with brain tumors, lymphoma/ with a higher number of alterations in the multivariable leukemia, thymoma, and NF were excluded, with P values analysis were alterations in TP53, PIK3CA, BRAF, KRAS, of < 0.05 for association between MET alterations and or MYC genes as well as the presence of metastasis in lymph node or bone metastases and increased number of the lymph node or adrenal tissue (Table 3). metastases. MET alterations and survival outcomes Multivariate analyses Overall survival In themultiplelogistic regressionmodel(thatincludedany Overall, 213 patients (213/438, 48.6%) had died at the parameters that were significant (P ≤ 0.05) in univariate time of our analysis (20/31 with MET alterations; 193/ analysis), the only variables that remained statistically 407 without MET alterations detected in their ctDNA). associated with MET alterations were aberrations in TP53 A log rank test demonstrated a poorer survival (time (P = 0.001) and PTEN (P = 0.003) genes, as well as an in- from diagnosis until death or last follow-up date) for pa- creased incidence of bone metastasis (P = 0.007) (Table 2). tients bearing MET alterations (30.6 months vs 58.4, Ikeda et al. Journal of Hematology & Oncology (2018) 11:76 Page 5 of 10 Table 2 Multivariable analysis of the variables associated with MET alterations* MET wild type, MET altered, Univariable Multivariable N = 407 N =31 Characteristics Wald Odds ratio (CI 95%) P value Wald Odds ratio (CI 95%) P value Type of cancer Brain (n = 109)** 107 (26.3%) 2 (6.5%) 4.94 0.19 (0.05–0.83) 0.026** –– – Lung (n = 102) 90 (22.1%) 12 (38.7%) 4.26 2.22 (1.04–4.76) 0.039 –– – Presence of metastasis or recurrence 284 (69.8%) 28 (90.3%) 5.13 4.0 (1.2–14.3) 0.024 –– – at the time of blood draw (n = 312) Genetic alteration type TP53 (n = 149) 129 (31.7%) 20 (64.5%) 12.2 3.85 (1.82–8.33) < 0.001 10.9 3.7 (1.7–8.3) 0.001 EGFR (n = 53) 43 (10.6%) 10 (32.4%) 11.2 4.0 (1.79–9.1) 0.001 –– – PIK3CA (n = 42) 35 (8.6%) 7 (22.6%) 5.93 3.1 (1.25–7.7) 0.015 –– – BRAF (n = 27) 22 (5.4%) 5 (16.1%) 5.14 3.3 (1.2–10.0) 0.023 –– – ARID1A (n = 19) 15 (3.7%) 4 (12.9%) 5.14 3.85 (1.2–12.5) 0.023 –– – ALK (n = 14) 11 (2.7%) 3 (9.7%) 3.94 3.85 (1.02–14.3) 0.047 –– – PTEN (n = 9) 6 (1.5%) 3 (9.7%) 7.72 7.1 (1.7–33.3) 0.007 9.12 11.1 (2.3–50.0) 0.003 Metastatic sites Lymph node (n = 139) 123 (30.2%) 16 (51.6%) 5.77 2.4 (1.2–5.3) 0.016 –– – Bone (n = 102) 88 (21.6%) 14 (45.2%) 8.26 2.9 (1.4–6.3) 0.004 7.34 2.9 (1.35–6.25) 0.007 No metastases (n = 147)** 145 (35.6%) 2 (6.5%) 7.96 0.13 (0.03–0.5) 0.005** –– – The Wald statistics test the unique contribution of each variable; the higher the Wald statistics, the higher the association/contribution in the model *Only variables that were significant in the univariable models (logistic regression) were included in the multivariable analysis, with the final model containing only significant covariates in the multivariable analyses (forward stepwise selection model) **“Brain tumors” and “no metastases” were negatively associated with MET alterations in univariate analysis. These variables were not significant in the final multivariate analysis model P = 0.013). In the multivariable analysis, only lung cancer most of the patients with both types of testing, the ctDNA as well as the presence of ARID1A, KRAS, ALK,and MYC test was performed after the tissue testing (16/18 cases). alterations and liver metastasis remained significant pre- Only two patients who had a MET alteration identified in dictors of a poorer survival (all P <0.50, Fig. 1b)(Table 4). their ctDNA also had a MET alteration found in their tis- sue testing (11.1%; MET amplification and MET Y501C, Time to metastasis/recurrence one patient each). In these two patients, the time interval Patients with MET alterations had a significantly shorter between the ctDNA and tissue biopsies was 1.8 and median time to metastasis/recurrence, with a median of 15.3 months, and both tissues used for the testing were 1.0 months (95%CI could not be computed) versus from the primary tumor. 10.4 months (95%CI 6.9–14.0) (P = 0.003) (Table 5). In only one patient, a MET amplification was de- MET alterations remained significantly associated with a tected in the tissue and not in ctDNA (1 of 263 total shorter time to metastasis/recurrence in a multivariable patients who had NGS tissue testing). The tissue test analysis (P = 0.044), along with the presence of liver was performed on a pancreatic tumor that was surgi- (P = 0.022) or lymph node metastases (P < 0.001). In- cally removed, and the ctDNA test was done more versely, breast cancer correlated with a longer time to than 1 year later (recent scans showed appearance of metastasis/recurrence (P < 0.001) (Fig. 1b and Table 5). new pulmonary and liver nodules 2 months prior the ctDNA testing). Comparison with tissue testing Overall, MET alterations were detected at a significantly As noted, 438 patients had ctDNA testing; 263 of these lower frequency in tissue (3/263 patients, 1.14%) com- patients also had tissue NGS performed. pared to ctDNA (31/438, 7.1%), P = 0.0002. Further, of Of the 31 patients with MET alterations in ctDNA, 18 eight patients who harbored MET alterations in ctDNA, also had tissue testing (Foundation Medicine see the and had ctDNA and tissue testing within 2 months of “Methods” section). The median time interval between each other, only one patient showed a similar MET alter- the blood draw and the tissue biopsy for these 18 patients ation in the tissue. Six of the seven patients with only was 6.1 months (95% CI (2–13.7); range (0.2–32.6)). In ctDNA positive for MET alterations had bone metastases. Ikeda et al. Journal of Hematology & Oncology (2018) 11:76 Page 6 of 10 Table 3 Association with the number of alterations* Variables Median N of alterations P value t statistic P value (95% CI) (range) (univariable) (multivariable) (multivariable) Overall 1 (1–1) (0–26) Tumor type Gastrointestinal 2 (1–2) (0–26) 0.031 –– Brain 0 (0–0) (0–5) < 0.001 –– Lung 2 (2–3) (0–21) < 0.001 –– Genomic alterations MET 4(3–6) (1–21) < 0.001 3.35 0.001 TP53 3(3–3) (1–21) < 0.001 6.86 < 0.001 EGFR 3(3–4) (1–26) < 0.001 –– PIK3CA 5(4–6) (1–26) < 0.001 7.31 < 0.001 BRAF 5(4–6) (1–26) < 0.001 6.54 < 0.001 KRAS 4(3–4) (1–26) < 0.001 6.01 < 0.001 MYC 5(3–6) (2–7) < 0.001 2.70 0.007 Metastatic/recurrence at the time of blood draw 2 (1–2) (0–26) < 0.001 –– Metastatic sites Lymph node 2 (2–3) (0–21) < 0.001 2.05 0.041 Bone 2 (2–3) (0–26) < 0.001 –– Liver 2 (2–3) (0–26) < 0.001 –– Lung 2 (1–3) (0–19) 0.010 –– Brain 2 (1–3) (0–21) < 0.001 –– Adrenal 2 (2–4) (0–20) < 0.001 2.06 0.040 *Only variables with ≥ 20 patients in the overall population were tested. Only significant variables in the univariable analysis (non-parametric Mann-Whitney U test) were included in the multivariable model (multiple linear regression), with the final model containing only significant covariates (stepwise modelselection). The t statistics test the unique contribution of each variable; the higher the t statistics, the higher the association/contribution in the model Discussion ); (ii) our study used blood-derived ctDNA, which This is the largest study interrogating the feasibility and could capture shed tumor DNA from multiple sites, while utility of MET alteration detection through blood-derived the previous reports used tissue-based testing, which ctDNA. Liquid biopsy is a non-invasive method to find would only detect aberrations in the piece of tissue biop- genomic aberrations and is increasingly utilized in the sied. Consistent with the above observations, our tissue clinical setting as reflected by the non-small cell lung can- NGS testing also showed significantly lower rates of MET cer National Comprehensive Cancer Network (NCCN) alterations than the ctDNA NGS: 3 of 263 patients (1.1%) guidelines [18, 19]. Our study demonstrated that MET (who also had ctDNA tests) were positive for MET alter- ctDNA alterations were detected in 7.1% of patients with ations in tissue versus 31 of 438 ctDNA-tested patients solid tumors. This detection rate is higher compared to (7%) being positive for MET alterations (P = 0.0002). Fur- previous tissue studies [5, 20, 21]. For instance, a study of ther, of 18 patients positive for MET ctDNA alterations MET tissue amplification determined by fluorescent in who also had tissue NGS, only 2 (11.1%) were also positive situ hybridization (FISH) demonstrated that 2.6% of 1115 for a MET alteration by tissue NGS. The biologic under- solid tumor specimens were positive . Furthermore, the pinnings of discordance between ctDNA and tissue NGS MSKCC-IMPACT study showed that 3% of patients had results have been previously documented and include MET tissue alterations [20, 21]. The differences in rate of spatial (intra-tumor and inter-tumor) and temporal het- MET alterations between our study and the other studies erogeneity in genomic anomalies in cancers along with could be due to following reasons: (i) our technology de- the fact that ctDNA is comprised of DNA that has leaked tected both single nucleotide substitutions and amplifica- into the circulation from diverse metastatic sites while tis- tions, while the previous investigation by Jardim et al. at sue NGS reflects only the tissue specimen analyzed . MD Anderson Cancer Center  discerned only amplifi- The question that arises is whether or not biological cations (though the MSKCC study would have discerned explanations can specifically account for the higher rate both single nucleotide substitutions and amplifications of MET alterations in ctDNA. Of interest in this regard, Ikeda et al. Journal of Hematology & Oncology (2018) 11:76 Page 7 of 10 Table 4 Overall survival analysis (N = 438 patients)* Univariable Multivariable Characteristics Median time (months) (std. error) P value Wald Hazard ratio P value (CI 95%) Type of cancer Gastrointestinal (n = 123) vs not 36.3 (4.5) vs 71.6 (6.5) 0.002 –– – Brain (n = 109) vs not** 95.0 (50.1) vs 51.5 (5.9) 0.003 –– – Lung (n = 102) vs not 31.8 (6.0) vs 61.0 (7.4) 0.002 9.59 1.7 (1.21–4.04) 0.002 Breast (n = 51) vs not 104.8 (20.3) vs 50.0 (5.6) 0.027 Genetic alteration type MET (n = 31) vs not 30.6 (6.2) vs 58.4 (6.3) 0.013 –– – TP53 (n = 149) vs not 42.1 (6.1) vs 66.5 (10.3) 0.003 –– – PIK3CA (n = 42) vs not 43.2 (10.6) vs (58.4 (6.3) 0.013 –– – ARID1A (n = 19) vs not 24.8 (6.2) vs 58.4 (6.8) 0.002 7.98 2.7 (1.25–4.0) 0.005 KRAS (n = 50) vs not 18.3 (7.3) vs 59.7 (6.5) < 0.001 15.68 2.1 (1.5–3.12) < 0.001 ALK (n = 14) vs not 31.3 (9.0) vs 58.3 (6.7) 0.015 5.76 2.04 (1.14–3.7) 0.016 SMAD4 (n = 12) vs not 21.7 (8.2) vs 58.3 (6.2) 0.007 –– – MYC (n = 20) vs not 22.2 (3.9) vs 59.1 (6.5) < 0.001 7.59 2.17 (1.24–3.77) 0.006 Metastatic/recurrence at the time of 53.6 (4.9) vs NR 0.011 –– – blood draw vs not Metastatic sites Liver (n = 111) vs not 49.3 (8.2) vs 59.1 (7.9) 0.023 4.85 1.4 (1.03–1.89) 0.028 Lymph node (n = 139) vs not 39.9 (7.4) vs 74.1 (11.1) < 0.001 –– – Only variables that were significant in the univariable models (log rank test) were included in the multivariable analysis (Cox regression model), with the final model containing only significant covariates in the multivariable analyses (forward stepwise selection model). The Wald statistics test the unique contribution of each variable; the higher the Wald statistics, the higher the association/contribution in the model NR not reached *Overall survival was defined as the time from diagnosis to death or last follow-up date **Brain tumors (N = 109) included n = 50 glioblastoma cases, while the rest of the tumors were lower grade astrocytomas or other lower grade brain tumors our study showed that bone metastases were independ- metastasis (HR 0.71, 95% CI 0.55–0.91) . Thesedatato- ently correlated with MET alterations (Additional file 1: gether with our results raise the possibility that cancer Table S2 and Table S4). Indeed, 14 of 31 patients (45%) clones with MET alterations preferentially localize to bone positive for MET ctDNA alterations had bone metastases and may therefore explain bone responses after MET in- (Additional file 1: Table S4). Of possible relevance in this hibitor therapy. Tissue biopsies are rarely performed regard, MET inhibitors are known to show efficacy in on bones (none of our 18 patients with MET alter- bone lesions [22, 23]. In the COMET-1 trial, 682 patients ations in ctDNA who also had tissue NGS had a bone with castrate-resistant prostate cancer who progressed biopsy). Of interest in this regard, MET is promin- after docetaxel and androgen modulators (abiraterone ently expressed (as determined by immunohistochem- and/or enzalutamide) were randomly assigned to either istry) at the site of bone metastases in renal cell cabozantinib (MET inhibitor) or prednisone. Although cancer . It is therefore conceivable that the high rates there was no difference in overall survival, 42% of cabo- of ctDNA positivity for MET alterations, which strongly zantinib-treated patients showed bone scan response com- and independently correlated with bone metastases in our pared to 3% of prednisone-treated patients (P < 0.001), study, reflect shed MET alteration-bearing DNA from albeit without prostate surface antigen (PSA) response . bone lesions (Additional file 1: Table S2 and Table S4). In the METEOR trial, 658 patients with advanced renal cell Our study also found that MET ctDNA alterations are carcinoma who progressed with at least one VEGFR small associated with poor prognosis, including decreased sur- molecule inhibitor were randomized to the MET inhibitor vival and shorter time to recurrence/metastasis (Tables 4 cabozantinib or the mTOR inhibitor everolimus . In a and 5, Fig. 1b). These results are consistent with those sub-group analysis, patients randomized to cabozantinib found by correlating MET alterations found in tissue arm with bone metastasis were associated with better over- NGS with outcome in specific malignancies, such as as- all survival (OS) (hazard ratio (HR) 0.54, 95% confidence trocytomas , breast cancers , genitourinary ma- interval (CI) 0.34–0.84) compared with non-bone lignancies , and ovarian  or gastric/esophageal Ikeda et al. Journal of Hematology & Oncology (2018) 11:76 Page 8 of 10 Table 5 Time to metastasis/recurrence from diagnosis (N = 438 patients)* Univariable Multivariable Characteristics Median time (months) (95% CI) P value Wald Hazard ratio P value (CI 95%) Type of cancer Gastrointestinal (n = 123) vs not 0.67 (0–1.9) vs 14.1 (8.3–19.8) < 0.001 –– – Brain (n = 109) vs not 42.8 (19.0–66.7) vs 2.5 (0.49–4.4) < 0.001 –– – Lung (n = 102) vs not 0.6 (0–1.3) vs 14.0 (9.5–18.6) < 0.001 –– – Breast (n = 51) vs not 30.4 (14.3–46.4) vs 7.0 (3.8–10.2) 0.001 40.03 0.33 (0.23–0.47) < 0.001 Genitourinary (n = 18) vs not 0 (−) vs 10.1 (6.8–13.4) 0.020 –– – Head and neck (n = 10) vs not 0 (−) vs 9.8 (6.2–13.4) 0.015 –– – Genetic alteration type MET (n = 31) vs not 1.0 (−) vs 10.4 (6.9–14.0) 0.003 4.05 1.47 (1.01–2.13) 0.044 TP53 (n = 149) vs not 4.1 (0.5–7.7) vs 12.7 (6.1–19.3) 0.001 –– – EGFR (n = 53) vs not 3.6 (0–8.2) vs 11.4 (7.9–14.8) 0.003 –– – ARID1A (n = 19) vs not 3.6 (0–10.1) vs 10.0 (6.3–13.7) 0.013 –– – KRAS (n = 50) vs not 0 (−) vs 11.4 (8.3–13.7) < 0.001 –– – SMAD4 (n = 12) vs not 0 (−) vs 10.1 (6.7–13.4) 0.040 –– – MYC (n = 20) vs not 0.2 (0–0.75) vs 10.4 (7.0–13.9) < 0.001 –– – Metastatic sites Adrenal (n = 29) vs not 0 (−–) vs 11.3 (8.0–14.6) 0.002 –– – Bone (n = 102) vs not 1.5 (0–6.7) vs 11.4 (8.2–14.5) 0.049 –– – Liver (n = 111) vs not 0.9 (0–2.6) vs 132.3 (9.7–16.9) < 0.001 10.04 1.49 (1.16–1.92) 0.022 Lymph node (n = 139) vs not 0.3 (−) vs 20.1 (14.3–25.9) < 0.001 13.41 1.54 (1.22–1.92) < 0.001 Peritoneal (n = 49) vs not 0 (−) vs 11.3 (8.1–14.5) < 0.001 –– – Brain (n = 63) vs not 3.6 (0–7.8) vs 10.4 (7.0–13.9) 0.045 –– – Lung (n = 90) vs not 1.2 (0–8.0) vs 11.3 (7.7–15.0) 0.003 –– – The Wald statistics test the unique contribution of each variable; the higher the Wald statistics, the higher the association/contribution in the model *Only variables that were significant in the univariable models (log rank test) were included in the multivariable analysis (Cox regression model), with the final model containing only significant covariates in the multivariable analyses (forward stepwise selection model). For some values, the 95% CI could not be computed (−) cancers . Specific genes that were co-altered with MET alterations are found in such biopsies, since ctDNA MET in multivariate analysis of our study participants MET alterations correlate independently with the pres- include PTEN and TP53 (Table 2); previously, another ence of bone metastases. Third, though our study with study has also shown the association between MET and 438 patients is the largest to date, the rarity of MET PTEN abnormalities . Overall, MET alterations signifi- alterations suggests that investigations of even greater num- cantly correlated with a higher number of alterations, bers ofpatients maybeworthwhileinorder to best under- which may explain—at least in part—the relatively lim- stand the biology and correlations of MET alterations. ited efficacy of MET inhibitors as single agents for the treatment of MET-altered advanced malignancies ob- Conclusions served in prior studies [5, 29–31]. Indeed, if patients In summary, our study demonstrated that assessment of with MET alterations generally have multiple genomic MET genomicaberrationsbyliquidbiopsyisfeasible. We abnormalities, it is likely that combination therapy, ra- found that MET ctDNA anomalies were associated with ther than monotherapy with a MET inhibitor, may be bone metastases, multiple genomic alterations, and a poorer necessary to achieve salutary effects. prognosis, including poorer overall survival and a shorter Our study has several limitations. First, only 60% of time to recurrence/metastases. Further studies are needed our patients (N = 263) with ctDNA analyses also had tis- to better understand the biologic relationship between MET sue sequencing and the median time between the two alterations and bone lesions, and next-generation trials with tests in the patients with MET-altered ctDNA was about MET inhibitors may require combinations of drugs that ad- 6 months. Second, our patients did not have bone biop- dress the genes such as PTEN that are frequently co-altered sies done, which would be of interest to determine if in these patients. Ikeda et al. Journal of Hematology & Oncology (2018) 11:76 Page 9 of 10 Additional file 3. Bottaro DP, Rubin JS, Faletto DL, Chan AM, Kmiecik TE, Vande Woude GF, et al. Identification of the hepatocyte growth factor receptor as the c-met proto-oncogene product. Science. 1991;251:802–4. Additional file 1: Table S1. 54-gene panel (N = 122 patients)— identifies 4. Birchmeier C, Birchmeier W, Gherardi E, Vande Woude GF. Met, metastasis, potential tumor-related genomic alterations within 54 cancer-related genes motility and more. Nat Rev Mol Cell Biol. 2003;4:915–25. including amplifications in ERBB2, EGFR,and MET. Only non-synonymous 5. Jardim DLF, Tang C, Gagliato DDM, Falchook GS, Hess K, Janku F, et al. Analysis alterations were analyzed. Table S2. 68-gene panel (N =272 patients), of 1,115 patients tested for MET amplification and therapy response in the MD comprising amplifications in 16 genes as well as some fusions and Anderson Phase I Clinic. Clin Cancer Res. 2014;20:6336–45. indels. Only non-synonymous alterations were analyzed. Table S3. 70-gene 6. Schmidt L, Duh FM, Chen F, Kishida T, Glenn G, Choyke P, et al. Germline and panel (N = 22 patients). Only non-synonymous alterations were analyzed. somatic mutations in the tyrosine kinase domain of the MET proto-oncogene Table S4. Comparison of clinical characteristics in 438 patients with in papillary renal carcinomas. Nat Genet. 1997;16:68–73. or without MET alterations (univariate analysis). (DOCX 20 kb) 7. Kawakami H, Okamoto I, Okamoto W, Tanizaki J, Nakagawa K, Nishio K. Targeting MET amplification as a new oncogenic driver. Cancers. 2014;6:1540–52. 8. Madoz-Gúrpide J, Zazo S, Chamizo C, Casado V, Caramés C, Gavín E, et al. Abbreviations Activation of MET pathway predicts poor outcome to cetuximab in patients 95% CI: 95% confidence interval; CAP: College of American Pathologists; with recurrent or metastatic head and neck cancer. J Transl Med. 2015;13:282. ctDNA: Circulating tumor DNA; FDA: Food and Drug Administration; 9. Kurzrock R, Sherman SI, Ball DW, Forastiere AA, Cohen RB, Mehra R, et al. FISH: Fluorescent in situ hybridization; GAB1: GRB2-associated binding Activity of XL184 (cabozantinib), an oral tyrosine kinase inhibitor, in patients protein 1; GRB2: Growth factor receptor-bound protein 2; HGF: Hepatocyte with medullary thyroid cancer. J Clin Oncol. 2011;29:2660–6. growth factor; HR: Hazard ratio; NCCN: National Comprehensive Cancer 10. Shaw AT, Kim DW, Nakagawa K, Seto T, Crinó L, Ahn M-J, et al. Crizotinib Network; NGS: Next-generation sequencing; NSCLC: Non-small cell lung versus chemotherapy in advanced ALK-positive lung cancer. N Engl J Med. cancer; OS: Overall survival; PREDICT: Profile Related Evidence Determining 2013;368:2385–94. Individualized Cancer Therapy; VAF: Variant allele fraction 11. Ross C, Sai-Hong IO, Shapiro G. Efficacy and safety of crizotinib in patients with advanced c MET-amplified non-small cell lung cancer (NSCLC). JCO. Acknowledgments 2014;15:suppl 8001. Abstract. Available from: http://meetinglibrary.asco.org/ This work was supported in part by the Joan and Irwin Jacobs Fund philanthropic record/92507/abstract. fund [no grant number is applicable—Razelle Kurzrock] and by the National 12. Schwaederle M, Patel SP, Husain H, Ikeda M, Lanman R, Banks KC, et al. Cancer Institute at the National Institutes of Health [grant P30 CA016672 (Razelle Utility of genomic assessment of blood-derived circulating tumor DNA Kurzrock, firstname.lastname@example.org)]. (ctDNA) in patients with advanced lung adenocarcinoma. Clin Cancer Res. 2017;23:5101–11. Availability of data and materials 13. Kato S, Krishnamurthy N, Banks KC, De P, Williams K, Williams C, et al. Utility All data generated or analyzed during this study are included in this published of genomic analysis in circulating tumor DNA from patients with carcinoma article [and its additional files]. of unknown primary. Cancer Res. 2017;77:4238–46. 14. Schwaederle M, Husain H, Fanta PT, Piccioni DE, Kesari S, Schwab RB, et al. Authors’ contributions Use of liquid biopsies in clinical oncology: pilot experience in patients. Clin SI, RK, and MS designed the study. MS analyzed the data to generate the Cancer Res. 2016;22:5497–505. results and created the figures and tables. SI and MS drafted the manuscript. 15. Bidard F-C, Madic J, Mariani P, Piperno-Neumann S, Rampanou A, Servois V, All authors edited and approved the final manuscript. et al. Detection rate and prognostic value of circulating tumor cells and circulating tumor DNA in metastatic uveal melanoma. Int J Cancer. 2014; Ethics approval and consent to participate 134:1207–13. Not applicable 16. Schwaederle M, Husain H, Fanta PT, Piccioni DE, Kesari S, Schwab RB, et al. Detection rate of actionable mutations in diverse cancers using a biopsy-free (blood) circulating tumor cell DNA assay. Oncotarget. 2016; Competing interests 7:9707–17. Dr. Razelle Kurzrock has research funding from Incyte, Genentech, Merck 17. Lanman RB, Mortimer SA, Zill OA, Sebisanovic D, Lopez R, Blau S, et al. Serono, Pfizer, Sequenom, Foundation Medicine, and Guardant as well as Analytical and clinical validation of a digital sequencing panel for quantitative, consultant fees from Sequenom, LOXO, and Actuate Therapeutics. Dr. Kurzrock highly accurate evaluation of cell-free circulating tumor DNA. PLoS One. 2015; is also a consultant for Genentech. She also has an ownership interest in 10:e0140712. CureMatch, Inc. The other authors declare that they have no competing interests. 18. Kwapisz D. The first liquid biopsy test approved. Is it a new era of mutation testing for non small cell lung cancer? Ann Transl. 2017;Med5:46. Publisher’sNote 19. Chabon JJ, Simmons AD, Lovejoy AF, Esfahani MS, Newman AM, Haringsma Springer Nature remains neutral with regard to jurisdictional claims in HJ, et al. Circulating tumour DNA profiling reveals heterogeneity of EGFR published maps and institutional affiliations. inhibitor resistance mechanisms in lung cancer patients. Nat Commun. 2016;7:11815. Author details 20. Cheng DT, Mitchell TN, Zehir A, Shah RH, Benayed R, Syed A, et al. Department of Medicine, Center for Personalized Cancer Therapy, Division Memorial Sloan Kettering-integrated mutation profiling of actionable of Hematology/Oncology, University of California, San Diego, Moores Cancer cancer targets (MSK-IMPACT): a hybridization capture-based next-generation Center, 3855 Health Sciences Drive, #0658, La Jolla, CA 92093-0987, USA. sequencing clinical assay for solid tumor molecular oncology. J Mol Diagn. 2 3 Tokyo Medical and Dental University, Tokyo, Japan. Department of Clinical 2015;17:251–64. Oncology, Hospital Sirio Libanes, Sao Paulo, Brazil. 21. Zehir A, Benayed R, Shah RH, Syed A, Middha S, Kim HR, et al. Mutational landscape of metastatic cancer revealed from prospective clinical sequencing Received: 1 March 2018 Accepted: 26 April 2018 of 10,000 patients. Nat Med. 2017;23:703–13. 22. Graham TJ, Box G, Tunariu N, Crespo M, Spinks TJ, Miranda S, et al. Preclinical evaluation of imaging biomarkers for prostate cancer bone metastasis and response to cabozantinib. J Natl Cancer Inst. 2014;106:dju033. References 1. Cooper CS,ParkM,Blair DG, Tainsky MA,Huebner K, CroceCM, etal. 23. McKay RR, Kroeger N, Xie W, Lee J-L, Knox JJ, Bjarnason GA, et al. Impact of Molecular cloning of a new transforming gene from a chemically transformed bone and liver metastases on patients with renal cell carcinoma treated with human cell line. Nature. 1984;311:b29–33. targeted therapy. Eur Urol. 2014;65:577–84. 2. Park M, Dean M, Kaul K, Braun MJ, Gonda MA, Vande WG. Sequence of MET 24. Smith M, De Bono J, Sternberg C, Le Moulec S, Oudard S, De Giorgi U, et al. protooncogene cDNA has features characteristic of the tyrosine kinase family Phase III study of cabozantinib in previously treated metastatic castration- of growth-factor receptors. Proc Natl Acad Sci U S A. 1987;84:6379–83. resistant prostate cancer: COMET 1. J Clin Oncol. 2016;34:3005–13. Ikeda et al. Journal of Hematology & Oncology (2018) 11:76 Page 10 of 10 25. Choueiri TK, Escudier B, Powles T, Tannir NM, Mainwaring PN, Rini BI, et al. Cabozantinib versus everolimus in advanced renal cell carcinoma (METEOR): final results from a randomised, open-label, phase 3 trial. Lancet Oncol. 2016;17:917–27. 26. Mukai S, Yorita K, Kawagoe Y, Katayama Y, Nakahara K, Kamibeppu T, et al. Matriptase and MET are prominently expressed at the site of bone metastasis in renal cell carcinoma: immunohistochemical analysis. Hum Cell. 2015;28:44–50. 27. Pierscianek D, Kim YH, Motomura K, Mittelbronn M, Paulus W, Brokinkel B, et al. MET gain in diffuse astrocytomas is associated with poorer outcome. Brain Pathol. 2013;23:13–8. 28. de Melo Gagliato D, Fontes Jardim DL, Falchook G, Tang C, Zinner R, et al. Analysis of MET genetic aberrations in patients with breast cancer at MD Anderson Phase I Unit. Clin Breast Cancer. 2014;14:468–74. 29. Jardim DLF, de Melo Gagliato D, Falchook G, Zinner R, Wheler JJ, Janku F, et al. MET abnormalities in patients with genitourinary malignancies and outcomes with c-MET inhibitors. Clin Genitourin Cancer. 2015;13:e19–26. 30. Tang C, Fontes Jardim DL, Falchook GS, Hess K, Fu S, Wheler JJ, et al. MET nucleotide variations and amplification in advanced ovarian cancer: characteristics and outcomes with c-Met inhibitors. Oncoscience. 2013;1:5–13. 31. Jardim DLF, Gagliato D de M, Falchook GS, Janku F, Zinner R, Wheler JJ, et al. MET aberrations and c-MET inhibitors in patients with gastric and esophageal cancers in a phase I unit. Oncotarget. 2014;5:1837–45.
Journal of Hematology & Oncology
– Springer Journals
Published: Jun 4, 2018