This is a module for the Insight Toolkit (ITK) that provides a generic, modular, and extensible architecture for lesion sizing algorithms in medical images as well as a reference algorithm for lung solid lesion segmentation in CT images.
For more information, see the Insight Journal article:
Liu X., Helba B., Krishnan K., Reynolds P., McCormick M., Turner W., Ibáñez L., Yankelevitz D., Avila R. Fostering Open Science in Lung Cancer Lesion Sizing with ITK module LSTK The Insight Journal. January-December. 2012. https://hdl.handle.net/10380/3369 https://www.insight-journal.org/browse/publication/869
Binary Python packages are available for Linux, macOS, and Windows. They can be installed with:
python -m pip install --upgrade pip python -m pip install itk-lesionsizingtoolkit
The project has extensively used the CT lesion data assembled by NIST for the Biochange 2008 Pilot Study. The data collection can be obtained via ftp from here.
The team participated in the Volcano'09 Challenge benchmark.
The project has also been used on the BioChange 2011 study. The clinical data used here was chosen from the publicly available Reference Image Database to Evaluate Therapy Response (RIDER) database and from the NIST-generated CT phantom series. 96 pairs of clinical datasets are present. Datasets have slice thickness varying from 0.63 to 2.5 mm. Kitware's results on the Biochange2011 challenge may be found here.
This software is distributed under the Apache 2.0 license. Please see the LICENSE file for details.
This work was supported by the Optical Society of America (OSA), the Air Force Research Laboratory (AFRL), and the National Library of Medicine (NLM).