Pulmonary Function Changes After Radiotherapy for Lung or Esophageal Cancer:A Systematic Review Focusing on Dose-Volume Parameters

Background. Despite technical developments in treatment delivery, radiation-induced lung toxicity (RILT) remains a crucial problem in thoracic radiotherapy. Clinically based RILT scores have their limitations, and more objective measures such as pulmonary functions tests (PFTs) might help to improve treatment strategies.Purpose. To summarize the available evidence about the effect of dose to the lung in thoracic radiotherapy on forced expiratory volume in one second (FEV1) and diffusion capacity (DLCO) in patients with lung and esophageal cancer treated with curative intent.Material and methods. A systematic review following the PRISMA guidelines was performed, using MEDLINE and including clinical studies using (chemo) radiotherapy (CRT) or stereotactic ablative radiotherapy (SABR) for lung or CRT for esophageal cancer that reported both lung dose-volume histogram (DVH) parameters and changes in PFT results. Search terms included lung and esophageal neoplasms, respiratory function tests, and radiotherapy.Results. Fifteen studies met the inclusion criteria. Seven out of 13 studies on lung cancer reported significant declines (defined as a p value <.05) in PFT results. Both esophageal studies reported significant DLCO declines. One SABR study found a correlation between low lung-dose parameters and FEV1 decline. Relations between decline of FEV1 (three studies) or decline of DLCO (five studies), respectively, and DVH parameters were found in eight studies analyzing CRT. Furthermore, a heterogeneous range of clinical risk factors for pulmonary function changes were reported in the selected studies.Conclusion. There is evidence that pulmonary function declines after RT in a dose-dependent manner, but solid data about lung DVH parameters predicting changes in PFT results are scarce. A major disadvantage was the wide variety of methods used, frequently lacking multivariable analyses. Studies using prospective high-quality data, analyzed with appropriate statistical methods, are needed.