What determines the outcome

For her partner, 8 September 2026. Answer to: is age and fitness, or the spread and size of the tumour, the bigger factor in outcome and recovery? Every claim is sourced in research thread 19 (53 papers, PubMed IDs checked). Figures are relative effects from published series, not predictions for the patient. Where something is my judgement it says so.

The ranking, strongest first

1. What the tumour is under the microscope. The biggest single factor, and the one that changes the disease rather than the odds. Conventional chordoma is a slow, local problem where surgery plus full-dose radiotherapy buys years. Dedifferentiated (a high-grade sarcoma growing alongside the chordoma, usually with a TP53 mutation and loss of brachyury) and poorly differentiated (INI1 lost) chordomas spread through the blood and are treated more like sarcomas. TP53 mutation is found in 4 of 5 aggressive cases versus 1 in 100 conventional ones (Baluszek 2025). This is not known until tissue. Two honest notes: partial loss of INI1 on staining is seen in a third of ordinary conventional chordomas with no effect on outcome, so the poorly differentiated label needs the full picture, not the stain alone (Righi 2023); and by age alone the odds favour conventional, because poorly differentiated chordoma is rare above 29 and dedifferentiated peaks around 54. Age is not proof; the biopsy is.

2. How much the surgeons can remove. The strongest factor anyone can act on. Across 807 patients, incomplete removal meant 3.8 times the recurrence and 5.9 times the deaths at five years (Di Maio 2011); across 1,754, five times the risk of death (Zou 2018). Complete or near-complete removal is also what lets the radiotherapy work.

3. Molecular markers inside conventional chordoma. Loss of chromosome 1p36 and deletion of 9p21 (the CDKN2A gene) were the two independent predictors in 105 prospectively tested clival chordomas, with a wide spread of risk between the best and worst combinations (Zenonos 2018). A high Ki-67 (5 to 6 percent or more) is adverse in two series. These come from the same tissue as item 1 and are worth asking for by name.

4. Spread, which acts through items 2 and 6. Growth into the cavernous sinus, wrapping of the carotids or basilar, growth through the dura, contact with the brainstem and extension down the lower clivus each reduce how much can be removed (Passeri 2023, 210 patients). Contact with the brainstem or the optic pathways also forces the radiotherapy dose down, and in the Boston series that alone cut tumour control by nearly half (Hug 1999), confirmed in four later series. The patient has cavernous sinus and vessel involvement; whether the tumour touches the brainstem or the optic apparatus is the key question for the neuroradiologist, and it is on the list for the report.

5. Size. Four centimetres or more carried four times the risk of death in 736 US patients (Soffer 2024); above 20 to 25 ml, complete removal and radiotherapy control both fall (Koutourousiou 2012; Hug 1999). At 3.8 cm the patient is at the line on every cut-off. Ask for the measured volume in ml, not just the longest dimension.

6. The radiotherapy dose. Seventy Gy or more, and protons or carbon ions rather than standard photons for the definitive setting, improve outcome (Palm 2019; adjuvant radiotherapy cut progression by 70 percent in Zou 2018). Protons and carbon are equivalent head to head (Mattke 2023; Iannalfi 2020). Whether she gets the full dose depends on item 4.

7. Being untreated. Recurrence-free survival after a second operation is well below that after a first (Tzortzidis 2006). The patient is untreated, which is in her favour, and is the reason to get the first operation done at the unit that removes most.

8. Age. Each extra year adds about 3 percent to the hazard (Zou 2018; Jones 2014). Skull base chordomas present youngest and do best of any site. At 34 the patient is on the good side of every adult cut-off. This is a modest effect, real but smaller than everything above it.

9. Fitness, weight and nutrition. No chordoma-specific data. In 11,510 brain tumour operations a low albumin nearly doubled the odds of death within 30 days and raised the odds of a complicated discharge (Dasenbrock 2017). Frailty scores predict long stays and readmission after skull base surgery but not surgical complications, and the patient scores zero on them anyway (Kshirsagar 2024; Olsson 2025). Performance status helped in one small chordoma series. A complicated surgical course itself worsened long-term outcome in one series (Li 2026). Fitness and nutrition do not change what the tumour is; they change whether the surgery goes cleanly, and that is what sets the recovery time.

10. Sex. The series contradict each other. Unresolved.

So, to the question

Spread matters more than size, and both matter mainly through what they do to the operation and the radiotherapy dose. Age is a smaller effect, in her favour. Fitness and nutrition are the smallest tumour factor and the largest recovery factor, and the only ones you can move this week. Above all of them sits the histology, which nobody knows yet.

For recovery time specifically (the detail is in the recovery timeline document): complications and staged surgery lengthen it most; then nutrition and age; size and spread only through the operation they force.

What this means for the week

1. Protein and calories every day (yesterday and today show it can be done), and a dietitian before surgery. Albumin on the blood form.
2. Ask the neuroradiologist, in writing: does the tumour touch or compress the brainstem; does it touch the optic nerves or chiasm; what is the volume in ml; what is the relationship to the carotids (abutting or encasing).
3. Ask the surgeons: how much do they expect to remove, and what limits it in her case.
4. Ask for the tissue to be tested for brachyury, INI1, Ki-67, TP53, and for 1p36 and 9p21 loss (the Pittsburgh panel), and for whole genome sequencing through the NHS.
5. Ask the proton centre, when the time comes, whether the full 74 Gy(RBE) is deliverable or whether the brainstem or optic constraint will limit it.