Evaluation of diffusion-weighted MRI and (18F) fluorothymidine-PET biomarkers for early response assessment in patients with operable non-small cell lung cancer treated with neoadjuvant chemotherapy

Dominic Carlin, Alexander Weller, Gem Kramer, Yan Liu, John C. Waterton, Arturo Chiti, Martina Sollini, Adrianus J. de Langen, Mary O'Brien, María Urbanowicz, Bart K. M. Jacobs, Nandita M. deSouzaBJR|Open 2019 · DOI

TL;DR

This study investigates whether diffusion-weighted MRI (DW-MRI) and 18F-fluorothymidine PET (FLT-PET) can provide early indicators of treatment response in non-small cell lung cancer (NSCLC) patients undergoing neoadjuvant chemotherapy. While these imaging biomarkers showed early changes that preceded tumor size reduction, they did not correlate with histopathological measures of response, suggesting their utility might lie in predicting clinical outcomes rather than direct histological assessment.

What Problem Does This Paper Address?

Non-small cell lung cancer (NSCLC) patients often receive neoadjuvant chemotherapy to shrink tumors before surgery. Current response assessment relies on RECIST criteria, which measure tumor size and typically take 2-3 months to show changes. This delay means patients might continue ineffective treatments for too long. The challenge is to find earlier, more sensitive biomarkers that can predict treatment effectiveness sooner, allowing for timely adjustments to therapy. This paper explores whether changes in diffusion-weighted MRI (DW-MRI) and 18F-fluorothymidine PET (FLT-PET) can serve as these early indicators by reflecting biological changes within the tumor.

Methods

The study employed a prospective, multicenter, single-arm imaging trial (EORTC-1217) involving 14 patients with operable NSCLC receiving platinum-based neoadjuvant chemotherapy.

  • Imaging Modalities:
    • DW-MRI: Used to measure the apparent diffusion coefficient (ADC), a marker of cellularity. Whole-tumor ADC histograms were generated, focusing on median ADC (ADCmedian), 25th and 75th percentiles, and interquartile range (IQR).
  • FLT-PET/CT: Used to assess tumor proliferation via 18F-fluorothymidine uptake. Standardized Uptake Value (SUV) metrics (SUVmean, SUVmax, SUVpeak) and total lesion proliferation (TLP) were calculated.
  • Timing: Imaging was performed at three time points: baseline (before chemotherapy), Day 14 post-treatment, and within one week of surgery.
  • Histopathology: Surgical specimens were analyzed to estimate the percentage of residual viable tumor/necrosis and the proliferation index (Ki67 staining).
  • Statistical Analysis: The study compared changes in imaging biomarkers (ADC and SUV) over time using Wilcoxon signed-rank tests. Correlations between imaging biomarkers and histopathological measures were assessed using Spearman rank correlation. Repeatability estimates for imaging metrics were used to determine if observed changes exceeded measurement variability.

Datasets

The study utilized data from 14 patients enrolled in the EORTC-1217 trial. Imaging data (DW-MRI and FLT-PET/CT) and histopathological data were collected from these participants. No new benchmark datasets were introduced.

Key Insights and Ideas

  • Early Biological Changes Precede Size Reductions: Both DW-MRI (ADC) and FLT-PET (SUV) showed significant changes by Day 14 of treatment, well before any RECIST-defined size changes would typically be observed. This highlights the potential of these techniques for early response assessment.
  • ADC Increases, FLT-SUV Decreases: Specifically, ADC median increased and SUV mean decreased significantly between baseline and Day 14, indicating increased cell death/necrosis and decreased proliferation, respectively.
  • No Direct Correlation with Histopathology: Despite early changes in imaging biomarkers, the study found no significant correlation between these changes and histopathological estimates of response (percentage of viable tumor or necrosis). This suggests that while these biomarkers reflect biological processes, they may not directly map to the specific histological features assessed in this study.
  • Repeatability Matters: The study emphasized the importance of comparing imaging changes against established measurement repeatability. Changes exceeding these limits were considered more likely to represent true biological effects.
  • Responders Show Larger Early Changes: Exploratory analysis suggested that patients who responded to treatment (based on RECIST criteria) generally exhibited larger increases in ADC and decreases in FLT-SUV at Day 14 compared to non-responders, though small sample sizes limited statistical power.

Key Results

  • Significant Early Changes: ADCmedian increased (p < 0.001) and SUVmean decreased (p < 0.001) significantly between baseline and Day 14.
  • Limited Correlation with Histology: No meaningful correlation was found between changes in ADC or SUV and the percentage of viable tumor cells or necrosis.
  • Changes Exceed Variability: Measurable increases in ADC (particularly ADC75th) and decreases in FLT-SUV metrics were observed in responders, often exceeding established limits of measurement variability.
  • Volume Changes: Tumor volumes generally decreased by Day 14 across all imaging modalities (DWI, FLT, and CT), though these changes were not always statistically significant or consistent.

Limitations and Future Work

  • Small Sample Size: The study was limited by a small number of patients (n=14), particularly in the responder group (n=3), which restricted the statistical power to demonstrate correlations and draw definitive conclusions.
  • Patient Recruitment Challenges: Stringent inclusion criteria requiring operability and issues like disease progression and toxicity significantly impacted recruitment.
  • Histopathology Limitations: The semi-quantitative nature of histopathological analysis and the potential for selected sections to not be representative of the entire lesion were noted as limitations. Future work could explore more robust measures like digital pathology for tumor cellularity.
  • Confounding Factors: Peritumoral atelectasis and inflammation could confound RECIST-based size measurements, especially when segmenting volumes.
  • Future Directions: The authors suggest that while these imaging biomarkers may not directly reflect histopathological necrosis, their utility in predicting clinical outcomes (e.g., progression-free survival, overall survival) or non-response based on tumor size warrants further investigation. Future studies should aim for larger sample sizes and potentially explore different histopathological endpoints.

Summary updated

Abstract

During treatment, ADC increase precedes size reductions, but does not reflect histopathological necrosis.
BibTeX
@inproceedings{carlin2019evaluation,
  author = {Dominic Carlin and Alexander Weller and Gem Kramer and Yan Liu and John C. Waterton and Arturo Chiti and Martina Sollini and Adrianus J. de Langen and Mary O'Brien and María Urbanowicz and Bart K. M. Jacobs and Nandita M. deSouza},
  title = {{Evaluation of diffusion-weighted MRI and (18F) fluorothymidine-PET biomarkers for early response assessment in patients with operable non-small cell lung cancer treated with neoadjuvant chemotherapy}},
  booktitle = {BJR|Open},
  year = {2019},
  doi = {10.1259/bjro.20190029},
}
Figure