A novel algorithm to differentiate between multiple primary lung cancers and intrapulmonary metastasis in multiple lung cancers with multiple pulmonary sites of involvement

2019 
Abstract Introduction Differentiating between multiple primary lung cancer (MPLC) and intrapulmonary metastasis (IPM) is critical for developing a therapeutic strategy to treat multiple lung cancers with multiple pulmonary sites of involvement. Methods We retrospectively included 252 lesions (126 pairs) from 126 patients with surgically resected multiple lung adenocarcinomas. Each pair was classified as MPLC or IPM based on histopathologic findings as the reference standard. A novel algorithm was established with four sequential decision steps based on the combination of computed tomography (CT) lesion types (Step 1), CT lesion morphology (Step 2), difference of maximal standardized uptake values on positron-emission tomography/CT (Step 3), and presence of N2/3 lymph node metastasis or distant metastasis (Step 4). The diagnostic accuracy of the algorithm was analyzed. Performances of eleven observers were assessed without and with knowledge of algorithm. Results Among 126 pairs, 90 (71.4%) were classified as MPLCs and 36 (28.6%) as IPMs. On applying the diagnostic algorithm, the overall accuracy for diagnosis of IPM among conclusive cases up to step 4 was 88.9%, and 65 and 44 pairs were correctly diagnosed based on Step 1 and Step 2, respectively. Specificity and positive predictive value for diagnosis of IPM increased significantly in all observers compared with reading rounds without the algorithm. Conclusions Application of the algorithm based on comprehensive information on clinical and imaging variables can allow differentiation between MPLCs and IPMs. When both of two suspected malignant lesions appear as solid predominant lesions without spiculation nor air-bronchogram on CT, IPM should be considered.
    • Correction
    • Source
    • Cite
    • Save
    • Machine Reading By IdeaReader
    33
    References
    13
    Citations
    NaN
    KQI
    []