Original title: Evaluating the predictive power of pump-probe imaging contrast of melanin for metastatic outcome: A 71-patient study
David Grass, Emma Bao, Martin C. Fischer, M. Angelica Selim +2 · September 2026
Skin cancer causes most deaths when it spreads, but current tests struggle to spot which original tumors will spread. Researchers tested if studying melanin, the pigment in skin, could help. As cancer grows, melanin structure changes. The team used a laser technique called pump-probe microscopya laser technique that uses two light pulses to measure how a material responds to light very quickly. to track how melanin absorbs and releases light energy in tumor samples. This reveals the pigment's molecular state without extra chemical stains. They found that melanin reacted much faster in original tumors than in cancer that had already spread. This laser method easily told the two kinds of tissue apart. However, it failed to tell which original tumors would later spread and which would not. Cancer within a single tumor is diverse, so testing just one thin slice did not give reliable predictions about a patient's future. The best computer models only achieved moderate accuracy.
Imagine sorting tomatoes for a salad. You can tell a ripe tomato from a rotten one by squeezing it, but you want to know which firm, ripe-looking tomatoes will spoil in two days. Squeezing tells you about the present, not the future. In this paper, the laser measurement is the squeezing: it clearly separated tumors that hadn't spread from ones that had, but it could not predict which of the former would later spread.
Most melanoma deaths come from metastasis, yet doctors cannot reliably tell which early tumors will spread. This study tested whether imaging melanin's light response in biopsy samples could predict that outcome. It could separate primary from metastatic tissue, but it could not predict which primary tumors would metastasize, so it is not ready to guide treatment decisions.
The abstract does not establish a reliable predictive model for metastatic outcome. The authors note that single-section measurements are confounded by heterogeneity, and the best classifier's accuracy plateaued at 74% (AUC 0.73). The findings are limited to the 71-patient cohort described.
Ask anything about this paper. Answers come from its abstract, and say so when the abstract doesn't cover your question.
Evaluating the predictive power of pump-probe imaging contrast of melanin for metastatic outcome: A 71-patient study — David Grass, Emma Bao, Martin C. Fischer 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.