Impact of EML4-ALK Variant on Resistance Mechanisms and Clinical Outcomes in ALK-Positive Lung Cancer

Abstract

Purpose

Advanced anaplastic lymphoma kinase (ALK) fusion-positive non–small-cell lung cancers (NSCLCs) are effectively treated with ALK tyrosine kinase inhibitors (TKIs). However, clinical outcomes in these patients vary, and the benefit of TKIs is limited as a result of acquired resistance. Emerging data suggest that the ALK fusion variant may affect clinical outcome, but the molecular basis for this association is unknown.

Patients and Methods

We identified 129 patients with ALK-positive NSCLC with known ALK variants. ALK resistance mutations and clinical outcomes on ALK TKIs were retrospectively evaluated according to ALK variant. A Foundation Medicine data set of 577 patients with ALK-positive NSCLC was also examined.

Results

The most frequent ALK variants were EML4-ALK variant 1 in 55 patients (43%) and variant 3 in 51 patients (40%). We analyzed 77 tumor biopsy specimens from patients with variants 1 and 3 who had progressed on an ALK TKI. ALK resistance mutations were significantly more common in variant 3 than in variant 1 (57% v 30%; P = .023). In particular, ALK G1202R was more common in variant 3 than in variant 1 (32% v 0%; P < .001). Analysis of the Foundation Medicine database revealed similar associations of variant 3 with ALK resistance mutation and with G1202R (P = .010 and .015, respectively). Among patients treated with the third-generation ALK TKI lorlatinib, variant 3 was associated with a significantly longer progression-free survival than variant 1 (hazard ratio, 0.31; 95% CI, 0.12 to 0.79; P = .011).

Conclusion

Specific ALK variants may be associated with the development of ALK resistance mutations, particularly G1202R, and provide a molecular link between variant and clinical outcome. ALK variant thus represents a potentially important factor in the selection of next-generation ALK inhibitors.

INTRODUCTION

Anaplastic lymphoma kinase (ALK) gene rearrangements encode driver fusion oncoproteins and are found in approximately 5% of non–small-cell lung cancers (NSCLCs).¹ Since crizotinib,^2-5 multiple second-generation (eg, ceritinib, alectinib, brigatinib)^6-14 and third-generation (eg, lorlatinib)¹⁵ ALK tyrosine kinase inhibitors (TKIs) have been developed for patients with ALK-positive NSCLC, all with higher potency and greater CNS penetration than crizotinib. Although these ALK inhibitors have dramatically expanded the therapeutic landscape of ALK-positive NSCLC, clinical outcomes in patients can vary widely, and the biologic mechanisms that underpin such heterogeneous outcomes are unknown. Moreover, at some point, essentially all patients experience a relapse while receiving TKI therapy as a result of acquired drug resistance.^1,4,5 The elucidation of resistance mechanisms in patients has proven critical in efforts to rationally select subsequent therapies.^1,16

Emerging data indicate that ALK fusion variants may have biologic and clinical implications in ALK-positive lung cancer. The predominant ALK fusion partner in NSCLC is the echinoderm microtubule–associated protein-like 4 (EML4) gene.^17,18 Among > 15 EML4-ALK variants identified to date, the most common are variant 1 (v1; exon 13 of EML4 fused to exon 20 of ALK [E13;A20]) and v3a/b (exon 6a/b of EML4 fused to exon 20 of ALK [E6a/b;A20]).^17,19-22 All variants retain the entire tyrosine kinase domain of ALK and the N-terminal coiled-coil region of EML4, which is necessary and sufficient for the dimerization and constitutive activation of ALK.¹⁷ Different variants may have different protein stabilities, which affects sensitivity to crizotinib in vitro.^23-26

Recent studies have suggested differential responses to crizotinib according to ALK variant in patients.^27,28 For example, longer responses to crizotinib were reported with v1 compared with non-v1²⁷ or with non-v3 compared with v3,²⁸ yet two other studies found no difference in clinical response to crizotinib on the basis of ALK variant, which highlights the need for additional investigation.^29,30 Moreover, the potential effect of ALK variants on the efficacy of next-generation ALK TKIs or the development of resistance mechanisms, which can influence responses to subsequent therapies, has not been examined. We evaluated the frequency and spectrum of ALK resistance mutations according to fusion variant in patients with ALK-positive NSCLC with acquired TKI resistance and clinical outcomes of these patients who received various generations of ALK inhibitors.

PATIENTS AND METHODS

Study Population

Between January 2008 and January 2017, 129 patients with ALK-positive NSCLC and known ALK variant were identified at the Massachusetts General Hospital (MGH; n = 113) and University of California, Irvine (n = 16; Data Supplement). The study was approved by the institutional review boards (IRBs) at each site.

In addition, a separate group of 577 patients with ALK-positive NSCLC and known ALK variant identified during routine clinical care from August 2012 to December 2016 with FoundationOne next-generation sequencing (NGS) assays at Foundation Medicine were analyzed for the frequency and distribution of ALK resistance mutations. Approval for the study of this cohort was obtained from the Western IRB (protocol no. 20152817).

Data Collection

For the 129 patients included in the main study cohort, data on clinicopathologic features and treatment histories were extracted from medical records. Progression-free survival (PFS) and overall survival (OS) outcomes were measured as detailed in the Data Supplement. Data were updated as of November 15, 2017.

Identification of ALK Variant

ALK fusion variants were detected by using the MGH fusion panel, which uses targeted RNA sequencing with anchored multiplex polymerase chain reaction (PCR) to detect fusion transcripts that involve known oncogenes, including ALK³¹; the FoundationOne platform³²; targeted NGS platforms at outside institutions as previously described^33,34; or reverse transcription PCR (commercial or as previously described³⁵) and Sanger sequencing of cDNA (Data Supplement).

Genotyping for ALK Resistance Mutation

Postprogression tumor biopsy specimens were analyzed for the presence of ALK resistance mutations under an IRB-approved tissue collection protocol. Methodologies to detect ALK resistance mutations included the MGH SNaPshot NGS platform (which uses anchored PCR to detect single-nucleotide variants and insertions/deletions in cancer-related genes, including ALK)³¹, the FoundationOne NGS³², and the OncoPanel NGS.³³ A subset of specimens underwent Sanger sequencing of cDNA for the entire ALK domain,³⁶ and one specimen underwent whole-exome sequencing as previously reported.¹⁶

Statistical Analysis

Detailed statistical methods are provided in the Data Supplement. PFS and OS curves were estimated using the Kaplan-Meier method, and the Cox proportional hazards regression model was used to estimate the hazard ratio (HR) to express PFS and OS differences between variant groups. All statistical analyses were performed with SAS 9.4 software (SAS Institute, Cary, NC).

RESULTS

Patient Characteristics

Baseline patient characteristics are listed in Table 1. The majority of patients received crizotinib (121 [94%] of 129)—67 (52%) as first-line, 31 (24%) as second-line, and 23 (18%) as third-line or higher-line treatment. Over the course of their disease, 45 patients (35%) received two ALK inhibitors, and 56 (43%) received three or more ALK inhibitors.

Table 1.

Clinical Characteristics According to ALK Variant

Among the 129 patients, 123 (95%) had an EML4-ALK fusion. The most frequent EML4-ALK variants were v1 in 55 patients (43%), and v3 in 51 patients (40%; Fig 1). No differences were found in clinicopathologic features between patients with v1 and v3 (Table 1). The remaining EML4-ALK fusions consisted of v2 (6% [E20;A20]), v5′ (4% [E18;A20]), v5 (2% [E2;A20]), and v7 (1% [E14;A20]). Among non–EML4-ALK fusions detected in six patients (5%), the fusion partner genes included HIP1 (n = 3),^37-39 KIF5B (n = 1),⁴⁰ PRKAR1A (n = 1),⁴¹ and MTA3 (n = 1; not previously reported). Comparable baseline characteristics also were observed across these variant groups (Data Supplement).

Fig 1.

Frequency of ALK variants in the study cohort (N = 129). Schematic key: blue, coiled-coil region of EML4; gold, tandem atypical propeller EML domain of EML4; gray, tyrosine kinase domain of ALK. Note that v3 and v5 exist as isoforms (v3a and v3b and v5a and v5b, respectively) generated by alternative splicing.²⁴ A, ALK exon; E, EML4 exon; v, variant.

ALK Resistance Mutations by Variant

To determine whether the ALK variant affects the development of molecular mechanisms of resistance, we identified patients in the study cohort who underwent a repeat biopsy after progression on a first- or second-generation ALK inhibitor. Seventy-seven patients (60%) had a postprogression tumor biopsy. Of these, 12 had two repeat biopsies, and two underwent three serial biopsies (Data Supplement). A total of 93 tumor biopsies were performed. The Data Supplement shows the distribution of post-TKI biopsies according to ALK fusion variant and timing of biopsy.

Because the number of patients with non-v1 and/or -v3 was small (Data Supplement; Fig 1), the most common variant groups, v1 and v3, were selected for additional analysis. Overall, 33 v1 and 44 v3 repeat biopsies were performed after progression on an ALK inhibitor. ALK resistance mutations were identified in 10 patients with v1 (30%) compared with 25 with v3 (57%; P = .023; Fig 2). Of note, the ALK G1202R solvent-front mutation, which causes steric interference with drug binding^16,36,42 and confers high-level resistance to first- and second-generation ALK inhibitors,¹⁶ was detected in zero (0%) of 33 patients with v1 versus 14 (32%) of 44 with v3 (P < .001).

Fig 2.

ALK resistance mutations in tumor biopsy specimens obtained after progression on an ALK inhibitor according to EML4-ALK variant.

Our group has previously shown that ALK resistance mutations as a whole and ALK G1202R specifically occur more frequently after second-generation ALK TKIs (50% to 60% and 20% to 40%, respectively) compared with crizotinib (20% to 30% and 2%, respectively).¹⁶ To address imbalances in postcrizotinib versus post–second-generation ALK TKI samples that could potentially confound the comparison of resistance mutations in v1 versus v3, we focused on the distribution of ALK mutations in biopsy specimens obtained after progression on a second-generation ALK inhibitor. ALK resistance mutations were more common in v3 (21 [66%] of 32) than v1 (eight [42%] of 19), although this difference was not statistically significant (P = .145). ALK G1202R was significantly more common in v3 (14 [44%] of 32) than in v1 (zero [0%] of 19; P = .001; Fig 3). Of note, no statistically significant difference was found between v1 and v3 in the prebiopsy second-generation ALK inhibitor administered, although more v3 than v1 specimens were postalectinib biopsies. Similarly, the cumulative number of prior TKIs was not significantly different (Data Supplement). The sequencing methods used to detect an ALK resistance mutation for v1 and v3 are shown in the Data Supplement.

Fig 3.

Distribution of ALK resistance mutations in tumor biopsy specimens obtained after disease progression on a second-generation ALK inhibitor by EML4-ALK variant. WT, wild-type (nonmutated) ALK.

Among postcrizotinib tumor biopsy specimens, ALK resistance mutations were detected in two (14%) of 14 samples of v1 and four (33%) of 12 samples of v3 (P = .365; Fig 4). None of the samples harbored an ALK G1202R mutation. The overall lower frequency of ALK resistance mutations (23%; P = .007) and specifically of ALK G1202R (0%; P = .002) after crizotinib versus a second-generation ALK inhibitor (57% and 27%, respectively) is consistent with our previous report.¹⁶

Fig 4.

Distribution of ALK resistance mutations in tumor biopsy specimens obtained after disease progression on crizotinib by EML4-ALK variant. WT, wild-type ALK.

Analysis of 577 ALK-Positive NSCLCs

To validate the associations between variant and ALK resistance mutations, we subsequently examined a database of ALK-positive NSCLCs with known ALK variants sequenced at Foundation Medicine. Among 577 patients with ALK-positive NSCLC, the most common variants were EML4-ALK v3 (n = 186 [32%]) and v1 (n = 182 [32%]) followed by v2 (n = 47 [8%]), other EML4-ALK variants (n = 69 [12%]), and non–EML4-ALK variants (n = 93 [16%]; Data Supplement). A total of 624 tumor tissue biopsy specimens were taken from these patients (v3, n = 201; v1, n = 199; v2, n = 49; other EML4-ALK, n = 74; non–EML4-ALK, n = 101). Although the frequencies of EML4-ALK v1 and v3 were almost identical, ALK resistance mutations (n = 30) and specifically ALK G1202R (n = 9) were significantly more common for v3 than for v1 (ALK mutation, 8% v 2% [P = .010]; G1202R, 3.5% v 0% [P = .015]; Data Supplement). Of note, clinical information, including prebiopsy treatment history, was not available for the patients included in the Foundation Medicine data set, and therefore, the distribution of patients who were treatment naive, or treated with crizotinib or next-generation ALK inhibitors, is not known. Nevertheless, these findings collectively support the notion that EML4-ALK v3 is associated with the development of ALK resistance mutations and, in particular, the highly refractory G1202R mutation.

Clinical Outcomes on ALK TKIs With v1 and v3

We next evaluated the effect of ALK variants on clinical responses to various ALK inhibitors. Again, we focused this analysis on the most common variants v1 and v3. Treatment histories of patients in the v1 and v3 cohorts were overall well balanced (Table 1). The median OS from the time of diagnosis of advanced disease was 5.0 years and 3.6 years for v1 and v3 cohorts, respectively (HR, 1.16; 95% CI, 0.67 to 2.01; P = .584; Data Supplement), similar to the previously reported OS for patients with ALK-positive NSCLC who received sequential ALK TKIs.⁴³ Of note, this OS analysis is relatively immature, with only 52 deaths (49%) at the time of data cutoff.

Among 99 patients treated with crizotinib as the first ALK inhibitor, no significant difference was found in PFS in patients with EML4-ALK v1 (n = 51) versus v3 (n = 48; HR, 1.30; 95% CI, 0.85 to 1.98; P = .229; Fig 5A). To address whether these results were confounded by prior lines of chemotherapy, we also examined PFS among 55 patients who received crizotinib as first-line therapy. Again, no difference in PFS between v1 (n = 27) and v3 (n = 28) were found (HR, 1.61; 95% CI, 0.84 to 2.75; P = .163; Fig 5B). Similarly, no significant difference was found in PFS after second-generation ALK TKIs (ie, ceritinib, alectinib, brigatinib) given as the second ALK inhibitor after crizotinib among 77 patients with v1 (n = 37) versus v3 (n = 40; HR, 1.45; 95% CI, 0.88 to 2.38; P = .141; Fig 5C). For patients in each cohort who experienced disease progression, the pattern of progression (ie, CNS only, both intra- and extracranially, or extracranially only) was not different between v1 and v3 (Data Supplement).

Fig 5.

Kaplan-Meier curves for progression-free survival (PFS) for (A) crizotinib as first ALK tyrosine kinase inhibitor (TKI) in patients with EML4-ALK variant 1 (v1; n = 51) versus v3 (n = 48), (B) crizotinib as first-line therapy in v1 (n = 27) versus v3 (n = 28), (C) second-generation ALK TKI administered as the second ALK inhibitor after crizotinib in v1 (n = 37) versus v3 (n = 40), and (D) lorlatinib administered after crizotinib and at least one second-generation ALK TKI in v1 (n = 12) and v3 (n = 17).

In an exploratory analysis of 12 patients with v1 and 17 with v3 who received the third-generation ALK TKI lorlatinib after experiencing treatment failure with both crizotinib and at least one second-generation ALK inhibitor, v3 was associated with significantly longer PFS than v1 (median, 11.0 v 3.3 months; HR, 0.31; 95% CI, 0.12 to 0.79; P = .011; Fig 5D). By univariable analysis, no other baseline features, including sex, age, race/ethnicity, smoking history, initial stage at diagnosis, presence of extrathoracic disease, and presence of CNS metastasis at the diagnosis of advanced disease, was significantly associated with PFS on lorlatinib. Lorlatinib has previously been shown to retain potent activity against all known crizotinib-resistant ALK mutations, including G1202R, in vitro^16,44 and has demonstrated activity in patients with ALK-positive NSCLC previously treated with two or more ALK inhibitors, including those with ALK G1202R.¹⁵ In the current cohort, six patients with v1 and 15 with v3 who received lorlatinib underwent a prelorlatinib tumor biopsy, and an ALK resistance mutation was present in one v1 (17%; G1269A) and 13 v3 (87%) samples, respectively (Data Supplement). Of the 13 v3 samples with a prelorlatinib ALK mutation, 12 harbored ALK G1202R. Finally, no difference in PFS was noted in patients with v1 versus v3 who received pemetrexed (HR, 1.11; 95% CI, 0.60 to 2.07; P = .742) or platinum and pemetrexed combination (HR, 0.84; 95% CI, 0.40 to 1.78; P = .649; Data Supplement).

DISCUSSION

To our knowledge, we present the largest analysis to date to examine the clinical effect of ALK variants in ALK-positive NSCLC and the first study to evaluate ALK resistance mutations according to EML4-ALK variant. We found that EML4-ALK v3 was significantly associated with the development of ALK resistance mutations, particularly ALK G1202R, which suggests that patients with v3 may be more likely to acquire distinct secondary ALK resistance mutations as they undergo treatment with ALK TKIs.

Previously, we demonstrated the unique spectrum of activity of each currently available ALK inhibitor against different ALK resistance mutations.¹⁶ Of note, the solvent-front ALK G1202R mutation conferred high-level resistance to all first- and second-generation ALK inhibitors but retained sensitivity to lorlatinib.¹⁶ The strong association between v3 and ALK G1202R suggests that the establishment of ALK variant status may have important therapeutic implications. Indeed, we observed a significantly longer PFS among patients with EML4-ALK v3 versus v1 who received lorlatinib. A priori knowledge of the ALK variant status thus could help to select patients more likely to achieve a durable response to lorlatinib. We speculate that the greater propensity of v3 to develop ALK mutations, particularly G1202R, likely underlies these different outcomes. Supportive of this notion, the activity of lorlatinib was found to correlate tightly with the presence of an ALK resistance mutation (which suggests continued ALK dependency) in a small series of ceritinib-resistant patient-derived cell lines.¹⁶ Several ongoing and completed studies of second- and third-generation ALK TKIs, including the phase III study of first-line lorlatinib versus crizotinib (ClinicalTrials.gov identifier: NCT03052608), are retrospectively evaluating the association between efficacy and ALK variant and/or preexisting ALK resistance mutations and may further enhance our understanding of the molecular variables that contribute to heterogeneous clinical outcomes.

We also examined the effect of ALK variants on clinical responses to first- and second-generation ALK inhibitors. In this study, no statistically significant difference in PFS was observed among patients with v1 versus v3 treated with crizotinib or a second-generation ALK TKI. However, the median PFS in all contexts was numerically shorter for v3 than for v1, which suggests that larger patient cohorts may be needed to achieve sufficient power to detect a statistically significant difference. In light of the Alectinib Versus Crizotinib in Untreated ALK-Positive NSCLC study that demonstrated superiority of front-line alectinib over crizotinib,¹² continuing the investigation of the effect of specific ALK variants on resistance patterns and clinical outcomes in patients treated with first-line alectinib will be of particular interest.

Four previously published studies evaluated clinical responses to ALK TKIs according to ALK variant. All were centered on crizotinib, and conflicting findings have been reported (eg, PFS longer for patients with v1 v non-v1,²⁷ longer for those with non-v3 v v3,²⁸ not different on the basis of variant^29,30). Several limitations may account for the discordant findings, including small sample size, lack of distinction in crizotinib line of therapy, and clustering of variants into groups that may not be biologically or clinically relevant. To address these limitations, we analyzed a larger cohort of patients with a known ALK variant, and focused on the most common variants v1 and v3 (to avoid the confounding effect of arbitrary variant clustering) and on crizotinib administered as the first ALK TKI or as first-line therapy (to minimize the effect of treatment lines).

Nonetheless, the current study had limitations. First, it was a retrospective study and still limited in the number of patients. Variant analysis in larger, prospective studies will be needed to validate and expand on our findings. Second, this study focused on the most common EML4-ALK variants v1 and v3, and therefore, the potential effect of non-v1/v3 variants on ALK TKI resistance mechanisms and clinical outcomes remains unknown. Moreover, the effect of ALK variants on OS of patients with ALK-positive NSCLC requires more investigation because the OS data in this study were not mature. Of note, as shown in the Data Supplement, some differences were found in the sequencing methods used to detect the ALK resistance mutations in v1 and v3 post-TKI biopsies. To our knowledge, the particular methods used in this study are not limited in the ability to detect specific ALK fusion variants or resistance mutations such as G1202R.

Future studies will need to explore the mechanisms that underlie the differential association between ALK variant and resistance mutations. Prior studies have shown that ALK variants retain varying portions of the EML4 tandem atypical propeller EML (TAPE) domain, which results in differential fusion protein stability in vitro.^25,26 In particular, shorter EML4-ALK variants (eg, v3) that lack the entire TAPE domain have been found to be more stable than longer variants (eg, v1) that retain a partial TAPE domain.^23-26 In theory, ALK-rearranged tumor cells with the more stable v3 could be more ALK addicted, which necessitates the development of potent on-target ALK mutations (eg, G1202R) to mediate acquired resistance. Alternatively, an assessment of whether certain variants such as v3 are more structurally vulnerable to developing particular ALK resistance mutations and whether mutations confer differential levels of resistance depending on the variant would be interesting.

In summary, the findings suggest that EML4-ALK v3 is associated with a significantly higher incidence of ALK resistance mutations, particularly G1202R, and provide a potential molecular link between variant and clinical outcome. Thus, ALK variant status may represent an important emerging factor in guiding the treatment strategy for ALK-positive NSCLC.

AUTHOR CONTRIBUTIONS

Conception and design: Jessica J. Lin, Jeffrey S. Ross, Alice T. Shaw, Sai-Hong Ignatius Ou

Financial support: Alice T. Shaw

Provision of study materials or patients: Jessica J. Lin, Alexa B. Schrock, Alice T. Shaw, Sai-Hong Ignatius Ou

Collection and assembly of data: Jessica J. Lin, Viola W. Zhu, Satoshi Yoda, Alexa B. Schrock, Nicholas A. Jessop, Ginger Y. Jiang, Long P. Le, Jeffrey S. Ross, Alice T. Shaw, Sai-Hong Ignatius Ou

Data analysis and interpretation: Jessica J. Lin, Satoshi Yoda, Beow Y. Yeap, Alexa B. Schrock, Ibiayi Dagogo-Jack, Long P. Le, Kyle Gowen, Philip J. Stephens, Jeffrey S. Ross, Siraj M. Ali, Vincent A. Miller, Melissa L. Johnson, Christine M. Lovly, Aaron N. Hata, Justin F. Gainor, Anthony J. Iafrate, Alice T. Shaw, Sai-Hong Ignatius Ou

Manuscript writing: All authors

Final approval of manuscript: All authors

Accountable for all aspects of the work: All authors

AUTHORS' DISCLOSURES OF POTENTIAL CONFLICTS OF INTEREST

Impact of EML4-ALK Variant on Resistance Mechanisms and Clinical Outcomes in ALK-Positive Lung Cancer

The following represents disclosure information provided by authors of this manuscript. All relationships are considered compensated. Relationships are self-held unless noted. I = Immediate Family Member, Inst = My Institution. Relationships may not relate to the subject matter of this manuscript. For more information about ASCO's conflict of interest policy, please refer to www.asco.org/rwc or ascopubs.org/jco/site/ifc.

Jessica J. Lin

Honoraria: Chugai Pharmaceutical

Consulting or Advisory Role: Boehringer Ingelheim

Viola W. Zhu

Stock or Other Ownership: TP Therapeutics (I)

Honoraria: AstraZeneca

Consulting or Advisory Role: AstraZeneca, Bayer AG, TP Therapeutics, Biocept

Speakers’ Bureau: AstraZeneca, Roche, Genentech, Takeda Pharmaceuticals

Satoshi Yoda

No relationship to disclose

Beow Y. Yeap

Stock or Other Ownership: SISCAPA Assay Technologies (I)

Consulting or Advisory Role: Abcodia (I)

Alexa B. Schrock

Employment: Foundation Medicine

Stock or Other Ownership: Foundation Medicine

Ibiayi Dagogo-Jack

Honoraria: Foundation Medicine

Consulting or Advisory Role: Boehringer Ingelheim

Nicholas A. Jessop

No relationship to disclose

Ginger Y. Jiang

Employment: Novartis (I)

Long P. Le

Stock or Other Ownership: ArcherDx

Consulting or Advisory Role: ArcherDx

Patents, Royalties, Other Intellectual Property: Anchored multiplex polymerase chain reaction patent licensed to ArcherDx

Kyle Gowen

Employment: Foundation Medicine

Stock or Other Ownership: Foundation Medicine

Travel, Accommodations, Expenses: Foundation Medicine

Philip J. Stephens

Employment: Foundation Medicine

Leadership: Foundation Medicine

Stock or Other Ownership: Foundation Medicine

Jeffrey S. Ross

Employment: Foundation Medicine

Leadership: Foundation Medicine

Stock or Other Ownership: Foundation Medicine

Siraj M. Ali

Employment: Foundation Medicine

Stock or Other Ownership: Exelixis, Otonomy, OncoSec Medical, Genocea Biosciences

Consulting or Advisory Role: Incysus

Patents, Royalties, Other Intellectual Property: Patents through Foundation Medicine, and patents through Seres Health on microbiomes in non-neoplastic disease (I)

Vincent A. Miller

Employment: Foundation Medicine

Leadership: Foundation Medicine

Stock or Other Ownership: Foundation Medicine

Consulting or Advisory Role: Revolution Medicine

Patents, Royalties, Other Intellectual Property: Receives periodic royalties related to T790M patent awarded to Memorial Sloan Kettering Cancer Center

Melissa L. Johnson

Consulting or Advisory Role: Astellas Pharma (I), Otsuka (I), Genentech (Inst), Roche (Inst), Celgene (Inst), Boehringer Ingelheim (Inst), AbbVie (Inst)

Research Funding: OncoMed Pharmaceuticals (Inst), BerGenBio (Inst), Eli Lilly (Inst), EMD Serono (Inst), Kadmon Holdings (Inst), Janssen Pharmaceuticals (Inst), Mirati Therapeutics (Inst), Genmab (Inst), Pfizer (Inst), AstraZeneca (Inst), Genentech (Inst), Roche (Inst), Stemcentrx (Inst), Novartis (Inst), Checkpoint Therapeutics (Inst), Array BioPharma (Inst), Regeneron Pharmaceuticals (Inst)

Christine M. Lovly

Honoraria: Novartis, Sequenom, QIAGEN, Pfizer, National Comprehensive Cancer Network, Takeda Pharmaceuticals

Consulting or Advisory Role: ARIAD Pharmaceuticals, Clovis Oncology, Genoptix, Novartis, ARIAD Pharmaceuticals, Foundation Medicine, Cepheid

Research Funding: AstraZeneca, Novartis

Travel, Accommodations, Expenses: Pfizer, Takeda Pharmaceuticals, Roche, Cepheid

Aaron N. Hata

Research Funding: Amgen, Novartis (Inst), Relay Therapeutics

Justin F. Gainor

Honoraria: Merck, Incyte, ARIAD Pharmaceuticals, Pfizer, Novartis

Consulting or Advisory Role: Genentech, Bristol-Myers Squibb, Theravance Biopharma, Loxo Oncology, Takeda Pharmaceuticals

Research Funding: Merck (Inst), Novartis (Inst), Genetech (Inst), Bristol-Myers Squibb (Inst), Adaptimmune (Inst), AstraZeneca (Inst), ARIAD Pharmaceuticals (Inst), Jounce Therapeutics (Inst), Moderna Therapeutics (Inst)

Travel, Accommodations, Expenses: Affymetrix

Anthony J. Iafrate

Stock or Other Ownership: ArcherDx

Consulting or Advisory Role: Debiopharm Group, Constellation Pharmaceuticals, Chugai Pharmaceutical, Roche

Research Funding: Blueprint Medicines

Patents, Royalties, Other Intellectual Property: ArcherDx exclusive license to AMP technology

Alice T. Shaw

Honoraria: Pfizer, Novartis, Roche, Genentech, Foundation Medicine

Consulting or Advisory Role: Pfizer, Novartis, Genentech, Roche, ARIAD Pharmaceuticals, Ignyta, Blueprint Medicines, Daiichi Sankyo, EMD Serono, Taiho Pharmaceutical, KSQ Therapeutics, Natera

Research Funding: Pfizer, Novartis, Roche, Genentech

Sai-Hong Ignatius Ou

Honoraria: Novartis, Pfizer, Roche, Genentech, ARIAD Pharmaceuticals, Takeda Pharmaceuticals, AstraZeneca

Consulting or Advisory Role: ARIAD Pharmaceuticals, Takeda Pharmaceuticals, Pfizer, Novartis, AstraZeneca, Roche, Genentech

Speakers’ Bureau: AstraZeneca, Genentech, Takeda Pharmaceuticals, ARIAD Pharmaceuticals

Research Funding: AstraZeneca (Inst), Pfizer (Inst), Roche (Inst), AstraZeneca (Inst), MedImmune (Inst), Clovis Oncology (Inst), ARIAD Pharmaceuticals (Inst), Ignyta (Inst), Peregrine Pharmaceuticals (Inst), GlaxoSmithKline (Inst), Astellas Pharma (Inst), Chugai Pharmaceutical (Inst), TP Therapeutics (Inst), Novartis (Inst), Blueprint Medicines (Inst)