Association of Machine Learning-Based Assessment of Tumor-Infiltrating Lymphocytes on Standard Histologic Images With Outcomes of Immunotherapy in Patients With NSCLC.

Importance: Currently, predictive biomarkers for response to immune checkpoint inhibitor (ICI) therapy in lung cancer are limited. Identifying such biomarkers would be useful to refine patient selection and guide precision therapy. Objective: To develop a machine-learning (ML)-based tumor-infiltrating lymphocytes (TILs) scoring approach, and to evaluate TIL association with clinical outcomes in patients with advanced non-small cell lung cancer (NSCLC). Design, Setting, and Participants: This multicenter retrospective discovery-validation cohort study included 685 ICI-treated patients with NSCLC with median follow-up of 38.1 and 43.3 months for the discovery (n = 446) and validation (n = 239) cohorts, respectively. Patients were treated between February 2014 and September 2021. We developed an ML automated method to count tumor, stroma, and TIL cells in whole-slide hematoxylin-eosin-stained images of NSCLC tumors. Tumor mutational burden (TMB) and programmed death ligand-1 (PD-L1) expression were assessed separately, and clinical response to ICI therapy was determined by medical record review. Data analysis was performed from June 2021 to April 2022. Exposures: All patients received anti-PD-(L)1 monotherapy. Main Outcomes and Measures: Objective response rate (ORR), progression-free survival (PFS), and overall survival (OS) were determined by blinded medical record review. The area under curve (AUC) of TIL levels, TMB, and PD-L1 in predicting ICI response were calculated using ORR. Results: Overall, there were 248 (56%) women in the discovery cohort and 97 (41%) in the validation cohort. In a multivariable analysis, high TIL level (≥250 cells/mm2) was independently associated with ICI response in both the discovery (PFS: HR, 0.71; P = .006; OS: HR, 0.74; P = .03) and validation (PFS: HR = 0.80; P = .01; OS: HR = 0.75; P = .001) cohorts. Survival benefit was seen in both first- and subsequent-line ICI treatments in patients with NSCLC. In the discovery cohort, the combined models of TILs/PD-L1 or TMB/PD-L1 had additional specificity in differentiating ICI responders compared with PD-L1 alone. In the PD-L1 negative (<1%) subgroup, TIL levels had superior classification accuracy for ICI response (AUC = 0.77) compared with TMB (AUC = 0.65). Conclusions and Relevance: In these cohorts, TIL levels were robustly and independently associated with response to ICI treatment. Patient TIL assessment is relatively easily incorporated into the workflow of pathology laboratories at minimal additional cost, and may enhance precision therapy.

Increased cortical porosity and reduced cortical thickness of the proximal femur are associated with nonvertebral fracture independent of Fracture Risk Assessment Tool and Garvan estimates in postmenopausal women.

The Fracture Risk Assessment Tool (FRAX) and Garvan Calculator have improved the individual prediction of fracture risk. However, additional bone measurements that might enhance the predictive ability of these tools are the subject of research. There is increasing interest in cortical parameters, especially cortical porosity. Neither FRAX nor Garvan include measurements of cortical architecture, important for bone strength, and providing independent information beyond the conventional approaches. We tested the hypothesis that cortical parameters are associated with fracture risk, independent of FRAX and Garvan estimates. This nested case-control study included 211 postmenopausal women aged 54-94 years with nonvertebral fractures, and 232 controls from the Tromsø Study in Norway. We assessed FRAX and Garvan 10-year risk estimates for fragility fracture, and quantified femoral subtrochanteric cortical porosity, thickness, and area from computed tomography images using StrAx1.0 software. Per standard deviation higher cortical porosity, thinner cortices, and smaller cortical area, the odds ratio (95% confidence interval) for fracture was 1.71 (1.38-2.11), 1.79 (1.44-2.23), and 1.52 (1.19-1.95), respectively. Cortical porosity and thickness, but not area, remained associated with fracture when adjusted for FRAX and Garvan estimates. Adding cortical porosity and thickness to FRAX or Garvan resulted in greater area under the receiver operating characteristic curves. When using cortical porosity (>80th percentile) or cortical thickness (<20th percentile) combined with FRAX (threshold >20%), 45.5% and 42.7% of fracture cases were identified, respectively. Using the same cutoffs for cortical porosity or thickness combined with Garvan (threshold >25%), 51.2% and 48.3% were identified, respectively. Specificity for all combinations ranged from 81.0-83.6%. Measurement of cortical porosity or thickness identified 20.4% and 17.5% additional fracture cases that, were unidentified using FRAX alone, and 16.6% and 13.7% fracture cases unidentified using Garvan alone. In conclusion, cortical parameters may help to improve identification of women at risk for fracture.