Purpose: To evaluate the results of a new technique of dynamic contrast enhanced Magnetic Resonance (MR) imaging subtraction in the assessment of osteosarcoma response to chemotherapy.
Methods: 24 patients with high grade osteosarcoma, treated with preoperative neo-adjuvant chemotherapy, underwent MR at high field strength (1.5 T). Both unenhanced conventional SE T1- and T2-weighted sequences in the coronal and axial plane and dynamic Gd-DTPA-enhanced SE T1-weighted sequences in the coronal plane were performed. Image postprocessing included subtraction of unenhanced image from enhanced images (arbitrary called "angiographic subtraction") and subtraction of each enhanced image from the last-enhanced image (arbitrary called "pathologic area" subtraction). The early enhancing areas detected in the angiographic subtraction and the pathologic areas detected in the pathologic area subtraction were correlated with histopathological findings on histological macrosections obtained from the resected specimen. The sensitivity, specificity, accuracy, positive and negative predictive value of both the subtraction techniques were calculated.
Results: The early enhancing areas of angiographic subtraction were related not only to viable tumor but also to the host reactions such as flogosis and granulation tissue. The pathologic areas detected at the pathologic area subtraction correlated in most cases with viable tumor, while in 3 cases they did not correspond to viable tumor tissue and in 1 case a small area of residual viable tumor was missed. In assessing response to chemotherapy, pathologic areas subtraction had an accuracy of 95% (specificity: 100%, sensitivity: 93%, PPV: 100%, NPV: 88%), whereas angiographic subtraction had an accuracy of 79% (specificity: 37%, sensitivity: 100%, PPV: 76%, NPV: 100%).
Conclusions: Pathologic area subtraction may be a useful technique for assessing the response of osteosarcoma to chemotherapy and for detecting residual viable tumor tissue.