ADVERTISEMENT
Percentage of residual fibrosis versus Rubbia-Brandt classification as predictors of overall survival in metastatic colorectal cancer. Which is better?
Background
Pathologic response to preoperative chemotherapy for colorectal liver metastases (CRLM) has been associated with survival after hepatectomy in some publications. Rubbia-Brandt L et al stablished a five pathological categories classification with prognostic value in this setting. Poultsides GA et al, in 2012, published that fibrosis is the predominant chemotherapy-related pathologic alteration driving prognostic information. There are no publications comparing both pathologic prognostic classifications in a Cox regression model including other clinical variables that influence prognosis.
Methods
The primary end-point of this study is to demonstrate that anti-Epidermal Growth Factor Receptor treatment (anti-EGFR) (cetuximab or panitumumab) associated to chemotherapy induces more fibrosis (>40%) than bevacizumab in patients with CRLM before hepatectomy. Secondary endpoints are to evaluate differences in median overall survival (OS) related to clinical (tumor location, TNM stage, CEA levels, treatment schedule...) and pathological data (fibrosis, necrosis, Rubbia-Brandt categories, liver toxicity...) in a Cox Regression Model; to stablish its prognostic value and the magnitude of this effect. This is a retrospective study (N sample calculated 108 patients). The protocol was approved by the research ethics committee in 2019. The Fisher exact test was used to compare 2x2 proportions and the Cox Regression Model was used to analyse and compare survival data. Statistical analysis was performed with SPSS package v20.0.
Results
We report the second interin analysis results after reviewing 77 patients. 23 patients (59%) in the anti-EGFR group showed fibrosis>40% compared to 10 patients (26.3%) in the bevacizumab group, Fisher exact test p=0.006. In the survival multivariate analysis only tumor location (right vs left; HR 0.516, IC95 0.272-0.979), primary tumor removed (no vs yes; HR 0.169, IC95 0.062-0.460) and tumor fibrosis (40%; HR 0.418, IC 0.229-0.761) showed significant prognostic value. The Rubbia-Brandt Categorized Classification (TR3,4,5 vs TR1,2; HR 0.573, IC95 0.320-1.026), treatment schedule (anti-EGFR vs bevacizumab), pT or pN stage didn’t provided significant prognostic information in the uni and multivariated analysis.
Conclusions
Treatment with anti-EGFR before hepatectomy in CRLM induces significant more residual fibrosis when compared to bevacizumab based-treatment. Fibrosis percentage (more or less 40%) provide more prognostic information than Rubbia-Brandt Categorized Classifications in our serie.
Legal entity responsible for the study
The author.
Funding
Has not received any funding.
Disclosures
X. Hernández-Yagüe: Advisory / Consultancy: SANOFY-AVENTIS S.A. All other authors have declared no conflicts of interest.
