Aims
This secondary analysis of a multicentre HPVOPC study aimed to identify patient-reported trajectories during and following CRT.
Methods
All 182 TROG12.01 patients were included. HNSS and HNSI (MDASI-HN), HRQL (FACT-G) and emotional distress (HADS) were assessed before, during and after CRT. Latent class growth mixture modelling (LCMM) was performed to identify trajectories.
Results
The HNSS model identified 4 trajectory classes (HNSS1-4) distinguished by differences at baseline, at the peak of treatment symptoms and during early (to 9w post CRT) and intermediate recovery (9w to 12m post CRT). All trajectories were stable beyond 12m. The reference trajectory (HNSS4, n=74) score was 0.1 (CI 0.1-0.2) at baseline, peaking at 4.6 (CI 4.2-5.0), with rapid early recovery (1.1, CI 0.8-2.2) and improvement to 12m (0.6, CI 0.5-0.8). HNSS2 (“high baseline", n=30) reported higher baseline scores (1.4, CI 0.8-2.0) but was otherwise similar, HNSS3 (“low acute”, n=53) reported reduced acute symptoms (2.5, CI 2.2-2.9) with stable scores beyond 9w post CRT (1.1, CI 0.9-1.4). HNSS1 (“slow recovery”, n=25) had slower recovery from an acute peak of 4.9 (CI 4.3-5.6) to 0.9 (CI 0.6-1.3) at 12m. Compared to HNSS4, the high baseline group were younger (mean 53.6 vs 57.4), with fewer ECOG0 (87 vs 99%) or with higher degree (38 vs 69%), with higher baseline anxiety (HADS, mean 6.8 vs 4.4). The low acute group were older (mean 60.1), had fewer ECOG0 patients (87%) and were more likely to have received cetuximab (68% vs 43%). The slow recovery group had fewer with a higher degree (52%).
LCMM also identified 4 HNSI, 2 HRQL, 3 anxiety and 4 depression trajectories, which will be presented.
Conclusion
LCMM identified 4 distinct HNSS trajectories which in combination with variations in patient characteristics may be clinically useful in identifying patients requiring increased support during and after CRT.