Accuracy of clinical estimates of tumor dimensions. A clinical-pathologic correlation study of posterior uveal melanomas.

The authors reviewed the clinical records of 50 patients who underwent enucleation for posterior uveal malignant melanoma to determine the correlation between clinical estimates of tumor size made prior to enucleation and pathologic measurements made during gross examination. Linear regression of the pathologic measurements of several tumor dimensions on the corresponding clinical estimates documented correlation coefficients ranging from 0.81 for minimal basal diameter to 0.92 for tumor thickness. Although these values suggest a high level of linear correlation, wide confidence limits indicate a significant level of random variation in the data points. Further, nonzero intercepts and nonunity slopes of the linear regression lines indicate a systematic variation between clinical estimates and gross pathologic measurements of tumor size.