Hier finden Sie wissenschaftliche Publikationen aus den Fraunhofer-Instituten.

Software support for combined staging of lung cancer in CT, functional MRI and pathology

: Laue, H.O.A.; Kohlmann, P.; Lotz, J.; Sedlaczek, O.; Müller, B.; Breuhahn, K.; Grabe, N.; Warth, A.; Hahn, H.

Journal of thoracic oncology : JTO 10 (2015), Nr.9, Supplement 2, S.S467
ISSN: 1556-0864
ISSN: 1556-1380
Fraunhofer MEVIS ()

Background: Treatment and diagnosis of lung cancer is an interdisciplinary challenge. New treatment options can benefit from refined information on biological processes in the tumor. Two diagnostic disciplines, pathology and radiology, can provide oncology with valuable information allowing to select of the most appropriate treatment. Linking radiology and pathology on the imaging level requires sophisticated software. Therefore, we developed a computer tool combining quantitative functional MRI and CT with state-of the-art whole-slide and 3D reconstructed histology.
Methods: We selected a model lung cancer patient to investigate the requirements and possibilities for a combined software. The patient had a squamous cell carcinoma. Prior to excision, functional imaging using MRI as well as highly resolved CT and MRI volumes were acquired. These included a dynamic contrast-enhanced (DCE)- and a diffusion-weighted imaging (DWI)- MR scan. After imaging, the tumor was resected and a macroscopic slice of 5 mm thickness was resected from the center region of the tumor. This slice was further divided into 11 blocks, which were then cut into microscopic (~ 2 μm) slices. Staining was applied to these slices according to the requirements of the pathologist and then scanned by a whole-slide histological scanner (Hamamatsu, Japan). The obtained images of the microscopic slices were automatically reconstructed into histological 3D volumes by means of registration. Functional MRI and morphological CT data was imported into the software prototype. The tumor volume was determined in the CT image by a two-click automatic segmentation method. Quantitative parametric maps of the extended general kinetic model (eGKM) for vascular information and the apparent diffusion coefficient (ADC) for cell density were calculated. Afterwards, the histological blocks were aligned such that the blocks were located correctly on a photo of the macroscopic slice. They were then manually aligned to the morphologic contrast enhanced T1 image showing the best contrast for structures visible in the macroscopic slice. Finally, the pathological data were imported into the software for direct comparison of pathology and functional imaging in human lung tumors.
Results: We were able to combine histological and radiological information into a software solution which provids an integrated and improved research opportunity for pathologists, radiologists and biologists. It allows correlation of findings on molecular and cell level with findings from in-vivo functional imaging.
Conclusion: We demonstrated that a combined evaluation of functional MRI and pathology can be facilitated. New studies will show the usage in more common lung cancers and larger numbers of patients.