Abstract
BACKGROUND: UV-related DNA damage signature (UVsig) is highly specific for cutaneous cancers. The prevalence of UVsig among tumors without a primary site and tumors of extracutaneous origin were previously reported, suggesting potential misclassification of cancers. Our study aims to assess if the knowledge of UVsig at diagnosis would change first-line treatment recommendation. METHODS: The main outcome was the potential clinical impact (PCI) of UVsig. High PCI was defined as UVsig leading to change in diagnosis and first-line therapy. Medium PCI was a change in diagnosis, but appropriate therapy was offered. Low PCI group had diagnosis modified by clinicians and treated as cutaneous cancer independently of UVsig. RESULTS: Among 5565 cases, 650 (12%) were positive for a UVsig. In the cancer of unknown primary group: 20 (49%), 9 (22%), and 12 (29%) cases were categorized in the high, medium, and low PCI group, respectively. In the cancer of extracutaneous origin cohort: 22 (54%), 15 (36%), and 4 (10%) cases were high, medium, and low PCI, respectively. The diagnosis would have changed in 14% of Veterans with UVsig positive tumor. Among all high PCI cases, 37 (88%) received chemotherapy that was not indicated based on a UVsig-informed diagnosis of cutaneous malignancy. CONCLUSION: Our study suggested that UVsig would lead to revision of the working clinical diagnosis and significantly alter the first-line treatment in at least half of cancers of unknown primary or extracutaneous origin with UVsig. Knowledge of UVsig could lead to more effective and less toxic therapy for patients with cancer.