Article Text
Abstract
Introduction/objectives The majority of patients are diagnosed with lung cancer following an abnormal chest X-ray (CXR) ordered in response to respiratory symptoms. However, CXR will miss some symptomatic lung cancers, and such patients may be expected to present with lung cancer in the months after a non-diagnostic CXR. Using information from a large dataset of patients presenting for self-request CXR (SR-CXR), we sought to determine outcomes from lung cancers diagnosed at different time intervals following an initial SR-CXR performed for symptoms as described by NICE lung cancer guidelines (CG24).
Methods We reviewed 8,986 SR-CXRs carried out between 2011 and 2015. We identified 143 patients diagnosed with lung cancer within 24 months of SR-CXR. Patients were grouped according to the time interval between SR-CXR and lung cancer diagnosis (0–3, 4–12 and 13–24 months) and cancer outcomes (stage, treatment and survival) were compared.
Results 64, 38 and 42 patients respectively were diagnosed with lung cancer during these three time intervals. The proportions of patients with early stage disease (I+II combined) were 22%, 42% and 38%, respectively. The proportion of patients undergoing radical treatment (combined surgical and radical oncology) were 35%, 53% and 41% respectively.
Survival curves are shown in figure 1, with no indication of a difference in survival between the three groups. Using survival of those patients diagnosed within the first 3 months as reference, the hazard ratio for the 4–12 month group was 0.78 (95% CI 0.50 to 1.23) and for the 13–24 months group was 1.14 (95% CI 0.72 to 1.73).
Conclusions There was no obvious association between lung cancer outcomes and the length of time between SR-CXR and lung cancer diagnosis. Making the (untested) assumption that lung cancers presenting within three months represent a true positive CXR result, and those diagnosed between 4 and 12 months represent a false negative CXR result, there do not appear to be worse outcomes for those missed on initial CXR. Further work will seek to categorise each individual CXR result to allow a more definitive assessment of CXR sensitivity and specificity for lung cancer.