PRECISION AND RELIABILITY FOR MEASUREMENT OF CHANGE IN MRI LESION VOLUME IN MULTIPLE-SCLEROSIS - A COMPARISON OF 2 COMPUTER-ASSISTED TECHNIQUES

Objective-The serial quantification of MRI lesion load in multiple scl erosis provides an effective tool for monitoring disease progression a nd this has led to its increasing use as an outcome measure in treatme nt trials. Segmentation techniques must display a high degree of preci sion and reliability if they are to be responsive to small changes ove r time. This study has evaluated the performance of two such technique s, the manual outlining and contour methods, in serial lesion load qua ntification. Methods-Sixteen patients with clinically definite multipl e sclerosis were scanned at baseline and after two years. Scan analysi s was performed twice, independently by three observers using each tec hnique. Results-For the absolute lesion volumes the median intrarater coefficient of variation (CV) was 3.2% for the contour technique and 7 .6% for the manual outlining method (p<0.005), the interrater CVs were 3.8% and 6.1% respectively (p<0.01) and the reliability of both techn iques was very high. For the change in lesion volume the intrarater an d interrater repeatability coefficients were respectively 2.6 cm(3) an d 2.8 cm(3) for the contour technique, and 3.3 cm(3) and 3.7 cm(3) for the manual outlining method (lower values reflect higher precision). The values for intrarater and interrater reliability for measuring cha nge in lesion volume were respectively, 0.945 and 0.944 for the contou r technique, and 0.939 and 0.921 for the manual outline method (perfec t reliability = 1.0). Conclusions-With such high values for reliabilit y, the impact of measurement error in lesion segmentation on sample si ze requirements in multiple sclerosis treatment trials is minor This s tudy shows that a change in lesion volume can be measured with a highe r level of precision and reliability with the contour technique and th is supports its further application in serial studies.