The clinical management of patients with RCC remains difficult, and the development of adequate diagnostic, prognostic and therapeutic tools is still required. In this issue of the BJUI, Zhang et al. [1] publish the results of a study on the prognostic role of carbonic anhydrase IX (CAIX), an antigen almost invariably present in clear-cell RCC (ccRCC), but not in the normal kidney. CAIX is a Von-Hippel-Lindau (VHL) protein-mediated anhydrase, governed by hypoxia-inducible factor (HIF)-1α, and thought to be engaged in the regulation of hypoxia-induced cell proliferation. The distribution of CAIX (or G250) was first described in 1986 in an immunohistochemical study assessing tumour specificity of the monoclonal antibody G250. Later, Liao et al. [2] reported the presence of higher CAIX levels in ccRCC compared with other subtypes. Since then, a vast investigative interest in CAIX has emerged, which has also led to the development of treatment and diagnostic strategies targeting CAIX.

Zhang et al. now provide long-term follow-up data of a study cohort previously analysed in 2007, including 730 cases of surgically treated ccRCC. CAIX was expressed in 708 (97.0%) of the specimens with 163 tumours (22.3%) exhibiting low (≤85%) and 567 (77.7%) exhibiting high (>85%) expression. A strong univariate inverse correlation was observed for CAIX expression with nuclear grade and tumour necrosis. Low expression was further associated with increased cancer-specific mortality (hazard ratio:1.62; P < 0.001) and a higher risk of developing metastasis (hazard ratio:1.65; P = 0.001); however, after adjustment forFuhrman grade and tumour necrosis, low CAIX expression neither significantly affected cancer-specific mortality nor risk of metastasis.

Some results of this study by Zhang et al. partially contradict previous studies on this topic. Studies examining tissue microarrays (TMAs) in series of 100–359 patients with ccRCC led to the conclusion that low CAIX expression (≤85%) is an independent adverse predictor of outcome and is associated with a higher risk of lymphatic spread and higher nuclear grade [3, 4]. Another study showed that inclusion of biomarkers, such as CAIX, can improve prognostic accuracy compared with the TNM system and the UCLA Integrated Staging System [5]. In addition, CAIX was considered to predict response to interleukin-2 immunotherapy [6]. Based on these studies, a recent review article by Stillebroer et al. [4] favoured the use of CAIX expression as an independent prognostic marker and also suggested it represented an attractive target for ccRCC imaging and treatment.

Discordance between the results of available studies may at least partially be explained by the variable numbers of patients and specimens examined. The present study has evaluated the largest group of consecutive patients with a relatively long median follow-up of 13.8 years [1]. In addition, immunohistochemical assessment was performed on whole-tissue sections, while previous studies used TMAs instead. Although Stillebroer et al. [4] state that most ccRCC lesions express CAIX homogeneously, in the present study a significant heterogeneity within whole-tissue sections was detected, suggesting unreliability of TMAs for CAIX expression assessment [1]. Some differences might furthermore be explained by adjustment to varying parameters in multivariable analysis.

To date, the reliability and applicability of molecular markers such as CAIX in RCC has not been discussed. Indeed, such markers are clearly required, especially because there is a wide variability of histological and prognostic patterns in RCC, which often hampers accurate prognosis prediction based only on classical prognostic variables such as TNM classification and Fuhrman grade. Furthermore, the use of new anti-angiogenic treatment strategies requires the development of reliable prognostic and predictive markers. Nevertheless, currently we cannot rely on the significance of CAIX. Amongst other factors, this problem might also be based on a lack of clear definition with regard to CAIX expression. Various questions remain. What actually is a positive diagnostic finding and what is not? Is immuno-histochemical assessment the way to go, or do we need to apply different methods? Which threshold should be applied to define positive vs negative expression? What kind of specimens should be used? And finally, which patients should we focus on: patients with localized RCC to assess tumour aggressiveness and recurrence risk or patients with metastasis to predict treatment response or both?

Currently the role of CAIX in prognosis and treatment prediction and as a diagnostic and therapeutic tool remains undetermined. Although high CAIX expression has been considered a predictor of response to interleukin-2 immunotherapy, CAIX expression status was neither predictive for the clinical benefit of treatment with sorafenib nor prognostic after cytokine therapy, based on the results of a recent study [7]. Regarding treatment strategies using CAIX as a therapeutic target, a recently finished phase III trial has not yet proven its efficacy in the adjuvant setting, despite promising results from phase I/II studies [8]. By contrast, studies using immuno-positron emission tomography as an imaging method for ccRCC lesions with monoclonal antibody cG250 targeting CAIX have shown promise for the use of CAIX as a potential diagnostic tool for undetermined renal masses and for staging purposes [8]. Finally, further improvements in our ability to predict RCC prognosis will probably rely on the integration of molecular and genetic marker panels, including CAIX and further possible prognostic markers (e.g. microvascular density, vascular endothelial growth factor, VHL and HIF) into established prognostic models.


  1. Top of page
  2. References