KEY POINTS
- POLCAR evaluated a simple symptom score in 61 lung cancer patients aged ≥65 years undergoing thoracic RT. Median age was 73 years; 50 patients had NSCLC and 11 had SCLC, and median RT dose was 62 Gy.
- The patient-completed score assigns 0–3 points each for cough, dyspnea, and temperature, producing a total score from 0–9. Symptoms were recorded during RT and for up to 24 weeks afterward.
- Eight patients (13.1%) developed radiation pneumonitis, at a median of 12.5 weeks after RT. Their median maximum symptom score was 5 versus 3 in patients without pneumonitis (p=0.001).
- The absolute symptom score showed good discrimination, with an AUC of 0.85 (95% CI 0.73–0.97). A threshold of ≥4 points provided 87.5% sensitivity, 66.0% specificity, 28.0% positive predictive value, and 97.2% negative predictive value.
- Raising the threshold to ≥5 points improved specificity to 90.6% and positive predictive value to 50%, but sensitivity fell to 62.5%. The authors therefore favored four points for screening and five points when fewer false-positive alerts are preferred.
- Change from baseline also performed well. An increase of ≥2 points detected all eight pneumonitis cases, with 100% sensitivity and 56.6% specificity; its AUC was 0.89, but this was not significantly better than the absolute score (p=0.265).
- Patient acceptance was high: only 3/56 respondents (5.4%) were classified as dissatisfied. However, threshold estimates were derived from only eight events in the same cohort in which performance was evaluated, making external validation essential.
CLINICAL TAKEAWAY
A very simple cough–dyspnea–temperature score could help older lung cancer patients flag possible pneumonitis between scheduled follow-up visits. A score of ≥4 should trigger clinical assessment, not diagnose pneumonitis, and the low event count means this threshold should remain provisional until independently validated.