The incidence of primary melanoma has increased by 55% in the past decade, and roughly 7000 to 8000 patients will die this year in the US of advanced metastatic melanoma according to the American Cancer Society. Despite progress in molecular diagnostics and prognostics for many other cancers, the prognosis of primary melanomas is still based solely upon histopathological factors (thickness and ulceration), and assessment of regional nodal disease, variables that basically have not changed for decades.
In our present work, we aimed to identify an objective molecular marker that could be easily applied in the clinics and improve the prognostication of primary melanomas. Melanomas are known to have a high mutational burden and generate an adaptive immune response in part via their mutation-derived neoantigens. In the past, several groups invested the prognostic role of tumor infiltrating lymphocytes (TILs) but reported conflicting results.
We sought to determine whether features of the T-cell repertoire predict recurrence in patients with primary melanomas and applied high-throughput sequencing of the TCR beta-chain (TCRB) to over 300 archival formalin fixed paraffin embedded (FFPE) primary excision samples.
A dermatoscopically suspected melanoma is usually diagnosed with a complete excision biopsy with narrow margin (2mm), unless there are compelling clinical reasons for doing otherwise. While for the majority of human cancers primary tumors are obtained through inpatient surgical procedures, melanoma is different. Most primary melanomas are diagnosed and removed by dermatologists or surgeons in an outpatient setting, and patients are often followed in a decentralized ambulatory setting rather than a cancer center. This makes sample collection and clinical follow up much more challenging than in other cancers.
In our cohort of primary melanomas, patients with a T cell fraction (TCFr, the percentage of nucleated cells that are T cells) of less than 20% were 2.5 times more likely to experience disease recurrence than patients with high TCFr. Astonishingly, independent of T stage, ulceration status, mitotic rate and nodal disease, a high TCFr was always protective of disease progression. Our hypothesis at the outset of the study was that T cell receptor clonality would indeed associate with prognosis; thus, we were surprised to see that only the T cell fraction, and not T cell clonality, associated with improved progression-free survival (PFS). The composition of the T cell infiltrate was unique for each tumor and highly variable regarding the abundance of CD4+, FoxP3+, CD8+, and putative tumor specific CD8+ T cells (CD39+ and CD39+CD103+). We were intrigued that it appears that the number of T cells, rather than their phenotype or clonality mediated the protective effect. It is likely that the majority of T cells within a tumor are not tumor antigen specific [1,2], so that the odds of a tumor harboring an antigen specific T cell is increased as a function of T cell number.
Our study suggests that a successful T cell-mediated antitumor response can be present in primary melanomas and can be assessed in part by TCRB sequencing. When we applied a machine-learning technique known as gradient boosting to characterize the relative prognostic importance, we were astounded that TCFr was the second most powerful predictor for PFS after tumor thickness. In this way the objective intra- as well as peritumoral measurement of T cells is a molecular assessment that could help to identify patients at high risk of metastatic disease for potential adjuvant therapy.
This study is a true example of transcontinental collaboration. We feel very fortunate to have been able to conduct this study with our multidisciplinary team of dermatologists, pathologists, immunologists and biostatisticians across three continents – Australia, North America and Europe.
Written by Wiebke Pruessmann and Thomas S. Kupper
1. Scheper, W., et al. Low and variable tumor reactivity of the intratumoral TCR repertoire in human cancers. Nature medicine 25, 89-94 (2019).
2. Rosato, P.C., et al. Virus-specific memory T cells populate tumors and can be repurposed for tumor immunotherapy. Nature communications 10, 567 (2019).