Diagnostica per immagini – Oncologia

SPECIAL CONTRIBUTION 18

Recommendations on the Use of F-FDG

PET in Oncology

1 2 2 3 4 5

James W. Fletcher , Benjamin Djulbegovic , Heloisa P. Soares , Barry A. Siegel , Val J. Lowe , Gary H. Lyman ,

5 6 7 8 1

R. Edward Coleman , Richard Wahl , John Christopher Paschold , Norbert Avril , Lawrence H. Einhorn ,

9 10 11 12 13

W. Warren Suh , David Samson , Dominique Delbeke , Mark Gorman , and Anthony F. Shields

1 2

Indiana University School of Medicine, Indianapolis, Indiana; H. Lee Moffitt Cancer Center at University of South Florida, Tampa,

3 4 5

Florida; Washington University School of Medicine, St. Louis, Missouri; Mayo Clinic, Rochester, Minnesota; Duke University

6 7

Johns Hopkins Medical Center, Baltimore, Maryland; US Oncology, Newport News,

Medical Center, Durham, North Carolina;

8 9

Virginia; Queen Mary’s School of Medicine and Dentistry, London, United Kingdom; Brigham and Women’s Hospital, Boston,

10 11

Massachusetts; Blue Cross and Blue Shield Association, Washington, DC; Vanderbilt University Medical Center, Nashville,

12 13

Tennessee; National Coalition for Cancer Survivorship, Washington, DC; and Karmanos Cancer Institute, Detroit, Michigan

P

The rationale was to develop recommendations on the use of ET is an imaging technique that provides unique

18

F-FDG PET in breast, colorectal, esophageal, head and neck, information about the molecular and metabolic changes

lung, pancreatic, and thyroid cancer; lymphoma, melanoma, associated with disease. The technology has existed for more

and sarcoma; and unknown primary tumor. Outcomes of interest than 30 years but has been used clinically for only the last

18

included the use of F-FDG PET for diagnosing, staging, and 10–15 years. In this period, dramatic improvements in

detecting the recurrence or progression of cancer. Methods: A technology, the routine availability of medical cyclotrons (to

search was performed to identify all published randomized con-

trolled trials and systematic reviews in the literature. An additional produce the necessary short-lived positron emitters), and

search was performed to identify relevant unpublished systematic favorable reimbursement decisions in the late 1990s have led

reviews. These publications comprised both retrospective and to a tremendous increase in the use of this technology. The

prospective studies of varied methodologic quality. The antici- major area of clinical application is currently in oncology,

pated consequences of false-positive and false-negative tests with some application in cardiology and neurology.

18

when evaluating clinical usefulness, and the impact of F-FDG PET requires the use of molecules (radiopharmaceuticals)

PET on the management of cancer patients, were also reviewed.

18

Results and Conclusion: F-FDG PET should be used as an that are labeled with radioactive nuclides. The amounts

imaging tool additional to conventional radiologic methods of radiolabeled material administered are extremely small

such as CT or MRI; any positive finding that could lead to a clin- 26 29

–10 g) and have essentially no pharmacologic effect.

(10

ically significant change in patient management should be con- In this regard, PET has the unique ability to assess molecular

firmed by subsequent histopathologic examination because of alterations associated with disease without perturbing or

18

the risk of false-positive results. F-FDG PET should be used altering the fundamental underlying molecular and biochem-

in the appropriate clinical setting for the diagnosis of head and ical processes. Although the number of molecular probes that

neck, lung, or pancreatic cancer and for unknown primary tumor.

PET is also indicated for staging of breast, colon, esophageal, can be radiolabeled with positron emitters is extremely large,

head and neck, and lung cancer and of lymphoma and mela- and clinical investigational uses number in the thousands,

18

noma. In addition, F-FDG PET should be used to detect recur- clinical practice has been limited principally to the use of a

rence of breast, colorectal, head and neck, or thyroid cancer and 18 F-FDG.

glucose analog labeled with the positron emitter

of lymphoma. 18 F-FDG was first synthesized in 1978 (1) and has become

18

Key Words: oncology; PET; F-FDG PET the most commonly used radiopharmaceutical for PET

J Nucl Med 2008; 49:480–508 studies of cancer and also for the study of normal functions

DOI: 10.2967/jnumed.107.047787 and diseases of the brain and heart. In March 2000, the Food

18 F-FDG to

and Drug Administration approved the use of

assist in the evaluation of malignancy in patients with known

or suspected abnormalities found by other testing methods or

in patients with an existing diagnosis of cancer.

Received Oct. 1, 2007; revision accepted Nov. 20, 2007.

For correspondence or reprints contact: James W. Fletcher, MD, The fact that cancer cells exhibit an increased rate of

Department of Radiology, Indiana/Purdue University, Indiana University 18 F-

glycolysis has been known since the 1920s (2), and

School of Medicine, University Hospital, Room 0655, 550 N. University

Blvd. Indianapolis, IN 46202-5253. FDG PET is able to assess a fundamental alteration in the

E-mail: jwfletch@iupui.edu cellular metabolism of glucose that is common to all

ª

COPYRIGHT 2008 by the Society of Nuclear Medicine, Inc.

T J N M • Vol. 49 • No. 3 • March 2008

480 HE OURNAL OF UCLEAR EDICINE

neoplasms. Increased cellular glucose uptake is one of the National Oncologic PET Registry to allow for assessment of

key alterations associated with the high glycolytic rate of the impact of PET on intended patient management.

cancer cells. 18

GENERAL LIMITATIONS OF DEDICATED F-FDG PET

HISTORY 18

There are inherent limitations of F-FDG PET that can

result in false-negative and false-positive findings. False-

The first medical application of positron emitters was positive findings are most commonly associated with uptake

reported more than 50 years ago in 1951 by Sweet at 18 18

of

Massachusetts General Hospital (3). This application in- F-FDG in infectious or inflammatory tissue (10). F-

volved a simple probe that used coincidence detectors to FDG has been reported to accumulate in various inflamma-

18

localize tumors in the brain. The first published PET images tory processes (11–13). Infection imaging with F-FDG

were acquired using a ring tomograph with the filtered PET relies on the fact that granulocytes and mononuclear

backprojection algorithm and included images of oxygen cells use glucose as an energy source during and only during

15

metabolism with their metabolic burst (14,15), which takes place when acti-

O-oxygen and glucose metabolism with

11 18 18

vated by local triggers. It is therefore not surprising that

C-glucose, as well as F-fluoride bone images (4,5). This F-

publication occurred in 1976, almost 25 years after Sweet’s FDG accumulates in many types of inflammatory tissue. For

18

work at Massachusetts General Hospital. Significant subse- example, F-FDG uptake can be seen in tissue after radiation

quent advances in PET technology were associated with the therapy. Inflammatory changes after radiation therapy can be

identification of bismuth-germanium-oxide as a scintillator protracted and a potential source of false-positive findings if

18

material in 1977 (6) and the successful synthesis of the history, timing, and volume of tissue irradiated are not

F-FDG

18 18

considered at the time of interpretation.

by Ido et al. at Brookhaven in 1978 (1). The first F-FDG F-FDG uptake can

scans were obtained at the University of Pennsylvania in vary widely in normal tissue, and regions of discrete uptake in

18

1979 by Phelps et al. using areas such as the ureters, bowel, lymphatic tissue, thymus,

F-FDG that was synthesized at brown fat, and muscle—so called normal variants—can be

Brookhaven National Laboratory in Long Island (7–9). The interpreted in error as abnormal or can confound the correct

most recent technical innovation, which has been available interpretation of the findings. Mildly to moderately increased

for only the last few years, is the integration of PET and CT 18

systems. These dual-modality systems offer an advantage F-FDG uptake can also be seen in a variety of benign

over dedicated PET in that they can concurrently provide processes, many of which represent inflammatory or hyper-

both metabolic and structural or anatomic images that are plastic conditions (e.g., villous adenomas, thyroid adenomas,

automatically fused and overcome some limitations of ded- Graves disease, adrenal adenoma, Paget’s disease, and fi-

icated PET. brous dysplasia), and familiarity with the behavior of these and

Reimbursement for PET procedures was not available other conditions is important in diminishing false-positive

through much of the 1990s, and adoption of the technology results. 18

was slow. In 1995, the Food and Drug Administration Weaknesses of F-FDG PET for cancer imaging include

18

approved its limited reconstructed spatial resolution of 4–10 mm in

F-FDG for brain imaging in patients with epi- available commercial systems. Negative scan findings cannot

lepsy. This approval paved the way for Health Care Financing exclude the presence of a small tumor or microscopic tissue

Administration reimbursement of PET in January 1998 for involvement, and precise anatomic localization of the signal

lung cancer and cardiovascular disease in Medicare benefi- can be difficult in certain anatomic regions (e.g., the head and

ciaries. This coverage was expanded by the Health Care neck). Tumors with a low metabolic rate (e.g., bronchoalve-

Financing Administration in 1999 to include restricted indi- olar carcinoma and mucinous adenocarcinoma) may show

cations for colorectal cancer, melanoma, and lymphoma. In 18

minimal uptake of

the following year, the Food and Drug Administration gave F-FDG, and certain tumors are known to

18 18

broad approval for have poor avidity for

F-FDG in all cancers and cardiovascular F-FDG (prostate carcinoma and

18

disease. Near the end of 2000, the Health Care Financing hepatocellular cancer). F-FDG PET is also generally con-

18

Administration expanded coverage for broad use of sidered to not be useful in the assessment of possible cerebral

F-FDG metastases from known primary neoplasms. High levels of

PET in lung, colorectal, head and neck, and esophageal 18

cancers as well as lymphoma and melanoma. Since that time, F-FDG are normally present in the cerebral cortex and

18

indications have been added for breast cancer and thyroid substantially limit the utility of F-FDG PET in this appli-

cancer. In February 2006, the Centers for Medicare and cation. For this reason, most clinical examinations are of the

Medicaid Services (the new agency name for the Health Care patient’s torso and include the area from the base of the brain

Financing Administration) announced that it would provide to the mid thigh.

18

coverage for use of F-FDG PET in essentially all other

cancers in accordance with its ‘‘coverage with evidence RATIONALE FOR THE RECOMMENDATIONS

development’’ program. For Medicare beneficiaries under-

going PET as part of this program, referring physicians and The adoption of PET has been variable, but despite

18

PET facilities will be required to provide certain data to the limitations in the published literature, F-FDG PET is

18

U F-FDG PET O • Fletcher et al. 481

SE OF IN NCOLOGY

18

rapidly becoming an integral part of oncology practice in F-FDG PET be used? Recommendations were

should

the United States, Europe, and other countries. developed to assist practitioner and patient decisions about

For these reasons, a multidisciplinary expert panel of health care for specific clinical circumstances (16). It is

important to realize, however, that recommendations can-

oncologists, radiologists, and nuclear physicians with exper- not always account for individual variation among patients.

tise in PET/CT convened to develop recommendations on the The recommendations are not intended to supplant physi-

18

use of F-FDG PET in oncology practice and to determine cian judgment with respect to particular patients or special

18

the suitability of F-FDG PET in the management of cancer. clinical situations.

The multidisciplinary panel was initially convened by the

American Society of Clinical Oncology with members from

the Society of Nuclear Medicine, American College of

Radiology, American Cancer Society, Blue Cross and Blue MATERIALS AND METHODS

Shield Association (BCBSA), National Coalition of Cancer Panel Composition

Survivorship, US Oncology, and American Society for The panel comprised experts in clinical oncology or hematol-

Therapeutic Radiology and Oncology to evaluate the status ogy, radiology or nuclear medicine (specializing in PET), and

of the published literature on PET in oncology and to outcomes or health services researchers with expertise in evidence-

determine whether recommendations on PET could be de- based medicine. Both academic and community practitioners

veloped for referring oncology physicians. The SNM subse- were included. A patient representative was also included on the

quently assumed the responsibility for reviewing and panel.

evaluating the outcome of the panel’s efforts and recommen-

dations. On July 13, 2007, the SNM Board of Directors Process Overview

approved publication of the panel’s findings as this special In evaluating evidence on the role of PET, the panel was guided

contribution to the Journal of Nuclear Medicine. by the process established by the GRADE (Grades of Recom-

Most studies that the panel reviewed included PET with- mendations, Assessment, Development and Evaluation) Working

out CT augmentation. However, the panel realizes PET/CT Group (17). This process follows the principle that systematic

use is increasingly common and expects PET/CT to further reviews of the totality of research evidence represent the scientific

improve the utility of PET. foundation for development of clinical recommendations (18,19).

18

The use of F-FDG PET in the following types of cancer Therefore, the panel first attempted to identify all systematic

was assessed: breast, colorectal, esophageal, head and neck, 18

reviews on the use of F-FDG PET oncology and used these to

lung, pancreas, and thyroid cancer; lymphoma, melanoma, assess the quality of primary research evidence (Tables 1 and 2).

In doing so, the panel soon clearly saw that the systematic reviews

and sarcoma; and unknown primary tumor. The goal was to themselves were of varying quality and that a separate assessment

provide practitioners with recommendations on the appro- of the quality of the systematic reviews was required (Table 3). It

priate use of PET in the management of these cancers and to also became clear that no systematic review was performed using

identify gaps in knowledge that may affect future research. evidence from randomized controlled trials (RCTs). Because, in

Other neoplasms that have been reported and generally general, evidence obtained in RCTs is considered the most reliable

18

recognized as non– F-FDG-avid (e.g., renal, prostate, and (17,20) (Table 4), the panel decided to perform an additional

hepatocellular cancer) were not addressed. search for randomized evidence and perform its critical appraisal.

18

Two principal questions on the appropriateness of F- Therefore, the final recommendations were based on the system-

FDG PET for the management of cancer were addressed: atic review of available randomized evidence and an overview

18

For what cancers should F-FDG PET be used in clinical (systematic review) of the existing systematic reviews addressing

practice, and under what specific clinical circumstances clinical indications of interest (Table 5).

TABLE 1

Definition of Grade of Evidence for Primary Studies and Systematic Reviews

Quality Definition

High Further research is unlikely to change confidence in estimate of effect of intervention.

No serious limitations were noted.

Moderate Further research is likely to have important impact on confidence in estimate of effect of

intervention and may change estimate. Few serious limitations were noted.

Low Further research is very likely to have important impact on confidence in estimate of effect

of intervention and is likely to change estimate. Typically, more than 2 serious limitations were noted.

Unclear Any estimate of effect is uncertain. Evidence is either lacking or was not described well enough to

allow critical appraisal or make any estimate.

Adopted from recommendations of GRADE Working Group (17).

T J N M • Vol. 49 • No. 3 • March 2008

482 HE OURNAL OF UCLEAR EDICINE TABLE 2

GRADE Definitions to Assess Primary Studies

Quality of evidence Study design Decrease if. Increase if.

. . . .

High Randomized trial Study limitations exist Association is. . .

Moderate Serious limitations (21) Strong, with no plausible confounders

Low Observational study Very serious limitations (22) (11)

Very low Any other evidence Important inconsistency is Very strong, with no major threats

present (21) to validity (12)

Directness (generalizability) is Evidence exists of a dose–response

uncertain gradient (11)

Some uncertainty (21) All plausible confounders would have

Major uncertainty (22) reduced the effect (11)

Data are sparse or

imprecise (21)

Probability of reporting bias is

high (21)

Literature Review and Data Collection The evidence profiles were distributed to the panel members, who

used them during the final panel meeting to make their judgments on

Pertinent systematic reviews and RCTs from the published the use of PET for each indication.

literature were retrieved and reviewed for the development of these

recommendations. Searches of MEDLINE (National Library of

Medicine) and other databases (Institute for Clinical Evaluative Consensus Development Based on Evidence

Sciences, Blue Cross Blue Shield Technology Evaluation Center, The entire panel met 3 times. At the first meeting, the panel

and the NHS Health Technology Assessment Program) for pertinent identified the topics of the recommendations, developed a strategy

articles were done using strategies developed by Montori et al. (21) for completion of the recommendations, and did a preliminary

and Mijnhout et al. (22). The search was repeated on June