Gastrointestinal Imaging Original Research

Transcription

1 Gastrointestinal Imaging Original Research Chung et al. Response Evaluation in Patients With Colorectal Liver Metastases Gastrointestinal Imaging Original Research Woo-Suk Chung 1 Mi-Suk Park 1 Sang Joon Shin 2 Song-Ee aek 1 Yeo-Eun Kim 1 Jin Young Choi 1 Myeong-Jin Kim 1 Chung WS, Park MS, Shin SJ, et al. Keywords: colorectal cancer, liver metastasis, modified CT criteria, RECIST 1.1, Response Evaluation Criteria in Solid Tumors (RECIST), treatment response, tumor density, tumor size DOI:1214/JR Received September 6, 2011; accepted after revision February 24, This study was supported by a grant from the National Research Foundation of Korea ( ). 1 Department of Diagnostic Radiology, Institute of Gastroenterology, Research Institute of Radiological Science, Severance Hospital, Yonsei University College of Medicine, Seodaemun-ku, Shinchon-dong 134, Seoul , Republic of Korea. ddress correspondence to M. S. Park (radpms@yuhs.ac). 2 Department of Internal Medicine, Severance Hospital, Yonsei University College of Medicine, Seoul, Republic of Korea. JR 2012; 199: X/12/ merican Roentgen Ray Society Response Evaluation in Patients With Colorectal Liver Metastases: RECIST Version 1.1 Versus Modified CT Criteria OJECTIVE. Our retrospective study compared Response Evaluation Criteria in Solid Tumors (RECIST) version 1.1 with modified CT criteria to determine their respective utilities as a prognostic indicator. Using both sets of criteria, we compared treatment responses of patients with colorectal liver metastases treated with bevacizumab-containing chemotherapy with those of patients treated with chemotherapy alone. MTERILS ND METHODS. Fifty-nine patients who received bevacizumab-containing chemotherapy (n = 30, group 1) or chemotherapy alone (n = 29, group 2) for the treatment of colorectal carcinoma underwent contrast-enhanced CT before treatment and 2 months after treatment. Two radiologists determined changes in tumor size and density between the pretreatment and 2-month follow-up CT images. RECIST 1.1 assesses responses on the basis of changes in tumor size, and the modified CT criteria assesses responses on the basis of changes in tumor density and size. Responses were correlated with time to tumor progression by log rank test. RESULTS. ccording to RECIST 1.1, nine of 30 patients (30%) in group 1 and 12 of 29 patients (41%) in group 2 were good responders. ccording to the modified CT criteria, 23 of 30 patients in group 1 (77%) and 23 of 29 patients in group 2 (79%) were good responders. s assessed by the modified CT criteria, good responders in both groups had significantly longer time to tumor progression than poor responders (p < 5). s assessed by RECIST 1.1, good responders in group 1 had significantly longer time to tumor progression than poor responders (p = 4), but there was no difference in group 2. CONCLUSION. Evaluating treatment response with tumor size and density changes on CT was a better predictor of time to tumor progression than changes in tumor size alone in both groups. C olorectal carcinoma is one of the most common cancers worldwide. pproximately 60% of patients with colorectal carcinoma eventually develop lymph node or distant metastasis, with the liver as the initial site in 30% of distant metastases [1]. In patients with unresectable metastases, chemotherapy is the treatment of choice, either as palliative therapy or to render metastases resectable [2 5]. lthough prolonged overall survival is the ultimate proof of the effectiveness of anticancer drugs, response evaluation during treatment is needed to guide treatment and to reassure the patient. Tumor shrinkage can be readily evaluated and shows that the selected therapy is effective. The therapeutic response of solid tumors has traditionally been evaluated by serial tumor size measurements, as described by the Response Evaluation Criteria in Solid Tumors (RECIST) [1, 6 9]. How- ever, some studies have reported that change in tumor size is only weakly associated with overall survival [5, 7]. Meanwhile, the RECIST Working Group has recently written revised RECIST guidelines that is, RECIST version 1.1 to simplify, optimize, and standardize the original criteria [8]. The effectiveness of new targeted drugs that lack intrinsic direct cytotoxic activity further challenges the concept of tumor size alone as a good indicator of efficacy [5]. Choi et al. [6] proposed new CT response criteria to assess the response of gastrointestinal stromal tumors (GISTs) to the tyrosine kinase inhibitor imatinib mesylate (Gleevec, Novartis) (Table 1). Their new CT response criteria for GIST patients ( 10% decrease in tumor size or % decrease in tumor density on CT) correlated better with time to tumor progression than RECIST. Choi et al. [6] reported that small changes in tumor size or density JR:199, October

2 Chung et al. on CT were sensitive and specific criteria for assessing the GIST response to imatinib, whereas RECIST was insensitive. In patients with advanced metastatic colorectal cancer, bevacizumab (vastin, Genentech), a recombinant human monoclonal antibody against vascular endothelial growth factor α, has recently been used with cytotoxic chemotherapy such as 5-fluorouracil, leucovorin, irinotecan, and oxaliplatin [3, 10, 11]. The addition of bevacizumab to chemotherapy is associated with improved survival in patients with stage IV colorectal cancer and higher pathologic response rates in patients undergoing resection of colorectal liver metastases [3, 10 12]. In a number of cancer types, the new molecular targeted agents often inhibit growth rather than shrink tumors, resulting in low objective response rates as assessed by RECIST, but improved survival. Therefore, the accuracy of the current response criteria should be reexamined with the goal of establishing simple, objective, and quantitative response evaluation criteria that are well correlated with long-term response. The purpose of this study was to compare the response of colorectal cancer patients with liver metastases treated with chemotherapy alone with the response of patients treated with chemotherapy combined with bevacizumab using both RECIST 1.1 and the modified CT criteria proposed by Choi et al. [6]. We aimed to determine which criteria were a better prognostic indicator for predicting time to tumor progression. Materials and Methods This retrospective study received institutional review board approval and was exempted from informed consent requirements. Patients We identified 347 consecutive patients with colorectal liver metastases treated from March 2006 to ugust Of these patients, 223 were treated with chemotherapy that consisted of either 5-fluorouracil, leucovorin, and oxaliplatin (or FOLFOX ) or 5-fluorouracil, leucovorin, and irinotecan (or FOLFIRI ); 124 patients received chemotherapy combined with bevacizumab. Whether to receive bevacizumab was the patient s choice. Patients were excluded from the study if they underwent hepatic resection or radiofrequency ablation after treatment (n = 19), had unenhanced CT images only (n = 3), or did not undergo regular follow-up CT examinations performed during the same portal venous phase (n = 266). Our study population consisted of 59 patients with regular follow-up CT who underwent contrast-enhanced CT of the abdomen within 1 month before treatment and 2 months after treatment and then underwent 2 3 months of follow-up CT until disease progression. Thirty patients (16 men, 14 women; mean age, 54.7 years; age range, years) received chemotherapy with bevacizumab and 29 patients (17 men, 12 women; mean age, 64 years; age range, years) received chemotherapy alone. The average follow-up periods for the bevacizumab-containing chemotherapy group and the chemotherapy alone group were 6.9 months (range, months) and 7.6 months (range, months), respectively. MDCT Technique CT examinations were performed on a 16- MDCT scanner (Somatom Sensation 16, Siemens Healthcare) or a 64-MDCT scanner (Somatom Sensation 64, Siemens Healthcare). ll patients fasted for at least 4 hours before the examination and received a 2 ml/kg IV dose (total volume, < 0 ml; 3 ml/s) of nonionic contrast material (iopromide [Ultravist 300, ayer HealthCare]) by power injector (EnVision CT, Medrad). fter unenhanced CT images were obtained, portal venous phase images were acquired seconds after the administration of contrast medium. The scanning parameters were as follows: collimation, 16 rows 0.75 mm or 64 rows mm; gantry rotation speed, 0.5 second; section thickness, 3 mm; image reconstruction increment, 1 mm; 120 kv; and effective tube current exposure time product, ms. Image nalysis Transverse portal venous phase images were displayed on a PCS (Integrated Imaging Solution, GE Healthcare) and evaluated in consensus by two radiologists with 8 and 5 years of clinical experience interpreting CT images, respectively. The radiologists were informed that the patients had colorectal liver metastases but were blinded to treatment. They measured tumor size and density for each patient at both time points. Tumor size Tumor size at each time point (i.e., before treatment and 2 months after treatment) was measured at the longest cross-sectional dimension on a PCS. If the patient had multiple colorectal liver metastases, the largest five lesions more than 1 cm in diameter were selected for target lesions. The target lesions were selected on pretreatment CT. If the lesion was seen on follow-up CT, it was considered a target lesion. However, if it was not seen on follow-up CT, we had to change the target lesion because we could not evaluate the size and density change. The longest dimensions of each patient s target lesions were added together. The absolute change and percentage change of the sum were then calculated for each patient. CT attenuation coefficient of tumors On the same PCS, the CT attenuation coefficient (density) of each metastasis was measured in Hounsfield units by drawing a maximum region of inter- TLE 1: Response Evaluation Criteria in Solid Tumors (RECIST) Version 1.1 and Modified CT Criteria [6] Response RECIST 1.1 Modified CT Criteria CR Disappearance of all tumors Disappearance of all tumors PR 30% decrease in target lesion size 10% decrease in target lesion Size a or % decrease in tumor density (in Hounsfield units) of target lesion on CT No obvious progression of nonmeasurable disease SD Does not meet the criteria for CR, PR, or PD Does not meet the criteria for CR, PR, or PD PD 20% increase in the sum of the longest diameters of target tumors 10% increase in target lesion size and does not meet criteria of PR by tumor density (in Hounsfield units) on CT New lesions New lesions Note CR = complete response, PR = partial response, SD = stable disease, PD = progressive disease. a The sum of longest diameters of target lesions as defined in RECIST JR:199, October 2012

3 Response Evaluation in Patients With Colorectal Liver Metastases est (ROI) inside the tumor margin at the level of the longest cross-sectional dimension of tumor at each time point (pretreatment and 2 months posttreatment). Each ROI was measured using freehand polygonal ROI to encompass the entire lesion as much as possible excluding the pseudocapsule if one was present (Fig. 1). The tumor density measurements of all lesions were combined, and the mean value (in Hounsfield units) of the five target lesions for each patient was calculated. The absolute and percentage changes in CT density were calculated for each patient. Time to tumor progression Time to tumor progression was based on the following CT findings: the appearance of new lesions or an increase of more than 20% in the sum of the longest diameters of the target tumors. Response evaluation using RECIST 1.1 RECIST 1.1 defines tumor response as complete response (CR), partial response (PR), stable disease (SD), or progressive disease (PD) (Table 1). To correlate the response according to RECIST with time to tumor progression, CR and PR were reclassified as good response and SD and PD, as poor response. Response evaluation using modified CT criteria Response by modified CT criteria [6] were defined as CR (disappearance of all lesions), PR ( 10% decrease in target lesion size or % decrease in tumor density), SD (none of the CR, PR, or PD criteria met), or PD ( 10% increase in target lesion size and does not meet tumor density criteria of PR density) (Table 1). To correlate this response with time to tumor progression, CR and PR were reclassified as good response and SD and PD, as poor response. Response evaluation was also performed separately for size ( 10% decrease in target lesion size) and for density ( % decrease in tumor density of target lesion) of modified CT criteria. Fig year-old woman with multiple colorectal liver metastases treated with chemotherapy alone. Regions of interest (ROIs) were manually drawn using freehand polygonal ROI. rrowheads point to average density measurements of ROIs in Hounsfield units. = area of ROI, Max = maximum density measurement of ROIs., Pretreatment CT scan shows low-attenuation lesion (arrow) in segment V of liver., CT scan obtained 2 months after treatment shows that lesion (arrow) has become more hypodense without significant decrease of tumor size. Response by Response Evaluation Criteria in Solid Tumors (RECIST) version 1.1 (percentage change of tumor size: increase of 0.7%) was classified as poor, whereas criteria modified by Choi et al. [6] classified response as good (percentage change of tumor density: decrease of 18.8%). Time to tumor progression was 17.9 months. Statistical nalysis The demographic and clinical data of the two patient groups were compared using a chi-square test for categoric data and by independent samples Student t test for continuous data. Treatment response in each group, as assessed by both RECIST 1.1 and the modified CT criteria, was correlated with time to tumor progression. Time to tumor progression curves were generated using the Kaplan-Meier method, and differences were evaluated with the TLE 2: aseline Patient Characteristics and Characteristics of Tumors and Treatments Characteristic evacizumab and Chemotherapy (n = 30 Patients [107 Lesions]) Chemotherapy lone (n = 29 Patients [111 Lesions]) p ge, y 08 Mean (range) 54.7 (35 77) 64 (47 78) Sex, no. (%) of patients 85 Female 14 (47) 12 (41) Male 16 (53) 17 (59) Primary tumor, no. (%) of patients 64 Colon 18 (60) 19 (66) Rectum 12 (40) 10 (34) Sites of metastasis, no. (%) of patients 88 Liver only 14 (47) 13 (45) Liver and extrahepatic 16 (53) 16 (55) Liver metastasis, no. (%) of patients 86 Solitary 3 (10) 8 (28) Multiple 27 (90) 21 (72) Size of hepatic metastasis a (cm) 70 Mean (range) 3.45 ( 12.4) 3.4 (0.5 1) Chemotherapy, no. of patients FOLFOX N 29 FOLFIRI with bevacizumab 9 N FOLFOX with bevacizumab 21 N Note FOLFOX = infusional 5-fluorouracil, leucovorin, and oxaliplatin; N = not applicable; FOLFIRI = infusional 5-fluorouracil, leucovorin, and irinotecan. a ased on the largest tumor diameter on pretreatment CT. JR:199, October

4 Chung et al. log-rank test. Interobserver agreement between two readers was evaluated using the intraclass correlation coefficient (ICC) for pretreatment CT and for 2-month posttreatment CT. n ICC of greater than 0.75 was considered to represent good agreement. ll statistical analyses were performed with SPSS software (version 17.0, SPSS), and a p value of less than 5 was considered significant. Results Comparison of Group Treated With evacizumab-containing Chemotherapy and Group Treated With Chemotherapy lone One hundred seven metastatic target lesions in 30 patients who received bevacizumabcontaining chemotherapy and 111 metastatic target lesions in 29 patients who received chemotherapy alone were evaluated. There was no significant difference in demographic and clinical characteristics between the bevacizumab-containing chemotherapy group and chemotherapy alone group (Table 2). ccording to RECIST 1.1, there was no significant difference in the ratio of goodversus-poor responses between the bevacizumab-containing chemotherapy group (9/30 patients [30%] vs 21/30 [70%], respectively) and the chemotherapy alone group (12/29 [41%] vs 17/29 [59%]) (p = 0.366). ccording to the modified CT criteria, there was no significant difference in the ratio of good-versus-poor responses between the bevacizumab-containing chemotherapy group (24/30 patients [80%] vs 6/30 [20%]) and the chemotherapy alone group (23/29 patients [79%] vs 6/29 [21%]) (p = 0.948). TLE 3: Comparison of Response Evaluation Using Response Evaluation Criteria in Solid Tumors (RECIST) Version 1.1 Versus Modified CT Criteria Treatment and Outcome RECIST 1.1 Modified CT Criteria p evacizumab and chemotherapy < 01 a, no. (%) of patients 9 (30) 24 (80), no. (%) of patients 21 (70) 6 (20) Chemotherapy alone < 03 a, no. (%) of patients 12 (41) 23 (79), no. (%) of patients 17 (59) 6 (21) a There was a significant difference in responses between RECIST 1.1 and modified CT criteria in both groups. Comparison of Response Evaluation Using RECIST 1.1 Versus the Modified CT Criteria In the bevacizumab-containing chemotherapy group, there was a significantly different level of response between using RECIST 1.1 (30% of patients with good response) and the modified CT criteria (80% of patients with good response) (p < 01) (Table 3 and Fig. 2). In the chemotherapy alone group, there was a significantly different level of response between using RECIST 1.1 (41% of patients with good response) and using the modified CT criteria (79% of patients with good response) (p = 03) (Table 3). Correlation of Response Evaluation With Time to Tumor Progression In the bevacizumab-containing chemotherapy group, there was a significant difference in time to tumor progression between the patients with good response and those with poor response based on RECIST 1.1 (p = 211 [Fig. 3]) and on modified CT criteria (p = 2 [Fig. 3]). In the chemotherapy alone group, there was no significant difference in time to tumor progression between good and poor responses based on RECIST 1.1 (p = 946 [Fig. 4]). However, there was a significant difference in time to tumor progression between good and poor responses based on the modified CT criteria (p = 22 [Fig. 4]). nalysis of tumor size only by the modified CT criteria did not show a significant difference in time to tumor progression between good and poor responders in both groups (bevacizumab-containing chemotherapy group, p = 695 [Fig. 5]; chemotherapy alone group, p = 423 [Fig. 5]). Similarly, analysis of tumor density only by the modified CT criteria did not show a significant difference in time to tumor progression between good and poor responders in both groups (bevacizumab-containing chemotherapy group, p = Fig year-old man with multiple colorectal liver metastases treated with bevacizumab-containing chemotherapy., Pretreatment CT scan shows low-attenuation lesion (arrow) in segment VIII of liver., CT scan obtained 2 months after treatment shows that lesion (arrow) has become more hyperdense with increase in tumor size. Response was classified as poor by both RECIST 1.1 (percentage change of tumor size: increase of 57%) and modified CT criteria (percent change of tumor density: increase of 2.6%). Time to tumor progression was 2.8 months. 812 JR:199, October 2012

5 Response Evaluation in Patients With Colorectal Liver Metastases (censored) (censored) [Fig. 5C]; chemotherapy alone group, p = 697 [Fig. 5D]). Interobserver greement of Tumor Density Measurement Interobserver agreement of tumor density measurement was good for pretreatment CT and 2-month posttreatment CT in both groups, with ICCs ranging from to Discussion In 1979, the World Health Organization first published objective tumor response criteria that were adopted around the world [9]. ecause of variability in the interpretation and application of these criteria, a task force proposed RECIST as a simplified set of the standardized criteria in 2000 [13]. The RECIST Working Group recently revised the RECIST guidelines (version 1.1) to simplify, optimize, and standardize the original criteria [8]. RECIST has been the current standard for assessing the response of solid tumors to anticancer drugs; however, sporadic reports have (censored) (censored) 20 (censored) (censored) shown a weak relationship between long-term prognosis and tumor response as assessed by RECIST [5 7]. RECIST was reported to be insensitive in assessing the response of GISTs to imatinib: The time to tumor progression of good responders did not differ from that of poor responders [6]. Similarly, Chun et al. [7] reported that response evaluation based on RECIST was not correlated with overall survival in patients with colorectal cancer liver metastasis treated with bevacizumab-containing chemotherapy. Consistent with these studies, our results showed that response evaluation based on RECIST 1.1 did not predict time to tumor progression in patients with colorectal liver metastases treated with cytotoxic chemotherapy alone. These results raise questions regarding the appropriateness of tumor shrinkage as a standard measure of treatment response. The introduction of new molecular target agents also encourages reexamination of the current response criteria. Unlike cytotoxic chemotherapy, these antiangiogenic agents often act by inhibiting growth rather than shrinking (censored) (censored) 20 Fig. 3 Time to tumor progression in patients with colorectal liver metastases treated with bevacizumab-containing chemotherapy classified as good responders and those classified as poor responders according to Response Evaluation Criteria in Solid Tumors (RECIST) version 1.1 and modified CT criteria [6]., When tumor response was evaluated at 2-month follow-up CT using RECIST 1.1, significant difference was observed in long-term prognosis between good and poor responders (p = 211) for up to 17 months., When tumor response was evaluated at 2-month follow-up CT using combination of tumor sizes and tumor density, significant difference was observed in long-term prognosis between good and poor responders (p = 2) for up to 17 months. Fig. 4 Time to tumor progression in patients with colorectal liver metastases treated with chemotherapy alone classified as good responders and those classified as poor responders according to Response Evaluation Criteria in Solid Tumors (RECIST) version 1.1 and modified CT criteria [6]., When tumor response was evaluated on 2-month follow-up CT using RECIST 1.1, no significant difference was observed in long-term prognosis between good and poor responders (p = 946) for up to 18 months., When tumor response was evaluated on 2-month follow-up CT using combination of tumor size and tumor density, significant difference was observed in long-term prognosis between good and poor responders (p = 22) for up to 18 months. tumors. To evaluate the antiangiogenic effects, perfusion imaging, such as perfusion CT or dynamic contrast-enhanced MRI, may be more appropriate [14 17]. Perfusion imaging directly measures tumor vascularity, which may be a good quantitative biomarker for evaluating antiangiogenic treatment [16, 17]. However, perfusion imaging requires specialized equipment and is a relatively long procedure. Therefore, this method does not easily lend itself to routine evaluation for all cancer treatments. Chun et al. [7] proposed the use of CT-based qualitative morphologic criteria characterized by overall attenuation, tumor-liver interface, and peripheral rim of enhancement; they suggested that changing from heterogeneous masses with ill-defined margins into homogeneous hypoattenuating lesions with sharp borders was an indication of a good response. Their morphologic criteria for evaluating colorectal liver metastasis treated with bevacizumab-containing chemotherapy was significantly associated with pathologic response and with overall survival [7]. However, their morphologic criteria JR:199, October

6 Chung et al. (censored) (censored) (censored) (censored) are difficult to use worldwide because they are not objective or quantitative. Modified RECIST was proposed for hepatocellular carcinoma (HCC) and assesses tumor response using the measurement of the longest diameter of a viable tumor [18]. However, the modified RECIST can be applied only in cases of typical HCCs, such as a measurable lesion with intratumoral arterial enhancement on contrast-enhanced CT or MRI [18]. Unlike HCC, colorectal liver metastasis shows variable enhancement patterns, such as peripheral enhancements, that make it difficult to measure the longest viable tumor diameter. Thus, we thought that tumor density would be more suitable than longest viable tumor diameter for assessing tumor response in patients with colorectal cancer liver metastases. The CT attenuation value of colorectal liver metastases shows a linear correlation with the concentration of contrast agents, which directly reflects tumor vascularity, although other factors may influence attenuation as well. Therefore, changes in tumor density C (censored) (censored) 20 (censored) (censored) 20 in the same phase and protocol of contrastenhanced CT indicate changes in tumor vascularity. The modified CT criteria proposed by Choi et al. [6] include changes in both the size and density of the tumor. Our results show that these modified CT criteria could differentiate good responders from poor responders regardless of treatment. However, when tumor size only or tumor density only was evaluated on CT, time to tumor progression did not differ between good responders and poor responders in both groups, which suggests that the combination of tumor vascularity and tumor size is important for predicting the effectiveness of both antiangiogenic agents and cytotoxic agents. The modified CT criteria can be quantified objectively, can be measured readily in daily practice, and provide an additional measure of response to therapy. esides the weaker correlation with longterm response in patients treated with chemotherapy alone, RECIST 1.1 significantly underestimated treatment response compared D Fig. 5 Time to tumor progression in good and poor responders as determined by CT assessment of tumor size only or tumor density only in bevacizumabcontaining chemotherapy group ( and C) and chemotherapy alone group ( and D). and, When response was evaluated at 2-month follow-up CT using tumor size ( 10% decrease in tumor size), long-term prognosis associated with good response did not differ from that of poor response in both groups (, bevacizumab-containing chemotherapy group, p = 695;, chemotherapy alone group, p = 423). C and D, When tumor response was evaluated at 2-month follow-up CT using tumor density ( % decrease in tumor density), long-term prognosis associated with good response did not differ from that of poor response in both groups (C, bevacizumabcontaining chemotherapy group, p = ; D, chemotherapy alone group, p = 697). with the modified CT criteria regardless of treatment. In patients treated with bevacizumab-containing chemotherapy, good response was found in 30% when assessed using RECIST 1.1, but good response was found in 80% when evaluated using the modified CT criteria. In patients treated with chemotherapy alone, good response was found in 41% according to RECIST 1.1 and in 79% according to the modified CT criteria. These results are consistent with those reported by Choi et al. [6], suggesting that RECIST 1.1 for treatment response are too strict and should be combined with tumor vascularity changes. Our study had some unsolved issues. First, we hypothesized that the modified CT criteria (tumor density in addition to tumor size) could work more effectively than RECIST 1.1 (tumor size only) for the evaluation of antiangiogenic agents. However, our results showed that the modified CT criteria worked similar to RECIST 1.1 in patients who had received bevacizumab-containing chemotherapy. One possible explanation for our results could be that the patients underwent combined therapy, antiangiogenic plus cytotoxic agents, not monotherapy of an antiangiogenic agent, which results in a synergic effect and in further reduction in size as well as vascularity. The standard therapeutic regimen of antiangiogenic agents is combined therapy, not monotherapy, so our results could be useful in clinical practice. nother possible reason could be because of the pharmacokinetics of antiangiogenic agents. The rationale for combined therapy of antiangiogenic plus cytotoxic agents is, according to Jain et al. [19], normalizing tumor vasculature before its destruction to improve the delivery of drugs and oxygen [19, 20]. Therefore, changes in density or a reduction of vascularity could not be as prominent 814 JR:199, October 2012

7 Response Evaluation in Patients With Colorectal Liver Metastases as we had expected. perfusion CT study in a rat model after antiangiogenic therapy reported that blood flow, blood volume, and the permeability surface area product were significantly higher in tumor treated with an antiangiogenic agent than in control tumors [14]. These results should be clarified by further research with serial change of tumor vascularity after antiangiogenic treatment. Second, our results showed that there were different levels of response between RECIST 1.1 and the modified CT criteria without a definite standard reference. RECIST 1.1 showed an underestimating response relative to the modified CT criteria, and the modified CT criteria showed an overestimating response relative to RECIST 1.1. However, based on the results that response evaluation by modified CT criteria better predicted time to tumor progression than that by RECIST 1.1, we suggest that RECIST 1.1 might show underestimation. Third, we included only the liver metastasis for uniformity of lesions and excluded metastases to any other organs or structures, such as the peritoneum. Further study to evaluate the disease as a whole without limiting to certain organs is needed. In conclusion, response evaluation using changes in tumor size and density on CT better predicted time to tumor progression than changes in tumor size only in patients with colorectal liver metastases treated with bevacizumab-containing chemotherapy or chemotherapy alone. On the basis of our results, we suggest adding tumor density changes to new criteria for evaluating treatment response regardless of the anticancer treatment agents or tumor type. These new criteria should be tested further in large-scale studies of other tumors treated with various therapies. References 1. Rubbia-randt L, Giostra E, rezault C, et al. Importance of histological tumor response assessment in predicting the outcome in patients with colorectal liver metastases treated with neo-adjuvant chemotherapy followed by liver surgery. nn Oncol 2007; 18: dam R, Pascal G, Castaing D, et al. Tumor progression while on chemotherapy: a contraindication to liver resection for multiple colorectal metastases? nn Surg 2004; 240: ; discussion, dam R, Wicherts D, de Haas RJ, et al. Complete pathologic response after preoperative chemotherapy for colorectal liver metastases: myth or reality? J Clin Oncol 2008; 26: lazer DG 3rd, Kishi Y, Maru DM, et al. Pathologic response to preoperative chemotherapy: a new outcome end point after resection of hepatic colorectal metastases. J Clin Oncol 2008; 26: Grothey, Hedrick EE, Mass RD, et al. Responseindependent survival benefit in metastatic colorectal cancer: a comparative analysis of N9741 and VF2107. J Clin Oncol 2008; 26: Choi H, Charnsangavej C, Faria SC, et al. Correlation of computed tomography and positron emission tomography in patients with metastatic gastrointestinal stromal tumor treated at a single institution with imatinib mesylate: proposal of new computed tomography response criteria. J Clin Oncol 2007; 25: Chun YS, Vauthey JN, oonsirikamchai P, et al. ssociation of computed tomography morphologic criteria with pathologic response and survival in patients treated with bevacizumab for colorectal liver metastases. JM 2009; 302: Eisenhauer E, Therasse P, ogaerts J, et al. New Response Evaluation Criteria in Solid Tumours: revised RECIST guideline (version 1.1). Eur J Cancer 2009; 45: van Persijn van Meerten EL, Gelderblom H, loem JL. RECIST revised: implications for the radiologist a review article on the modified RECIST guideline. Eur Radiol 2010; 20: Klinger M, Tamandl D, Eipeldauer S, et al. evacizumab improves pathological response of colorectal cancer liver metastases treated with XE- LOX/FOLFOX. nn Surg Oncol 2010; 17: Ribero D, Wang H, Donadon M, et al. evacizumab improves pathologic response and protects against hepatic injury in patients treated with oxaliplatin-based chemotherapy for colorectal liver metastases. Cancer 2007; 110: Javle M, Hsueh CT. Recent advances in gastrointestinal oncology: updates and insights from the 2009 annual meeting of the merican Society of Clinical Oncology. J Hematol Oncol 2010; 3: Therasse P, rbuck SG, Eisenhauer E, et al. New guidelines to evaluate the Response to Treatment in Solid Tumors: European Organization for Research and Treatment of Cancer, National Cancer Institute of the United States, National Cancer Institute of Canada. J Natl Cancer Inst 2000; 92: Kan Z, Phongkitkarun S, Kobayashi S, et al. Functional CT for quantifying tumor perfusion in antiangiogenic therapy in a rat model. Radiology 2005; 237:1 8. Park MS, Klotz E, Kim MJ, et al. Perfusion CT: noninvasive surrogate marker for stratification of pancreatic cancer response to concurrent chemo- and radiation therapy. Radiology 2009; 250: Provenzale JM. Imaging of angiogenesis: clinical techniques and novel imaging methods. JR 2007; 188: Sabir, Schor-ardach R, Wilcox CJ, et al. Perfusion MDCT enables early detection of therapeutic response to antiangiogenic therapy. JR 2008; 191: Lencioni R, Llovet JM. Modified RECIST (mrecist) assessment for hepatocellular carcinoma. Semin Liver Dis 2010; 30: Jain RK. Normalizing tumor vasculature with anti-angiogenic therapy: a new paradigm for combination therapy. Nat Med 2001; 7: Tong RT, oucher Y, Kozin SV, et al. Vascular normalization by vascular endothelial growth factor receptor 2 blockade induces a pressure gradient across the vasculature and improves drug penetration in tumors. Cancer Res 2004; 64: JR:199, October