Identify the source and localization of the metastatic predictive signal

Determine whether the molecular melanin signal predictive of metastasis is acquired only after disaggregation is well underway at the metastatic site, or whether the relevant features are present in primary tumors but are not captured by the fitted pump-probe representation, using volumetric imaging of entire tumors and substantially larger patient cohorts to distinguish these possibilities.

Background

The study finds that pump-probe microscopy-derived melanin excited-state dynamics distinguish primary melanoma tissue from metastatic tissue at the cohort level, but do not robustly predict whether an individual primary tumor will later metastasize. The authors propose a needle-in-the-haystack explanation: the metastatic-origin region may occupy only a small fraction of the primary tumor, causing its signal to be diluted by surrounding tissue, or the imaged sections may not contain that region.

The authors explicitly leave unresolved whether the predictive molecular signal is generated only during later metastatic progression or is already present in primary tumors but lost because the biexponential fit-parameter representation does not capture it. They indicate that volumetric imaging and larger patient cohorts may be needed to resolve this uncertainty.