Sepideh Azarianpour-Esfahani, Ph.D. of Case Western Reserve University discusses an abstract from ASCO 2020 Computerized features of spatial arrangement of tumor-infiltrating lymphocytes from H&E images predicts survival and response to checkpoint inhibitors in gynecologic cancers
Context:
Immune checkpoint inhibitors (ICI) in solid tumors have proved effective. Except for MSI-H endometrial cancer (~ 50 percent), the response rate is still poor in gynecologic cancers (GC) (~10-15 percent). Present biomarkers (e.g., expression of PDL1) have limited usefulness in recognizing ICI benefits in GC. In this research, the ability of H&E slide images to predict overall survival ( OS) and response to ICI in ovarian, cervical and endometrial cancers was evaluated by computational measurements of the spatial arrangement of infiltrating tumor lymphocytes (TIL).
Approaches:
The study included 151 patients, including 102 ovarian carcinomas treated with surgery and chemotherapy (D1) and another collection (D2) of patients treated with various ICI agents (Pembrolizumab, Nivolumab, Ipilimumab, Avelumab) in the second-line setting, including n=49 patients (n=14 ovarian, n=27 endometrial and n=8 cervical). Progressors and non-progressors in D2 have been graded by RECIST based on clinical advancement and radiological evaluation. To classify tumor regions on the diagnostic slides of D1 and D2, a machine learning method was used and then used to automatically identify TILs inside the tumor regions. Machine learning was subsequently used to describe TIL clusters based on TIL proximity, and graph network theory was used to capture measurements relating to TIL cluster spatial arrangement. The multivariable model of Cox regression (MCRM) was trained to predict OS on n=51 patients from D1 and then independently tested in predicting (1) OS on hold-out n=51 patients from D1 and (2) response and progression-free survival (PFS) from D2.
Reviews:
Statistical research identified 7 prognostic features related to the cancer nuclei interaction between TIL clusters. MCRM was predictive of OS in D1 (hazard ratio (HR)=2.06, 95 % confidence interval [1.04-4.07], p=0.008) and predictive of PFS in D2 (HR=2.24, CI=[1.13-4.44], p=0.03) for n=51 patients. In predicting progression in D2, the AUC for MCRM was 82%.
Conclusions Therein:
In three gynecological cancers, computerized features of spatial arrangement of TILs on H&E images were prognostic of OS and PFS and predicted ICI response. In broader, multi-site validation sets, these results need to be confirmed.