Purpose: The aim of this study was to assess the efficacy of a Computer-Aided Detection (CAD) system in the identification of lung metastases and to compare the volumetric CAD measurements with unidimensional observer measurements in the evaluation of treatment response in oncology patients.
Materials and methods: Two observers (A and B) evaluated nine patients undergoing lung computed tomography (CT) just before and immediately after treatment with chemotherapy. Multislice CT scans were performed before and after the injection of contrast material with a high-resolution protocol (collimation 4x1 mm, 100 mAs, 120 kV). Response Evaluation Criteria in Solid Tumours (RECIST) criteria were used to consider the disease as stable, increased or decreased. Subsequently, target lesions (most significant lesions identified before and after chemotherapy) were evaluated with a computerised system (CAD) to establish volumetric measurements. Observers' unidimensional measurements and CAD volumetric measurements were analysed for comparison.
Results: Twenty-four nodules (diameter: 5-18 mm in the first study and 4-20 mm in the follow-up study) were included. Observers agreed in the assessment of therapy response in 21 nodules: eight were considered to have increased in size, and 13 were judged stable. Observer and CAD measurements disagreed in three nodules: two were considered stable by radiologists and increased by CAD; one was considered increased by radiologists and stable by CAD.As regards patient response, radiologists disagreed in two cases. CAD and observers did not agree in one case.
Conclusions: Our preliminary data suggest that volumetric measurements can modify the diagnostic and therapeutic evaluation of oncology patients under chemotherapy.CAD volumetric measurements allow an easy and objective evaluation, reducing interobserver variability in the evaluation of chemotherapy response.