Digital Pathology Scoring of Immunohistochemical Staining Reliably Identifies Prognostic Markers and Anatomical Associations in a Large Cohort of Oral Cancers

Julius Moratin, Andreas Möck, Sonja Obradovic, Karl Metzger, Christa Flechtenmacher, Karim Zaoui, Stefan Fröhling, Dirk Jäger, Jürgen Krauß, Jürgen Hoffmann, Kolja Freier, Dominik Horn, Jochen Heß, Christian FreudlspergerFrontiers in Oncology 2021 · DOI

TL;DR

This study introduces a digital pathology approach using QuPath for objective quantification of immunohistochemical staining in oral cancers. This method reliably identifies prognostic biomarkers and their anatomical associations, leading to a combined score that significantly impacts patient survival, paving the way for more personalized cancer treatments.

What Problem Does This Paper Address?

Oral squamous cell carcinoma (OSCC) is a significant global health challenge, with survival rates stagnating despite advances in treatment. Accurate biomarkers are crucial for stratifying patients, predicting tumor progression, and tailoring therapy. Traditionally, assessing immunohistochemical staining relies on manual scoring, which is subjective, time-consuming, and prone to observer variability. This lack of standardization hinders the translation of promising biomarkers into clinical practice. The paper addresses this by proposing and validating a digital pathology approach to objectively quantify key biomarkers in OSCC.

Methods

The researchers employed a digital pathology workflow using QuPath, an open-source software for digital image analysis. This involved:

  • Tissue Microarrays (TMAs): TMAs were created from a cohort of 222 OSCC patient samples to efficiently analyze multiple markers across many samples.
  • Immunohistochemistry (IHC): TMAs were stained for several key biomarkers: PD-L1, EGFR, COX-2, CD8+ T cells, and NK cells (CD335).
  • Digital Scoring: QuPath was used for semi-automatic digital quantification. This involved:
  • TMA dearrayer: Identifying the grid of tissue cores within the TMA.
  • Manual exclusion: Removing artifacts and invalid samples.
  • Stain vector estimation: Automatically determining staining characteristics for each slide.
  • Positive cell detection: Quantifying the number of positive cells for each marker.
  • Manual Scoring: For comparison, CD8, CD335, and PD-L1 were also scored manually by three independent observers using an established immunoreactivity score (IRS) system.
  • Spatial Mapping: The location of tumors and the expression levels of biomarkers were mapped onto an anatomical illustration of the oral cavity to investigate anatomical associations.
  • Survival Analysis: Kaplan-Meier curves and Cox proportional hazard models were used to assess the prognostic significance of the identified biomarkers for overall survival (OS) and progression-free survival (PFS).

Datasets

The study utilized a dataset of 222 primary oral squamous cell carcinoma (OSCC) patient samples. These samples were collected retrospectively from patients treated at the University Hospital Heidelberg between 2010 and 2016. The dataset included demographic, clinical, and pathological data. No new benchmark datasets were introduced; rather, the study leveraged existing patient samples for analysis.

Key Insights and Ideas

  • Digital Pathology as a Reliable Standard: The study demonstrates that digital pathology, specifically using QuPath, can reliably quantify immunohistochemical staining for prognostic biomarkers in OSCC, offering an objective alternative to manual scoring.
  • Concordance with Manual Scoring: High levels of concordance were observed between digital and manual scoring for CD8+ T cells, NK cells, and PD-L1, validating the digital approach.
  • Anatomical Associations Matter: The spatial mapping revealed that PD-L1 expression was significantly lower in tumors located in the oropharynx compared to other oral subsites, suggesting that tumor location can influence biomarker expression.
  • Combined Prognostic Score: A key finding is the development of a combined score integrating EGFR expression, neck node status, and immune cell signatures (CD8/CD335 ratio). This combined score proved to be a powerful predictor of both overall and progression-free survival.
  • Immune Microenvironment is Key: The study reinforces the importance of the tumor immune microenvironment, highlighting the prognostic significance of CD8+ T cells and NK cells, particularly their ratio, in oral cancer.

Key Results

  • Digital Scoring Success: Digital scoring successfully quantified expression levels of CD8 (p<0.001), NK cells (p=0.002), and PD-L1 (p=0.026).
  • High Concordance: High concordance was observed between digital and manual scoring for CD8 (p<0.001), NK cells (p=0.002), and PD-L1 (p=0.026).
  • Prognostic Biomarkers Identified:
    • Overall Survival (OS): Nodal status and the CD8/CD335 ratio were independent prognostic markers.
  • Progression-Free Survival (PFS): Nodal status, the CD8/CD335 ratio, and EGFR status were independent prognostic markers.
  • Combined Score Impact: A combined score integrating EGFR, neck node status, and immune cell signatures showed a significant impact on both overall and progression-free survival (p<0.001).
  • Specific Findings:
    • Higher CD8/CD335 ratio (T cell dominance) and higher PD-L1 expression were associated with shorter OS.
  • Higher CD8/CD335 ratio and positive EGFR status were associated with shorter PFS.

Limitations and Future Work

  • TMA Limitations: The use of TMAs, while efficient, may not fully capture the intratumoral heterogeneity of certain biomarkers, such as PD-L1.
  • Validation Needed: The authors suggest that further studies using whole tissue slides are warranted to validate these findings.
  • Prospective Studies: The translation of these findings into clinical practice would benefit from prospective studies.
  • Cost-Effectiveness: The clinical applicability of biomarkers is dependent on factors like prognostic value, reliability, and cost-effectiveness.
  • Future Directions: The study contributes to the development of digital pathology as a standard procedure for biomarker identification and validation, aiming to enhance risk stratification and personalize tumor therapy for oral cancer patients. The authors emphasize the potential of combining multiple biomarkers to improve clinical applicability.

Summary updated

Abstract

Utilizing digital pathology algorithms for the objective quantification of immunohistochemical staining, this study aimed to identify robust prognostic biomarkers for oral cancer. Tissue microarrays with specimens of a large cohort of oral squamous cell carcinoma (n=222) were immunohistochemically stained to determine the expression of PD-L1, EGFR, and COX-2 and the amount of infiltrating NK cells and CD8-positive T cells. Immunoreactivity scores were assessed using both a classical manual scoring procedure and a digital semi-automatic approach using QuPath. Digital scoring was successful in quantifying the expression levels of different prognostic biomarkers (CD8: p<0.001; NK cells: p=0.002, PD-L1: p=0.026) and high levels of concordance with manual scoring results were observed. A combined score integrating EGFR expression, neck node status and immune cell signatures with a significant impact on overall and progression-free survival was identified (p<0.001). These data may contribute to the ongoing research on the identification of reliable and clinically relevant biomarkers for the individualization of primary and adjuvant treatment in oral cancer.
BibTeX
@inproceedings{moratin2021digital,
  author = {Julius Moratin and Andreas Möck and Sonja Obradovic and Karl Metzger and Christa Flechtenmacher and Karim Zaoui and Stefan Fröhling and Dirk Jäger and Jürgen Krauß and Jürgen Hoffmann and Kolja Freier and Dominik Horn and Jochen Heß and Christian Freudlsperger},
  title = {{Digital Pathology Scoring of Immunohistochemical Staining Reliably Identifies Prognostic Markers and Anatomical Associations in a Large Cohort of Oral Cancers}},
  booktitle = {Frontiers in Oncology},
  year = {2021},
  doi = {10.3389/fonc.2021.712944},
}
Figure