Objectives: Late recognition of lung cancer is the major factor contributing towards its unsuccessful treatment. We conducted a prospective study to define any significant relationship of presenting symptoms with the diagnosis of lung cancer with a view to develop a model to identify those at high risk.
Methods: A consecutive series of 587 patients referred to our rapid access chest clinic with the suspicion of lung cancer were included. The presenting symptoms, chest x-ray findings and final diagnosis of all the patients were recorded. Chi-square and t-test were used for univariate analysis. A model was generated from logistic regression analysis and the discriminatory power of the model was assessed using area under receiver operator characteristic curve.
Results: Univariate analysis demonstrated that smoking, anorexia, weight loss and voice change were significantly more common in patients with lung cancer (p<0.05). Cough, expectoration and hemoptysis were significantly less common (p<0.05). Regression analysis qualified age, weight loss and smoking as significant predictors of lung cancer.
Conclusion: Only few of the historically accepted symptoms demonstrated a strong association with lung cancer and the model developed on these can form basis for a scoring tool that can perhaps help identify those at higher risk of cancer. Further refinement of the tool is required to accommodate cases presenting at primary care level.
Kaleem Ullah Toori, Ali Zohair Nomani, Maxwell Winson, Mati ur Rehman. (2015) Is there light at the end of the tunnel; symptoms and chest x-ray help identify patients at high risk of lung cancer, , Volume-40, Issue-1.
