Background: Mitotic rate is a strong, independent prognostic factor in patients with melanoma. However, incorporating it into the melanoma staging system has proved challenging.
Methods: The prognostic impact of mitotic rate was assessed in a melanoma cohort comprising 5050 patients from 2 geographically distinct populations. Computer-generated cut points for mitotic rate were constructed to determine its impact on melanoma-associated survival using Kaplan-Meier and multivariate regression analyses. The impact of mitotic rate also was assessed in randomly split training and validation sets.
Results: Mitotic rate had a nonlinear impact on survival, as evidenced by unequally spaced cut points. An index incorporating these cut points that was constructed from one population produced significantly more accurate predictions of survival in the other population than using the entire scale of mitotic rate. An index constructed from the combined cohort was found to be independently predictive of survival, with an impact comparable to that of ulceration. Optimal high-versus-low cut points for mitotic rate were generated separately for each T category (<2 mitoses/mm2 vs ≥2 mitoses/mm2 for T1 melanoma, <4 mitoses/mm2 vs ≥4 mitoses/mm2 for T2 melanoma, <6 mitoses/mm2 vs ≥6/mitoses/mm2 for T3 melanoma, and <7 mitoses/mm2 vs ≥7 mitoses/mm2 for T4 melanoma). Using Kaplan-Meier analysis, elevated mitotic rate was found to have an impact on survival comparable to that of ulceration within each T category. Application of the index for mitotic rate that was constructed from the training data set demonstrated an independent impact in the validation data set, with a significance similar to that of ulceration.
Conclusions: The results of the current study demonstrated the comparable prognostic impact of mitotic rate and ulceration, providing support for its reincorporation into the T category.
Keywords: melanoma; mitotic rate; prognostic modeling; staging.
© 2020 American Cancer Society.