TIMES: A ctDNA tumor fraction based and externally validated nomogram to predict survival in cancer patients referred for early phase trials

Typical entry criteria in early phase studies include expected life expectancy greater than 3 months. However, this evaluation is quite subjective and more objective and reproducible tools are needed to improve patient selection for phase I trials. The quantification of ctDNA shed by measuring tumor fraction (TF) has been associated with prognosis in patients with solid tumors. Our objective was to develop and validate a nomogram integrating TF to predict overall survival in cancer patients referred for early phase studies. All consecutive adult patients with an advanced solid tumor included between December 2020 and December 2021 in two precision medicine studies (BIP, NCT02534649, sponsor: Institut Bergonié, Bordeaux, France) and STING (NCT04932525, sponsor: Institut Gustave Roussy, Villejuif, France). All patients underwent comprehensive genomic profiling with the 324-gene FoundationOne®. TF was analyzed as a binary variable, indicating whether a specimen had TF ≥10% or TF <10%. Other variables included previously validated prognostic scores such as the Royal Marsden (RMHs: albumin, LDH, number of metastatic sites ) and the GRIm (LDH, neutrophil/lymphocyte ratio, albumin) scores as well as their individual components. TF > 10, the GRIM score and the RMH score were independent prognostic factors in the training dataset (BIP study n= 965 patients). We generated a nomogram based on TF, number of metastatic sites, albumine level and neutrophil/lymphocyte ratio. To determine the clinical usefulness of our nomogram, we compared it with the RMH and GRIM scores. For our nomogram, the area under the receiver operating characteristic curve (AUC) was 0.79 as compared with 0.65 and 0.68 for the RMH and GRIM scores respectively. In the validation dataset (STING study, n= 947), the AUC was 0.76 as compared with 0.60 and 0.64 for for the RMH and GRIM scores respectively. The nomogram was well-calibrated in both the training and validation datasets. This is the first prospective analysis confirming that TF is a strong prognostic factor in patients with advanced solid tumors. The TIMES nomogram represent an helpful tool in the process of patient selection for phase I trial entry.