Abstract

BACKGROUND: Conventional MRI is unable to differentiate tumor/non-tumoral enhancing tissues on conventional T1-MRI (radionecrosis/pseudoprogression). We have applied delayed contrast MRI for calculating high resolution (1mm3) treatment response assessment maps (TRAMs) clearly differentiating tumor/non-tumoral tissues in brain tumor patients and validated the TRAMs histologically. We further assessed the sensitivity and positive predictive value (PPV) of rCBV/KTrans (DSC/DCE MRI) to tumor. METHODS: 162 patients with primary/metastatic tumors were recruited/followed on study. The maps were validated by comparing pre-surgical maps of 51 patients with histology. Histological evaluation consisted of blinded pathological reports and non-blinded evaluation of 62 biopsies and 8 en-block samples. The sensitivity/PPV of rCBV to tumor was studied for 46 resected patients, where rCBV was compared with histology. The sensitivity/PPV of rCBV/KTranse was studied separately for 49/42 non-resected patients where these parameters were compared with the TRAMs. RESULTS: Histological validation confirmed that regions of contrast agent-clearance > 1hr post contrast injection represent morphologically active tumor while regions of contrast accumulation represent non-tumor tissues. Based on the blinded histological reports, the TRAMs showed sensitivity/PPV to active tumor of 100%/92%. The addition of the non-blinded data resulted in sensitivity/PPV of 99%/95%. Significant correlation was found between tumor volumes in the TRAMs and in the en-block samples (r2 = 0.84, p < 0.0001). Comparison of rCBV with histology resulted in sensitivity/PPV of 51%/96%. When comparing rCBV/KTrans to the TRAMs we found a sensitivity of 23%/24% and PPV of 100%/90% to tumor regions in the TRAMs. The lower sensitivity of rCBV to tumor in the non-resected patients may be attributed to the larger percentage of smaller metastatic tumors that tend towards lower perfusion. CONCLUSIONS: Despite the high PPV of rCBV/KTrans to tumor, the low sensitivity suggests that 49%-76% of the tumors will remain undetected. Our data suggest that the TRAMs provide both high PPV and high sensitivity to tumor tissues.

Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call