Lee et al.
Journal of Clinical Oncology 2024
30 ROC Trial (NCT03323463)

Background

  • HPV-related oropharyngeal cancers have favorable outcomes but cause significant long-term toxicity
  • Standard treatment: 70 Gy chemoradiotherapy results in dysphagia, xerostomia, dental complications
  • Previous de-escalation trials failed:
    • Cetuximab substitution → worse outcomes (RTOG 1016, De-ESCALaTE)
    • Dose reduction to 60 Gy → inferior oncologic outcomes (NRG HN-002)
  • Tumor hypoxia causes radioresistance through reduced free radical production
  • FMISO PET can measure tumor hypoxia

De-escalation Trials in HPV+ OPC: Learning from Our Failures

Why De-escalate?

  • HPV+ OPC has 85-95% cure rates with standard therapy
  • Younger patients living decades with treatment toxicity
  • Dysphagia, xerostomia, dental complications affect QOL
  • Goal: Maintain cure rates while reducing morbidity

Three Main De-escalation Strategies:

1. Replace Cisplatin

  • RTOG 1016
  • De-ESCALaTE
  • Result: FAILED

2. Reduce RT Dose

  • NRG HN-002 (Phase II)
  • NRG HN-005
  • Result: MIXED

3. Surgery + Reduced Adjuvant

  • ECOG 3311
  • MC1675
  • PATHOS
  • Result: PROMISING

Key Learning: Not all de-escalation strategies are equal. Biology-based selection (like 30 ROC) may succeed where unselected de-escalation failed.

RTOG 1016 & De-ESCALaTE: Cetuximab Cannot Replace Cisplatin

RTOG 1016 (n=849)

Design: 70 Gy (6 weeks) + Cisplatin vs Cetuximab

5-year OS: 85% vs 78% (p=0.01)

5-year PFS: 78% vs 67%

5-year LRF: 10% vs 17%

Toxicity: Similar (no benefit)

De-ESCALaTE (n=304)

Design: 70 Gy + Cisplatin vs Cetuximab

2-year OS: 98% vs 89%

2-year recurrence: 6% vs 16%

Toxicity: Similar (no benefit)

Note: Excellent 94% control with cisplatin

Clear Conclusions:

  • Cetuximab is inferior to cisplatin - period
  • No toxicity benefit despite hopes
  • Standard chemoRT achieves 90% local control
  • Message: Keep cisplatin for HPV+ disease

The 60 Gy Story: Context for 30 ROC's Bold Design

Timeline: The Optimistic Era of Dose Reduction

  • 2017: HN-002 shows promising 60 Gy results → optimism builds
  • 2017: 30 ROC opens (NCT03323463) - betting on biology over empiric reduction
  • 2018: HN-005 launches with 60 Gy arm based on HN-002
  • 2024: HN-005 60 Gy arm closed for futility

The 60 Gy Hope (2017 Perspective)

HN-002 Phase II Results:

  • Selected patients (T1-2, ≤10 PY)
  • 2-yr PFS 90.5% with 60 Gy
  • Met acceptability criteria
  • 14% dose reduction seemed safe

"If 60 Gy works, why not go further?"

The 30 ROC Vision (2017)

Different Philosophy:

  • Don't reduce dose empirically
  • Use biology (hypoxia) to select
  • 57% reduction needs safety net
  • Test mechanism, not just dose

"Some tumors need 70 Gy, some don't"

Two Parallel Philosophies Emerge:

Empiric Dose Reduction (HN-005)

  • Treat everyone the same
  • Hope selection criteria work
  • Modest 14% reduction
  • Result: Failed (selection? dose?)

Biology-Driven Selection (30 ROC)

  • Test each tumor's biology
  • Personalize based on hypoxia
  • Radical 57% reduction if safe
  • Result: 95% local control

The Key Insight:

30 ROC investigators understood something fundamental: successful de-escalation requires knowing WHY you can de-escalate, not just hoping a lower dose works.

HN-005's struggles (patient selection? dose? both?) vindicate the 30 ROC approach of using tumor biology rather than clinical criteria alone.

Study Hypothesis

  • Tumors without hypoxia can be treated with significantly lower radiation doses
  • FMISO PET can identify patients eligible for de-escalation from 70 Gy to 30 Gy
  • 30 Gy dose based on curative dose for HPV-related anal cancer
  • Pilot study (n=19): 87% pathologic response with 30 Gy in non-hypoxic tumors

Methods: Study Design

  • Phase II, single-arm trial
  • Primary endpoint: 2-year locoregional control
  • Non-inferiority design with 7% margin (historical control 95%)
  • 158 patients enrolled (152 eligible)
  • Multi-site within Memorial Sloan Kettering network

Eligibility Criteria

  • HPV-positive oropharyngeal carcinoma (p16+ or HPV RNA ISH+)
  • T0-2, N1-2c, M0 (AJCC 7th edition)
  • Age ≥18 years
  • ECOG 0-2
  • Able to receive high-dose cisplatin or carboplatin/5-FU

What is FMISO?

  • 18F-Fluoromisonidazole - a 2-nitroimidazole radiotracer
  • Most validated PET tracer for hypoxia imaging
  • Half-life: 109.8 minutes (ideal for delayed imaging)
  • Moderate lipophilicity allows cellular uptake

Mechanism of Action

  • Enters cells via passive diffusion
  • In normoxic cells (pOâ‚‚ >10 mmHg):
    → Re-oxidized and washed out
  • In hypoxic cells (pOâ‚‚ <10 mmHg):
    → Undergoes nitroreductase-mediated reduction
    → Forms reactive intermediates
    → Covalently binds to macromolecules
    → Trapped in viable hypoxic cells

Imaging Protocol

  • Patient prep: 4-6 hour fast, good hydration
  • Dose: 2-4 MBq/kg (150-400 MBq typical)
  • Critical: 2-4 hour delay before imaging
    → Allows clearance from normoxic tissue
    → Maximizes hypoxic contrast
  • 10-20 minute acquisition per bed position

Interpretation (30 ROC Trial)

  • Hybrid method (visual + quantitative)
  • Tumor-to-background ratio (TBR) >1.3
  • Visual assessment of 4 image characteristics
  • Visual prevails if disagreement
  • Excellent inter-observer agreement (Îş = 0.859)

Key Point: FMISO specifically accumulates in viable hypoxic cells (not necrotic tissue), providing a functional map of radioresistant tumor regions that can guide personalized dose escalation or de-escalation strategies.

Treatment Allocation

  • 30 Gy cohort (84% of patients):
    • Baseline FMISO negative OR
    • Hypoxia resolution on repeat FMISO
    • 15 fractions with concurrent chemotherapy
  • 70 Gy cohort (16% of patients):
    • Persistent hypoxia on repeat FMISO
    • 35 fractions with concurrent chemotherapy
  • Chemotherapy: Cisplatin 100 mg/m² or Carboplatin/5-FU
  • All patients had surgical removal of primary tumor

Radiation Planning: MSKCC Contouring & Dose Specifications

Target Volume Definitions

CTV Primary (Postoperative bed):
  • Post-op cavity + 5-10 mm margin
  • Include pre-op GTV extent
  • Tonsil: Include ipsilateral pterygoid plates
  • BOT: Include pre-epiglottic space
  • Unknown primary: Entire oropharyngeal axis

CTV Nodal:

  • Gross nodes + 5 mm (no additional margin if >3cm)
  • Entire involved nodal level
  • No intentional ECE coverage required

CTV Elective:

  • Bilateral levels II-IV (standard)
  • Level IB if oral cavity involvement
  • Retropharyngeal if posterior pharyngeal wall
  • Well-lateralized T1-2 tonsil + single node → ipsilateral only

Dose Prescription

ALL PATIENTS - First 30 Gy:

  • PTV_3000: All volumes (primary, nodes, elective)
  • 2 Gy Ă— 15 fractions = 30 Gy
  • Delivered regardless of hypoxia status

HYPOXIA NEGATIVE/RESOLVED (84%):

  • Stop at 30 Gy total dose
  • 3 weeks total treatment time

PERSISTENT HYPOXIA (16%):

  • PTV_7000: Hypoxic gross nodes only
  • Additional 2 Gy Ă— 20 fractions = 40 Gy boost
  • Total 70 Gy to hypoxic nodes
  • 7 weeks total treatment time

Planning Specifications

  • Technique: IMRT (VMAT acceptable)
  • PTV margins: 3-5 mm from CTV
  • Coverage: 95% of PTV receives 95% of Rx dose
  • Hot spots: <107% within PTV, <110% point max
  • Plan review: 2 radiation oncologists (N.Y.L. and N.R.)

Critical Innovation: Universal elective dose of 30 Gy (vs historical 50-63 Gy) dramatically reduces toxicity while maintaining excellent regional control. This represents a paradigm shift in elective nodal treatment.

Patient Characteristics

  • Median age: 59 years (90% male)
  • Primary site: Tonsil 55%, BOT 27%, Unknown 18%
  • Stage: 89% stage IVA (AJCC 7th)
  • Never smokers: 53%
  • Baseline hypoxia: 72% (110/152 patients)
  • Hypoxia resolution: 78% (86/110 patients)
  • Smokers 3x more likely to have persistent hypoxia (p=0.02)

Primary Outcome: Locoregional Control

 

  • 2-year LRC: 94.7% (95% CI: 89.8-97.7)
  • Met primary objective (>88% lower bound)
  • Median follow-up: 38.3 months

Secondary Outcomes

  • 30 Gy cohort (n=128):
    • 2-year PFS: 94%
    • 2-year OS: 100%
    • 8 nodal recurrences (all salvaged with neck dissection)
    • 1 distant metastasis
  • 70 Gy cohort (n=24):
    • 2-year PFS: 96%
    • 2-year OS: 96%
    • No locoregional failures
  • No primary site recurrences in either cohort

Acute Toxicity

  • Grade 3-4 acute toxicity: 32% (30 Gy) vs 58.3% (70 Gy), p=0.02

Late Toxicity and PROs

  • Grade 3-4 late toxicity only in 70 Gy cohort (4.5% dysphagia)
  • Grade 2 late toxicity (entire cohort):
    • Xerostomia: 2.6%
    • Dysgeusia: 5.2%
    • Dysphagia: 1.4%
  • Modified barium swallow at 1 year: No moderate dysphagia in 30 Gy cohort
  • MDADI scores improved from baseline to 1 year

Late Toxicity and PROs

Exploratory Biomarkers

  • MRI parameters (ADC, volume) did not predict hypoxia status
  • ctDNA detectable in 64% at 2 weeks → cannot replace FMISO PET
  • Financial toxicity: 63% reduction in healthcare costs (30 Gy vs 70 Gy)

Surgery in HPV+ OPC: Why 30 ROC Went Against the Grain

The Radiation Oncology View of Surgery:

  • ORATOR2 terminated early: 4% surgical mortality (2/35 patients died from post-op hemorrhage)
  • ECOG 3311: Only 0.2% mortality (1/495) - but was this cherry-picked excellence?
  • Definitive chemoRT: Zero surgical mortality risk, 85-95% cure rates
  • The verdict: Why risk surgery when radiation works perfectly?

Evidence Against Surgery:

  • ORATOR2: PFS 100% (RT) vs 84% (surgery)
  • Positive margins: 8% requiring adjuvant RT anyway
  • Variable surgical quality across centers
  • 78% still needed radiation after surgery
  • Two treatments = two sources of morbidity

Why Some Still Pursue Surgery:

  • ENTs: "ORATOR2 doesn't represent good surgery"
  • High-volume centers claim better outcomes
  • Pathologic staging information
  • Potential to avoid chemotherapy (30-40%)
  • Patient preference for "cutting it out"

So why did 30 ROC require surgery? Not because Nancy Lee believes in surgery for HPV+ disease, but because 30 Gy was too radical to test without removing the primary first.

The Context of Dose De-escalation Failures:

  • RTOG 1016 & De-ESCALaTE: Cetuximab failed → stick with cisplatin
  • NRG HN-002: 60 Gy looked promising in Phase II
  • NRG HN-005: 60 Gy arm closed for inferiority → need 70 Gy
  • Lesson learned: Even modest de-escalation (60 Gy) can fail

The 30 ROC Gamble:

  • Standard: 70 Gy (proven)
  • Failed attempt: 60 Gy (14% reduction)
  • 30 ROC: 30 Gy (57% reduction!)
  • 4x more aggressive than failed trials
  • Based on anal cancer data

Why Surgery Made It Possible:

  • Primary site failure → catastrophic
  • Re-irradiation to 70+ Gy → severe toxicity
  • Surgical salvage of primary → high morbidity
  • But nodal salvage → feasible
  • IRB approval required this safety net

The Investigators' True Intent:

This was NOT a surgery trial. This was a hypoxia-directed dose trial.

  • Surgery = enabling tool, not the intervention being tested
  • FMISO PET = the real innovation
  • 84% could get 30 Gy based on biology
  • Future goal: Apply to intact primaries once proven safe

"Given that 60 Gy failed, testing 30 Gy on intact primaries would have been unethical. Surgery wasn't philosophically preferred - it was pragmatically necessary."

What 30 ROC Teaches Us (Despite the Surgery):

Biological Insights:

  • FMISO PET reliably identifies hypoxia
  • 84% of HPV+ tumors lack persistent hypoxia
  • Non-hypoxic disease needs much less dose
  • 30 Gy controls non-hypoxic nodes perfectly
  • Biology > clinical factors for selection

Technical Advances:

  • 30 Gy to elective volumes is safe
  • Dramatic toxicity reduction achieved
  • FMISO interpretation standardized
  • Multi-site feasibility demonstrated
  • Financial toxicity reduced 63%

The Path to Non-Surgical Implementation:

Next Steps for Radiation Oncology:

  1. Dose finding for primaries: Test 50-60 Gy for FMISO-negative intact tumors?
  2. Selective boost strategy: 50 Gy whole field + 20 Gy to hypoxic volumes only?
  3. Combine biomarkers: FMISO + ctDNA + MRI for better selection?
  4. Start conservatively: Small T1N0-1 primaries first?
  5. EORTC "Best of" results: Will definitive answers on surgery vs RT debate

What NOT to Take from 30 ROC:

• Surgery is necessary for HPV+ OPC
• TORS should be routine
• 30 Gy is ready for intact primaries
• This validates surgical approaches

What TO Take from 30 ROC:

• Biology-based selection works
• Massive de-escalation is possible
• FMISO PET is clinically ready
• Future is personalized dose

Bottom Line: 30 ROC used surgery as a necessary safety measure, not an endorsement. The real advance is proving hypoxia-directed therapy works. Now we must adapt this to our non-surgical patients.

Strengths

  • First personalized radiotherapy trial using functional imaging
  • Dramatic dose reduction (57%) with maintained efficacy
  • Excellent inter-observer agreement for FMISO interpretation
  • Comprehensive toxicity assessment (PRO + objective)
  • Median follow-up >3 years
  • Multi-site feasibility demonstrated

Limitations

  • Single-arm design (no randomized comparison)
  • Selected population (surgical candidates, T0-2)
  • Primary tumor removed (limits generalizability)
  • FMISO PET not widely available
  • Limited evidence for FMISO as prognostic biomarker in HPV+ disease
  • Surgical morbidity not accounted for

Discussion Points

  1. Is the 7% non-inferiority margin appropriate for this dramatic dose reduction?
  2. How does primary tumor resection impact the generalizability?
  3. What are the barriers to implementing FMISO PET in clinical practice?
  4. Should we wait for Phase III results or consider early adoption?
  5. How does this compare to other de-escalation strategies (immunotherapy, protons)?
  6. What is the role of salvage therapy planning in de-escalation trials?