NCBI Bookshelf. A service of the National Library of Medicine, National Institutes of Health.

Brush J, Boyd K, Chappell F, et al. The Value of FDG Positron Emission Tomography/Computerised Tomography (PET/CT) in Pre-Operative Staging of Colorectal Cancer: A Systematic Review and Economic Evaluation. Southampton (UK): NIHR Journals Library; 2011 Sep. (Health Technology Assessment, No. 15.35.)

Cover of The Value of FDG Positron Emission Tomography/Computerised Tomography (PET/CT) in Pre-Operative Staging of Colorectal Cancer: A Systematic Review and Economic Evaluation

The Value of FDG Positron Emission Tomography/Computerised Tomography (PET/CT) in Pre-Operative Staging of Colorectal Cancer: A Systematic Review and Economic Evaluation.

Show details

7FDG PET/CT for the pre-operative staging of metastatic colorectal cancer

Approximately 37,000 CRC cases are diagnosed each year and 17,000 deaths are attributed to the disease,2 mainly due to metastatic disease. At presentation, approximately 20–25% of patients have clinically detectable metastases and a further 40–50% of patients subsequently develop metastases after resection of the primary tumour, most commonly within the first 3 years of follow-up.

The liver is often the first site of metastatic disease. It has been postulated that the principal mode of tumour dissemination to the liver is via the portal system and therefore that surgical resection of isolated hepatic metastases from CRC may be curative. However, the natural history of metastatic CRC is variable. Median survival without treatment is < 8 months from presentation, but the prognosis is better for those patients with isolated hepatic metastases. Patients with a limited number of metastases or those with disease confined to one lobe of the liver have a longer duration of survival than those with more advanced disease.

Thus, some 20–30% of patients with metastatic CRC have disease that is confined to the liver and is potentially resectable. For the UK, this equates to approximately 3500 patients per year.72 Following the seminal series reported by Scheele and colleagues in 1995,73 several large series on resection for colorectal liver metastases have reported 5-year survival rates ranging from 25% to 44%, with operative mortality of approximately 2.5%.74

The key issue after hepatic resection for metastatic disease is the high relapse rate and/or metastatic disease elsewhere. There are a number of predictors of relapse including the number of intrahepatic metastases and extrahepatic disease.75 To address this, pre-operative chemotherapy is increasingly used and shown to reduce the risk of progression in resected patients.76 Furthermore, modern combinational chemotherapy regimens may downstage unresectable liver disease to resectable in approximately 15% of cases.77

The lung is the second most common site for metastases. In modern oncological practice it is not uncommon to consider patients with isolated lung metastases for metastatectomy, including patients undergoing resection for hepatic metastases, with moderate success rates.78 However, as the overall number of cases of patients undergoing pulmonary metastatectomy is small, the lung is not the primary site of interest in this section of the report.

Given the high morbidity and potential mortality associated with hepatic resection, and the need to predict cases most likely to benefit from such major surgery, the rationale for accurate pre-operative staging becomes clear.25

Systematic reviews to evaluate FDG PET against standard imaging techniques used in staging metastatic CRCs have been conducted and FDG PET was shown to compare favourably with helical CT, non-helical CT and MRI in a meta-analysis of data from 1058 patients, with a sensitivity of 94.6% (95% CI 92.5% to 96.1%). The authors also found PET to be most accurate in a meta-analysis of overall diagnostic performance based on lesion-level data, with a sensitivity of 75% (95% CI 61.1% to 86.3%).40

In a second systematic review that compared FDG PET with CT scanning for the detection of hepatic and extrahepatic liver lesions, PET was shown to be superior to CT in 1843 patients, with a sensitivity of 88.0% (95% CI 88.0% to 98.0%) and a specificity of 96.1% (95% CI 70.4% to 104.3%).44 Integrated FDG PET/CT equipment has not been evaluated in a systematic review, and the benefits or harms associated with this new technology are unclear.

This systematic review considers diagnostic test accuracy of integrated or hybrid FDG PET/CT for the staging of metastatic CRC.

Aim

The aim of this systematic review is to compare the performance of combined functional and anatomical imaging with integrated FDG PET/CT scanning with standard imaging modalities in the investigation of patients with suspected metastatic disease.

Objectives

To evaluate the diagnostic accuracy of FDG PET/CT for the detection of extrahepatic and intrahepatic lesions.

Results

Our search did not identify any systematic reviews to evaluate the diagnostic test accuracy of integrated FDG PET/CT in the pre-operative staging of metastatic CRC.

Number of included studies

Sixteen studies24,62,63,6871,7987 evaluating the diagnostic test accuracy of combined functional and anatomical imaging with FDG PET/CT for the detection of colorectal metastases were identified. The majority of the studies identified investigated metastatic disease to the liver. However, some studies also evaluated extrahepatic disease including local recurrence.62,63,68

Study characteristics and study designs

The study characteristics are shown in Table 7 and accuracy data are shown in Table 8.

TABLE 7. Study characteristics.

TABLE 7

Study characteristics.

TABLE 8. Accuracy data.

TABLE 8

Accuracy data.

The majority of studies used a retrospective design,24,62,63,68,71,79,80,82,85,86 with only five studies using a prospective design70,81,83,84,87 and one being unclear whether the study was retrospective or prospective.69 The majority of reports did not reveal the manner in which sample patients were recruited; only six reported taking a consecutive approach24,62,83,84,86,87 and one82 reported that the sample was not recruited consecutively.

Study setting and country in which the research was conducted

Most studies were conducted in institutes of nuclear medicine, university hospitals and specialist cancer centres in Europe62,63,7981,8385 and the USA.24,68,82,86,87 One was conducted in a university hospital in India,70 one in a university hospital and China71 and one in a teaching and research medical centre in Israel.69

Patient populations

The 16 included studies reported outcomes for 890 patients. When gender proportions were reported, men were usually in the majority. The mean ages of patients when reported were 58–65 years, with a range of 31–92 years. Two studies70,79 included people with different types of cancer in whom FDG PET/CT was used to detect metastases. The age and gender of the patients with CRC were not reported separately by the authors.

Indication for FDG PET/CT

In 13 studies,62,63,6870,7985,87 patients were indicated for an FDG PET/CT scan when hepatic metastases were suspected from other follow-up examinations, imaging tests and blood tests. In three studies,62,63,68 patients were investigated for a recurrence of CRC, but metastatic disease was detected.

FDG PET/CT equipment and patient preparation

Full details of all of the equipment used in the primary studies can be found in Table 7. FDG PET/CT equipment was manufactured by one of four companies: Philips Medical Systems, GE Healthcare CPS Innovations or Siemens Medical Solutions. The fasting duration prior to the scan was at least 4 hours in all studies that reported these data.

A wide range of injected FDG doses was reported, with units ranging from 296 to 740 MBq, and patients were scanned between 60 and 120 minutes after the administration of the radioactive tracer. In some studies patients received either oral68,69,84 or intravenous63,8385 contrast agents with the CT component of the FDG PET/CT test.

Image interpretation

Differences in interpretation were identified between studies, with most studies involving a radiologist working with an expert in nuclear medicine. Some studies24,68,82,83 also used a qualitative judgement scoring grade [using scores of 0 to equate to no lesion, 1 = definitely benign, 2 = probably benign, 3 = possibly benign, 4 = possibly malignant, 5 = probably malignant, 6 = definitely malignant (1–3 benign, 4–6 malignant)], while others used focal increased FDG uptake compared with background lesions.63,69,79,84,85,87

Only one study71 reported using SUVmax in the assessment of FDG PET/CT images. Lesions with a SUVmax of > 2.5 in the early imaging and a change in the SUVmax of > 20% in the delayed imaging were considered positive.

Anatomical delineation data from the CT component of the FDG PET/CT scan were used in the overall assessment in four studies,62,63,84,85 and lung nodules and sclerotic bone lesions on CT images even in the absence of increased FDG uptake were considered malignant, but otherwise abnormalities on CT without corresponding increased uptake were considered benign, in one report.68 No information regarding the methods of assessment was given in five studies.70,71,80,81,86

Reference standard

The studies reported the use of various methods as the reference standard, including surgically resected specimens, biopsy, other imaging modalities and clinical follow-up. None of the studies attributed a specific reference standard test to individual patients' index tests. Most reported a period of follow-up, and this ranged from 1 month to 30 months.62,63,6870,7985

Data synthesis – diagnostic performance

FDG PET/CT versus contrast-enhanced CT

The accuracy of FDG PET/CT versus contrast-enhanced CT was compared in six studies69,70,7981,87 involving 362 patients. Only two reported exclusion criteria: patients who had received chemotherapy < 3 months before FDG PET/CT was performed,80 and intact primary tumour or previous treatment of liver metastases by radiotherapy or surgery.81

Estimates of diagnostic accuracy (sensitivity and specificity) for the detection of liver metastases were published with both patients and lesions as the unit of analysis.

Patient-level accuracy estimates – hepatic metastases

Estimates of sensitivity for FDG PET/CT ranged from 87% to 100%, and estimates of specificity ranged from 75% to 100% in four studies70,7981 that reported data with patients as the unit of analysis. The sensitivity of contrast-enhanced CT ranged from 75% to 98% and specificity from 25% to 100%. In two of the four studies, FDG PET/CT appeared to demonstrate greater diagnostic accuracy than contrast-enhanced CT;70,79 in one, FDG PET/CT and contrast-enhanced CT were found to be equally accurate in the detection of liver metastases;80 and a fourth found that FDG PET/CT detected two lesions that proved to be FPs, and that contrast-enhanced CT detected two lesions that were confirmed to be FNs.81

Lesion-level accuracy estimates – hepatic metastases

Estimates of sensitivity of FDG PET/CT in the detection of hepatic metastases with lesions as the unit of analysis ranged from 49% to 98%, and estimates of specificity ranged from 76% to 100%. For contrast-enhanced CT, sensitivity ranged from 55% to 100% and specificity from 65% to 99%.

Patient-level accuracy estimates – extrahepatic metastases

One study87 presented accuracy estimates for FDG PET/CT and contrast-enhanced CT in the diagnosis of colorectal metastases that were not confined to the liver. FDG PET/CT had a sensitivity of 46% (95% CI 34% to 57%) and a specificity of 67% (95% CI 30% to 90%), and contrast-enhanced CT demonstrated a sensitivity of 37% (95% CI 26% to 50%) and a specificity of 92% (95% CI 68% to 98%) with the patient as the unit of analysis.

FDG PET/CT versus contrast-enhanced FDG PET/CT and MRI

One study82 compared FDG PET/CT with contrast-enhanced FDG PET/CT and MRI in the detection of colorectal metastases and presented lesion-level data showing the sensitivity of FDG PET/CT to be 67% (95% CI 57% to 75%) and the specificity to be 60% (95% CI 31% to 83%). Contrast-enhanced FDG PET/CT had a sensitivity of 85% (95% CI 76% to 90%) and a specificity of 100% (95% CI 72% to 100%), but MRI demonstrated greatest accuracy with a sensitivity of 98% (95% CI 93% to 99%) and a specificity of 100% (95% CI 72% to 100%).

FDG PET/CT versus MRI

One study84 compared FDG PET/CT with MRI including unenhanced single-shot spin-echo echo planar imaging (SS SE-EPI) and superparamagnetic oxide (SPIO) enhancement. FDG PET/CT had a sensitivity of 96% (95% CI 80% to 99%), MRI (SS SE-EPI) 100% (95% CI 86% to 100%) and MRI (SPIO) 100% (95% CI 86% to 100%) based on estimates with patients as the units of analysis. Sensitivities calculated from hepatic metastatic lesion-level data were 61%, 100% and 89% (95% CI not calculable) for FDG PET/CT, MRI (SS SE-EPI) and MRI (SPIO), respectively.

FDG PET/CT versus PET versus CT versus MRI

One study83 compared images taken with integrated FDG PET/CT equipment with the PET image alone, the CT image alone and MRI in 35 patients. Patient-level estimates of accuracy from the FDG PET/CT integrated equipment were reported to be superior to all other comparisons: FDG PET/CT sensitivity 93% (95% CI 77% to 98%) and specificity 100% (95% CI 43% to 100%); PET alone sensitivity 82% (95% CI 64% to 92%) and specificity 100% (95% CI 43% to 100%); CT alone sensitivity 100% (95% CI 88% to 100%) and specificity 33% (95% CI 6% to 79%); MRI sensitivity 100% (95% CI 88% to 100%) and specificity 33% (95% CI 61% to 79%).

FDG PET/CT versus FDG PET/CT plus intra-operative ultrasound

A study evaluating FDG PET/CT versus FDG PET/CT plus intra-operative ultrasound (IOUS) for colorectal liver metastases stratified patients into those who had received chemotherapy and those who had not.85 The accuracy estimates were a sensitivity of 63% (95% CI 44% to 78%) (specificity not calculable) for FDG PET/CT and a sensitivity of 92% (95% CI 77% to 97%) and a specificity of 25% (95% CI 4% to 69%) for FDG PET/CT plus IOUS for data based on 31 patients as the units of analysis.

Patient-level data collected from 16 patients without pre-operative chemotherapy were as follows: FDG PET/CT sensitivity 77% (95% CI 50% to 92%) (specificity not calculable); FDG PET/CT plus IOUS sensitivity 100% (95% CI 75% to 100%) (specificity not calculable).

Studies with mixed indications for FDG PET/CT (primary, recurrent and metastatic disease)

FDG PET/CT versus PET alone

One study24 compared FDG PET/CT scans with solitary PET images from the PET component of the integrated equipment. The sensitivity of FDG PET/CT was 86% (95% CI 77% to 91%) and the specificity was 67% (95% CI 44% to 83%), which was slightly superior to the sensitivity [88% (95% CI 80% to 92%)] and specificity [56% (95% CI 33% to 75%)] of PET alone in detecting all lesions. Where patients had multiple liver lesions, a maximum of five were included in the analysis.

FDG PET/CT was shown to possess greater accuracy than PET alone in detecting extrahepatic intra-abdominal disease: FDG PET/CT had a sensitivity of 86% (95% CI 49% to 97%) and a specificity of 93% (95% CI 78% to 98%), while the sensitivity of the PET component was 71% (95% CI 36% to 92%) and the specificity was 90% (95% CI 74% to 96%).

FDG PET/CT alone

A study71 with no comparator reported PET/CT to have a sensitivity of 95% (95% CI 85% to 98%) and a specificity of 83% (95% CI 55% to 95%) in patients with a diagnosis of recurrent and metastatic disease undergoing postoperative follow-up.

FDG PET/CT versus FDG PET/CT plus dedicated CT

A study86 including patients with both primary and metastatic disease found that the diagnostic accuracy of FDG PET/CT was improved by the addition of a dedicated CT scan: FDG PET/CT sensitivity 91% (95% CI 82% to 96%) and specificity 63% (95% CI 45% to 78%) compared with FDG PET/CT plus CT sensitivity 97% (95% CI 91% to 99%) and specificity 100% (95% CI 87% to 100%).

Three studies in which the indication was suspicion of recurrent disease

FDG PET/CT versus multidetector CT

In one study62 involving 67 patients with a suspicion of recurrent CRC, FDG PET/CT was compared with multidetector CT and both tests showed a sensitivity of 100% (95% CI 81% to 100%) and a specificity of 100% (95% CI 92% to 100%) in the diagnosis of liver metastases. Eight patients received a diagnosis of lung metastases, and in these individuals FDG PET/CT demonstrated a sensitivity of 75% (95% CI 40% to 92%) compared with a sensitivity of 100% (95% CI not calculable) for multidetector CT.

FDG PET/CT versus PET (component of the integrated equipment)

In a study63 involving 84 patients, FDG PET/CT demonstrated a sensitivity of 88% (95% CI 73% to 95%) and a specificity of 94% (95% CI 84% to 97%) in diagnosing the presence of intra-abdominal and/or extrahepatic metastases compared with a sensitivity of 81% (95% CI 66% to 91%) and a specificity of 88% (95% CI 76% to 94%) for the use of the PET component of the integrated equipment by itself. For extra-abdominal and/or extrahepatic metastases, the sensitivity of FDG PET/CT was 94% (95% CI 75% to 99%) and the specificity was 100% (95% CI 94% to 100%), while PET alone showed a sensitivity of 74% (95% CI 51% to 88%) and a specificity of 88% (95% CI 77% to 93%).

In a second study68 comparing FDG PET/CT with the PET component of an integrated scanner, FDG PET/CT demonstrated a sensitivity of 87% (95% CI 62% to 96%) and a specificity of 100% (95% CI 89% to 100%), while PET alone had a sensitivity of 87% (95% CI 62% to 96%) and a specificity of 97% (95% CI 84% to 99%).

The accuracy of FDG PET/CT in detecting liver metastases

The accuracy data for FDG PET/CT in detecting liver metastases are derived from seven studies70,79,80,8385,87 including 281 subjects (Figure 3). Only four reported exclusion criteria: patients who had received chemotherapy < 3 months before FDG PET/CT was performed,80 diabetes,83 contraindications for the comparison test (MRI) or > 3 months,84 second primary tumour.85

FIGURE 3. Accuracy of FDG PET/CT in the detection of hepatic metastases based on patient-level data.

FIGURE 3

Accuracy of FDG PET/CT in the detection of hepatic metastases based on patient-level data.

The bivariate HSROC method was not possible as the data did not allow adequate estimation of all the model parameters; therefore, two separate univariate meta-analyses for sensitivity and specificity were used. There was little evidence of heterogeneity in the sensitivity estimates, so fixed-effects meta-analysis was used. The overall estimate of sensitivity is 91% (95% CI 87% to 94%). There was evidence of some heterogeneity in the specificity estimates, so a random effects model was used and the overall estimate of specificity is 76% (95% CI 58% to 88%).

Quality assessment of included studies

Fourteen items from the QUADAS checklist were used to assess the methodological quality of the results, and the findings from this process are shown in Tables 9 and 10.

TABLE 9. Quality assessment of studies evaluating the diagnostic test accuracy of FDG PET/CT in the detection of metastases.

TABLE 9

Quality assessment of studies evaluating the diagnostic test accuracy of FDG PET/CT in the detection of metastases.

TABLE 10. Quality assessment results for studies in which the indication was suspected recurrence but metastatic disease was diagnosed.

TABLE 10

Quality assessment results for studies in which the indication was suspected recurrence but metastatic disease was diagnosed.

In most studies it was unclear whether consecutive series of patients or a random sample of adults were undergoing staging for metastatic CRC; this was reported in only five studies.24,62,84,86,87 We considered 6 weeks to be the time limit after which disease progression might occur, and the time between the reference standard and the index test was reported to be ≤ 6 weeks in only two studies.69,83

A reference standard of surgically resected specimen, biopsy and/or clinical imaging follow-up of at least 6 months was reported in six studies.62,63,68,69,83,87 Most studies reported that a whole or a random selection of the sample received verification using a reference standard of diagnosis. But FDG PET/CT results were verified using a variety of reference standards and, although these were independent of the index test in all studies, the index tests were interpreted without knowledge of the reference standard in only two studies83,84 and this item was not clearly reported in 14 studies.24,62,63,6871,7982,8587

The majority of studies did not clearly report whether the reference standard results were interpreted without knowledge of the index test results or vice versa. Only one study reported uninterpretable test results84 and this information was not clear in all of the others.62,63,6871,7987

The validity of the conclusions from these studies was found to be compromised by the spectrum of patients, disease progression, differential verification and review bias.

Summary

  • The poor quality of the studies means that their conclusions should be interpreted cautiously, although overall it is clear that PET/CT is capable of identifying the small proportion of distant disease that is not detectable by conventional imaging modalities. The pooled accuracy data show FDG PET/CT to have a sensitivity of 91% (87% to 94%) and a specificity of 76% (95% CI 58% to 88%).
  • There are threats to the validity of these conclusions arising from retrospective (case series, audits) study designs and several types of bias. A major threat to the validity of these conclusions arises from the variation in the types of reference standard used (differential verification bias), which undermine the estimates.
  • Data to allow a cross-tabulation of results of different tests for patients contributing to the same study were not available and significance testing for differences between the sensitivity and the specificity of individual tests was not carried out.
  • FDG PET/CT has been shown to be only slightly more accurate than PET alone in detecting metastases in patients with an indication of recurrent disease and in patients with an indication of metastatic disease. However, it has been shown to increase diagnostic certainty as to the type and site of disease.
© 2011, Crown Copyright.

Included under terms of UK Non-commercial Government License.

Bookshelf ID: NBK99950

Views

  • PubReader
  • Print View
  • Cite this Page
  • PDF version of this title (1.7M)
  • Disable Glossary Links

Other titles in this collection

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...