The Molecular Landscape of Recurrent and Metastatic Head and Neck Cancers: Insights From a Precision Oncology Sequencing Platform

Authors

Luc G. T. Morris

Raghu Chandramohan

Lyndsay West

Ahmet Zehir

Debyani Chakravarty

David G. Pfister

Richard J. Wong

Nancy Y. Lee

Eric J. Sherman

Shrujal S. Baxi

Ian Ganly

Bhuvanesh Singh

Jatin P. Shah

Ashok R. Shaha

Jay O. Boyle

Snehal G. Patel

Benjamin R. Roman

Christopher A. Barker

Sean M. McBride

Timothy A. Chan

Snjezana Dogan

David M. Hyman

Michael F. Berger

David B. Solit

Nadeem Riaz

Alan L. Ho

Doi

10.1001/jamaoncol.2016.1790

PMID: 27442865 · DOI: 10.1001/jamaoncol.2016.1790 · Journal: JAMA Oncology (2017)

TL;DR

Morris et al. report the first large clinical-grade molecular landscape of recurrent and metastatic head and neck cancers, sequencing 151 advanced tumors at Memorial Sloan Kettering Cancer Center between January 2014 and July 2015 with the MSK-IMPACT 410-gene panel. Next-generation sequencing guided therapy in 21 of 151 patients (14%; 25% of HNSC cases), refined diagnoses in several rare cancer types (notably reclassifying acinic cell carcinomas to MASC via ETV6-NTRK3 fusions), and revealed that recurrent/metastatic HPV-positive HNSC frequently acquires an HPV-negative-like genotype enriched for TP53 mutation, whole-genome duplication and 3p deletion. Activating NOTCH1 mutations were enriched in metastatic adenoid cystic carcinoma (ACYC), and TERT promoter mutations were strikingly common in advanced HPV-negative HNSC, CSCC, and BCC.

Cohort & data

  • 151 patients (95 men, 56 women; mean age 61.8, range 17–100) with advanced, treatment-resistant head and neck tumors profiled at MSK between January 2014 and July 2015 (IRB protocol NCT01775072; 224 enrolled, 151 with ≥6-month follow-up).
  • Disease distribution: 53 HNSC, 56 salivary carcinomas (36 ACYC plus 20 other salivary), 27 cutaneous carcinomas (21 CSCC, 4 BCC, 2 skin adnexal), 8 NPC, 3 HNNE, 2 odontogenic carcinomas (AMBL/CCOC), and 2 olfactory neuroblastomas (ONBL).
  • Site distribution: 66 patients with locoregional recurrence, 106 with distant metastases, 6 with locally advanced surgically unresectable tumors.
  • Assay: MSK-IMPACT targeted NGS of 410 cancer-relevant genes (IMPACT410) plus genome-wide copy number, with matched normal DNA; median coverage 600× (range 82×–1165×); CLIA-approved.
  • Dataset: publicly released as hnc_mskcc_2016 (cBioPortal study).
  • Analyses: FACETS for allele-specific copy number and clonality; an SVM classifier trained on TCGA HNSC for HPV-status genotyping; mutational signature decomposition (APOBEC, tobacco, UV light); enrichment comparison versus the TCGA HNSC primary cohort and primary ACYC sequencing from Ho et al. 2013.

Key findings

  • Therapy guidance. NGS guided therapy in 21/151 patients (14%) overall and 13/53 (25%) with HNSC (PMID:27442865).
  • Actionable alterations. 28/135 patients (21%) harbored alterations with potential investigational implications: 7 (5%) with an FDA-approved biomarker–drug pair in another indication, 21 (16%) with clinical evidence linking the alteration to drug response.
  • HNSC landscape. Whole-genome duplication in 12/48 cases (25%) with informative copy number; mismatch-repair gene mutations in 4 tumors (MLH1 truncating, MSH2/MSH6/POLD1 missense) carried significantly elevated mutation counts (17.3 vs 4.5; P < .001).
  • TERT promoter enrichment in advanced HPV-negative HNSC. Mutations in 16/30 (53%) HPV-negative tumors versus 12/70 (17%) primary HPV-negative tumors (OR 5.5; P < .001); strikingly 10/11 (91%) HPV-negative tongue SCCs harbored TERT promoter mutations. Zero TERT promoter mutations in HPV-positive HNSC, consistent with E6-driven telomerase activation as a parallel mechanism.
  • HPV+ versus HPV− differences attenuated in advanced disease. No difference in mutation counts (4.4 vs 5.9; P = .42) or fraction of genome copy-number-altered (0.19 vs 0.18; P = .78) between HPV-positive (n = 21) and HPV-negative (n = 30) recurrent/metastatic HNSC, in contrast to primary tumors.
  • Genes enriched in HPV-negative recurrent/metastatic HNSC versus HPV-positive: TP53 (72% vs 15%; P < .001), TERT promoter (55% vs 0%; P = 9 × 10⁻⁵), CDKN2A mutation/deletion (37% vs 5%; P = .016).
  • “HPV-negative-like” HPV-positive recurrent/metastatic HNSC. An SVM classifier trained on TCGA correctly categorized 29/31 (91%) HPV-negative MSK-IMPACT cases but only 12/21 (57%) HPV-positive cases (P = .007); 43% of advanced HPV-positive tumors had genotypes resembling primary HPV-negative tumors and trended toward worse survival (HR 2.3; P = .19).
  • Alterations enriched in recurrent/metastatic HPV-positive HNSC versus TCGA primary HPV-positive: whole-genome duplication 6.7-fold (P = .024), 3p deletion 9.1-fold (P = .002), 3p deletion + TP53 co-occurrence 7.6-fold (P = .03); higher TP53 mutation rate (15% vs 3%; enrichment 4.6, P = .11) and lower PIK3CA mutation (10% vs 36%; enrichment 0.21, P = .06).
  • TP53 in HPV-positive tumors is often subclonal: 3 of 4 (75%) HPV-positive TP53 mutations were subclonal versus 4 of 27 (15%) HPV-negative (P = .03); all TP53 mutations in HPV-positive tumors were truncating.
  • Confirmation in cervical cancer. TP53 mutation frequency markedly higher in MSK-IMPACT recurrent/metastatic CESC (4/16, 25%) than primary TCGA cervix (5/170, 3%; OR 8.3; P = .008).
  • Chromosomal instability: TP53/RB1-mutated HPV-positive tumors had significantly more copy-number-altered genome than wildtype tumors (P = .009).
  • NOTCH1 enrichment in metastatic ACYC. NOTCH1 altered in 12/33 (33%) of recurrent/metastatic ACYC; activating mutations or amplifications in 8/36 (22.2%) versus 2/60 (3.3%) in primary ACYC (OR 8.3; P = .005); three samples had recurrent truncating alterations at serine 2467, mirroring leukemia activating mutations.
  • TERT promoter mutations identified in 5 ACYC (14%) — the first report of TERT promoter mutations in salivary cancers.
  • Salivary duct carcinoma (SDCA) actionable findings: PIK3CA, HRAS, ERBB2, and AR alterations plus structural variants in NF1 and ROS1.
  • MASC reclassification by ETV6-NTRK3 fusion changed the diagnosis of 2 cases originally called acinic cell carcinoma; both subsequently enrolled on a TRK inhibitor basket study with major / near-complete responses.
  • Cutaneous carcinomas. UV-light mutational signature in 18/21 (86%) CSCC, associated with higher mutation counts (37.9 vs 4.3; P = .008); 3p deletion + TP53 co-occurrence in 6/19 CSCC (32%) — the first report in cutaneous SCC. Three of 4 BCC had UV signature; all 3 hedgehog-pathway-inhibitor-responsive BCCs carried PTCH1 mutations.
  • Odontogenic carcinomas. Both profiled cases had EWSR1 gene fusions: an EWSR1-FLI1 fusion in an “ameloblastic carcinoma” reclassified as Ewing sarcoma with epithelial differentiation, and an EWSR1-ATF1 fusion in a clear cell odontogenic carcinoma (CCOC) — not previously reported in odontogenic tumors.
  • Germline finding. One HNSC patient harbored a germline heterozygous FANCA deletion plus a somatic FANCA stopgain — a “double hit” with potential cisplatin sensitivity.

Genes & alterations

  • TP53 — mutated in 21/29 (72%) HPV-negative versus 3/20 (15%) HPV-positive recurrent/metastatic HNSC; all HPV-positive TP53 mutations truncating and predominantly subclonal; co-occurrence with 3p deletion enriched 7.6-fold in advanced HPV-positive HNSC; 25% of recurrent/metastatic CESC (versus 3% TCGA primary).
  • TERT — promoter mutations in 53% of HPV-negative HNSC (91% of HPV-negative tongue SCC), 52% of CSCC, 75% of BCC, and 14% of ACYC (first report in salivary cancers); zero in HPV-positive HNSC.
  • NOTCH1 — altered in 33% of ACYC with activating mutations/amplifications in 22.2% (8.3-fold enriched versus primary ACYC); recurrent S2467 truncating events; nominates γ-secretase inhibitors.
  • PIK3CA — mutated in 10% of recurrent/metastatic HPV-positive HNSC (down from 36% in primary HPV+); 4 HNSC and 1 NPC and 1 ACYC patient enrolled on PI3K-inhibitor trials; also altered in SDCA.
  • PTEN — alteration prompted PI3K-inhibitor trial enrollment.
  • CDKN2A — mutation/deletion in 37% of HPV-negative versus 5% of HPV-positive recurrent/metastatic HNSC.
  • HRAS — mutated in 1 HNSC (treated on farnesyl-transferase-inhibitor trial); G13V in a mucoepidermoid carcinoma (MUCC) treated similarly; recurrent in SDCA.
  • RB1 — alterations in 3/20 (15%) recurrent/metastatic HPV-positive HNSC; co-mutation with TP53 associated with elevated chromosomal instability.
  • MAPK1 — mutation guided treatment with an ERK inhibitor on single-patient IND.
  • SMARCB1 — deletion led to enrollment on an EZH2 inhibitor trial.
  • FGFR3 — mutation in 1 ACYC patient treated with an FGFR inhibitor.
  • PIK3R1 — mutation in 1 ACYC patient treated with PI3K inhibitor.
  • MDM2 — amplification in 1 ACYC patient treated with MDM2 inhibitor.
  • PDGFRA, KDR, KIT — co-amplification in 4 ACYC patients treated with regorafenib.
  • ERBB2, AR — actionable alterations in SDCA.
  • NF1, ROS1 — interesting structural variants in SDCA.
  • PTCH1 — mutations in 3 of 4 BCC, all responsive to hedgehog-pathway inhibitors before NGS.
  • MLH1, MSH2, MSH6, POLD1 — mismatch repair / proofreading defects in 4 hypermutated HNSC tumors.
  • ETV6-NTRK3 — fusion identified in 2 tumors initially diagnosed as acinic cell carcinoma, reclassifying them as MASC; both responded to TRK inhibitor on basket trial.
  • EWSR1-FLI1 — fusion in an odontogenic case reclassified as ES with epithelial differentiation.
  • EWSR1-ATF1 — fusion in a clear cell odontogenic carcinoma (CCOC); novel in odontogenic tumors.
  • FANCA — germline heterozygous deletion + somatic stopgain in one HNSC patient (potential cisplatin-sensitivity biomarker).

Clinical implications

  • Sequencing recurrent/metastatic disease matters: the genetic profile of advanced head and neck tumors is distinct from primary disease (notably for TERT promoter, TP53 and NOTCH1 enrichment, and HPV-positive tumors drifting toward an HPV-negative-like genotype).
  • Basket studies are the principal vehicle for therapeutic gain in this rare-cancer-rich population — basket trials accounted for half of clinical-trial enrollments in this cohort. Notable responses include MASC patients on TRK inhibitors and ACYC patients matched to PI3K, FGFR, MDM2, and γ-secretase inhibitors based on actionable alterations.
  • Diagnosis refinement: NGS reclassified 2 acinic cell carcinomas as MASC (via ETV6-NTRK3) and 1 ameloblastic carcinoma as ES (via EWSR1-FLI1), enabling appropriate targeted/cytotoxic therapy choice.
  • Negative biomarker → immunotherapy: in 6/53 (12%) HNSC patients, the lack of actionable alterations on MSK-IMPACT prompted enrollment on immunotherapy trials; one hypermutated BCC (122 mutations) was directed to immunotherapy on the basis of mutation load.
  • Prognostic signal: “HPV-negative-like” HPV-positive HNSC trended toward worse survival (HR 2.3; P = .19) — relevant to ongoing chemoradiotherapy de-escalation strategies in HPV-positive disease.
  • Hypothesis-generating predictive biomarkers for immune checkpoint inhibitors enumerated by the authors: intratumor heterogeneity (low ITH), high mutational load, and tobacco/UV-light mutational signatures.
  • HNSC therapeutic landscape context: the authors note that no new targeted therapies have been approved for HNSC since cetuximab in 2006 and there are no validated predictive biomarkers — this work motivates routine NGS to expand options.

Limitations & open questions

  • Single-institution cohort with modest sample sizes for individual rare-cancer subtypes (e.g., 4 BCCs, 2 odontogenic carcinomas, 2 olfactory neuroblastomas), limiting broad conclusions; the authors call for multi-institutional pooling.
  • Targeted 410-gene panel does not capture exome-wide mutational load with the resolution of WES; deep sequencing may artifactually inflate sensitivity for low-allele-fraction variants like TERT promoter mutations.
  • Cross-cohort comparisons (MSK-IMPACT versus TCGA) are confounded by differences in sequencing platform, depth, analytic pipeline, and patient selection — enrichment estimates require independent confirmation.
  • The “HPV-negative-like” subset of advanced HPV-positive HNSC (43%) has only a trend toward poorer survival (P = .19); larger cohorts are needed to confirm prognostic significance and to test whether de-intensified therapy is appropriate.
  • TP53 mutations in HPV-positive tumors are often subclonal, raising the question of whether they are late events driven by tobacco exposure or selected under treatment pressure — biology and timing are unresolved.
  • No standard-of-care biomarkers exist for head and neck cancer; only 21% of cases harbored potentially actionable alterations, leaving the majority of patients without molecularly matched options.
  • HPV status was not available for all 53 HNSC cases (51 with known HPV status); some downstream comparisons are limited by sample dropout.
  • Biological role of TERT promoter mutations in ACYC is undefined — first description in salivary cancer; clinical consequences unknown.
  • Generalizability of activating NOTCH1 mutations in ACYC as a predictive biomarker for γ-secretase inhibitors awaits prospective testing.

Citations from this paper used in the wiki

  • “Next-generation sequencing ultimately guided therapy in 21 of 151 patients (14%) (13 of 53 [25%] of patients with HNSCC) by refining diagnoses and matching patients to specific therapies, in some cases with dramatic responses on basket studies. Molecular alterations were potentially actionable in 28 of 135 patients (21%).” — Abstract, PMID:27442865.
  • “There were high rates of TERT promoter mutation in recurrent and metastatic HPV-negative HNSCC (13 of 30 tumors [43%]), cutaneous SCC (11 of 21 tumors [52%]), basal cell carcinoma (3 of 4 tumors [75%]), and ACC (5 of 36 tumors [14%]). Activating NOTCH1 mutations were enriched in metastatic ACCs (8 of 36 tumors [22%]).” — Abstract, PMID:27442865.
  • “The promoter region of TERT harbored mutations in 16 of 30 HPV-negative tumors (53%) … In this recurrent and metastatic population, mutations in the TERT promoter were significantly more frequent than rates reported in primary HPV-negative HNSCCs (12 of 70 [17%]) (odds ratio (OR), 5.5; P < .001) … Remarkably, 10 of 11 (91%) HPV-negative tongue SCCs had TERT mutations.” — Results, p. 4.
  • “We used a support vector machine algorithm to categorize tumors as more similar to HPV-positive or HPV-negative profiles … The classifier correctly categorized 29 of 31 (91%) HPV-negative MSK-IMPACT cases, but only 12 of 21 (57%) HPV-positive cases (P = .007) … 43% of advanced HPV-positive HNSCC tumors had genotypes that resembled primary HPV-negative tumors.” — Results, pp. 5–6.
  • “In the recurrent and metastatic HPV-positive tumors, there was significant enrichment for whole-genome duplication (6.7-fold; P = .024), 3p deletion (9.1-fold; P = .002), and 3p deletion-TP53 mutation (7.6-fold; P = .03).” — Results, p. 6.
  • “We identified significant enrichment of NOTCH1 activating alterations (8 of 36 tumors [22.2%] vs 2 of 60 tumors [3.3%]; OR, 8.3; P = .005) … Activating NOTCH1 mutations confer sensitivity to γ-secretase inhibitors, nominating these drugs as potentially active in NOTCH1-mutated ACCs.” — Results, p. 7.
  • “Sequencing identified an ETV6-NTRK3 fusion … changing the diagnosis to MASC. The patient was then enrolled on a TRK inhibitor basket study, with a clinical near-complete response.” — Results, p. 8.
  • “Complete genomic and clinical data from this study are available in searchable form at http://www.cbioportal.org/study?id=hnc_mskcc_2016.” — Results, p. 9.

This page was processed by crosslinker on 2026-05-14.