A graph-theoretic approach for the analysis of lesion changes and lesions detection review in longitudinal oncological imaging

•The identification of missed and wrongly identified lesions in radiological scans is related to suspicious patterns of lesion changes and requires the examination of longitudinal patient sequences.•A new generic model-based method for the volumetric analysis of lesions and their changes in longitudinal scans based on lesion matching, classification of changes in individual lesions, and detection of patterns of lesion changes.•A new workflow that guides clinicians in the detection of missed and wrongly identified lesions in manual and computed lesion annotations using the analysis of lesion changes.•A new heuristic method for the automatic revision of ground truth lesion annotations in longitudinal scans based on the workflow.•Experimental results on patient studies with ≥3 examinations of metastatic lesions in lung, liver, and brain studies (67 patients, 190 CT and MRI scans, 2295 lesions) yielded a precision of 0.92–1.0, recall of 0.91–0.99 for lesion marching, and an accuracy of 0.87–0.97 for changes in individual lesions and 0.80–0.94 for patterns of lesions changes. Radiological follow-up of oncology patients requires the detection of lesions and the quantitative analysis of lesion changes in longitudinal imaging studies of patients, which is time-consuming and requires expertise. We present here a new method and workflow for the analysis and review of lesions and volumetric lesion changes in longitudinal scans of a patient. The generic graph-based method consists of lesion matching, classification of changes in individual lesions, and detection of patterns of lesion changes computed from the properties of the graph and its connected components. The workflow guides clinicians in the detection of missed lesions and wrongly identified lesions in manual and computed lesion annotations using the analysis of lesion changes. It serves as a heuristic method for the automatic revision of ground truth lesion annotations in longitudinal scans. The methods were evaluated on longitudinal studies of patients with three or more examinations of metastatic lesions in the lung (19 patients, 83 CT scans, 1178 lesions), the liver (18 patients, 77 CECT scans, 800 lesions) and the brain (30 patients, 102 T1W-Gad MRI scans, 317 lesions) with ground-truth lesion annotations. Lesion matching yielded a precision of 0.92–1.0 and recall of 0.91–0.99. The classification of changes in individual lesions yielded an accuracy of 0.87–0.97. The classification of patterns of lesion ch