Development of Radiomic Signatures in Renal Cell Carcinoma
Posted: Monday, January 13, 2020
Radiomic signatures from computed tomography scans may help to predict CD8-positive T-cell infiltration and PD-L1 expression in clear cell renal cell carcinoma. Simply, radiomics is the translation of standard medical imaging data into quantitative data. Haris Zahoor, MD, MS, of the University of Southern California, presented the work he completed with colleagues at the 2019 International Kidney Cancer Symposium in Miami.
The team performed a retrospective analysis wherein samples from 50 patients with localized clear cell renal cell carcinoma who had undergone surgical resection of their tumors between 2009 and 2018 were analyzed. Patients were included if they had preoperative multiphase computed tomography scans and tumor tissue available. The radiomics panel comprised 1,708 metrics, and the radiomic signatures were created from a voxel-wise data analysis code developed in MATLAB and tumor segmentation in Synapse 3D. CD8 staining was also performed.
Of the patients included in the study, 32% had PD-L1–positive disease, and 68% had PD-L1–negative disease. There were 48 patients (90%) available for CD8 staining; of them, 52% had high levels of CD8 and 48% had low levels of CD8. A radiomic signature of CD8 T cells was created and was able to discriminate CD8-high and CD8-low tumors with an AUC of 0.9.
The researchers noted that they cannot rule out that this discovery is simply due to chance, as their sample size was rather small. However, the distribution of predicted probability using histogram analysis showed a complete separation between CD8-high and CD8-low tumors. A similar radiomic signature has been previously developed that can discriminate between PD-L1–positive and PD-L1–negative tumors. Studies to validate these early findings are ongoing.
Disclosure: The full disclosures of the study authors can be found at kidneycancer.org.