Predicting immunotherapy outcomes in patients with MSI tumors using NLR and CT global tumor volume

Younes Belkouchi, Laëtitia Nebot-Bral, Littisha Lawrance, Michèle Kind, C. David, Samy Ammari, Paul-Henry Cournède, Hugues Talbot, Perrine Vuagnat, Cristina Smolenschi, Patricia Kannouche, Nathalie Chaput, Nathalie Lassau, Antoine HollebecqueFrontiers in Oncology 2022 · DOI

TL;DR

This study investigates new ways to predict how well patients with MSI (microsatellite instability) tumors will respond to immunotherapy. By combining CT scan-based tumor volume measurements with blood-based Neutrophil-to-Lymphocyte Ratio (NLR), researchers developed a more accurate predictive score than using either method alone, offering a promising tool for personalized treatment decisions.

What Problem Does This Paper Address?

Immunotherapy, particularly immune checkpoint inhibitors (ICIs) like anti-PD-(L)1 antibodies, has revolutionized cancer treatment. These therapies work by unleashing the patient's own immune system to fight cancer. They are particularly effective for tumors with microsatellite instability (MSI), a genetic characteristic that leads to a high number of mutations and makes tumors more visible to the immune system. However, the response to ICIs in MSI tumors is highly variable, meaning some patients benefit greatly while others see little to no effect. This heterogeneity poses a significant challenge, as it's difficult to predict who will respond. Current methods, like RECIST (Response Evaluation Criteria In Solid Tumors), primarily focus on the size of a few target lesions and may not fully capture the extent of disease or predict the unique responses seen with immunotherapy, such as pseudo-progression. Therefore, there's a critical need for better biomarkers that can predict ICI response early on, allowing for more tailored treatment strategies.

Methods

The researchers conducted a retrospective study analyzing CT scans and blood samples from 61 patients with MSI tumors treated with anti-PD-(L)1 therapy.

  • Tumor Volume Measurement: Instead of relying solely on RECIST, which limits the number of lesions considered, this study employed a more comprehensive approach. Expert radiologists meticulously annotated all visible lesions on baseline and first-evaluation CT scans. This allowed for the calculation of the total tumor volume for each patient, providing a more complete picture of disease burden.
  • Neutrophil-to-Lymphocyte Ratio (NLR): Blood samples were analyzed to determine the NLR, a known indicator of systemic inflammation and a potential marker of immune response.
  • Parameter Variation: The study focused not only on baseline values but also on the change in these parameters (tumor volume and NLR) between the baseline scan and the first evaluation (typically around 2 months). This dynamic assessment was hypothesized to be more predictive of early treatment response.
  • Concordance Score: A novel "concordance score" was developed to combine the variations in tumor volume and NLR. This score aimed to capture how these two parameters changed together (e.g., both increasing, both decreasing, or one increasing while the other decreased), providing a more nuanced predictive metric.
  • Statistical Analysis: Kaplan-Meier curves and Cox regression models were used to assess the association between these parameters (baseline and variation) and patient survival outcomes (Overall Survival - OS, and Progression-Free Survival - PFS).

Datasets

The study utilized a dataset of 61 patients with metastatic MSI tumors treated with anti-PD-(L)1 therapy at Gustave Roussy. The data included:

  • Baseline and first-evaluation CT scans for each patient.
  • Blood samples for NLR measurement at both time points.
  • Clinical outcome data (OS and PFS).

No new benchmark datasets were introduced; the study relied on existing patient data.

Key Insights and Ideas

  • Total Tumor Volume is a Stronger Prognosticator than RECIST: The study found that the total tumor volume at baseline was significantly correlated with both OS and PFS, outperforming the traditional RECIST sum of target lesion diameters. This highlights the limitation of RECIST in fully capturing the extent of disease for immunotherapy prediction.
  • Dynamic Changes Matter: The variation in tumor volume and NLR between baseline and the first evaluation proved to be highly predictive of survival outcomes. This suggests that how the disease and inflammatory markers change early in treatment is crucial.
  • NLR as an Early Indicator: The variation in NLR (NLRvariation) emerged as an early predictive biomarker, aligning with previous research suggesting its role in immunotherapy response.
  • Synergy of Multiple Biomarkers: Combining tumor volume and NLR variation into a "concordance score" significantly improved the ability to stratify patients by survival compared to using either parameter alone. This underscores the power of multi-modal assessment.
  • Beyond Target Lesions: The study's emphasis on total tumor volume implicitly addresses the limitations of RECIST, which only considers a maximum of 5 target lesions. This broader view is likely more representative of the overall tumor burden and its impact on treatment response.

Key Results

  • Baseline Tumor Volume: Significantly correlated with OS (p=0.005) and PFS (p<0.001).
  • Variation in Parameters:
    • RECIST variation was highly correlated with OS and PFS (p<0.001 for both).
  • Volume variation was significantly correlated with OS (p=0.006) and PFS (p<0.001).
  • NLR variation was significantly correlated with OS (p<0.001) and PFS (p=0.007).
  • Concordance Score: A concordance score combining Volumevariation and NLRvariation showed better stratification of patients for both OS and PFS compared to individual parameters. The log-rank test p-values for OS and PFS were 0.033 and 0.041 respectively for the combined score, outperforming individual parameters.

Limitations and Future Work

  • Retrospective Nature: The study is retrospective, which inherently limits the ability to establish causality and may be subject to selection bias.
  • Single-Center Study: Data was collected from a single institution, which may limit the generalizability of the findings.
  • NLR Influences: The authors acknowledge that NLR can be influenced by factors like infection or drug uptake, which were not assessed in this study.
  • Future Directions:
    • Multi-center Prospective Studies: Larger, prospective, multi-center studies are needed to validate these findings and confirm the predictive power of the proposed biomarkers.
  • Combination Therapies: The authors suggest that the findings could inform treatment strategies, potentially combining ICIs with other agents (like anti-CTLA-4) for patients identified as high-risk by these new biomarkers.
  • Radiomics and AI: The potential of radiomics to extract more detailed radiographic features and the application of AI algorithms are proposed as future avenues for developing even more sophisticated non-invasive biomarkers.

Summary updated

Abstract

Total tumor volume appears to be a prognostic biomarker of anti-PD-(L)1 response to immunotherapy in metastatic patients with MMRD tumors. Combining tumor volume and NLR with a simple concordance score stratifies patients well according to their survival and offers a good predictive measure of response to immunotherapy.
BibTeX
@inproceedings{belkouchi2022predicting,
  author = {Younes Belkouchi and Laëtitia Nebot-Bral and Littisha Lawrance and Michèle Kind and C. David and Samy Ammari and Paul-Henry Cournède and Hugues Talbot and Perrine Vuagnat and Cristina Smolenschi and Patricia Kannouche and Nathalie Chaput and Nathalie Lassau and Antoine Hollebecque},
  title = {{Predicting immunotherapy outcomes in patients with MSI tumors using NLR and CT global tumor volume}},
  booktitle = {Frontiers in Oncology},
  year = {2022},
  doi = {10.3389/fonc.2022.982790},
}
Figure