Impaired Geriatric 8 Score is Associated with Worse Survival after Radiotherapy in Older Patients with Cancer

Clin Oncol (R Coll Radiol). 2021 Apr;33(4):e203-e210. doi: 10.1016/j.clon.2020.09.002. Epub 2020 Sep 22.


Aims: To investigate whether the Geriatric 8 (G8) score and the Timed Get Up and Go Test (TGUGT), together with clinical and demographic patient characteristics, are associated with survival and late toxicity after (chemo)radiation therapy, administered with curative intent in older patients with cancer.

Materials and methods: Four hundred and two patients aged ≥65 years (median age 72 years, range 65-96 years), diagnosed with either breast, non-small cell lung, prostate, head and neck, rectal or oesophageal cancer, and referred for curative (chemo)radiation therapy, took part in a multicentre prospective cohort study in eight radiotherapy centres in the Netherlands. The G8 and TGUGT scores were assessed before starting treatment. Other potential predictors and late toxicity were also recorded. Survival status and date of death, if applicable, were ascertained at the Dutch national death registry.

Results: After 2.5 years, the overall survival was 83%. Survival was 87% for patients with high G8 scores and 55% for patients with low G8 scores (Log-rank P value < 0.0001). Survival was 77% for patients with good TGUGT results and 50% for patients with poor TGUGT results (Log-rank P value < 0.001). In multivariable analysis, in addition to age and type of primary tumour, the association of the G8 score with overall survival remained, with a hazard ratio of 2.1 (95% confidence interval 1.2-3.8) for low versus high scores.

Conclusions: G8 was associated with overall survival in older patients with cancer irradiated with curative intent. This association was independent of the predictive value of age and primary tumour.

Keywords: Cancer; G8; TGUGT; elderly; radiotherapy; survival.

MeSH terms

  • Aged
  • Aged, 80 and over
  • Geriatric Assessment*
  • Humans
  • Male
  • Neoplasms* / epidemiology
  • Netherlands / epidemiology
  • Proportional Hazards Models
  • Prospective Studies