Liver metastases occur in 30% of all patients with colorectal cancer (CRC), and are responsible for death in at least two thirds of CRC patients. Approximately 80% of patients with colorectal liver metastasis (CLM) present with unresectable disease at initial diagnosis (1, 2). Profound improvements in CLM outcome seem to be associated with advancements in medical therapy, and the increased use of hepatic resection (3). Additionally, a disease-free interval longer than 5 years can now be achieved in 16% of patients with initially unresectable CLM, resected after downsizing chemotherapy. Therefore, the importance of CLM tumor response evaluation following chemotherapy has gradually increased (4).

Therefore, an imaging test is necessary for the prognostic prediction of the response of hepatic metastasis to chemotherapy, in order to identify patients likely to respond poorly to a chemotherapeutic regimen. This may allow for changes to the treatment plan that avoid unnecessary drug toxicity and maximize chances of tumor regression.

Accurate evaluation of therapeutic responses to treatment could have considerable clinical benefit for patients who receive chemotherapy. Although target therapies are increasingly being used in patients with advanced cancer, it is clear that standard tumor response evaluation methods are of limited value for assessment of treatment modality efficacy. Because of their varying mechanisms of action, it follows that the response patterns with target therapies differ from those seen with cytotoxic treatment. According to a previous renal cell carcinoma trial, the Response Evaluation Criteria in Solid Tumors (RECIST) often underestimate the effect of targeted therapy, because necrosis or hemorrhage frequently occur without any change in size (5, 6). In addition, it remains to be determined whether these one-dimensional criteria can sufficiently reflect treatment responses to combined targeted biological therapy, although RECIST criteria is widely used for the evaluation of responses to cytotoxic chemotherapy (7, 8).

In previous studies, volumetric measurement has been preferred over RECIST criteria, to predict tumor progression (9, 10). Volumetric analysis has the ability to account for three dimensions, as opposed to RECIST criteria, in which only a single axial dimension is analyzed. Therefore, we hypothesized that volumetric analysis would be superior to RECIST criteria, for the evaluation of tumor response after targeted therapy. The aim of this study was to compare RECIST and volumetric measurements of tumor response and outcome in CLM treated with a targeted agent plus folinic acid, fluorouracil, and irinotecan (FOLFIRI).

Our study demonstrates that volumetric measurement may have the potential to predict overall survival in patients with unresectable CLM treated with target therapy. Accurate and early detection of treatment response of liver metastases is important for optimal intervention planning (12). Early, reliable prognostic information may help physicians to develop proper treatment plans for individual patients, and allow for more timely attempts of alternative therapies for treatment-resistant tumors.

RECIST criteria is easy to use, allows for rapid classification, and is highly reproducible and therefore, is widely used in clinical trials and clinical practice (13). However, increases in the sophistication of imaging instrumentation has led to the development of markedly diverse and complex oncologic therapies (14). For example, volumetric image assessment allows precise three-dimensional measurement of tissue volumes (15). In this study, the nonprogressing group showed longer overall survival than the progressing group based on volumetric criteria, while the RECIST criteria was not significantly different between them. Three patients were additionally assessed as PD according to volumetric criteria and SD according to RECIST criteria. We hypothesized that volumetric evaluation allowed precise measurement of the geometric tumor and reflected smaller changes more than the traditional criteria (16, 17). In addition, there was no significant difference between the responder and non-responder groups according to the volumetric and RECIST criteria. Molecular targeted therapy is often associated with potential intratumoral hemorrhage and peritumoral edema; this is attributed to increasing tumor size despite good clinical response. Because of the apparent increase in lesion size, SD may often be misinterpreted as PD (18, 19).

The National Comprehensive Cancer Network and European Society for Medical Oncology guidelines (20, 21) suggest an active combination regimen doublet chemotherapy plus a targeted agent (e.g. bevacizumab or cetuximab for KRAS wild-type tumors). Bevacizumab, a monoclonal antibody against vascular endothelial growth factor (VEGF), prevents VEGF from binding to its functional receptor, leading to the inhibition of tumor vessel growth, neovascularization, and decreased permeability in the surviving vasculature. Cetuximab, a monoclonal antibody against epidermal growth factor receptor, inhibits downstream signaling pathways, leading to the inhibition of cell proliferation and angiogenesis (22). Targeted therapies have different mechanisms of action than cytotoxic chemotherapy. Some agents induce apoptosis, whereas other agents stop progression. Due to these differences in the mechanisms of action, tumors treated with targeted therapies may not demonstrate radiographic findings similar to that of tumors treated with conventional cytotoxic therapies (23). Therefore, traditional diameter–based criteria can lead to improper classification of treatment response for tumors treated with targeted therapies.

Our results further support the necessity for revised RECIST criteria that includes the use of volume measurements, in the response assessment of metastatic tumor burden in the liver after treatment with target agents. Similarly, another group has reported the need for revised RECIST criteria, adopting the concept of volumetry superiority over unidimensional measurements (24). Conversely, volumetric measurements are cumbersome and time consuming, and thus, are unlikely to be widely adopted in clinical practice. However, recent technological advancement in medical imaging software has promised to provide solutions for rapid, easy, and accurate volume measurements (25, 26).

Response evaluation with RECIST criteria is limited by the use of a unidimensional measurement as a surrogate marker for three-dimensional growing tumors. Conventional oneor two-dimension tumor measurements are made based on the assumption that lesion diameter is correlated to lesion volume. This assumption is due to the inaccurate belief that tumors grow and shrink in a spherical pattern. Volumetric measurements may more accurately reflect actual tumor size and shape compared to conventional measurements (8).

Advances in CT technology allow for the acquisition of isotropic data with accurate and reproducible tumor volumetric data (15). Semi-automated segmentation techniques use software algorithms to measure tumor volume based on ROI outlines drawn by the reader. Recently, the feasibility of volumetric tumor measurement as well as semiautomatic quantification of viable tumor (enhancing portion), has been shown for hepatocellular carcinoma (HCC) after transcatheter arterial chemoembolization (27). Volumetric quantification of HCC necrosis has been shown to be more reproducible than quantification using EASL criteria (28). As RECIST is primarily based on one-dimensional measurements of the longest diameters of the tumors in a transverse image, it is often difficult for radiologists to determine the longest diameters in cases in which the target lesions split into smaller tumor, or merge to form a conglomerate mass after intervention (9, 29). A previous study used RECIST to assess responses to imatinib in gastrointestinal stromal tumors, and found that the time to tumor progression did not differ between good and poor responders, suggesting that RECIST is not effective for this assessment. Similarly, Chun et al. (30) demonstrated that response assessment based on RECIST criteria was not correlated to overall survival in patients with CLM treated with target agents. Instead, interest and acceptance of volumetric evaluation as an alternative assessment method has increased (31, 32). An important theoretical advantage of volumetric evaluation is that it offers more accurate measurement and reflection of tumor burden, with less inter-observer variability, and is superior to measurements of the longest diameters from up to five target lesions per organ (15). Previous studies have also reported that pretreatment tumor volume of HCC is a potential predictor of posttransplant recurrence (33, 34).

Our study has several limitations. First, our study had a small number of patients. A larger series is needed to clinically validate our results. Second, our analysis was a retrospective assessment of dominant index lesions and non-target lesions were not included. However, several studies established that prognostic use of the dominant lesion, is sufficient to predict outcome after treatment (30, 35, 36). This access is based on the assumption that the longest tumors can reflect tumor burden and, represent the most aggressive condition of the cancer. Third, we enrolled CLM patients treated with different target agents (bevacizumab and cetuximab). However, both bevacizumab and cetuximab are now routinely used molecular target agents in metastatic CRC. Thus, this study could reflect real-time clinical practice.

In conclusion, volumetric assessment of liver metastases may become the preferred method to detect tumor progression, and predict overall survival for patients with unresectable liver metastases from CRC treated with a targeted agent.