Interstitial lung diseases with concomitant lung cancer: a data mining approach

Among interstitial lung diseases, idiopathic pulmonary fibrosis (IPF) represents a condition rapidly progressive to respiratory failure and patients’s death. Notably, the incidence of lung cancer (LC) in patients already affected by IPF is significantly higher than in general population. Moreover these cases are often neglected and deprived of active oncologic treatments. We here aim at identifying ILD variables predictive of patient outcome by analysing present exhaustive data from 42 patients featuring LC and co-existing IPF. The entire dataset was then analysed by using the JMP partition algorithm which is able to search all possible splits of best response predictors and variable associated to outcome (survival). The row overall survival curve showed differences based on patients’ gender, being women survival lower than men’ one (Fig 1, a). The distribution of probabilities of the value “survival” has been to the five groups, representing the different first line therapeutic strategy. This analysis allows to point out that highest probability of survival occurs in the sub-population who underwent surgery against cancer. (Fig. 1, panel B-C). The co-existence of previous diagnosis of COPD and the absence of circulating autoantibodies act as protective factors being associated with higher survival probability (Fig. 2). Quite unexpectedly this approach showed that IPF related features as well as respiratory performance are associated to patients’ outcome (prediction) irrespective of those associated with the concomitant tumor (histology and stage).