BILL CD8 – A Multivariable Survival Model as a Simple and Clinically Useful Prognostic Tool to Identify High-risk Cutaneous Melanoma Patients
Przemysław Biecek , Piotr Donizy , Agnieszka Halon , Rafal Matkowski
AbstractTo develop a multivariable survival model based on histopathological parameters that would provide the best possible estimate of the risk of death in cutaneous melanoma patients. Materials and Methods: Breslow thickness and Infiltrating Lymphocytes (BILLCD8) is based on two key parameters, namely Breslow thickness and tumor-infiltrating lymphocytes (TILs) shown to be the most important prognostic factors using logistic regression model. The next step was to apply a decision tree method to analyze the impact of the percentage of CD8+ lymphocytes on the risk of cancer death. Results: A patient is given two points if Breslow thickness is >2 mm and one point if TILs are not found in the tumor. The probability of death over a 4-year follow-up is 75% if the patient′s score is 3 points; in patients with 0 point score the risk of death is 0%. A significant positive impact of high percentage of CD8+ lymphocytes on the risk of death was shown in the subgroup of melanoma patients with Breslow thickness >2 mm where the risk of death is only 28%. In the case of low percentage of CD8+ lymphocytes in melanomas with Breslow thickness over 2 mm, the risk is as high as 67%. Conclusion: Breslow thickness and TILs, with the evaluation of the CD8+ lymphocytes, may be used for precise prognostic stratification of patients and may help intensify the follow-up, diagnose a relapse sooner and apply proper treatment.
|Journal series||Anticancer Research, ISSN 0250-7005, (A 20 pkt)|
|Publication size in sheets||0.5|
|Keywords in English||Melanoma, survival models|
|ASJC Classification||; ;|
|Abstract in Polish||Konstrukcja modelu określającego ryzyka niepowodzenia leczenia u pacjentów z melanoma|
|Score|| = 20.0, 04-09-2019, ArticleFromJournal|
= 20.0, 04-09-2019, ArticleFromJournal
|Publication indicators||= 0; = 0; : 2016 = 0.669; : 2016 = 1.937 (2) - 2016=1.925 (5)|
* presented citation count is obtained through Internet information analysis and it is close to the number calculated by the Publish or Perish system.