Procedure risks and the cost of delay

(Agent report, 2026-09-03. Reviewed by Kya at synthesis.)

EEA biopsy-only series

Zhang 2023 (11): 10/11 dx, 1 epistaxis return to theatre. Liu&Di 2009 (10): 10/10, 1 ICA injury (balloon), 1 transient CSF leak. Kelley 1999 (7): 5/6, 1 CSF leak, 1 bleed. Graillon 2014 (3): 3/3, no complications, mean stay 4 d. Palejwala 2018 Meckel's cave (6): 6/6. Pooled ~37: non-dx ~9%, ICA 1/37, CSF leak 2/37, death 0.

EEA resection of clival chordoma — meta-analyses

• Arend 2026 Oper Neurosurg (18 studies/1197): GTR 48.3%; CSF leak 12.4%; CN palsy 7.5%; infarction 1.2%; major haemorrhage 0.3%.
• Komotar 2011 (766): endoscopic vs open GTR 61 vs 48%; CN deficits 1.3 vs 24.2%; meningitis 0.9 vs 5.9%; mortality 4.7 vs 21.6%.
• Cannizzaro 2021: CSF leak 10.3% endo. Gupta 2022 (363): 10.1%. Brown 2023 (2453 all approaches): CSF leak 5.4%, periop mortality 2.5%.
• Cavallo 2021 Italian multicentre (182): complications 14.3%; CSF leak 4.9%; ICH 3.8%; meningitis 2.7%; ischaemia 2.7%.
• Zoli 2018 (65): 15.1% complications, 0 deaths, ICA injury 2.5% (coiled, no deficit), permanent deficit 3.8%.
• Koutourousiou 2012 Pittsburgh (60): CSF leak 20%, new CN 6.7%, 0 mortality. 2017 CS tumours (234; 59 non-adenoma): new permanent CN palsy 11.9% non-adenoma; ICA injury 4, 1 death (0.4%).
• Perry 2019 (11,149 TSS): ICA injury 0.4% EEA; "Operator inexperience may be a more important risk factor than choice of surgical technique."
• Lai 2014: meningitis 1.8% overall; 13% if CSF leak, 0.1% if not.
• Soffer 2024 NCDB (736): EEA LOS 2.1 d shorter than open; size >=4cm HR 4.03 for survival.

Alternatives

• Percutaneous foramen ovale: Dellaretti 2017 SR 65/75 dx, 3 permanent deficits; Messerer 2012 86% productive; Graillon: "only provides cell analysis results, and not tissue analysis". May not reach clival-centred lesion.
• Open craniotomy: no biopsy-only series; Carrabba 2008 open resection neuro worsening 33% vs EEA 6%.

Yield

~90-95% first attempt. Brachyury IHC works on small biopsies (Jambhekar 2010: 46/51 chordomas +, 0/58 non-chordoma).

Cost of delay

• GTR limits: volume >20 cm3 (Koutourousiou 2012), >=40 cm3 (Wang 2017), paramedian/lateral OR 6.25 against GTR (Kim 2018). Abdallah 2021: low-risk tumours mean 0.98cc, GTR 24/25.
• Fiore 2024 meta-analysis (3871): 5-yr PFS 52% overall, 74% with GTR; OS 73% -> 84%; first surgery 2.13x disease-free.
• Di Maio 2011 (807): incomplete resection 3.83x recurrence, 5.85x death at 5 yr.
• Zou 2018 (1754): pre-op visual deficit HR 2.77 PFS; non-total resection HR 2.01 PFS / 5.16 OS.
• Proton LC vs GTV: McDonald 2016 GTV <=20 cm3 5-yr LC 81%; Fung 2018 GTV >25 mL HR 2.22; Ares 2009 PSI GTV >25 mL & brainstem compression worse; Weber 2016 PSI GTV independent; Potluri 2011 30 cm3 threshold.
• Passeri 2021: TGR >=10.12%/month -> lower PFS.
• Freeman 2018 MD Anderson: PFS 64 vs 25 months initial mgmt at MDT centre; "Biopsy followed by attempted radical resection at a dedicated center does not affect PFS and, therefore, represents a reasonable first step in management."
• Model: 2x3cm ≈ 6-8 cm3 now; at 10%/month doubles in ~7 months, crosses 20 mL in ~10 months.

Liquid biopsy

None validated. Mattox 2021 ctDNA needed tissue WES first. Zuccato 2022 plasma methylome AUROC 0.84 (CI 0.52-1.00), research only. Gill 2020: "Obtaining a sample for pathological analysis is key for a final diagnosis and should be done prior to the initiation of treatment efforts".

Observation

No primary-chordoma guidance for observing a symptomatic lesion; only incidental non-enhancing BNCT-type lesions. Stanford: MRI usually sufficient to suspect → proceed to biopsy+resection same sitting (i.e. still tissue).