A method for characterizing sarcoma tumor boundary scanned with T2-weighted MRI

Abstract

Purposes: Analysis of the tumor boundary could play an important role in increasing the understanding of tumor relationships with normal structures for surgical resection planning and risk assessment for local recurrence in sarcoma.  The objective of this paper is to present analytical methods that were developed to characterize the tumor boundary using the signal from T2-weighted MRI in a test set of images from patients with sarcomas.

Patients and Methods: The study MRI parameters were compared to histopathological data from the resected tumors.  T2-weighted scans were acquired from 20 subjects from patients diagnosed with sarcoma.   Computer software was developed to characterize the MR intensity of the tumor boundary as function of distance along probing radii.

Results: There were significant differences in the MRI tumor boundary parameters between tumors resected with positive or negative margins.  The slope of the T2 signal change at the tumor/normal tissue boundary (p=0.007) and boundary distance (p=0.001) were significantly correlated with tumor margin status.

Conclusions:  These data demonstrate that this image analysis algorithm which characterizes the change in MR signal at the tumor boundary could be useful in the clinical evaluation of the tumor.  The spatial position of this MR signal slope parameter that correlates with positive tumor margins can be identified precisely while the tumor is still in the body of the patient prior to tumor removal. This boundary analysis method which is different than simple edge detection reports on the tumor-normal tissue interface.