Rades Dirk, Werner Elisa M, Glatzel Esther, Bohnet Sabine, Schild Steven E, Tvilsted Søren S, Janssen Stefan
Department of Radiation Oncology, University of Lubeck, 23562 Lubeck, Germany.
Department of Pulmonology, University of Lubeck, 23562 Lubeck, Germany.
Cancers (Basel). 2023 Jan 4;15(2):326. doi: 10.3390/cancers15020326.
Radiotherapy of lung cancer may cause pneumonitis that generally occurs weeks or months following therapy and can be missed. This prospective trial aimed to pave the way for a mobile application (app) allowing early diagnosis of pneumonitis. The primary goal was the identification of the optimal cut-off of a score to detect pneumonitis of grade ≥2 after radiotherapy for lung cancer. Based on the severity of symptoms (cough, dyspnea, fever), scoring points were 0−9. Receiver operating characteristic (ROC)-curves were used to describe the sensitivity and specificity. The area under the ROC-curve (AUC) was calculated to judge the accuracy of the score, Youden-index was employed to define the optimal cut-off. Until trial termination, 57 of 98 patients were included. Eight of 42 patients evaluable for the primary endpoint (presence or absence of radiation pneumonitis) experienced pneumonitis. AUC was 0.987 (0.961−1.000). The highest sensitivity was achieved with 0−4 points (100%), followed by 5 points (87.5%), highest specificity with 5−6 points (100%). The highest Youden-index was found for 5 points (87.5%). The rate of patient satisfaction with the symptom-based scoring system was 93.5%. A cut-off of 5 points was identified as optimal to differentiate between pneumonitis and no pneumonitis. Moreover, pneumonitis was significantly associated with an increase of ≥3 points from baseline (p < 0.0001). The scoring system provided excellent accuracy and high patient satisfaction. Important foundations for the development of a mobile application were laid.
肺癌放疗可能会导致肺炎，这种肺炎通常在治疗后数周或数月出现，且可能被漏诊。这项前瞻性试验旨在为一款能够实现肺炎早期诊断的移动应用程序（app）铺平道路。主要目标是确定一个分数的最佳临界值，以检测肺癌放疗后≥2级肺炎。根据症状（咳嗽、呼吸困难、发热）的严重程度，评分范围为0 - 9分。采用受试者工作特征（ROC）曲线来描述敏感性和特异性。计算ROC曲线下面积（AUC）以判断分数的准确性，使用约登指数来确定最佳临界值。直到试验结束，98名患者中有57名被纳入。42名可评估主要终点（放射性肺炎的有无）的患者中有8名发生了肺炎。AUC为0.987（0.961 - 1.000）。0 - 4分的敏感性最高（100%），其次是5分（87.5%）；5 - 6分的特异性最高（100%）。5分的约登指数最高（87.5%）。基于症状的评分系统的患者满意度为93.5%。确定5分为区分肺炎和无肺炎的最佳临界值。此外，肺炎与基线时分数增加≥3分显著相关（p < 0.0001）。该评分系统具有出色的准确性和较高的患者满意度。为移动应用程序的开发奠定了重要基础。
