The Radiation Oncology Feed For radiation oncologists · physicists · RTTs · dosimetrists
Worth Attention toxicity mucositis dysphagia

Clinicians disagreed substantially on maximum acute toxicity during head and neck radiotherapy

Two blinded oncologists often assigned different peak toxicity grades, including grade 3 mucositis rates of 40.5% versus 17.6%.

RadOncToday · ·1 min read

KEY POINTS

  • This predefined prospective subgroup analysis was embedded within an ongoing randomized trial. Of the first 100 enrolled patients, 74 were evaluable after excluding 22 with insufficient dual-observer assessments, two who did not complete treatment, and two who withdrew.
  • Patients had head and neck squamous cell carcinoma treated with curative-intent VMAT. 56.8% received concurrent chemotherapy, 51.4% were treated postoperatively, and the commonest primary sites were oral cavity (52.7%) and oropharynx (25.7%).
  • Two radiation oncologists with more than five years of head and neck experience independently examined every patient weekly, blinded to each other's scores and within 24 hours of one another. Dermatitis, oral mucositis, dysphagia, and xerostomia were graded using CTCAE v5.0.
  • For maximum grade ≥2 toxicity, observer-reported incidence was 55.4% versus 45.9% for dermatitis, 91.9% versus 70.2% for mucositis, 95.9% versus 91.9% for dysphagia, and 55.4% versus 58.1% for xerostomia. Corresponding κ values were only 0.33, 0.35, 0.41, and 0.23.
  • The most striking discrepancy involved grade ≥3 oral mucositis: one observer classified 40.5% of patients as grade ≥3, versus only 17.6% by the other, despite both working at the same centre. Agreement remained only moderate at κ = 0.41.
  • Grade ≥3 agreement was better for dermatitis (κ = 0.79) and dysphagia (κ = 0.63) but poor for xerostomia (κ = 0.25). Low event frequencies for dermatitis and xerostomia make these κ estimates less stable.
  • Pooled weekly grade ≥2 assessments were considerably more consistent, with κ ranging from 0.49 to 0.89 and the highest concordance during weeks 5–6. However, these weekly estimates included repeated observations from the same patients and were therefore treated as descriptive rather than formal reliability statistics.

CLINICAL TAKEAWAY

“Maximum grade 3 toxicity” looks objective in a trial table but may depend substantially on who performs the assessment. Trials using toxicity as a primary endpoint should consider longitudinal scoring, objective clinical events, standardized assessment procedures, and patient-reported outcomes alongside peak clinician-assigned grade.

SOURCE

International Journal of Radiation Oncology, Biology, Physics