Comparing CT perfusion with oxygen partial pressure in a rabbit VX2 soft-tissue tumor model

Chuannan Sun, C. Li, Lv Hai-bo, Chao Zhao, Jinming Yu, Guanghui Wang, Yunxiu Luo, Y. Li, Xiao Ming-qing, Jun Yin, Jinyi LangJournal of Radiation Research 2013 · DOI

TL;DR

This study explores using CT perfusion imaging as a non-invasive method to assess tumor hypoxia, a critical factor in radiotherapy resistance. Researchers found a significant positive correlation between CT perfusion values (specifically blood flow) and direct oxygen partial pressure measurements in a rabbit VX2 tumor model, suggesting CT perfusion could be a valuable tool for identifying hypoxic targets in cancer treatment.

What Problem Does This Paper Address?

Hypoxia, or low oxygen levels, within solid tumors is a major challenge in cancer treatment. Hypoxic cells are notoriously resistant to radiotherapy and chemotherapy, leading to poorer patient outcomes, increased risk of recurrence, and difficulty in achieving complete remission. While direct measurement of tumor oxygen partial pressure (PO2) using oxygen electrodes is considered the "gold standard," it's an invasive procedure, often limited to superficial tumors, and can be cumbersome. Existing non-invasive methods have their own limitations. The paper addresses the need for a more accessible, non-invasive technique to accurately assess tumor hypoxia, which is crucial for effective radiotherapy planning and treatment. Specifically, it investigates whether CT perfusion imaging, a technique that measures blood flow and related parameters, can serve as a proxy for tumor oxygenation.

Methods

The researchers utilized a rabbit VX2 soft-tissue tumor model, a well-established preclinical model that mimics aspects of human head and neck cancers, liver cancer, and other solid tumors.

  • Tumor Induction: VX2 tumor cells were implanted into the hind legs of New Zealand white rabbits.
  • Oxygen Measurement: The "gold standard" oxygen microelectrode method (OxyLab pOM instrument) was used to directly measure the partial pressure of oxygen (PO2) at specific points within the tumor.
  • CT Perfusion Imaging: A 64-slice spiral CT scanner (Philips Brilliance) was employed. After PO2 measurement, the same region of interest (ROI) was targeted for CT perfusion. A contrast agent was injected intravenously, and dynamic scanning was performed to capture the contrast enhancement over time.
  • Perfusion Parameter Calculation: A commercial perfusion software package (Brilliance perfusion 2.1.1) was used, employing the maximum-slope (MS) method. This method derives parameters like blood flow (BF), blood volume (BV), time to peak (TTP), and peak enhancement intensity (PEI).
  • Statistical Analysis: Pearson correlation analysis was used to assess the relationship between CT perfusion parameters and PO2. ROC curve analysis was performed to determine diagnostic cutoff values for hypoxia.

Datasets

The study used data from 45 successfully scanned rabbits with induced VX2 tumors. No new benchmark datasets were introduced; rather, the study generated paired measurements of PO2 and CT perfusion parameters from the same tumor regions within these animal models.

Key Insights and Ideas

  • CT Perfusion as a Hypoxia Indicator: The core idea is that better blood flow (perfusion) should, in principle, lead to better oxygen delivery to tumor tissues. This study aimed to validate this assumption in a preclinical model.
  • Direct Correlation is Key: The study sought to establish a direct, quantitative link between a non-invasive imaging technique (CT perfusion) and a direct physiological measurement (PO2).
  • Blood Flow is the Strongest Predictor: Among the CT perfusion parameters evaluated, blood flow (BF) showed the most significant positive correlation with tumor PO2. This aligns with the intuitive understanding that increased blood flow can deliver more oxygen.
  • Implications for Radiotherapy: If CT perfusion can reliably indicate hypoxia, it could help clinicians identify "hypoxic targets" within tumors. These areas are less responsive to radiation, and precisely delineating them could allow for more targeted or escalated radiotherapy to improve treatment efficacy.

Key Results

  • Positive Correlation: A significant positive correlation was found between CT perfusion blood flow (BF) values and tumor PO2 (Pearson correlation coefficient r = 0.696, P < 0.0001).
  • Other Parameters Less Correlated: Blood volume (BV), peak enhancement intensity (PEI), and time to peak (TTP) showed weaker or non-statistically significant correlations with PO2.
  • Hypoxia Cutoff: A BF cutoff value of 8.9 ml/min/ml was identified as a potential indicator for diagnosing hypoxia (PO2 < 5 mmHg), achieving a sensitivity of 0.88 and specificity of 0.80.
  • Diagnostic Accuracy: The study reported an overall accuracy of 0.84 for using the perfusion value cutoff to diagnose hypoxia.

Limitations and Future Work

  • Limited Coverage: The 64-slice CT scanner used had a limited z-axis coverage (4 cm per rotation), meaning the entire tumor perfusion could not always be captured. This restricts its application for assessing overall tumor hypoxia percentage (HP). Newer scanners with wider coverage could overcome this.
  • Point-to-Point Measurement: The study performed point-to-point analysis, correlating a single PO2 measurement with the perfusion value of the exact same spot. This doesn't allow for assessing the ratio of perfusion values across the entire tumor or comparing them with overall tumor hypoxia metrics (like HP2, HP5, HP10).
  • Lack of Immunohistochemistry: The study acknowledges that tumor hypoxia was not further proven using immunohistochemical markers, which could provide additional validation.
  • Preclinical Model: The findings are from a rabbit model and need to be validated in human clinical trials.
  • Future Directions: The authors suggest that advancements in CT technology (e.g., 320-slice scanners) can enable full tumor perfusion imaging, making it more practical for assessing tumor HP and delineating hypoxic targets for radiotherapy. Further research is needed to translate these findings into clinical practice.

Summary updated

Abstract

The aim of this study was to evaluate the oxygen partial pressure of the rabbit model of the VX2 tumor using a 64-slice perfusion CT and to compare the results with that obtained using the oxygen microelectrode method. Perfusion CT was performed for 45 successfully constructed rabbit models of a VX2 brain tumor. The perfusion values of the brain tumor region of interest, the blood volume (BV), the time to peak (TTP) and the peak enhancement intensity (PEI) were measured. The results were compared with the partial pressure of oxygen (PO2) of that region of interest obtained using the oxygen microelectrode method. The perfusion values of the brain tumor region of interest in 45 successfully constructed rabbit models of a VX2 brain tumor ranged from 1.3-127.0 (average, 21.1 ± 26.7 ml/min/ml); BV ranged from 1.2-53.5 ml/100g (average, 22.2 ± 13.7 ml/100g); PEI ranged from 8.7-124.6 HU (average, 43.5 ± 28.7 HU); and TTP ranged from 8.2-62.3 s (average, 38.8 ± 14.8 s). The PO2 in the corresponding region ranged from 0.14-47 mmHg (average, 16 ± 14.8 mmHg). The perfusion CT positively correlated with the tumor PO2, which can be used for evaluating the tumor hypoxia in clinical practice.
BibTeX
@article{sun2013comparing,
  author = {Chuannan Sun and C. Li and LV Hai-bo and Chao Zhao and Jinming Yu and Guanghui Wang and Yunxiu Luo and Y. Li and Xiao Ming-qing and Jun Yin and Jinyi Lang},
  title = {{Comparing CT perfusion with oxygen partial pressure in a rabbit VX2 soft-tissue tumor model}},
  journal = {Journal of Radiation Research},
  year = {2013},
  doi = {10.1093/jrr/rrt092},
}
Figure