Histomorphometry in the assessment of prognosis in stage II malignant melanoma

The prognosis for patients with node-spread melanoma is often based on the number of nodes that contain tumor. This predicts survivors with 85% accuracy but melanoma deaths with only 40% accuracy. This overall predictive accuracy of 61% may be due, in part, to node-counting that takes no account of the volume of tumors in the nodes. We examined the capacity of morphometric techniques (planimetry, image analysis, and measurement of tumor diameter by ocular micrometer) to improve predictive accuracy. A combination of the number of tumor positive lymph nodes (less than or equal to 3 nodes positive = favorable; greater than or equal to 4 nodes positive = unfavorable) and the aggregate micrometer measured tumor diameter as a percentage of aggregate node diameter (less than or equal to 15% = favorable; greater than or equal to 15% = unfavorable) significantly increased predictive accuracy. Favorable outcome was accurately predicted for 12 of 13 patients (92%), and death from melanoma was accurately predicted in 12 of 15 patients (80%), an overall predictive accuracy of 86%.