Original title: DNA Methylation Profiling in Melanoma: From Lesion Classification to Therapeutic Stratification
Jana T. Winterstein, Lukas Heinlein, Günter Raddatz, Carina Nogueira Garcia +33 · August 2026
Skin cancer like melanoma can be difficult to identify and track accurately. Cells alter their DNA with chemical tags called DNA methylation, which switch genes on or off without altering the genetic code itself. As normal cells turn into tumors, these tags change. Researchers wanted to see if these patterns could diagnose skin lesions and guide treatment. They tested computer models on DNA methylation data taken from skin tissue. The models read the specific chemical tags on the DNA to categorize the samples. They found that a model using specific methylation sites successfully separated harmless moles, early noninvasive melanomas, and dangerous invasive melanomas. The noninvasive tumors sat directly in the middle chemically, explaining why they are tricky to diagnose. Meanwhile, a model using broader biological clues best predicted which treatment group a patient needed. This shows DNA tags can help identify and stage melanoma.
You are baking a cake and need to decide if it's done. Usually you stick a toothpick in the center—if it comes out clean, the cake is ready. But to know how long to let it cool before frosting, you need to know more: the oven temperature, the pan size, and whether you used a moist ingredient like applesauce. In this paper, one set of DNA markers works like the toothpick for classifying skin lesions, while a broader set of biological features works like the extra details for predicting the best treatment.
Melanoma misdiagnosis can lead to unnecessary surgery or missed cancers. This work shows that reading the chemical tags on a tumor's DNA can accurately classify skin lesions and even suggest the right treatment approach. It brings precision medicine one step closer to routine skin cancer care.
The diagnostic model was tested on an external cohort, but the treatment prediction model's performance (MAE 0.627) may not be sufficient for clinical use. The models were built on samples from German hospitals only. The study does not show that using these models improves patient outcomes. Prospective validation is still required.
Ask anything about this paper. Answers come from its abstract, and say so when the abstract doesn't cover your question.
DNA Methylation Profiling in Melanoma: From Lesion Classification to Therapeutic Stratification — Jana T. Winterstein, Lukas Heinlein, Günter Raddatz et al.
Abstract licence: CC0
Explanation generated by Synthesis from the abstract — it is not the original paper. This is a preprint, so it hasn't necessarily been peer reviewed.