Machine Learning Analysis of PD-L1 and CD8 Expression Patterns in Penile Squamous Cell Carcinoma: Implications for Immunotherapy
Listed in
This article is not in any list yet, why not save it to one of your lists.Abstract
Introduction
In 2023, the US reported an estimated 2,050 new penile squamous cell carcinoma (SCC) cases, with mortality rates rising to 23%. Despite therapeutic advancements, the need to identify molecular and immunotherapeutic targets, especially PD-L1 and CD8, has become increasingly evident as immunotherapy emerges as a promising treatment strategy.
Methods
Tissue samples from 108 SCC patients were examined using four tissue microarrays (TMA). Alongside the assessment of histologic features and PD-L1 and CD8 expressions, a machine learning model was employed. This model, constructed in Python, leveraged various statistical tests and methodologies to discern the predictive relationships between PD-L1/CD8 expression and pathologic features, with potential implications for immunotherapy selection.
Results
Histologic subtype was predominant in PD-L1 expression in tumor cells, explaining 13.5% variability, whereas combined with host response, it reached 20%. For intratumoral lymphocytes, host response stood out, explaining 11.2% of PD-L1 and 15% of CD8 variability. In peritumoral lymphocytes, host response again dominated, accounting for 21.1% of CD8 variability. These findings align with recent studies showing the importance of immune markers in predicting treatment response and prognosis.
Conclusions
The study highlighted the crucial relationship between penile SCC’s pathologic features and PD-L1/CD8 expression, emphasizing their potential as both prognostic and predictive biomarkers. By integrating traditional histopathology with machine learning, we provide valuable insights for patient stratification and personalized immunotherapeutic strategies in penile SCC.
