Objective(s) of the proposed study:

• The evaluation of the efficiency of 18F deoxyglucose-Positron Emission Tomography (FDG-PET) in staging patients eligible for hepatic resection of colorectal liver metastases in a randomized clinical multicentre setting.

Research questions of the proposed study:

• What are the effects and costs for patients with liver metastases of colorectal cancer indicated for potentially curative hepatic resection, using the conventional diagnostic strategy with computed tomography (CT) scan in comparison to the experimental diagnostic strategy incorporating FDG-PET scan (CT + FDG-PET scan), based on a health care perspective and a time horizon of 9 months.

More specifically:

• Does the experimental diagnostic strategy which includes FDG-PET scan in the diagnostic work-up of patients eligible for potentially curative hepatic resection of colorectal liver metastases lead to a better disease-free survival at 9 months after hepatic resection in comparison to the conventional diagnostic strategy using CT scan without FDG-PET scan.
• What are the costs of diagnostic and therapeutic care for the two diagnostic strategies for patients eligible for potentially curative hepatic resection of colorectal liver metastases.
• What is the effect of including the FDG-PET scan in the diagnostic work-up of patients eligible for potentially curative hepatic resection of colorectal liver metastases after hepatic resection, expressed as disease-free survival at 9 months adjusted for quality of health (Q-TWIST), in comparison to the use of CT scan only.

Study design:

Prospective randomized study. In both arms conventional diagnostic tests will be performed. In arm A (conventional strategy) no FDG-PET scan will be performed. In arm B (experimental strategy) an FDG-PET scan will be made and the results will be incorporated in decisions for further clinical management by the referring surgeon.

Motivation study design:

Randomization in one strategy with and one strategy without FDG-PET scan will allow to truly estimate the impact of including FDG-PET scan in the diagnostic work-up. More specifically this study design allows to determine:

1. The impact of FDG-PET on disease-free survival after hepatic resection.
2. The amount and costs of diagnostic and (non)-therapeutic procedures of both strategies.
3. The average number of months of disease-free survival adjusted for quality of health (Q-TWIST) for the two strategies.

Scientific basis of study proposal:

Although the added value of FDG-PET in the diagnostic work-up is well established, it is unclear from current studies, whether more sensitive and meticulous preoperative staging has an impact on patient outcome. Moreover, prospective comparative economic evaluations on strategies of diagnostic work-up with and without FDG-PET are lacking.

Patients are considered eligible for the trial when they fulfill all of the following criteria:

1. One to four colorectal liver metastases on spiral CT, judged potentially resectable by an experienced liver surgeon from the institution participating in the trial.
2. No evidence of extrahepatic disease as demonstrated by spiral CT scan of chest and abdomen with oral and intravenous contrast with contiguous reconstruction algorithm. In case of previous rectal cancer spiral CT should include the pelvic area. CT readings should be performed by an experienced radiologist from the institution participating in the trial.
3. No signs of recurrent or second colorectal carcinoma on barium enema or colonoscopy.
4. Absence of any other previous malignancy other than adequately treated in situ carcinoma of the cervix or non-melanoma skin cancer (unless there has been a disease-free interval of at least 10 years).
5. Absence of major hepatic insufficiency.
6. Absence of active infection and diabetes mellitus
7. WHO performance status 0, 1 or 2.
8. Age 18-75 years
9. Written informed consent

Description of the intervention:

• Randomization: patients will be centrally randomized with random diagnostic strategy allocation stratifying by institution.
• FDG-PET scan; Image acquisition: Imaging should be performed with dedicated PET-scanners only. Patients will be fasted for at least 6 hours before PET-scanning. Immediately prior to the procedure, patients will be hydrated with 500 ml of water. Blood glucose levels will be measured. Thereafter, FDG (370 MBq when using full-ring systems, 220 MBq when using a half-ring system) and 20 mg of furosemide will be injected intravenously. Image acquisition will start 60-75 minutes after FDG injection (emission and transmission images of the area between proximal femora and the base of the skull). The images will be corrected for attenuation and reconstructed using the ordered-subsets expectation maximization (OSEM) algorithm.
• Image interpretation:

The FDG-PET studies will be read on site. The final result of FDG-PET alone will be communicated to the referring surgeon on a confidence scale as (a) normal or benign disease, (b) probably no malignancy, (c) malignant/benign unclear (d) definitely malignant disease. Further evaluation will take place by joint reading of FDG-PET and CT of the chest and abdomen, the combined reading will again be scored on the same confidence scale to allow estimation of increased confidence in the co-interpreted procedures. The combination of these tests will form the basis for the surgeons' decision making. In case of concordant evidence of extra-hepatic disease the patient is considered to be non-resectable. In case of discordance between CT and FDG-PET, it is at the referring surgeon's discretion to request additional diagnostic tests, to reject the option of laparotomy or to attempt surgical resection of the liver metastases.

• Laparotomy is performed within 4 weeks after randomization. At laparotomy careful examination of the abdominal cavity is performed to exclude extra-hepatic disease. In case of any doubt of extra-hepatic disease samples are taken for frozen sections. When positive, the operation is terminated. Intra-operative ultrasound should be performed to detect and localize all metastatic lesions. Lesions within 1 cm of each other are considered as satellite lesions and are counted as one lesion. In case intra-operative examination shows more than 4 lesions and complete resection of all lesions is still possible, resection is allowed within the study proposal. Complete resection of all metastatic lesions should always be achieved, if possible with a safety margin of > 1cm of normal parenchyma. The type of liver resection (anatomical, wedge or combination) is at the surgeon's discretion. In case complete resection of all lesions seems not possible the operation is terminated. The use of local ablative techniques like radiofrequency or cryosurgery is not allowed within the study. The post-operative course and complications are recorded as well as duration of hospital stay.
• Follow-up should concentrate on the detection of recurrent disease and the registration of sequela and financial consequences of the medical interventions performed in the protocol. At 3, 6 and 9 months after surgery, the following follow-up items will be performed: history, physical examination, plasma CEA-level and CT of the abdomen and chest. All CT readings during follow-up should be performed by an experienced radiologist in the field.

The diagnosis of tumor recurrence should be made only by one of the criteria defined below:

1. Objective radiological recurrence on radiological CT imaging. That is the appearance of one or more new lesions on radiological imaging or at least 20% increase of a suspected lesion on earlier investigations.
2. Positive cytology or histology (in case of ambiguous radiological imaging). The documented date of recurrence will be the date of confirmation of the recurrence using one of the methods of diagnosis. An elevated CEA level, as a solitary finding, will not be considered as acceptable evidence of colon cancer recurrence. In case of elevation of serum CEA on two successive time points further examinations must be performed searching for disease recurrence.

Primary outcome parameters

• Disease-free survival 9 months after hepatic resection. Disease free survival defined as the absence of recurrent disease on conventional diagnostic imaging (CT liver/abdomen, CT chest). Comment: The study proposal assumes that patients unnecessarily operated in the conventional arm (patients who would have had a positive PET for extrahepatic disease and negative CT for extrahepatic disease) will show extrahepatic disease on conventional CT imaging within 9 months of randomization. This assumption is confirmed in a preliminary study conducted at our institute (Langenhoff 2001a).

Secondary outcome parameters

• The total costs of diagnostic procedures and potential therapeutic medical interventions (hepatic resection, laparotomy) and their sequences in strategy A and B, in a time horizon of 9 months after randomization.
• The cost-effectiveness of strategy A and B expressed in costs per month disease-free survival for patients undergoing hepatic resection.· The number of operations not resulting in hepatic resection in strategy A and B.
• The diagnostic yield of FDG-PET versus CT imaging for evaluating hepatic resectability and detecting extra-hepatic disease.
• Quality of life of the patient groups analyzed by strategy A and B.

Primary analysis:

Based on publications [Fong 1997, Scheele 1991] the following percentages are the expected disease-free survival after 9 months follow-up: for the group of patients who underwent the conventional CT strategy the expected disease-free survival rate is 70% and for the group of patients in the experimental strategy (CT + FDG-PET) the expected disease-free survival rate is 95%. To detect this difference between the two groups of 25% with a power of 80% and an alpha of 0.05 (1-sided), 25 patients are required for each group (Cohen 1977). To anticipate for patients not eligible for laparotomy or hepatic resection 43 patients are needed in each arm to perform the above mentioned analysis. Dropouts of patients is not anticipated. Patient accrual for this part of the analyses will be sufficient after 2 years of patient inclusion.

Additional analyses:

Furthermore, an intention-to treat analysis will be performed and in both arms both the proportion of resected patients and the 3 year survival will be determined (the latter by continued follow-up of all recruited - resected and nonresected - patients beyond the formal conclusion of the present proposal). To detect a difference in resection percentage between both arms of 15% with a power of 87% (60% in CT+PET group vs 75% in CT-only group, a=0.2, one-sided) 75 patients are needed in each study arm. The number of 150 recruited patients (2x75), is a realistic estimate of the number of patients that can be included in a 3 year period. This means that at the time point of study closure we will be able to detect [A] a difference in disease free survival of 25% (power 80%, a=0.05) between patients resected in both study arms and [B] a difference in resection percentage between both arms of 15 % (power 87%, a=0.2).

After the formal closure of the study after 3 years the intention to treat analysis will be continued. With the number of included patients (2x75) we will be able to detect a difference in 3 years survival of 20% or less (power 80%, a=0.05, one-sided) between resected patients of group A and B. Definite conclusions on equal overall survival in arm A and B of the study will obviously be limited.

The following data will be collected:

• Demographic characteristics.
• Clinical outcomes: primary tumor (localization, resection date, histology, TNM), CEA level, number, size and localization metastases, resectability, histology.
• Diagnostic outcomes: CT (baseline and 3, 6 and 9 months after surgery): number, size and localization of metastases, extra-hepatic disease, recurrence; FDG-PET (baseline): number, localization and size of metastases, extra-hepatic disease.
• Follow-up: complaints, morbidity, mortality, presence of recurrent disease and site (CT), CEA, QoL.
• Costs: see elsewhere. To measure the separate effect of the two diagnostic strategies and their related treatment regimes (primary outcome) the proportion of respondents who are still disease-free after hepatic resection (CT diagnosis) 9 month after treatment will be estimated by applying life-table analysis (Kaplan-Meier). These estimated proportions will be used in the cost-effectiveness analysis (see section: Economic evaluation). A similar analysis, however, extended with quality of health outcomes (utilities: measured by the EQ-5D) will be performed for the 3 distinct clinical states that can be distinguished: 1) no laparotomy, no cure, 2) laparotomy + hepatic resection, possible cure, 3) laparotomy + no hepatic resection, no cure. In order to perform this so-called Q-TWIST analysis the EQ-5D will be measured at baseline and each month during follow-up. The results of this aggregated analysis will be expressed as the number of quality-adjusted months survival for the two treatment strategies. Another contribution of this approach is that all the information can be depicted in one single figure for each of the two strategies, clearly dividing the impact of the three distinct states and their occurrence in time.Quality-adjusted life years (QALYs) will be computed in order to perform a conventional cost-utility analysis for the two diagnostic strategies. Utilities (EQ-5D) will be based on the mean values for the two groups of patients at the 10 occasions (baseline, month 1-9). Contrary to the above-mentioned Q-TWIST analysis, no distinct states will be recognized, whereas cost will be incorporated.

Economic evaluation:

General considerations:

The basic assumption for this study is that the experimental diagnostic strategy based on CT + FDG-PET compared to the conventional diagnostic strategy based on CT leads to a more accurate diagnosis about operability of liver metastases. Therefore, the economic evaluation is based on the principles of a cost-effectiveness analysis. Using a differential approach, patients are followed up for a period of 9 months. In addition, patients suffering from recurrence of the disease will be followed during this period in order to be able to evaluate the total costs involved in the care of these patients. Because of the specific patient population in this study, non-medical costs are expected not to be different between the patient groups. Therefore, a health care perspective is the basis for the analyses, indicating that only health care costs are subject to study. A problem with basing cost estimates on data gathered as part of a clinical trial is the extent to which one is capturing resource use associated with the trial per se (i.e. costs of doing scientific research) rather than the costs of providing the diagnostic and therapeutic care [Drummond 1997]. These so called protocol driven costs will be excluded from the cost analyses. Regarding the cost-effectiveness analysis, for each patient the total costs will be split into diagnostic and therapeutic costs. Focusing on the effectiveness measure 'proportion of accurate diagnoses', only the diagnostic costs will be used to calculate the incremental ratio comparing the FDG-PET strategy versus the conventional strategy. This results in an estimate of the additional diagnostic costs per extra accurate diagnosis. Concerning the effectiveness measure 'proportion of patients with disease-free survival', the total costs until recurrence of the disease or to the end of follow up will be related to the mean time of disease free survival. This results in an estimate of the additional medical costs per month of disease free survival. In addition, a cost-utility analysis (CUA) will be carried out to incorporate the preferences of the participants considering their own health status at the start of the treatment and after follow-up. Based on these preferences combined with life expectancy figures, calculation of so-called quality-adjusted life years (QALYs) is applicable. A general measure to express the benefits for the two different treatment strategies will be calculated by computing the area under the curve of the utility x measurement moment (from baseline through the assessment at month 9) curve. These QALYs for the 9-month period can be combined with the costs of each strategy to arrive at the cost per QALY gained.

Economic evaluation/cost analysis:

The cost analysis exists of two main parts. First, on the patient level, volumes of diagnostic and therapeutic care will be measured prospectively using standardized Case Report Forms (CRF). In each participating hospital research nurses will register in the CRF the number of times a patient visits the out patient department, the number of days in hospital (normal care and intensive care distinguished), the number and kind of diagnostic procedures (CT, FDG-PET, ultrasonography, histological examination), the type and number of operations (intra-operative ultrasonography, endoscopy, resection of liver metastases), the duration of the specific operations, the number and duration of therapeutic sessions, the amount and kind of pharmaceutical therapies (e.g. pain killers), days in hospital (normal care and intensive care distinguished) and so on. This part of the CRF will be designed and pilot tested in the first months of the study. The second part of the cost analysis consists of determining the cost prices for each volume parameter to use these for multiplying the volumes registered for each participating patient. The Dutch guidelines for conducting pharmaco-economic studies (CVZ, 1999) and the guidelines for cost analyses will be used [Oostenbrink 2000]. Because of the health care perspective of the study approximations of real integral costs will be used in the cost analysis. As a basis for the cost price calculations for each participating hospital a unique cost price, based on standardized calculation methods will be determined. For the baseline analysis a weighed average (using the number of patients included in the study by each participating hospital) will be used. Mean, median, range and standard deviation of the total medical costs per patient will be determined for each of the two patient groups. Skewness of the distribution will be examined to determine whether parametric or non-parametric statistical techniques will be used to test a possible statistical difference between the groups. Analysis will be performed on the basis of intention to treat. Besides this statistical analysis, the impact of deterministic variables, such as the cost prices used for the different volume parameters, will be investigated using sensitivity analyses on the basis of the range of extremes [Briggs 1994]. For definition of the range of extremes to be used in this sensitivity analysis, the range of the cost prices as calculated in the participating hospitals and the national guidelines will be used.

Economic evaluation/patient outcome analysis:

Patient outcome analysis will be primarily based on the 'proportion of disease-free patients after 9 months follow-up' (cost-effectiveness analysis). As a secondary analysis a cost-utility analysis will be carried out. For this purpose the EQ-5D instrument will be used to estimate utilities for each patient at the different measurement moments. In addition, a generic (SF-36) and a disease-specific (EORTC QLQ-C30) descriptive health-related quality of life questionnaire will be applied.

Inclusion Criteria:

• One to four colorectal liver metastases on spiral CT, judged potentially resectable by an experienced liver surgeon from the institution participating in the trial.
• WHO performance status 0, 1 or 2.
• Age 18-75 years
• Written informed consent

Exclusion Criteria

• Evidence of extrahepatic disease as demonstrated by spiral CT scan of chest and abdomen with oral and intravenous contrast with contiguous reconstruction algorithm. In case of previous rectal cancer spiral CT should include the pelvic area. CT readings should be performed by an experienced radiologist from the institution participating in the trial.
• Signs of recurrent or second colorectal carcinoma on barium enema or colonoscopy.
• Any other previous malignancy other than adequately treated in situ carcinoma of the cervix or non-melanoma skin cancer (unless there has been a disease-free interval of at least 10 years).
• Major hepatic insufficiency.
• Active infection and diabetes mellitus