Samuel G Armato1, Ping Li2, Aliya N Husain3, Christopher Straus2, Ashoke Khanwalkar2, Hedy L Kindler4, Wickii T Vigneswaran5. 1. Department of Radiology, The University of Chicago, Chicago, IL, USA. Electronic address: s-armato@uchicago.edu. 2. Department of Radiology, The University of Chicago, Chicago, IL, USA. 3. Department of Pathology, The University of Chicago, Chicago, IL, USA. 4. Section of Hematology/Oncology, Department of Medicine, The University of Chicago, Chicago, IL, USA. 5. Department of Surgery, The University of Chicago, Chicago, IL, USA.
Abstract
OBJECTIVES: Tumor volume promises to become a more important factor in patient management. Mesothelioma, with its unique morphology and complex growth pattern, presents a challenging target for tumor volumetrics derived from computed tomography (CT) scans. This study evaluated the validity of image-based mesothelioma tumor volume against the physical volume of the tumor bulk captured by the images. MATERIALS AND METHODS: Twenty-eight patients underwent CT scanning prior to pleurectomy/decortication with an intent to achieve a macroscopic complete resection. A radiologist manually outlined the tumor border in all CT sections in which tumor appeared in the pre-surgery scan. CT-based tumor volume was computed as the number of image pixels enclosed by all tumor outlines across all sections in the scan multiplied by the physical dimensions of the voxel of tissue captured by each image pixel. The gross tumor specimen volume was measured ex vivo through a water-displacement technique. Correlation between CT volume and pathology volume was calculated. RESULTS: A correlation coefficient r-squared value of 0.66 was found between CT-based tumor volume and gross tumor specimen volume. Differences between the mean volume (either CT volume or pathology volume) across tumors of different T stages did not achieve statistical significance. CONCLUSION: Despite a modest correlation between CT-based tumor volume and gross tumor specimen volume, image-based volumetry for mesothelioma is not straightforward-perhaps, in part, due to the challenges of distinguishing tumor borders from adjacent structures and perhaps, in part, due to a complex pathologic reference standard.
OBJECTIVES:Tumor volume promises to become a more important factor in patient management. Mesothelioma, with its unique morphology and complex growth pattern, presents a challenging target for tumor volumetrics derived from computed tomography (CT) scans. This study evaluated the validity of image-based mesothelioma tumor volume against the physical volume of the tumor bulk captured by the images. MATERIALS AND METHODS: Twenty-eight patients underwent CT scanning prior to pleurectomy/decortication with an intent to achieve a macroscopic complete resection. A radiologist manually outlined the tumor border in all CT sections in which tumor appeared in the pre-surgery scan. CT-based tumor volume was computed as the number of image pixels enclosed by all tumor outlines across all sections in the scan multiplied by the physical dimensions of the voxel of tissue captured by each image pixel. The gross tumor specimen volume was measured ex vivo through a water-displacement technique. Correlation between CT volume and pathology volume was calculated. RESULTS: A correlation coefficient r-squared value of 0.66 was found between CT-based tumor volume and gross tumor specimen volume. Differences between the mean volume (either CT volume or pathology volume) across tumors of different T stages did not achieve statistical significance. CONCLUSION: Despite a modest correlation between CT-based tumor volume and gross tumor specimen volume, image-based volumetry for mesothelioma is not straightforward-perhaps, in part, due to the challenges of distinguishing tumor borders from adjacent structures and perhaps, in part, due to a complex pathologic reference standard.
Authors: Samuel G Armato; Kevin G Blyth; Jane J Keating; Sharyn Katz; Selina Tsim; Johan Coolen; Eyjolfur Gudmundsson; Isabelle Opitz; Anna K Nowak Journal: Lung Cancer Date: 2016-09-05 Impact factor: 5.705