Positron Emission Tomography (FDG)

CMS NCA document | source_status=Closed | review_type=New | public_comment_open=False | document_id=CAG-00065N

Conditions and limitations for coverage : We do not believe that it is reasonable and necessary to cover specific clinical indications for which adequate scientific data demonstrate that PET does not provide medical benefit. When such evidence exists, use in these indications will be specifically excluded from coverage. For use in oncologic diagnosis: PET is covered in clinical situations in which the PET results may assist in avoiding an invasive diagnostic procedure, or in which the PET results may assist in determining the optimal location to perform an invasive diagnostic procedure. PET is not covered for other diagnostic uses, and is not covered for screening (testing of patients without specific symptoms). For staging and restaging: Coverage for PET is subject to 2 conditions: 1) the stage of the cancer remains in doubt after completion of a standard diagnostic workup, including conventional imaging, and 2) clinical management of the patient would differ depending on the stage of the cancer identified. Use of PET would also be considered reasonable and necessary if it could potentially replace one or more conventional imaging studies. We consider restaging to include both restaging in the setting of recurrence and restaging following completion of a therapeutic regimen or to assess whether a complete response has been achieved. Use of PET to monitor tumor response during the planned course of therapy (i.e. when no change in therapy is being contemplated) is not covered. Prior to obtaining an FDG PET study, the physician ordering this imaging procedure will be required to document in the patient’s chart the specific clinical question that will be answered by the imaging study. The ordering physician will thereby be certifying the medical necessity of the study according to the conditions described above. This documentation is necessary in order for HCFA to be able to reliably review the appropriateness of use of FDG PET under the expanded coverage described in this document. HCFA plans to conduct a review within the first year following the effective date of this new coverage, and will use the results of this review to determine whether there is any need for further review and to decide whether revisions to the coverage policy would be indicated.

Decision Memo: This decision memorandum does not constitute a national coverage determination (NCD). It states CMS's intent to issue an NCD. Prior to any new or modified policy taking effect, CMS must first issue a manual instruction giving specific directions to our claims-processing contractors. That manual issuance, which includes an effective date, is the NCD. If appropriate, the Agency must also change billing and claims processing systems and issue related instructions to allow for payment. The NCD will be published in the Medicare Coverage Issues Manual. Policy changes become effective as of the date listed in the transmittal that announces the Coverage Issues Manual revision. To: File: FDG Positron Emission Tomography (PET) CAG-00065N From: Sean Tunis, MD, MSc Director, Coverage and Analysis Group Mary Stojak Health Insurance Specialist, Coverage and Analysis Group Samantha Richardson Health Insurance Specialist, Coverage and Analysis Group Mitchell Burken, MD Medical Officer, Coverage and Analysis Group Michael Londner, MD, MPH Medical Officer, Coverage and Analysis Group Madeline Ulrich, MD Medical Officer, Coverage and Analysis Group Shana Olshan Technical Advisor, Coverage and Analysis Group Subject: National Medicare Coverage Decision on FDG PET Date: December 15, 2000 In this memorandum we: 1) describe FDG PET scans; 2) review the history of Medicare's coverage policy on PET scans and give an explanation of the coverage guidelines; 3) present and analyze the relevant scientific data including the literature submitted by the requestor; and 4) delineate the changed national coverage policy and HCFA's reason for the coverage decision policy. A summary of the new coverage policy can be found in the last section of this document prior to the appendices. Description and Background of FDG Positron Emission Tomography (PET) PET is a noninvasive diagnostic imaging procedure that assesses the level of metabolic activity and perfusion in various organ systems of the human body. Images are obtained from positron-emitting radioactive tracer substances (radiopharmaceuticals) that are usually administered intravenously to the patient Positron-emitting radioisotopes were first discovered in the 1930's. FDG PET has been evaluated for several decades in pre-clinical models, and is premised on basic research in biochemisty and biology that have established the basis of glucose metabolism in normal cell function, and it's alteration in diseases like cancer, ischemic heart disease and some neurological disorders. The first PET scanners were developed in the United States in the 1970's with the first scan of a human reported in 1978. Through the early 1980's, PET scans were used primarily in research and predominantly focused on the neurosciences because scanners were typically only large enough for head studies. Due largely to the emergence of two major commercial suppliers in the mid-1980's, PET scanners have become capable of whole body imaging and increased computer processing capability. Improvements in the technology have had a significant impact on the quality of PET's image reconstruction and display. PET's Ability to Identify Pathophysiology Most of the disease-specific indications addressed in this coverage determination are related to PET use for various types of malignancies. As a group many of these diseases, which frequently are life-threatening, involve uncontrolled reproduction and spread of abnormal malignant cells. In adults, normal cells in most tissues divide only infrequently to replace worn-out or dying cells and to repair injuries. Malignant cells, which are both structurally and functionally abnormal, compete with and destroy normal cells and may spread throughout the body. They may aggregate in solid masses referred to as tumors. The spread of malignancy to a new site is called metastasis. Classification of cancer by its appearance under a microscope and the part of the body in which it began, is important because different types of cancer vary in growth rates; how they spread through out the body, and in their susceptibility to various anticancer therapies. An accurate diagnosis of where the cancer originated in the body and its type is necessary so that the physician can determine the appropriate clinical management of the patient. As a molecular diagnostic imaging modality, PET can detect rates of biological activity, as contrasted other imaging modalities such as x-ray films, computed tomography (CT), and magnetic resonance imaging (MRI), which depict the anatomical location of both normal and abnormal structures in the body. Malignancies can cause abnormalities of blood flow or metabolism before anatomic changes are apparent. Thus, disease can be detected by PET when anatomic imaging studies are still normal, and may be informative in differentiating benign from malignant processes. PET evaluation of tissue metabolism can indicate the probable presence or absence of malignancy based on observed differences of biologic activity, whereas anatomic imaging depends on the size and radiographic characteristics of lesions to determine the likelihood of malignancy. In addition, whole body imaging with PET provides the means to examine all organ systems for both primary and metastatic disease in a single procedure. Safety of PET and Approval by the Food and Drug Administration (FDA) of FDG for PET Scans The safety of PET is usually discussed in terms of the safety of the positron emitting radiopharmaceuticals or tracers. Silberstein (1998) conduced a study of 22 PET centers to determine what adverse reactions to the pharmaceuticals were observed retrospectively from the date the centers opened until 1994, and prospectively from 1994 to 1997. No negative effects were observed. In 1972, FDA first approved a new drug application (NDA) for sodium fluoride F 18 injection as a bone imaging agent to define areas of altered osteogenic activity. Marketing of this product ceased in 1975. Another tracer, Rubidium chloride 82 injection was approved in 1989 for assessing regional myocardial perfusion in the diagnosis and localization of myocardial infarction. The last tracer approved prior to 2000, was for the use of FDG injection for identification of regions of abnormal glucose metabolism associated with foci of epileptic seizures, in 1989. On March 12, 2000, the FDA published a notice in the Federal Register that expanded approval of FDG for new indications. FDA concluded in that notice that a 10-millicuries (mCi ) dosage (for adults) of FDG is safe and effective for oncological and cardiac applications. For cancer, FDG was specifically approved for assessing abnormal glucose metabolism to assist in evaluating malignancy in patients with known or suspected abnormalities found by other testing modalities or in patients with an existing diagnosis of cancer. This approval was based on 2 well designed studies of the use of FDG PET for specific oncologic applications, and 10 additional supporting studies of lower methodologic quality. For cardiac applications, FDG was specifically approved for imaging of patients with coronary artery disease and left ventricular dysfunction, and when used together with myocardial perfusion imaging for identification of left ventricular myocardium with residual glucose metabolism and possible reversible loss of systolic function. Summary of the History of Medicare's Coverage of PET Scans and an Explanation of the Coverage Guidelines Medicare has reviewed the scientific literature regarding PET scans over a number of years, and has established coverage for six usesone in 1995, two in 1998, and three in 1999. All but the first use FDG as the tracer. PET Scans using Rubidium 82 (Rb 82) for the Imaging of Perfusion of the Heart and Management of Patients with Known or Suspected Coronary Artery Disease For services performed on or after March 14, 1995, Medicare first covered PET Scans using Rubidium 82 (Rb 82) done at rest or with pharmacological stress for the imaging of perfusion of the heart and management of patients with known or suspected coronary artery disease when: Used in place of, but not in addition to, a single photon emission computed tomography (SPECT), or The PET scan, whether rest alone or rest with stress, is used following a SPECT that was found inconclusive. In these cases, the PET scan must have been considered necessary in order to determine what medical or surgical intervention is required to treat the patient. (For purposes of this requirement, an inconclusive test is a test(s) whose results are equivocal, technically uninterpretable, or discordant with a patient's other clinical data.) The coverage policy did not allow for PET scans using Rubidium 82 for the screening of asymptomatic patients, regardless of the number and severity of risk. Staging of Non-Small Cell Lung Carcinoma (NSCLC) Starting in January 1998, FDG PET scans were covered when used for the initial staging of suspected metastatic NSCLC in thoracic (mediastinal) lymph nodes in patients who have a confirmed primary lung tumor, but for whom extent of disease has not yet been established. The primary purpose of such staging is to determine the progress and extent of the disease, and based on that information to plan future management for the patient. Evidence of primary tumorA surgical pathology report is necessary to document the presence of an NSCLC. Whole body PET scan results and results of concurrent computed tomography (CT) and follow-up lymph node biopsyPET scans must be properly coordinated with other diagnostic modalities. The following reports are required to verify testing: the results of concurrent thoracic CT, which is necessary for anatomic information the results of any lymph node biopsy performed to finalize whether the patient will be a surgical candidate. A lymph node biopsy is not covered in the case of a negative CT and negative PET, where the patient is considered a surgical candidate, given the presumed absence of metastatic NSCLC unless medical review supports a determination of medical necessity of a biopsy. A lymph node biopsy is covered in all other cases, i.e., positive CT+ positive PE T; negative CT+ positive PET; positive CT+ negative PET. Coverage of FDG PET Scans for Characterization of Solitary Pulmonary Nodules Also beginning in 1998, FDG PET scans were covered when used for the characterization of suspected solitary pulmonary nodules (SPNs). The primary purpose of such characterization should be to determine the likelihood of malignancy, in order to plan future management and treatment of the patient subject to the following conditions: Evidence of primary tumor Evidence of the initial detection of a SPN, usually by computed tomography (CT), is required. When other concurrent imaging techniques are also used, the results must be included on the claim. In the case of serial evaluation of SPNs using both CT and regional FDG PET chest scanning such PET scans will not be covered if repeated within 90 days following a negative PET scan. In 1999, coverages of FDG PET for evaluation of recurrent colorectal cancer in patients with rising levels of carcinoembryonic antigen (CEA), for staging of lymphoma (both Hodgkin's and non-Hodgkin's) when the PET scan substitutes for a Gallium scan, and for the detection of recurrent melanoma were added. Determining the Location of Recurrent Colorectal Tumors when Indicated by Rising Levels of CEA In 1999, FDG PET was covered when used for determining the location of recurrent colorectal tumors when such tumors were indicated by rising levels of CEA. The use of FDG PET was limited to locating such tumors for the purpose of making a decision as to whether surgical intervention is warranted. However, the use of FDG PET to stage colorectal carcinoma was not covered under this national coverage decision. The provisions of the coverage policy were designed to limit coverage of PET to those situations in which it is effective in determining the course of future patient treatment. Determining the medical effectiveness of a service based on its utility in determining the course of treatment, is generally applied by Medicare to diagnostic modalities that are used as a substitute, or are intended to replace, other diagnostic modalities. The following conditions were also required: Evidence of documented previous colorectal carcinoma. Use of results of concurrent computed tomography (CT) and/or other diagnostic modalities when they are necessary for additional anatomic information. Frequency limitation of once every 12 months, unless medical necessity documentation supports a separate re-elevation of CEA within this period. Staging of Lymphoma when Used as an Alternative to a Gallium Scan Also determined in 1999, FDG PET scans became covered when used for staging lymphoma as an alternative to a Gallium scan when the following conditions are met. Evidence of disease Before the FDG PET scan is performed, a pathologic diagnosis of lymphoma must have already been made. When other concurrent imaging techniques are also used, the results must be included on the claim. Assurance that the FDG PET scan is an alternative to a Gallium scan. Limitation on use PET scans are not allowed any sooner than 50 days following the last PET scan or Gallium scan. Whole body FDG PET scans are covered only once every 12 months unless medical necessity documentation supports the specific need for localization of possible recurrent tumor within this period. Evaluation of Recurrence of Melanoma Prior to Surgery and as an Alternative to a Gallium Scan The last medical condition that became covered in 1999, was for the evaluation of melanoma prior to surgery in situations under the following conditions: Evidence of disease The patient must have previously been diagnosed with melanoma. When other concurrent imaging techniques are also used, the results must be included on the claim. Assurance that the PET scan is an alternative to a Gallium scan. Limitation on use – PET scans are allowed no sooner than 50 days following the last PET scan or Gallium scan. Full body PET scans are covered only once every 12 months unless medical necessity documentation supports the specific need for localization of possible recurrent tumor within this period. Current FDG PET Scan Coverage Request On July 10, 2000, HCFA received a request for broad coverage of FDG PET scans from Drs. Michael Phelps and Sam Gambhir. A list of 22 diseases was included in the request which covered various oncological conditions, myocardial viability, and neurological conditions. We determined that the appropriate benefit category fell under §1861(s)(3) diagnostic services. Due to volume of the evidence submitted by the PET community, we requested assistance from the Agency for Health Research and Quality (AHRQ). AHRQ had an Evidence-based Practice center (EPC) perform a validation check of the entire FDG PET submission. The New England Medical center of Tufts University provided this validation. The EPC performed a literature search of the Medline and Biosis Previews databases for each of the clinical conditions listed in the PET request. The search was done to identify the universe of scientific evidence on the PET conditions submitted in the context of comparing the submitted material against a master bibliographic profile. The EPC conducted a search for potentially relevant PET scientific articles that dated from 1990 –2000. They located over 500 articles that were potentially relevant to the usage of PET scanners. The NEMC was not required to further analyze the data because HCFA did not request a full technology assessment. The EPC concluded that the PET request was not presented as a standard systematic literature review, but represented a large bibliographic compilation of the literature. The NEMC report raised some questions about relevant studies that might not have been included in the PET request, and identified several errors in the data that were abstracted from the studies to create the summary tables. HCFA concluded that it would be necessary to conduct independent systematic reviews of the FDG PET literature in order to produce appropriate coverage policy. In order to assure a full and open public discussion of the scientific and clinical issues raised by FDG PET, we requested advice from the Medicare Coverage Advisory Committee (MCAC) on October 17 th . The Executive Committee of the MCAC met on November 7, 2000 to consider guidelines for the evaluation of diagnostic tests in general and to consider selected issues (i.e. colorectal cancer management, differential diagnosis of dementia, and lung cancer diagnosis and staging) from the PET coverage request. After an overview of PET presented by Dr. Phelps, the Executive Committee chairman, Dr. Harold Sox, presented the Working Framework for Evaluating Diagnostic Tests, found in Appendix B. The Guidelines were discussed by the panelists, but not subjected to a formal vote. It was the sense of the panel that one should consider first whether the evidence is sufficient to establish that a test under consideration provides diagnostic information that is at least as effective as standard alternatives. The Committee then made suggestions of issues future panels might want to consider in assessing the impact on health outcomes of particular diagnostic tests. Following public comments, the Committee discussed application of the guidelines to some of the proposed new uses for PET. Although there was no formal vote, generally, the Committee suggested that PET had benefit in assessing recurrent colorectal cancer and that there was some evidence that might be generalized to the use of PET in other applications. It was noted that the performance of PET may differ depending on the specific cancer being evaluated and the physical location of the cancer and any possible metastases. It was further suggested that the MCAC Diagnostics Panel might look into the details of extending coverage for other oncologic indications based on the evidence related to colorectal cancer. Some Committee members also expressed concern that HCFA's policy might prevent coverage for PET use for certain cancers because their rarity precluded performance of necessary studies. Quality of Studies Evaluating Diagnostic Technology Over the past decade, the characteristics of high quality studies for evaluating diagnostic tests have been well-documented in a number of committee reports and peer-reviewed publications. Most of these documents come to similar conclusions about the study design characteristics that are helpful in reducing bias, and ensuring that the reported results are an accurate reflection of the performance of the test. These characteristics are similar to those included in the following chart. Experimental Design Features That Enhance Scientific Rigor of Diagnostic Test Evaluations* Design Feature Comments Defining the problem and hypotheses Helps to clarify the clinical problem Inclusion and exclusion criteria are defined to reduce confounding variables Adequate patient sample size for sufficient statistical power Depends on the expected magnitude of effect and whether all patients have both competing imaging tests Patient referral sources that include a clearly defined broad spectrum of disease presentation and severity Reduces referral bias (spectrum bias)** Clearly defined patient groups based on pre-test probability estimates Allows reader to judge generalizability of findings to his/her practice Offsets referral bias Consider adequate sample size for each subgroup analysis All patients have comparison tests and similar follow-up Reduces work-up bias*** Randomized, independent, blinded reading of competing tests Avoids test review bias**** Consider blinding test interpreters to clinical information, other tests, and final diagnosis Should develop methods to reduce interobserver variation Expert interdisciplinary gold standard panel and determination of true diagnosis Diagnosis determined both with and without test results allow measurement of the degree of diagnostic review bias (incorporation bias)***** in result Outcomes analysis Data on operating test characteristics are gathered using a research protocol Data on consequences of diagnostic and treatment choices on patient outcomes are obtained from the literature * Adapted from Veterans Health Administration Report (1997) ** referral bias relates to the differences among patient populations in the spectrum of disease presentation and severity *** work-up bias most commonly occurs when results from one test determines inclusion or exclusion from the study or from further work-up **** test review bias occurs when the final diagnosis or results of the comparison test are used in planning or interpreting the test under study ***** diagnostic review bias occurs when the gold standard diagnosis is influenced by results of the imaging test These principles of study design were incorporated into the Proposed Guidelines for Evaluating Diagnostic Tests discussed by the MCAC Executive Committee (see Appendix D). The following chart from that document illustrates how failure to follow established principles of scientific investigation can weaken study results. Ideal study Usual study Effect of Usual Study The study subjects are consecutive patients seen in a typical clinical setting with a chief complaint. Subjects selected because they have had the diagnostic gold standard. Overestimates sensitivity and underestimates specificity. All patients who get the index test also get the reference test. Patients with negative results on the index test often don't get the diagnostic gold standard. Overestimates sensitivity and underestimates specificity. The person who interprets the index test is blinded to all other information. The person who interprets the index knows the clinical history and the results of the diagnostic gold standard. Overestimates sensitivity and specificity. The person who interprets the reference test is blinded to all other information. The person who interprets the diagnostic gold standard knows the clinical history and the results of the index test. Overestimates sensitivity and specificity. The reference test is a valid measure of the disease state. The diagnostic gold standard imperfectly measures the disease state. The measured test performance could either be worse or better than the true performance. Analysis of the Relevant Scientific Data For this coverage request, we have supplemented the requestor's submission with technology assessments published in 2000 by Blue Cross Blue Shield, the Report of the Commonwealth Review of Positron Emission Tomography (also published in 2000), and additional analysis on lung and esophageal cancers using material from the requestor's submission. In the next section of this memorandum, we outline a number of disease-specific indications for use of FDG PET. Our coverage NCDs are based on the use of those assessments, the requestor's submission, and our limited literature review using the same basic questions that the MCAC used in their Working Diagnostic Guidelines. In addition to the published data reviewed above, HCFA also considered other forms of evidence, including extensive consultation with clinical experts in oncology, nuclear medicine, cardiology, neurology, and other relevant clinical disciplines. We also took into account the basic biology and biochemistry of disease upon which PET imaging technology is based. All of this information was helpful in interpreting the direct empirical studies that have been performed to evaluate the test performance and clinical utility of PET. The relevant body of evaluation literature is briefly summarized in this section for the subset of clinical applications of PET addressed in this coverage decision memo. Lung Cancer (Non-Small Cell) Background HCFA coverage has already been provided for evaluating solitary pulmonary nodules, as well as staging non-small cell carcinoma of the lung (NSCLC), making it logical to inquire about evidence which might support applying FDG PET to detecting residual or recurrent NSCLC. The submitted package by UCLA includes a relatively large diagnostic trial by Bury et al. (1999) which supports this application. In a group of 126 consecutive patients, divided into 58 who were in an early curative group and 68 in an early palliative group, there was considerably higher sensitivity for PET (100%) vs. CT (72%). Please note that specificities were both equivalent (> 90%), and the same performance trends were found in each patient subgroup. Application of working diagnostic guidelines: Is this study of PET accuracy sufficiently free of bias to permit conclusions about the accuracy of PET as a diagnostic imaging test? Other than a reported absence of blinding, there was no strong source of bias, given the relatively large sample size and use of consecutive patients to minimize selection bias. What is the potential impact of PET accuracy upon health outcomes? Per the relatively strong study by Bury, one may surmise that PET provides an appropriate degree of diagnostic accuracy. However, subsequent outcomes data are not furnished. Recommendation: There is evidence to support the role of PET detecting residual or recurrent tumor after treatment of NSCLC. Rationale: The Bury study provides an evidentiary rationale for supporting this particular application of FDG PET. Esophageal Cancer Background Esophageal cancer is a relatively rare but lethal type of cancer, which is newly diagnosed in approximately 10,000 Americans each year. This tumor has been considered synonymous with squamous cell carcinoma. However, adenocarcinoma is now more common in the United States, and has a rising incidence rate. Although the overall 5-year survival rate has remained steady at about 5%, patient management may be greatly assisted by diagnostic techniques that can properly assign patients to a curative subgroup, where extension of disease has not already disqualified patients for surgery. The 5-year survival rate with surgical intervention alone is approximately 30%. Role of PET in Pre-Surgical Staging The below scenarios can be described, whereby PET, if it is demonstrated to have added diagnostic benefit, could be used as an adjunct to conventional imaging (CI), such as computed tomography (CT) and ultrasound (US), in primary staging: If CI is negative and PET is positive for metastatic disease, then it is likely that the patient has unresectable disease and curative surgery is not applicable; If CI is negative and PET is negative, then, conversely, it is likely that the patient is a candidate for curative surgery; If CI is positive, then it may not be likely that PET is even needed since the patient has already been demonstrated to have unresectable disease. In tandem with full-length articles provided by the PET request package, a supplemental Medline search (Ovid) was conducted for the textwords "esophageal cancer" and "PET," with limitations to human studies in English, published from 1997-2000. The following inclusion criteria were applied such that eight studies were selected for further review: Study sample included at least 10 patients; Patient sample homogeneous with respect to type of primary cancer, and Study described correlation of FDG PET findings with data from an appropriate reference standard, for at least some of the patients in its sample. Three of these studies (Kole et al. 1998, Luketich et al. 1999, Flamen et al. 2000) report comparisons in the ability of CT to measure distant metastases versus FDG PET. The Kole study combines factors for overall resectability, demonstrating an accuracy of 65% for CT versus 88% for PET (p = 0.04, using McNemar test), but without mention of sensitivity and specificity values. The Luketich study demonstrates a sensitivity of 69% and specificity of 93% for PET versus 46% and 74%, respectively for CT. Finally, Flamen corroborates this favorable trend, by reporting a PET sensitivity/specificity of 74%/90% in detecting Stage IV disease versus 47%/78% for CT plus US. Thus, in all three studies, there is evidence of PET's additional diagnostic benefit for assessing metastatic disease. Four studies (Flanagan et al. 1997, Block et al. 1997, Luketich et al. 1997, Choi et al. 2000), in addition to Flamen and Kole, provided data on nodal evaluations, and there was at least comparable performance data for both conventional imaging and PET. Application of working diagnostic guidelines: Are the studies of PET accuracy sufficiently free of bias to permit conclusions about the accuracy of PET as a diagnostic imaging test? Of the three key studies used to support the relative benefit of PET in detecting metastatic disease, the Flamen article reveals no significant sources of bias. Although both Kole and Luketich et al. 1999, both used consecutive patients, each study did not apply all diagnostic tests to all patients, and the latter study also failed to demonstrate blinding. What is the potential impact of PET accuracy upon health outcomes? Yeung et al. 1999 and Flanagan have shown patient management changes of 14% and 17%, respectively, with respect to the use of PET. Furthermore, Luketich et al. 1999 used Kaplan-Meier survival analysis to demonstrate that patients with local disease on PET had a 30-month survival of 60% versus 20% survival for those who had distant disease on PET. These findings suggest that the use of PET has a positive effect upon health outcomes. Recommendation: Evidence is present to support the use of PET in pre-surgical staging of esophageal cancer. Rationale: Multiple studies provide the basis for such coverage. The clinical dilemma posed by the limitations of conventional imaging can be, in part, addressed by the further use of functional PET imaging. Role of PET in Monitoring Recurrence There is very limited data to suggest that PET can be a valuable tool for monitoring treatment; however, the Yeung study profiles 84/150 scans for this type of indication. Although there is combined data for both staging and recurrence in this data set, the overall superior performance of PET (80% sensitivity and 95% specificity) versus CT (68% and 81%, respectively) would suggest that some benefit may be conferred for this use of PET in managing esophageal cancer. Recommendation: Some evidence supports extension of PET coverage into this additional indication for esophageal cancer. Rationale: Unless strong negative evidence is present, HCFA can use this evidence to support broadening coverage within this tumor type. Colorectal Cancer Background Carcinoma of the large bowel is by far the most common and most curable carcinoma of the gastrointestinal tract, with approximately 140,000 new cases per year and 55,000 deaths per year. Males and females are affected equally, the mean age of incidence is 62 years. Different stages of tumor have been classified which depend upon whether: The tumor involves the wall of the bowel only, there is extension through the wall, there is lymph node metastatic disease, or there is distant metastatic involvement. Therefore, multiple patient management checkpoints will be evaluated where FDG PET may contribute useful diagnostic information: The ability of PET to differentiate local recurrence of tumor from postoperative scarring at the primary surgical site; The role of PET to provide additional benefit over conventional imaging for primary staging of hepatic and extrahepatic disease, before any surgery/therapy has been undertaken, and The role of PET for assessing recurrent colorectal cancer beyond simply where the tumor marker carcinoembryonic antigen (CEA) serves as a trigger for investigation, noting that current HCFA policy allows for PET evaluation only in the context of a rising CEA. Distinguishing Local Recurrence from Postoperative Scar In patients who have undergone primary resection for colorectal cancer, FDG PET may be instrumental in detecting whether tumor has recurred at the surgical site. The following management alternatives are faced by patients who present with this dilemma: Biopsy the area, or Perform a test, such as a PET scan, which may reduce the probability that an indurated area is recurrent cancer, such that, in turn: If the PET scan is negative, conduct watchful waiting, or If the PET scan is positive, proceed to biopsy. Beneficial outcomes (true negatives) occur when the PET scan correctly shows that a local lesion is a post-operative scar, and a biopsy procedure may be rendered unnecessary. Conversely, adverse outcomes (false negatives) occur when PET incorrectly suggests the area in question is postoperative scar, thus causing clinicians to forego biopsy which could have shown recurrent tumor. This incorrect imaging result could presumably result in a missed opportunity for curative resection. Consequently, PET would demonstrate greater clinical utility based upon its ability to generate a very high negative predictive value (NPV) in which there is a relatively low proportion of false negative results. Therefore, in the context of a high NPV, a patient might elect to forego a tissue sampling procedure and continue with less invasive monitoring. We chose six studies which were selected for review from the Blue Cross/Blue Shield TEC assessment, using the following inclusion criteria: Study published or accepted for publication as a full article in a peer-reviewed journal; Study sample included at least 10 patients; Patient sample homogeneous with respect to type of primary cancer; Study performed tomographic, not planar, imaging with FDG as the radiotracer, and Study described correlation of FDG PET findings with data from an appropriate reference standard, for at least some of the patients in its sample. Even though there was a high sensitivity = 96% and high specificity = 98%, the Bayesian estimate of NPV was 92%, given the unweighted pooled probability of local recurrence = 69%. This pooled NPV estimate of 92% means that the probability of occult local recurrence in patients with negative PET scans is 8%. Please note that if this prevalence of local recurrence had only been 5%, given the same values of sensitivity and specificity, the NPV would have been much higher at 99.8%. Application of working diagnostic guidelines: Are the studies of PET accuracy sufficiently free of bias to permit conclusions about the accuracy of PET as a diagnostic imaging test? The six studies provided useful diagnostic performance data, and this assertion was also confirmed by the MCAC Executive Committee panelists, subject to the following potential sources of bias: Consecutive patient enrollment was not required as a means of minimizing selection bias, none of the six studies successfully demonstrated blinding protocols, a gold standard reference test was not required for all study patients, and two out of six studies only had 15 and 18 patients, respectively. What is the potential impact of PET accuracy upon health outcomes? The TEC assessment postulated that patients and their physicians would be unlikely to forego histologic sampling, based upon PET scan findings, with a false negative rate as high as 8%, since this could likely cancel/delay re-operation, which has an approximate 20% chance of cure. The MCAC panelists expressed similar concerns that this reported false negative rate would impose such a barrier. However, since the panelists also surmised that using PET for suspected local recurrence could, in turn, pick up additional extra-pelvic metastases (see Schiepers et al. 1995), there was a majority opinion that PET imaging could have a favorable impact upon patient management/health outcomes. Thus, it may not be pertinent to consider PET scanning for recurrent tumor which occurs only at the local resection site, but PET would be helpful in detecting more widespread recurrent disease. Recommendation: Coverage is supported for FDG PET to help differentiate post-operative scar from the recurrence of colorectal carcinoma. Rationale: It appears that PET scanning has the ability to influence the post-test probabilities such that patients and their physicians can choose an appropriate biopsy strategy which, in turn, maximizes the opportunity for curative resection of recurrent colorectal carcinoma. Detecting Hepatic and Extrahepatic Metastases The detection of hepatic and extrahepatic metastases by clinicians can improve the selection of surgical candidates. Patients with non-resectable metastases can be more accurately identified, so that unnecessary surgery can be avoided. The logic of the Blue Cross/Blue Shield TEC causal chain is as follows, assuming that PET follows conventional imaging (CI): If CI demonstrates resectable disease, with cure potentially achievable in approximately 30% of patients, then either: CI and PET are concordant such that surgery is pursued, or CI and PET are discordant such that palliation is chosen in lieu of surgery. If CI demonstrates non-resectable disease, then either: Concordance of CI and PET avoids unnecessary surgery, or Discordance of CI and PET encourages the pathway of curative surgery. The previous July 1999 coverage instructions, which were issued after review of the presentations made at a January 1999 PET Town Hall Meeting, granted coverage for PET when recurrence is suspected as a result of a rising serum CEA level. Therefore, additional coverage deliberations on this issue should address the following two narrow questions: Does PET provide additional benefit over CI in primary staging of hepatic and extrahepatic disease, before any surgery/therapy has been undertaken? Should the assessment of recurrent colorectal cancer only be limited to situations where rising CEA serves as a trigger for investigation? Only one study by Abdel-Nabi et al. (1998) presents data on primary staging. With respect to hepatic metastases, this study showed a sensitivity of 38% for CT, as opposed to 88% for PET, and specificities of 97% and 100% for CT and PET, respectively. Regarding extrahepatic nodal metastases, both CT and PET had a sensitivity of 29%, compared to specificities of 85% and 96% for CT and PET, respectively. Application of working diagnostic guidelines for primary staging of metastatic lesions: Is this study of PET accuracy sufficiently free of bias to permit conclusions about the accuracy of PET as a diagnostic imaging test? Study strengths included consecutive recruitment of patients to minimize selection bias, Sample size nearly 50 (n = 48), and 44/48 patients subjected to a desirable surgical gold standard. The major study limitation was an unblinded design. What is the potential impact of PET accuracy upon health outcomes? When presented to the MCAC panelists, the post-test probability data for both primary staging, as well as for assessment of recurrence, received a positive response with respect to patient management changes/improved health outcomes. Based upon this general acceptance of the studies, one may infer a positive response for the use of PET related to can primary staging. Recommendation: Coverage is supported for the use of FDG PET when determining the presence of hepatic/extrahepatic metastases in the primary staging of colorectal carcinoma, prior to selecting a treatment regimen. Rationale: The relatively strong findings presented in the Abdel-Nabi study provide the evidentiary basis for this recommendation. Application of working diagnostic guidelines for evaluating recurrent hepatic/extrahepatic disease when there are indicators other than rising CEA: In 1999, Valk et al. presented data related to the issue of use of PET in the absence of rising CEA. Other studies have looked at rising CEA in combination with other indicators of suspected recurrence (e.g., abnormal CT scan). In the Valk study, a subgroup of 76 patients were referred for PET based solely upon positive CT findings (i.e., solitary recurrent lesion), as opposed to some admixture of rising CEA, CT, etc. PET results altered post-test probabilities in several ways: 47 patients had confirmed localized single recurrences with PET and proceeded to intended curative surgical follow-up; 23 patients were found to have unsuspected sites of recurrence, thus altering patient management, such that 10/23 patients did not undergo surgery; and 6 patients showed no tumor, causing 2 patients to defer surgery in favor of clinical follow-up (please note that both patients were free of recurrent tumor 14-32 months after PET scan). Is this study of PET accuracy sufficiently free of bias to permit conclusions about the accuracy of PET as a diagnostic imaging test? The Valk study's particular strengths is its recruitment of 155 consecutive patients, with relatively more complete blinding than many other PET studies. What is the potential impact of PET accuracy upon health outcomes? The above data provide evidence that rising CEA should not be viewed as the only trigger for evaluating recurrent disease. Although the change in eventual health outcomes may not be obvious from this limited data, there were documented patient management changes as a result of PET imaging under this clinical scenario. Recommendation: Coverage is supported for expanding the role of evaluating recurrent hepatic/extrahepatic colorectal cancer beyond the limited presentation of a rising CEA level. Rationale: Whereas rising CEA provides the most obvious trigger for evaluating colorectal cancer recurrence, the ability to tease out other potential risk factors was limited by several studies in which rising CEA was combined with multiple other factors. However, the Valk study presents convincing data on abnormal CT scans which, in turn, support a less restrictive approach to monitoring recurrence of colorectal cancer. Lymphoma Staging and restaging of both Hodgkin's and non-Hodgkin's disease have previously been approved for Medicare coverage. The recent Blue Cross/Blue Shield TEC Assessment supported that determination. Melanoma Background Malignant melanoma, which is a relatively aggressive cancer arising primarily in the skin, affected 44,000 new patients and resulted in 7,300 deaths in 1999 and this number continues to rise. Invasive melanoma is classified in four categories I – IV ranging from primary tumor, with thickness less than 1.5 mm, to extranodal metastastic involvement. These stages can be quantified to account for incidence of disease and mortality. Localized disease accounts for 82% of new disease and has a five-year survival of 87.7 % while distant disease accounts for 4% of new disease, but five-year survival is only 12.6%. The review of PET with regard to melanoma will include two indications: Detecting regional lymph node metastases in either initial staging or monitoring after primary treatment, and Detecting extranodal metastasis at initial staging or during follow-up after treatment. Detecting regional lymph node metastases during either initial staging or monitoring after primary treatment It is essential to first emphasize that HCFA already covers monitoring after primary treatment; therefore, the current discussion is limited to the detection of regional lymph node metastases during initial staging. This question addresses patients who have clinically localized disease with invasive cutaneous lesions of intermediate thickness (1.0-4.0 mm). For these patients, PET may be beneficial in determining the appropriateness of sentinel node biopsy (SNB). Traditionally, when patients are diagnosed with local disease, they undergo SNB to determine the need for elective lymph node dissection. If PET can be demonstrated to be as sensitive and specific as SNB in determining lymph node metastases, then these persons can be spared the possible adverse effects from SNB. When both PET and SNB are concordant, there is no change in management and no harm in a less invasive approach. The caveat arises when PET is falsely positive or negative. When PET is falsely negative, the patient forgoes or delays potentially beneficial lymph node dissection, and when PET is falsely positive, the patient undergoes an unnecessary dissection. Thus, the required study design compares the accuracy of PET to SNB, and there are four possible outcomes: PET positive and SNB positive (concordant true positive): In this instance both studies would recommend elective lymph node dissection. PET negative and SNB negative (concordant true negative): In this instance both pathways direct the safe avoidance of lymph node dissection. In fact if PET were equal or better than SNB, PET would avoid SNB as well. This is precisely the group PET looks to impact. PET negative and SNB positive (discordant false negative): This is the dangerous category since patients with true disease would forgo or delay elective lymph node dissection. PET positive and SNB negative (discordant false positive): These patients would get over treated with elective lymph node dissection. The following study selection criteria were used in the Blue Cross/Blue Shield TEC Assessment of this issue: Published or accepted for publication as a full article in a peer-reviewed journal; At least 10 patients; Patient sample homogeneous with respect to type of primary cancer (i.e., studies excluded if there were either patients with various tumor types or if there was a mixture of primary and metastatic lesions); Performed tomographic rather than planar imaging with FDG as the radiotracer, and Correlation of PET findings with data from an appropriate reference standard, for at least some of the patients in the standard. Of the seven studies included for review, only one addressed the use of PET in detecting lymph node metastases (Wagner et al. 1999). This "prospective blinded" study enrolled 74 patients, 70 of whom had assessable cutaneous lesions > 1 mm in depth. PET was positive in only three of the 18 patients with positive SNB, corresponding to a sensitivity of 17%. Although specificity was 96%, PET failed to capture 83% of patients with positive SNB. This is an unacceptable number of patients to forgo or delay necessary lymph node dissection. Application of working diagnostic guidelines: Is this study of FDG PET accuracy sufficiently free of bias to permit conclusions about the accuracy of PET as a diagnostic imaging test? The Wagner study has no obvious sources of bias, although it was not clear whether consecutive patients were recruited in an effort to minimize selection bias. What is the potential impact of FDG PET accuracy upon health outcomes? Based on the apparent lack of evidence presented above, coupled with the further lack of patient management and outcomes data, PET cannot replace SNB as a safe and less invasive method of detecting lymph node metastases. Recommendation: Coverage is not supported for using PET to evaluate regional lymph nodes. Rationale: It is clear that strongly negative studies should play an important role in providing requisite caveats. In this instance, where there is a lack of overwhelming evidence to the contrary, the Wagner study should be used to preclude PET coverage for regional lymph node evaluation. Detecting extranodal metastasis at initial staging or during follow-up after treatment As noted above, new coverage deliberations only apply to initial or primary, pre-treatment staging, given HCFA's reimbursement in monitoring for recurrent melanoma. This evaluation focuses upon the addition of PET to conventional imaging (CI) studies and whether PET offers benefit to clinical decision-making. This obviously would allow for more appropriate, directed therapy if PET is more accurate (than CI) with respect to disease quantification and localization. Conversely, if PET either under-or overestimates disease, these patients will be inadvertently mistreated. The potential impact of this new technology depends upon the extent of discordance between conventional imaging and PET imaging. When both agree regarding either localized or metastatic disease, the management will not presumably change, but when there is discordance, the patient is at risk for harm. When PET falsely underestimates the extent of extranodal disease, patients receive less than optimal therapy. Conversely, when PET overestimates extranodal disease, patients may receive unnecessary therapy and are exposed to greater treatment morbidity. The ideal study would prospectively categorize patients according to a reference standard stage of disease and compare the accuracy of identifying the stage with conventional imaging alone versus conventional imaging with PET. There were no studies available which were designed in this fashion. As an alternative, this review sought evidence which compared the diagnostic performance of PET and conventional imaging, whereby PET could demonstrate greater clinical utility if it was found to: Show better diagnostic performance; Be more often correct when discordant results are obtained; Accurately upstage or downstage patients, and Influence patient management NCDs. The following study selection criteria were used in the Blue Cross/Blue Shield TEC Assessment: Published or accepted for publication as a full article in a peer-reviewed journal; At least 10 patients; Patient sample homogeneous with respect to type of primary cancer (i.e., studies excluded if there were either patients with various tumor types or if there was a mixture of primary and metastatic lesions); Performed tomographic rather than planar imaging with FDG as the radiotracer, and Correlation of PET findings with data from an appropriate reference standard, for at least some of the patients in the standard. There were fifteen studies that met the selection criterion for melanoma. The most useful three include Rinne et al. (1998, n=100), Holder et al. (1998,