Nicolas Azzopardi(1), Sarah Lobet(2), David Ternant(3,4), Paul-Armand Dujardin(5), Thierry Lecomte(2,6).
(1) EA7501 GICC, University of Tours, Tours, France. (2) Inserm UMR 1069, N2C, University of Tours, Tours, France. (3) Department of clinical pharmacology, Tours University Hospital, Tours, France. (4) EA 4245 T2I, University of Tours, Tours, France. (5) Inserm CIC 1415, Tours University Hospital, Tours, France. (6) Department of Hepatogastroenterology, Tours University Hospital, Tours, France.
Objectives: The liver is the most frequent site of colorectal cancer (CRC) metastases [1]. The growth of a metastases is assured by its’ ability to stimulate neo-angiogenesis mainly by the secretion of vascular endothelial growth factor (VEGF). Clinical efficiency [2] and influence of the survival [3] of anti-VEGF bevacizumab (Avastin®) has been demonstrated in the treatment of solid tumours in association with chemotherapy. Contrast-enhanced ultrasound (CEUS) can be used for imaging the vascular network, thanks to the microbubbles contrast agent that remains strictly within the vessels [4] and thus can be used for evaluating the vascularisation at inclusion and the influence of anti-angiogenic treatments on it [5]. However, this evaluation remains qualitative in clinics due to the high inter-device, inter-operator and inter-occasion variabilities. Population modelling can be useful to bypass this limitation and to propose a homogeneous cross-centre quantitative comparison of metastatic vascularisation. Therefore, this work aimed at (i) propose a validated mechanistic model of metastatic liver CEUS using population compartmental analysis, (ii) investigating the relationship between initial metastasis vascularisation quantification and bevacizumab pharmacokinetics and survival.
Methods: STIC-Avastin was a French multicentre non-comparative, prospective, open-label, observational study (NCT00489697). Previously untreated metastatic CRC patients received four cycles of bevacizumab 5 mg/kg intravenously every 2 weeks in combination with chemotherapy. CEUS examination was performed at inclusion for 103 patients. After IV injection of the contrast agent, an 80-second sequence was acquired at the level of the target region. Then two regions of interest (ROI) were defined by the investigator. The first outlining the metastasis and the second the surrounding normal liver parenchyma for reference. The model described two observation compartments M and R, representing metastasis and reference tissue ROIs, respectively. The structure of the model was identical for M and R. A fraction of the contrast agent dose ρ entered the M sub-model and the complement 1-ρ entered the R sub-model. For each sub-model, the fraction of the dose was absorbed through a chain of 5 transit compartments characterised by mean absorption time MAT. The partial reperfusion of M and R by the contrast agent was described by a looping chain of 5 transit compartments characterised by a mean transit time MTT and a proportion lost ρlost. To work around the high IIV in observed intensity, total contrast agent dose was set to 1. The scale correction was assured by the estimation of a global gain G defined as OBS_M=M*10^(G/10) and OBS_R=R*10^(G/10). Parameters were estimated using nonlinear mixed effect model (Monolixsuite 2021R1, lixoft, antony, Frannce).
Results: Accurate description of CEUS intensity over time were obtained with the estimation of all parameter values and their variabilities with r.s.e. < 10%. The fraction ρ of the dose visible in M was 43%. MAT_M and MAT_R were 10.6 and 15.4 s, respectively. MTT_M and MTT_R were 10.2 and 6.33 s, respectively. The global gain G was estimated as a normal distribution with mean 35.1 dB and high variability ωG = 5.72 (r.s.e. = 7%). No significative correlation between estimated parameters and bevacizumab pharmacokinetic parameters [3] were observed. Patients with higher MAT_M than MAT_R had a significative poorer progression-free survival (PFS) (p = 0.023) and poorer overall survival (OS) (p = 0.036). Patients with higher MTT_M than MTT_R had also a significative poorer OS (p = 0.0016).
Conclusions: Our modelling approach provides robust and cross-centre estimation of metastatic vascularisation dynamics. Comparing mean absorption time of contrast agent in metastasis (MAT_M) to the mean time in reference tissue (MAT_R) can reflect the metastatic relative vascularisation capacity. This study quantified that CRC patients with excessive metastatic vascularisation derive more benefit from an anti-VEGF treatment. This quantification can be tested as early predictor of clinical efficacy of anti-angiogenic treatments in mCRC, which should be evaluated further in upcoming clinical trials.
References:
[1] Binder-Foucard F et al. Rev Epidemiol Sante Publique. 2014 Apr;62(2):95-108.
[2] Hurwitz H et al. N Engl J Med. 2004 Jun;350(23):2335-42.
[3] Caulet M et al. Clin Pharmacokinet. 2016 Nov;55(11):1381-1394.
[4] Krix M et al. Ultrasound Med Biol. 2003 Oct;29(10):1421-30.
[5] Tranquart F et al. Ultraschall in Med. 2018 Oct;39(5):544-558.
Reference: PAGE 30 (2022) Abstr 10214 [www.page-meeting.org/?abstract=10214]
Poster: Methodology - Other topics