Purpose: Molecular profiling may have prognostic and predictive value, and is increasingly used in the clinical setting. There are more than a dozen fibroblast growth factor receptor (FGFR) inhibitors in development. Optimal therapeutic application of FGFR inhibitors requires knowledge of the rates and types of FGFR aberrations in a variety of cancer types.

Experimental Design: We analyzed frequencies of FGFR aberrations in 4,853 solid tumors that were, on physician request, tested in a Clinical Laboratory Improvement Amendments (CLIA) laboratory (Foundation Medicine) using next-generation sequencing (182 or 236 genes), and analyzed by N-of-One.

Results: FGFR aberrations were found in 7.1% of cancers, with the majority being gene amplification (66% of the aberrations), followed by mutations (26%) and rearrangements (8%). FGFR1 (mostly amplification) was affected in 3.5% of 4,853 patients; FGFR2 in 1.5%; FGFR3 in 2.0%; and FGFR4 in 0.5%. Almost every type of malignancy examined showed some patients with FGFR aberrations, but the cancers most commonly affected were urothelial (32% FGFR-aberrant); breast (18%); endometrial (∼13%), squamous lung cancers (∼13%), and ovarian cancer (∼9%). Among 35 unique FGFR mutations seen in this dataset, all but two are found in COSMIC. Seventeen of the 35 are known to be activating, and 11 are transforming.

Conclusions: FGFR aberrations are common in a wide variety of cancers, with the majority being gene amplifications or activating mutations. These data suggest that FGFR inhibition could be an important therapeutic option across multiple tumor types. Clin Cancer Res; 22(1); 259–67. ©2015 AACR.

Translational Relevance

Cancer is fundamentally a disease of disordered genes. The paradigm of precision oncology hypothesizes that we understand the genetic abnormalities that drive cancer, that drugs successfully target the products of these abnormal genes, and that we can detect abnormal genes in individual patients. Considering cancer treatment development more broadly, the challenge lies in defining population(s) for which new therapeutics will be most effective. Currently, there are more than a dozen anti-fibroblast growth factor receptor (FGFR) drugs in development for cancer, but which patients or patient populations will benefit most from these drugs? To facilitate answering this question, we present an analysis of next-generation sequencing data from a very large database of nearly 5,000 cancers of diverse histologies. Our data provide robust evidence of frequencies and characteristics of FGFR aberrancies in cancer. These data will aid design of studies to further define the role of FGFR inhibitors in cancer.

Fibroblast growth factor receptors (FGFRs) are highly conserved, widely distributed transmembrane tyrosine kinase receptors. They are involved in development, differentiation, cell survival, migration, angiogenesis, and carcinogenesis (1). In humans, there are four such FGFRs that are typical tyrosine kinase receptors (FGFR1-4), and one that lacks an intracellular tyrosine kinase domain (FGFRL1 or FGFR5). There are also 18 human ligands for FGFRs, which are known as fibroblast growth factors (1). When FGFs bind to their cognate receptors, the receptors dimerize, leading to intracellular phosphorylation of receptor kinase domains, a cascade of intracellular signaling, and gene transcription (2). FGF/FGFRs signal through several intracellular pathways, including the Ras/Raf/MEK and the phosphatidylinositol-4,5-bisphosphate 3-kinase (PI3K)–Akt pathway (1). All four FGFRs share structural homology with vascular endothelial growth factor receptors (VEGFR), platelet-derived growth factor receptors (PDGFR), and other tyrosine kinase receptors, which has implications for pharmacologic therapy (2).

Specific FGFR aberrations have been observed in a proportion of certain cancers [e.g., FGFR3 mutations in bladder cancer (3) and FGFR1 amplification in squamous cell lung cancer (4)]. Some of these FGFR abnormalities are likely to be “driver” aberrations. There is also evidence that changes in specific FGFR expression may be related to prognosis or sensitivity to cancer treatments (5–7).

Because the majority of FGFR aberrations identified to date lead to gain-of-function, it is reasonable to hypothesize that targeting these cancers with FGFR inhibitors would be therapeutically beneficial (8). In vitro data suggest that this is indeed the case (9). Several tyrosine kinase inhibitors (TKI) have been identified as FGFR inhibitors, including ponatinib (AP24534), regorafenib (BAY 73-4506), lenvatinib (E7080), dovitinib (TKI258), lucitanib (E3810), nintedanib (BIBF 1120), and others. Some FGFR inhibitors also suppress VEGFRs and additional growth factor receptors, whereas others are more selective for FGFR inhibition (e.g., NVP-BGJ398, AZD4547, JNJ-42756493, etc.). At this time, four FGFR inhibitors are FDA approved for treatment of cancer. The most recently FDA-approved FGFR-inhibiting drug is lenvatinib, which is approved for iodine-refractory, well-differentiated thyroid carcinoma. Other FDA-approved FGFR-inhibiting drugs include regorafenib, approved for advanced colorectal carcinoma and drug-resistant gastrointestinal stromal tumors (GIST), ponatinib, approved for drug-resistant chronic myelogenous leukemia (CML) and Philadelphia chromosome-positive acute lymphocytic leukemia (ALL), and pazopanib, approved for renal cell carcinoma and sarcoma. None of these are FDA approved on the basis of targeting FGFR (or any other molecular phenotype). The hypothesis that selecting for FGF/FGFR aberration might increase response rates or other clinical benefits is being tested in several ongoing trials that require FGF/FGFR aberrations for eligibility. These and other clinical trials will shed light on the specific patient populations that would benefit from FGFR-inhibiting drugs and possibly on specific molecular aberrations that predict response to these drugs.

In fact, it is highly likely that optimizing the clinical utility of FGFR-targeting therapies will depend on appropriate selection of patient populations. To that end, developing a clear understanding of the landscape of FGFR aberrations in various cancer types is relevant and necessary for more effective use of these pharmaceutical agents. Next-generation sequencing technology makes rapid and accurate identification of these aberrations feasible. Herein, we describe the landscape of FGFR abnormalities, including mutations, amplifications, and rearrangements in 4,853 patient samples from diverse cancers.

We collected sequencing data from 4,853 cancers of various types (specific cancer types and numbers of cases are listed in Supplementary Tables S2 and S3) from patients whose formalin-fixed, paraffin-embedded (FFPE) tumor samples were submitted to a CLIA-certified lab for genomic profiling (Foundation Medicine). Samples were required to have a surface area ≥25 mm2, volume ≥1 mm3, nucleated cellularity ≥80%, and tumor content ≥20% (10). The methods used in this assay have been validated and previously reported (10–12). Briefly, 50 to 200 ng of genomic DNA was extracted and purified from the submitted FFPE tumor samples. This whole-genome DNA was subjected to shotgun library construction and hybridization-based capture before paired-end sequencing on the Illumina HiSeq2000 platform. Hybridization selection is performed using individually synthesized baits targeting the exons of 182 or 236 cancer-related genes and the introns of 14 or 19 genes frequently rearranged in cancer (Supplementary Table S1). The samples collected for this study were assayed between December 16, 2011 and November 14, 2013. Sequence data were processed using a customized analysis pipeline (10). Sequencing was performed with an average sequencing depth of coverage greater than 250×, with >100× at >99% of exons. This method of sequencing allows for detection of copy number alterations, gene rearrangements, and somatic mutations with 99% specificity, and >99% sensitivity for base substitutions at ≥5 mutant allele frequency and >95% sensitivity for copy number alterations. A threshold of ≥6 copies for gene amplification (except for ERRB2, which is considered amplified with ≥5 copies) was used. The submitting physicians provided specification of tumor types. The database was de-identified with only diagnosis available. Next-generation sequencing data were collected and interpreted by N-of-One. For this study, the dataset of 4,853 sequenced tumors was queried for alterations in FGFR1-4 and coaberrant genes. Data were analyzed in accordance with UCSD IRB guidelines. Here, we report on the prevalence and frequencies of these aberrations in human cancers.

Of the 4,853 cancers sequenced, we observed 360 FGFR aberrations in 343 cases (17 cancers had more than one FGFR alteration), for an overall frequency of 7.1%. FGFR1 alterations were more common than alterations in FGFR2–4 (Figs. 1 and 2). The majority of the FGFR aberrations were gene amplifications (66% of 360 FGFR aberrations), with gene mutations being less common (26%) and gene rearrangements rare (8%; Fig. 2). These proportions were similar across all four FGFRs (Supplementary Figs. S1–S4); however, FGFR1 and FGFR4 showed high rates of gene amplification (89% and 78% of all FGFR1 or FGFR4 alterations, respectively; Supplementary Figs. S1A and S4) and FGFR2 and FGFR3 had relatively more frequent gene rearrangements (16% and 19%, respectively; Supplementary Figs. S2A and S3A). The percentages of 343 patients with an aberrant FGFR that had any anomaly in FGFR1, FGFR2, FGFR3, and FGFR4 were 49, 19, 26 and 7, respectively (Fig. 1).

Figure 1.

All FGFR aberrations. Frequency of FGFR aberrations among all cases (left) and relative proportion of FGFR aberrations by FGFR gene among all cases with FGFR alterations (right). Sixteen cases had more than one aberration, so the total of the right chart is more than 100%.

Figure 1.

All FGFR aberrations. Frequency of FGFR aberrations among all cases (left) and relative proportion of FGFR aberrations by FGFR gene among all cases with FGFR alterations (right). Sixteen cases had more than one aberration, so the total of the right chart is more than 100%.

Close modal
Figure 2.

Frequencies and distributions of FGFR aberrations for all cancers with ≥75 cases analyzed. Within each cancer, the frequency of FGFR aberrations is reported as percentage of all cases of that cancer analyzed. The distributions of FGF receptors altered and types of alterations are normalized to 100% for each cancer type. See Supplementary Table S2 for additional information.

Figure 2.

Frequencies and distributions of FGFR aberrations for all cancers with ≥75 cases analyzed. Within each cancer, the frequency of FGFR aberrations is reported as percentage of all cases of that cancer analyzed. The distributions of FGF receptors altered and types of alterations are normalized to 100% for each cancer type. See Supplementary Table S2 for additional information.

Close modal

Frequencies of aberrations and relative distribution of types of aberration for histologies with ≥75 cases are shown in Fig. 2 and Supplementary Table S2. A summary of cases with FGFR aberration(s) in cancer types with fewer than 75 cases is presented in Supplementary Table S3. As expected, some cancer types had a higher frequency of FGFR alteration than others, and are discussed in greater detail below. It should be noted that no clinical data about the study population is available other than the submitting physician's indication of tumor type. For some tumor types, for example, urothelial carcinoma, frequencies, or types of FGFR aberrations may depend upon grade and/or stage of cancer. For instance, FGFR3 mutations are seen in >50% of bladder cancer cases with low-grade, noninvasive disease, but drop in frequency, once one looks at higher grade/stage (13). Because we do not have this information for our dataset, we are not able to provide analysis of this issue. For more information, see section “Urothelial Cancers.”

Non–small cell lung cancer

There were 675 cases of non–small cell lung cancer (NSCLC) in the dataset (Figs. 2 and 3, Supplementary Tables S2 and S6). There was a marked difference between squamous cell histology (N = 93), adenocarcinoma (N = 408), and other non–small cell types (e.g., large cell carcinoma). In particular, squamous cell lung carcinoma was most notable for its 9% frequency of FGFR1 amplification, which is in contrast to only 4% of lung adenocarcinomas harboring any FGFR abnormality (Fig. 3). Interestingly, 3% of squamous cell lung carcinomas had somatic FGFR2 and FGFR3 mutations identical to those seen in inherited dwarfism syndromes (14), including FGFR2P253R, FGFR2S252W, FGFR3G370C, FGFR3K650E, FGFR3R248C, and FGFR3S249C. See below for a discussion of the functional significance of these activating mutations.

Figure 3.

Relative frequencies of FGFR aberrations in non–small cell lung carcinoma. A, lung squamous cell carcinoma, 93 cases. Frequencies are reported as percentage of all 93 cases. B, lung adenocarcinoma, 408 cases. Frequencies are reported as percentage of all 408 cases. Non–small cell lung carcinomas not otherwise specified were excluded from this analysis.

Figure 3.

Relative frequencies of FGFR aberrations in non–small cell lung carcinoma. A, lung squamous cell carcinoma, 93 cases. Frequencies are reported as percentage of all 93 cases. B, lung adenocarcinoma, 408 cases. Frequencies are reported as percentage of all 408 cases. Non–small cell lung carcinomas not otherwise specified were excluded from this analysis.

Close modal

Urothelial cancers

There were 126 cases of urothelial cancers in the dataset (Figs. 2 and 4, Supplementary Tables S2 and S6), and urothelial (transitional cell) cancer of the bladder, renal pelvis, and ureter were represented. This dataset does not include cases of bladder sarcoma, small cell carcinoma, squamous cell carcinoma, or neuroendocrine carcinoma. In urothelial tumors, 15% of aberrations were somatic mutations in FGFR3 that are known to be activating (Supplementary Table S4). Another 7% of urothelial cancers had FGFR1 amplifications, 6% had gene fusions, and 3% had FGFR3 amplifications.

Figure 4.

Distribution of FGFR aberrancies in urothelial cancers. Cancers included urothelial carcinomas (transitional cell carcinomas) of the bladder, renal pelvis, ureter, and not specified. The majority of aberrations were activating mutations in FGFR3, including S249C (8 instances), R248C (6 instances), Y373C (2 instances), G370C (2 instances), and K650M (1 instance). Three of these FGFR3 mutations are also about to transform cells in vitro (S249C, S248C, Y737C; Supplementary Table S4). Frequencies are expressed as percentages of all 126 cases. There were 44 aberrations in 40 cases (4 cases had more than one aberration), so the total is greater than 100%.

Figure 4.

Distribution of FGFR aberrancies in urothelial cancers. Cancers included urothelial carcinomas (transitional cell carcinomas) of the bladder, renal pelvis, ureter, and not specified. The majority of aberrations were activating mutations in FGFR3, including S249C (8 instances), R248C (6 instances), Y373C (2 instances), G370C (2 instances), and K650M (1 instance). Three of these FGFR3 mutations are also about to transform cells in vitro (S249C, S248C, Y737C; Supplementary Table S4). Frequencies are expressed as percentages of all 126 cases. There were 44 aberrations in 40 cases (4 cases had more than one aberration), so the total is greater than 100%.

Close modal

Among the seven urothelial cases with gene fusions, six were fusions with FGFR3 and one was with FGFR1. The most common fusion was FGFR3–FGFR3–TACC3 (4 cases, 3%), which results from 4p16.3 rearrangements. The TACC3 gene is located within 48 kb of FGFR3 on 4p16.3, so this spatial proximity may support recombination. See below for discussions of these aberrations. All other FGFR gene fusions are listed in Supplementary Table S5.

The high prevalence of FGFR gene abnormalities in urothelial carcinomas not only suggests that anti-FGFR therapies may be effective for these patients, but also raises the question of whether there are coaberrant genes that could also be targeted by additional therapies. One such gene of interest is PIK3CA. The overall frequency of PIK3CA mutation among urothelial tumors was 20.6% (25 cases, 1 case had two mutations). Among the 32 urothelial tumors with FGFR3 abnormalities, 6 (24%) had coexisting PIK3CA gene abnormalities, suggesting that combination therapy with anti-FGFR and anti-PIK3CA drugs could be evaluated. The frequencies of aberration in these two genes is likely an independent occurrence (χ2P value = 0.86 in this dataset), which is in contrast to two other published studies (15, 16) in which PIK3CA mutation was positively correlated with FGFR aberration. However, it should be noted that in the first study, 77% were stage Ta/T1 and 57% were grade G1/G2, and in the second study 75% were stage Ta/T1 and grade G1/G2, and both studies reported a stronger correlation between PIK3CA and FGFR abnormalities in earlier stage and lower grade tumors. Other genes of interest that were coaberrant with FGFR3 amplification include CDKN2A (8 cases), TSC1 (5 cases), ARID1A (5 cases), and TP53 (4 cases). To facilitate exploration of coaberrant genes, we listed all urothelial and other tumor types from our dataset that had any FGFR aberration and all coexisting gene aberrations in Supplementary Table S6.

We grouped all urothelial carcinomas together for this analysis, although it is possible that there are differences in molecular phenotype according to where in the genitourinary tract the urothelial tumors arise. Among the 126 urothelial cancers we evaluated, 22 of 90 bladder carcinomas, 11 of 21 renal pelvis carcinomas, 3 of 6 ureteral carcinomas, and 4 of 9 urothelial carcinomas not otherwise specified (NOS) had FGFR aberrations. Although these data may suggest that FGFR gene aberrations are least frequent among urothelial carcinomas arising from the lower urinary tract, our dataset is not equipped to make this determination because of small numbers of patients in certain subsets. To avoid sample size bias, we chose to analyze only those tumor types with at least 75 representative cases, and bladder is the only site in the urothelial tract that meets this requirement with 90 cases (renal pelvis had 21, urothelial not otherwise specified 9, and ureter 6 cases).

Among urothelial tumors, FGFR3 mutations are very frequent among tumors of low-malignant potential, papillomas, low-grade, and low-stage tumors. Di Martino and colleagues showed that the most common FGFR3 mutations seen in urothelial cancers are able to transform NIH-3T3 cells, but have less potent effects on normal bladder cells (TERT-NHUC) (13). These data suggest that FGFR3 mutations may confer a selective proliferative advantage to early urothelial lesions, but that they may also have cell-type–specific effects that may explain the observed differences in mutation frequencies among urothelial tumors.

In our dataset, we do not have grade or stage information for any of the tumor samples, including the urothelial tumors. This means that we cannot say whether they are superficial or invasive nor whether they are low- or high-grade tumors. We therefore cannot draw conclusions about the significance of FGFR aberration frequencies in this dataset.

Breast cancer

There were 522 cases of breast cancer in the dataset (Fig. 2, Supplementary Tables S2 and S6), and included invasive ductal carcinoma, invasive lobular carcinoma, and invasive metaplastic carcinoma. Breast sarcoma, neuroendocrine breast cancer, and noninvasive breast cancer are not included. Estrogen receptor and progesterone receptor protein expression are not measured by the NGS assay used. In contrast, ERBB2 (Her2) amplification of ≥5-fold is detected by the assay. Only 4 of 72 breast cancer cases with any FGFR aberrations also had ERBB2 amplification measured in this fashion.

Eighteen percent of breast cancers analyzed had any FGFR aberration, the most common of which was FGFR1 amplification (∼14%), whereas amplification of FGFR2–4 was much less frequent (0.5%–2.3%). Because PIK3CA alterations are among the most commonly seen in breast cancer (17–19), it is interesting to note that 26.4% (19/72) of cases with FGFR1 amplification also harbored aberrations in the PIK3CA gene, nearly all of which are activating alterations: gene amplification (N = 2), PIK3CA–E545K (N = 7), PIK3CA–H1047R (N = 7), and one case each with PIK3CA–N345K, PIK3CA–E542K, PIK3CA–E545Q, PIK3CA–Q546K, and PIK3CA–M1043L (likely an activating mutation). The total number of PIK3CA aberrations listed is greater than 19 because one case had three distinct aberrations. This overall frequency of ∼30% is similar to what we observed in the entire set of 522 breast cancer cases (28.9%) and to the reported rates of PIK3CA mutations in breast cancer not selected for FGFR aberration, which range from 22% to 34.5% for hormone receptor positive and Her2-positive breast cancers (17–19), perhaps suggesting that there is no relationship to FGFR aberration. In fact, the χ2P value is 0.61, so it is very likely that these two genes are independently selected for in the breast cancer cases analyzed.

Other genes of interest that were coaberrant in the FGFR1-amplified subset of breast cancers include CCND1 amplification (21 cases), MYC amplification (21 cases), and mutations or loss of TP53 (31 cases). See Supplementary Table S6 for a list all co-aberrant genes for all cases with any FGFR abnormality.

Endometrial carcinoma

About 11% of 80 cases harbored FGFR abnormalities, most of which occurred in FGFR2 (Fig. 2, Supplementary Tables S2 and S6). These FGFR2 mutations included S252W (2 cases), P253R (2 cases), C382R and N549K (1 case each). All of these are gain-of-function mutations that are able to transform cells in vitro (Supplementary Table S4). Their functional significance is discussed below.

Non–lung squamous cell carcinomas

All non–lung squamous cell carcinomas were analyzed together. There were 273 cases, including esophageal, bladder, cervical, cutaneous, gallbladder, head and neck, ocular, penile, vulvar, vaginal, urethral, and rectal carcinomas; 5.1% of these cases harbored any FGFR aberration, and the majority (6 of 14 aberrations) were FGFR1 amplifications.

Other cancers

A wide variety of other cancers harbored FGFR aberrations in a subset of patients, ranging from 1% to 9% (Supplementary Tables S2 and S6). These cancers include, but are not limited to, ovarian cancer (∼9%), gliomas (∼8%), pancreatic, renal, colorectal cancer, and neuroendocrine (about 4% to 5% each), and sarcomas (∼4%).

Specific Aberrations and Their Functional Significance: Preclinical Work and Implications for Drug Development

FGFR1 amplification

FGFR1 amplification is one of the most common FGFR alterations seen in this dataset (Supplementary Fig. S1A). It was observed in 151 cases (3% of all cases, or about 42% of all observed FGFR aberrations). It was common in breast carcinoma (∼14% of patients with breast cancer), squamous cell lung carcinoma (∼9%), and ovarian carcinoma (∼5%), but was also seen in significant proportions of urothelial carcinoma (7% of cases; 20% of FGFR aberrations), gastric/gastroesophageal junction carcinoma (∼2% of cases; 25% of FGFR aberrations), colorectal carcinoma (2% of cases; 64% of FGFR aberrations), carcinoma of unknown primary (2% of cases; 23% of FGFR aberrations), and squamous non-lung tumors (2% of cases, 43% of FGFR aberrations). Presumably, FGFR1 (and other FGFR gene) amplification leads to protein overexpression and dependence on FGFR signaling. This assumption is borne out in preclinical models of squamous cell lung carcinoma in which FGFR1 amplification correlates with protein overexpression and increased sensitivity to FGFR-inhibiting drugs (20, 21). Similar findings are seen in breast cancer preclinical models for both FGFR1 (9, 22) and FGFR2 (23). These in vitro data suggest that FGFR1 and FGFR2 amplification could serve as biomarkers for efficacy of FGFR inhibiting drugs.

FGFR mutations

In our dataset, there were five unique FGFR1 mutations (Supplementary Table S4). All five have been reported previously in the COSMIC database (Catalogue of Somatic Mutations In Cancer, http://cancer.sanger.ac.uk/cosmic, accessed June 2015). The functional effects of three of them are unknown, but two of them (N546K and K656E) are known to be both activating and transforming. Both lie in the intracellular kinase domain. Lew and colleagues (24) showed that formation of the Fgfr1 monophosphorylated receptor is 25 times faster with N546K than with wild-type and that the mutant receptor is capable of transforming Rat-1 cells in vitro. The FGFR1 K656E mutation causes constitutively active receptor signaling in an analogous mutation in FGFR3 (25). This activating mutation in FGFR1 not only induces phosphorylation of downstream effectors, but is also capable of transforming NIH3T3 cells in vitro (25). These data suggest that both of these mutations are likely pathogenic in vivo and represent valid targets for drug development.

FGFR2 has a higher missense mutation rate in our dataset (12 unique mutations, all but one of which are reported in the COSMIC database; see Supplementary Table S4). Seven are known to be activating mutations. FGFR2 S252W, P253R, and N549K were the most commonly seen FGFR2 alterations. FGFR2 S252W and P253R lie in the receptor's extracellular linker region between the two immunoglobulin-like domains, a key site for ligand binding (26), and are thought to differentially increase ligand binding affinity, thereby increasing receptor signaling (27). Both are also capable of transforming NIH3T3 cells despite the fact that the mutant receptor is expressed at lower levels than the wild-type (26). Furthermore, knockdown of the S252W mutant receptor by specific shRNA inhibits both transformation and survival of MFE-280 cells in vivo (26), strongly suggesting that the FGFR2 S252W mutation and possibly also the P253R mutation are compelling targets for drug therapy. The FGFR2 N549 residue is associated with a “molecular brake” that keeps the kinase in an auto-inhibited state (28). The N549K mutation presumably disrupts this inhibition, leading to increased kinase activity. It also transforms NIH3T3 cells (26). Among the other FGFR2 mutations known to be activating (A315T, Y375C, C382R, and K659E), only C382R and K659E are known to transform NIH3T3 cells (26, 29). We are unaware of data regarding the transformational ability of the other FGFR2 mutations in our dataset.

FGFR3 also had a high rate of mutation, with 13 unique mutations identified in the dataset (Supplementary Table S4). All but one have been reported in the COSMIC database. Eight of them are known to be activating and four of them are able to transform cells in vitro. The most common missense mutations in FGFR3 were S249C (17 cases), R248C (9 cases), G370C (4 cases), K650E (4 cases), R399C (3 cases), and Y373C (3 cases). All other mutations were observed in single cases. The FGFR3 S249C mutation is both activating and transforming. It lies between the two extracellular immunoglobulin-like domains. In 293T cells, FGFR3 S249C induces ligand-independent dimerization and increased receptor basal phosphorylation (30) and leads to anchorage independent growth (31) and xenograft tumors in mice (32). Furthermore, gene knockdown of this mutant receptor abolishes transformation (33). The nearby FGFR3 R248C mutation, which is the second most common FGFR3 mutation in our dataset, is similarly activating and transforming. For both of these mutations, the creation of a new, unpaired cysteine residue results in formation of interreceptor disulfide bones, increased homodimerization and signaling (34). In vitro, FGFR3 R248C promotes increased cell numbers at confluence, induces proliferation, induces morphologic transformation, reduces apoptosis, and decreases attachment to fibronectin, but does not alter migration (34, 35). FGFR3 G370C lies in the extracellular juxtamembrane region. In 293T cells, it leads to ligand-independent dimerization and phosphorylation (30). We are unaware of data regarding the ability of this mutation to transform cells in vitro. FGFR K650E also shows ligand independent activation, although by undefined mechanism(s) (35, 36) and is able to transform NIH3T3 cells (37, 38). FGFR3 Y373C is also thought to induce disulfide bond formation causing constitutive activation of the receptor (32, 35). It is a strong inducer of transformation, which can be abrogated by siRNA-mediated knockdown or SU5402 (a potent FGFR inhibitor; refs. 32, 37), suggesting that this mutation represents a valid drug target. We are unaware of data regarding functional significance or transformational ability of FGFR3 R399C.

Among the five unique mutations observed in FGFR4, all were previously reported in the COSMIC database, but to our knowledge none of them have been characterized to date.

FGFR gene fusions

Fusions of FGFR genes with other genes or parts of genes were observed mostly with FGFR2 (10 cases) and FGFR3 (18 cases). By far, the most common fusion partner was TACC3 (Transforming Acidic Coiled-Coil Containing Protein 3; 12 cases). Other fusion partners included three cases with NPM1, two with TACC2, two with BICC1, and single cases with NTM, C10orf68, KIAA1598, NCALD, NOL4, PPAPDC1A, JAKMIP1, TNIP2, and WHSC1. Four of our cases that had gene fusions were urothelial carcinomas, two were glioblastomas, and the rest were single cases of cholangiocarcinoma, cervical adenocarcinoma, cervical squamous cell carcinoma, endometrial carcinoma, non–small cell lung carcinoma, pancreatic carcinoma, gallbladder carcinoma, renal cell carcinoma, and carcinoma of unknown primary. All gene fusions are listed in Supplementary Table S5.

Chromosomal translocations in cancers that lead to fusion proteins exert their oncogenic effects through overexpression of an otherwise normal gene or creation of a chimeric gene in which parts of two genes are fused together. In the case of FGFR3–TACC3, the entire FGFR3 kinase domain is fused with the TACC3 domain that mediates microtubule binding (31, 39). These fusion proteins activate the MAPK pathway when transfected into normal human urothelial cells, suggesting that they retain active signaling. Furthermore, cell lines harboring the fusion proteins are very sensitive to a selective FGFR inhibitor (PD173074), indicating that the fusion protein represents a valid therapeutic target in cancer cells (31). Similar FGFRTACC3 fusions that are also sensitive to PD173074 have been reported in glioblastoma (39).

Coexistent FGFR mutation and amplifications

Ten of the 17 tumors that had more than one FGFR gene aberration had amplifications concurrent with either mutation or fusion events, 8 of them involved FGFR3, and 2 involved FGFR2. The tumor types involved were urothelial carcinoma (N = 3), endometrial carcinoma (N = 2), and single cases of cervical carcinoma, gallbladder carcinoma, non–small cell lung carcinoma, pancreatic exocrine carcinoma, and renal cell carcinoma. All concurrent aberrations are listed in Supplementary Table S6.

This study represents a comprehensive overview of FGFR aberrations in a large cancer genomic database. About 7% of cancers had FGFR aberrations, with the most common abnormality being FGFR1 amplification. Overall, 5% of 4,853 patients had FGFR amplifications; 2% of patients had mutations; and 0.5% of patients had rearrangements. FGFR1 was affected in 3.5% of 4,853 patients; FGFR2 was affected in 1.5% of patients; FGFR3 was affected in 2.0% of patients, and FGFR4 was affected in 0.5% of patients (Fig. 1). Almost every histology included individuals who harbored FGFR aberrations, but the cancers most commonly affected were urothelial (32% FGFR-aberrant), breast (18%), endometrial (∼13%), squamous cell lung (∼13%; Fig. 3), ovarian (∼9%), carcinoma of unknown primary (∼8%), glioma (∼89%), and cholangiocarcinoma (7%; Fig. 2 and Supplementary Tables S2 and S6).

FGFR aberrations did not appear to segregate well by histology. However, some aberrations were found more frequently in certain cancers. For example, FGFR1 amplifications predominated in squamous cell lung, breast, ovarian, and urothelial cancers, observed in 5% to 14% of patients with these malignancies; FGFR3 mutations predominated in bladder and other urothelial tumors, observed in 15% of individuals. Others (20, 21, 40, 41) also reported high rates of FGFR1 amplification in squamous cell lung cancer (13%–22%). Squamous cell cancers originating in other organs were analyzed together and showed FGFR aberrations in 5.1% of cases (most frequently FGFR1 amplification). There were insufficient small cell lung cancers (43 cases) in our dataset to report.

Although therapies targeting the aberrant proteins produced by mutated EGFR or rearranged ALK have been applied successfully in lung adenocarcinoma and the FDA recently approved nivolumab for squamous cell lung carcinoma, no therapy based on molecular phenotype is currently approved for squamous cell or other non-adenocarcinoma types of lung cancers. However, FGFR inhibitors are being developed for NSCLC, including squamous cell carcinomas. For example, the results of at least two phase III clinical trials of the multikinase inhibitor nintedanib (which targets FGFR, VEGFR, and PDGFR) in NSCLC showed statistically significant, albeit modest, benefit (42, 43). These studies did not select for FGFR aberrations, so it would be of interest to determine the correlation between response and the presence of FGFR abnormalities.

Fifteen percent of urothelial malignancies (Fig. 4) also harbored somatic mutations in FGFR3, which are known to be activating and transforming. Because these activating mutations are easy to detect and are frequent, they represent attractive targets for drug development. There are several ongoing trials of FGFR inhibiting drugs in urothelial carcinoma, some of which are reporting early success. For example, preliminary analysis of a phase I trial of BGJ398, a potent, selective pan-FGFR inhibitor, showed tumor regression in four of five patients with urothelial carcinomas with FGFR3-activating mutations (with tumor reductions ranging from 27% to 48%; ref. 44).

Eighteen percent of breast cancers had an FGFR aberration, the most frequent being FGFR1 amplification (14% of cases), whereas amplification of FGFR2–4 was much less common (0.5–2.3%). FGFR1 amplification may be a strong independent predictor of overall survival in patients with breast cancer (45) and may also correlate with endocrine therapy resistance (6), suggesting prognostic value for assessing FGFR1 status. It is also possible that patients with FGFR1-amplified breast cancer might benefit from the administration of FGFR inhibitors. In fact, lucitanib, a dual kinase inhibitor (FGFR/VEGFR), has shown activity in FGFR1-amplified breast cancer, with an overall response rate of 50% (46), and other studies are ongoing.

About 11% of endometrial cancers were found to have FGFR abnormalities, mostly activating mutations in or amplification of FGFR2 (∼9% of patients). As described in the “Results,” the FGFR2–S252W mutation increases affinity for the FGF ligands, in particular FGF9, which may be especially important for endometrial cancer because it is found in abundance in the endometrial stroma (47). Also intriguing is that FGFR2 and KRAS mutations seem to be mutually exclusive in endometrial cancers, suggesting redundancy with regard to activation of the MAPK pathway (48).

We also noted FGFR gene fusions (usually involving FGFR2 or FGFR3) in a minority of the cases. The most common fusion partner was TACC3 (31), perhaps because FGFR3 and TACC3 are close together on chromosome 4p16 (39). Fusion of FGFR3 with TACC3 leads to ligand-independent signaling activation in glioblastoma and bladder cancer (31, 39, 49). Mice-harboring FGFR-TACC3–associated gliomas respond to administration of an FGFR inhibitor (39).

Stratifying by type of abnormality, FGFR1 amplification was one of the most common FGFR anomalies observed (Supplementary Fig. S1A; 3% of all cases, or approximately 42% of FGFR aberrations). FGFR1 amplification was frequent in breast carcinoma (∼14% of cases), squamous cell lung carcinoma (∼9%), urothelial carcinoma (∼7%), and ovarian carcinoma (∼5%), as well as other malignancies. Of interest, at this time there are at least five ongoing clinical trials of FGFR inhibitors that include FGFR1 amplification in their eligibility criteria (NCT01948141, NCT01283945, NCT01349296, NCT01202591, NCT02053636, see clinicaltrials.gov).

At this time, there are four FGFR inhibiting drugs approved by the FDA: ponatinib, regorafenib, pazopanib, and most recently lenvatinib. All are multikinase inhibitors, but there are also specific FGFR inhibitors in development as well. None of the approved FGFR inhibiting drugs were approved specifically for FGFR-selected populations, but several FGFR inhibiting agents are currently in clinical trials that require FGF/FGFR aberrations for eligibility. For example, dovitinib (TKI-258), a potent multikinase inhibitor (FGFR, VEGFR, and PDGFR), is being used in phase II trials for FGFR1-amplified squamous non–small cell lung cancer (NCT01861197), FGFR1- or FGFR2-amplified breast cancer (NCT01528345), and refractory urothelial carcinoma with FGFR3 mutations or overexpression (NCT01732107). Lucitanib (E-3810), a multikinase inhibitor (FGFR, VEGFR, and PDGFR), is being tested in a phase I/IIa trial in patients with solid tumors (NCT01283945), phase II studies of FGFR1-amplified lung cancer (NCT02109016) and FGFR1-amplified breast cancer (NCT02202746 and NCT02053636). Results of these and other trials will clarify the utility and safety of FGFR-inhibiting drugs, the advantages or disadvantages of drug specificity for FGFR, and refine appropriate biomarkers for response to these drugs.

There are some limitations to these data. First, the dataset was not annotated and therefore correlation with clinical characteristics (e.g., stage, phenotype, etc.) was not possible, which may have greater importance for some tumor types than for others (see “Urothelial cancers”). Second, the number of patients with each cancer was dependent on the number of cases submitted by physicians for next-generation sequencing analysis, which introduces the possibility of sample size bias. Finally, pathologic diagnosis was designated based on the determination of the submitting attending physician/pathologist.

Our observed frequency of primarily activating FGFR aberrations in diverse cancers, along with preclinical and early clinical data already reported suggest that targeting FGFR alterations with cognate inhibitors has therapeutic potential. There is also evidence that there are FGFR alterations that confer resistance to other types of cancer treatment (6, 7) and that some specific FGFR aberrations may demonstrate differential sensitivity/resistance to distinct FGFR inhibitors (50). Intriguingly, some FGFR2 and FGFR3 somatic mutations were identical to mutations that, in germline form, are associated with dwarfism. However, there are no published epidemiologic data to suggest that individuals with germline FGFR aberrations and dwarfism have an increased incidence of cancer, suggesting that developmental compensatory mechanisms can mitigate the oncogenic potential of these aberrations. FGFR may also have prognostic value. Indeed, in breast cancer, FGFR1 amplification was independently associated with poor survival (45). Further study will be needed to elucidate the impact of each of the FGFR aberrations on cancer phenotype, prognosis, and response to treatment. Because many FGFR changes appear to activate signaling, it is also important to characterize the clinically relevant effects of the many potent FGFR inhibitors that are currently in clinical trials. Based on the frequent finding of FGFR abnormalities in diverse malignancies, especially in urothelial, breast, ovarian, endometrial, and squamous lung cancers, molecular interrogation of patients for FGFR aberrations in the clinical research and practice setting is warranted.

S. Elkin and J. Carter have ownership interest (including patents) in N-of-One. R. Kurzrock reports receiving research funding from Foundation Medicine, Genentech, Merck Serono, and Pfizer; is a consultant/advisory board member for Sequenom, and has ownership interest (including patents) in RScueRX. No potential conflicts of interest were disclosed by the other authors.

Conception and design: T. Helsten, S. Elkin, J. Carter, R. Kurzrock

Development of methodology: T. Helsten

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): T. Helsten, S. Elkin, B.N. Tomson, R. Kurzrock

Writing, review, and/or revision of the manuscript: T. Helsten, S. Elkin, E. Arthur, B.N. Tomson, J. Carter, R. Kurzrock

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): T. Helsten, S. Elkin, E. Arthur, B.N. Tomson

Study supervision: R. Kurzrock

This work was supported in part by the Joan and Irwin Jacobs Fund and MyAnswerToCancer philanthropic fund.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

1.
Katoh
M
,
Nakagama
H
. 
FGF receptors: cancer biology and therapeutics
.
Med Res Rev
2013
;
34
:
280
300
.
2.
Hubbard
SR
,
Till
JH
. 
Protein tyrosine kinase structure and function
.
Ann Rev Biochem
2000
;
69
:
373
98
.
3.
Gust
KM
,
McConkey
DJ
,
Awrey
S
,
Hegarty
PK
,
Qing
J
,
Bondaruk
J
, et al
Fibroblast growth factor receptor 3 is a rational therapeutic target in bladder cancer
.
Mol Cancer Ther
2013
;
12
:
1245
54
.
4.
Heist
RS
,
Mino-Kenudson
M
,
Sequist
LV
,
Tammireddy
S
,
Morrissey
L
,
Christiani
DC
, et al
FGFR1 amplification in squamous cell carcinoma of the lung
.
J Thorac Oncol
2012
;
7
:
1775
80
.
5.
Donnem
T
,
Al-Shibli
K
,
Al-Saad
S
,
Busund
L-T
,
Bremnes
RM
. 
Prognostic impact of fibroblast growth factor 2 in non-small cell lung cancer: coexpression with VEGFR-3 and PDGF-B predicts poor survival
.
J Thorac Oncol
2009
;
4
:
578
85
.
6.
Turner
N
,
Pearson
A
,
Sharpe
R
,
Lambros
M
,
Geyer
F
,
Lopez-Garcia
MA
, et al
FGFR1 amplification drives endocrine therapy resistance and is a therapeutic target in breast cancer
.
Cancer Res
2010
;
70
:
2085
94
.
7.
Ware
KE
,
Hinz
TK
,
Kleczko
E
,
Singleton
KR
,
Marek
LA
,
Helfrich
BA
, et al
A mechanism of resistance to gefitinib mediated by cellular reprogramming and the acquisition of an FGF2–FGFR1 autocrine growth loop
.
Oncogenesis
2013
;
2
:
e39
.
8.
Dienstmann
R
,
Rodon
J
,
Prat
A
,
Perez-Garcia
J
,
Adamo
B
,
Felip
E
, et al
Genomic aberrations in the FGFR pathway: opportunities for targeted therapies in solid tumors
.
Ann Oncol
2013
;
25
:
552
63
.
9.
Gozgit
JM
,
Wong
MJ
,
Moran
L
,
Wardwell
S
,
Mohemmad
QK
,
Narasimhan
NI
, et al
Ponatinib (AP24534), a multitargeted pan-FGFR inhibitor with activity in multiple FGFR-amplified or mutated cancer models
.
Mol Cancer Ther
2012
;
11
:
690
9
.
10.
Frampton
GM
,
Fichtenholtz
A
,
Otto
GA
,
Wang
K
,
Downing
SR
,
He
J
, et al
Development and validation of a clinical cancer genomic profiling test based on massively parallel DNA sequencing
.
Nat Biotechnol
2013
;
31
:
1023
31
.
11.
Thomas
RK
,
Nickerson
E
,
Simons
JF
,
Jänne
PA
,
Tengs
T
,
Yuza
Y
, et al
Sensitive mutation detection in heterogeneous cancer specimens by massively parallel picoliter reactor sequencing
.
Nat Med
2006
;
12
:
852
5
.
12.
Wagle
N
,
Berger
MF
,
Davis
MJ
,
Blumenstiel
B
,
DeFelice
M
,
Pochanard
P
, et al
High-throughput detection of actionable genomic alterations in clinical tumor samples by targeted, massively parallel sequencing
.
Cancer Discov
2012
;
2
:
82
93
.
13.
Di Martino
E
,
Tomlinson
DC
,
Knowles
MA
. 
A decade of FGF receptor research in bladder cancer: past, present, and future challenges
.
Adv Urol
2012
;
2012
:
429213
.
14.
Tavormina
PL
,
Shiang
R
,
Thompson
LM
,
Zhu
YZ
,
Wilkin
DJ
,
Lachman
RS
, et al
Thanatophoric dysplasia (types I and II) caused by distinct mutations in fibroblast growth factor receptor 3
.
Nat Genet
1995
;
9
:
321
8
.
15.
López-Knowles
E
,
Hernández
S
,
Malats
N
,
Kogevinas
M
,
Lloreta
J
,
Carrato
A
, et al
PIK3CA mutations are an early genetic alteration associated with FGFR3 mutations in superficial papillary bladder tumors
.
Cancer Res
2006
;
66
:
7401
4
.
16.
Kompier
LC
,
Lurkin
I
,
van der Aa
MNM
,
van Rhijn
BWG
,
van der Kwast
TH
,
Zwarthoff
EC
. 
FGFR3, HRAS, KRAS, NRAS and PIK3CA mutations in bladder cancer and their potential as biomarkers for surveillance and therapy
.
PLoS One
2010
;
5
:
e13821
.
17.
Saal
LH
,
Holm
K
,
Maurer
M
,
Memeo
L
,
Su
T
,
Wang
X
, et al
PIK3CA mutations correlate with hormone receptors, node metastasis, and ERBB2, and are mutually exclusive with PTEN loss in human breast carcinoma
.
Cancer Res
2005
;
65
:
2554
9
.
18.
Stemke-Hale
K
,
Gonzalez-Angulo
AM
,
Lluch
A
,
Neve
RM
,
Kuo
W-L
,
Davies
M
, et al
An integrative genomic and proteomic analysis of PIK3CA, PTEN, and AKT mutations in breast cancer
.
Cancer Res
2008
;
68
:
6084
91
.
19.
O'Brien
C
,
Wallin
JJ
,
Sampath
D
,
GuhaThakurta
D
,
Savage
H
,
Punnoose
EA
, et al
Predictive biomarkers of sensitivity to the phosphatidylinositol 3′ kinase inhibitor GDC-0941 in breast cancer preclinical models
.
Clin Cancer Res
2010
;
16
:
3670
83
.
20.
Weiss
J
,
Sos
ML
,
Seidel
D
,
Peifer
M
. 
Frequent and focal FGFR1 amplification associates with therapeutically tractable FGFR1 dependency in squamous cell lung cancer
.
Sci Transl Med
2010
;
2
:
62ra93
.
21.
Dutt
A
,
Ramos
AH
,
Hammerman
PS
,
Mermel
C
,
Cho
J
,
Sharifnia
T
, et al
Inhibitor-sensitive FGFR1 amplification in human non-small cell lung cancer
.
PLoS One
2011
;
6
:
e20351
.
22.
Shiang
CY
,
Qi
Y
,
Wang
B
,
Lazar
V
,
Wang
J
,
Fraser Symmans
W
, et al
Amplification of fibroblast growth factor receptor-1 in breast cancer and the effects of brivanib alaninate
.
Breast Cancer Res Treat
2010
;
123
:
747
55
.
23.
Turner
N
,
Lambros
MB
,
Horlings
HM
,
Pearson
A
,
Sharpe
R
,
Natrajan
R
, et al
Integrative molecular profiling of triple negative breast cancers identifies amplicon drivers and potential therapeutic targets
.
Oncogene
2010
;
29
:
2013
23
.
24.
Lew
ED
,
Furdui
CM
,
Anderson
KS
,
Schlessinger
J
. 
The precise sequence of FGF receptor autophosphorylation is kinetically driven and is disrupted by oncogenic mutations
.
Sci Signal
2009
;
2
:
ra6
.
25.
Hart
KC
,
Robertson
SC
,
Kanemitsu
MY
,
Meyer
AN
,
Tynan
JA
,
Donoghue
DJ
. 
Transformation and Stat activation by derivatives of FGFR1, FGFR3, and FGFR4
.
Oncogene
2000
;
19
:
3309
20
.
26.
Dutt
A
,
Salvesen
HB
,
Chen
T-H
,
Ramos
AH
,
Onofrio
RC
,
Hatton
C
, et al
Drug-sensitive FGFR2 mutations in endometrial carcinoma
.
Proc Natl Acad Sci U S A
2008
;
105
:
8713
7
.
27.
Ibrahimi
OA
,
Zhang
F
,
Eliseenkova
AV
,
Itoh
N
,
Linhardt
RJ
,
Mohammadi
M
. 
Biochemical analysis of pathogenic ligand-dependent FGFR2 mutations suggests distinct pathophysiological mechanisms for craniofacial and limb abnormalities
.
Hum Mol Genet
2004
;
13
:
2313
24
.
28.
Chen
H
,
Ma
J
,
Li
W
,
Eliseenkova
AV
,
Xu
C
,
Neubert
TA
, et al
A molecular brake in the kinase hinge region regulates the activity of receptor tyrosine kinases
.
Mol Cell
2007
;
27
:
717
30
.
29.
Liao
RG
,
Jung
J
,
Tchaicha
J
,
Wilkerson
MD
,
Sivachenko
A
,
Beauchamp
EM
, et al
Inhibitor-sensitive FGFR2 and FGFR3 mutations in lung squamous cell carcinoma
.
Cancer Res
2013
;
73
:
5195
205
.
30.
Adar
R
,
Monsonego-Ornan
E
,
David
P
,
Yayon
A
. 
Differential activation of cysteine-substitution mutants of fibroblast growth factor receptor 3 is determined by cysteine localization
.
J Bone Miner Res
2002
;
17
:
860
8
.
31.
Williams
SV
,
Hurst
CD
,
Knowles
MA
. 
Oncogenic FGFR3 gene fusions in bladder cancer
.
Hum Mol Genet
2013
;
22
:
795
803
.
32.
Bernard-Pierrot
I
,
Brams
A
,
Dunois-Lardé
C
,
Caillault
A
, Diez
de Medina
SG
,
Cappellen
D
, et al
Oncogenic properties of the mutated forms of fibroblast growth factor receptor 3b
.
Carcinogenesis
2006
;
27
:
740
7
.
33.
Tomlinson
DC
,
Hurst
CD
,
Knowles
MA
. 
Knockdown by shRNA identifies S249C mutant FGFR3 as a potential therapeutic target in bladder cancer
.
Oncogene
2007
;
26
:
5889
99
.
34.
Hafner
C
,
Di Martino
E
,
Pitt
E
,
Stempfl
T
,
Tomlinson
D
,
Hartmann
A
, et al
FGFR3 mutation affects cell growth, apoptosis and attachment in keratinocytes
.
Exp Cell Res
2010
;
316
:
2008
16
.
35.
Bonaventure
J
,
Gibbs
L
,
Horne
WC
,
Baron
R
. 
The localization of FGFR3 mutations causing thanatophoric dysplasia type I differentially affects phosphorylation, processing and ubiquitylation of the receptor
.
FEBS J
2007
;
274
:
3078
93
.
36.
Raffioni
S
,
Zhu
YZ
,
Bradshaw
RA
,
Thompson
LM
. 
Effect of transmembrane and kinase domain mutations on fibroblast growth factor receptor 3 chimera signaling in PC12 cells. A model for the control of receptor tyrosine kinase activation
.
J Biol Chem
1998
;
273
:
35250
9
.
37.
Ronchetti
D
,
Greco
A
,
Compasso
S
,
Colombo
G
,
Dell'Era
P
,
Otsuki
T
, et al
Deregulated FGFR3 mutants in multiple myeloma cell lines with t(4;14): comparative analysis of Y373C, K650E and the novel G384D mutations
.
Oncogene
2001
;
20
:
3553
62
.
38.
Chesi
M
,
Brents
LA
,
Ely
SA
,
Bais
C
,
Robbiani
DF
,
Mesri
EA
, et al
Activated fibroblast growth factor receptor 3 is an oncogene that contributes to tumor progression in multiple myeloma
.
Blood
2001
;
97
:
729
36
.
39.
Singh
D
,
Chan
JM
,
Zoppoli
P
,
Niola
F
,
Sullivan
R
,
Castano
A
, et al
Transforming fusions of FGFR and TACC genes in human glioblastoma
.
Science
2012
;
337
:
1231
5
.
40.
Schildhaus
H-U
,
Heukamp
LC
,
Merkelbach-Bruse
S
,
Riesner
K
,
Schmitz
K
,
Binot
E
, et al
Definition of a fluorescence in-situ hybridization score identifies high- and low-level FGFR1 amplification types in squamous cell lung cancer
.
Mod Pathol
Nature Publishing Group
; 
2012
;
25
:
1473
80
.
41.
Kim
HR
,
Kim
DJ
,
Kang
DR
,
Lee
JG
,
Lim
SM
,
Lee
CY
, et al
Fibroblast growth factor receptor 1 gene amplification is associated with poor survival and cigarette smoking dosage in patients with resected squamous cell lung cancer
.
J Clin Oncol
2013
;
31
:
731
7
.
42.
Reck
M
,
Kaiser
R
,
Mellemgaard
A
,
Douillard
JY
,
Orlov
S
,
Krzakowski
M
, et al
Docetaxel plus nintedanib versus docetaxel plus placebo in patients with previously treated non-small-cell lung cancer (LUME- Lung 1): a phase 3, double-blind, randomised controlled trial
.
Lancet Oncol
2014
;
15
:
143
55
.
43.
Hanna
NH
,
Kaiser
R
,
Sullivan
RN
,
Aren
OR
,
Ahn
M-J
,
Tiangco
B
, et al
Lume-lung 2: A multicenter, randomized, double-blind, phase III study of nintedanib plus pemetrexed versus placebo plus pemetrexed in patients with advanced nonsquamous non- small cell lung cancer (NSCLC) after failure of first-line chemotherapy
.
ASCO Meeting Abstracts
2013
;
31
:
8034
.
44.
Sequist
LV
,
Cassier
P
,
Varga
A
,
Tabernero
J
,
Schellens
JH
,
Delord
J-P
, et al
Abstract CT326: phase I study of BGJ398, a selective pan-FGFR inhibitor in genetically preselected advanced solid tumors
.
Cancer Res
2014
;
74
:
CT326
6
.
45.
Elbauomy Elsheikh
S
,
Green
AR
,
Lambros
MBK
,
Turner
NC
,
Grainge
MJ
,
Powe
D
, et al
FGFR1 amplification in breast carcinomas: a chromogenic in situ hybridisation analysis
.
Breast Cancer Res
2007
;
9
:
R23
.
46.
Soria
J
,
De Braud
FG
,
Cereda
R
,
Bahleda
R
,
Delmonte
A
,
Angevin
E
, et al
First-in-man study of E-3810, a novel VEGFR and FGFR inhibitor, in patients with advanced solid tumors
.
ASCO Meeting Abstracts
2011
;
29
:
TPS149
.
47.
Tsai
S-J
,
Wu
M-H
,
Chen
H-M
,
Chuang
P-C
,
Wing
L-YC
. 
Fibroblast growth factor-9 is an endometrial stromal growth factor
.
Endocrinology
2002
;
143
:
2715
21
.
48.
Ahmad
I
,
Iwata
T
,
Leung
HY
. 
Mechanisms of FGFR-mediated carcinogenesis
.
Biochim Biophys Acta
2012
;
1823
:
850
60
.
49.
Parker
BC
,
Annala
MJ
,
Cogdell
DE
,
Granberg
KJ
,
Sun
Y
,
Ji
P
, et al
The tumorigenic FGFR3–TACC3 gene fusion escapes miR-99a regulation in glioblastoma
.
J Clin Invest
2013
;
123
:
855
65
.
50.
Byron
SA
,
Chen
H
,
Wortmann
A
,
Loch
D
,
Gartside
MG
,
Dehkhoda
F
, et al
The N550K/H mutations in FGFR2 confer differential resistance to PD173074, dovitinib, and ponatinib ATP-competitive inhibitors
.
Neoplasia
Elsevier
; 
2013
;
15
:
975–IN30
.