2
Tumor-Based Screening for Lynch Syndrome

In presenting the first clinical scenario, Marc Williams of Intermountain Healthcare’s Clinical Genetics Institute described how tumor screening and confirmatory genetic testing for mismatch repair gene mutations are being used to identify Lynch syndrome in individuals who are newly diagnosed with colorectal cancer.1 The intent is that family members of those with Lynch syndrome would then be screened, so that those identified as also having Lynch syndrome could take preventative measures in hopes of reducing their morbidity and mortality from colorectal cancer.

COLORECTAL CANCERAND LYNCH SYNDROME SCREENING

Colorectal cancer is the second leading cause of cancer death in the United States, accounting for about 50,000 deaths per year and affecting almost 150,000 people each year. One in every 19 people will be diagnosed with colorectal cancer in their lifetimes, and one person dies from the disease every nine minutes. About 5 to 10 percent of colorectal cancer cases are familial, and it is estimated that about 1 to 5 percent of cases are due to mutations in highly penetrant single genes. A subset of these mutations cause Lynch syndrome, sometimes referred to as hereditary non-polyposis colorectal cancer (HNPCC) because individuals with the syndrome tend to have a relatively small numbers of polyps. Lynch syndrome may account for as much as 2 to 4 percent of all colorectal cancers, and it also increases an

1

The complete scenario provided to workshop participants is available in Appendix C.



The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 5
2 Tumor-Based Screening for Lynch Syndrome In presenting the first clinical scenario, Marc Williams of Intermoun- tain Healthcare’s Clinical Genetics Institute described how tumor screening and confirmatory genetic testing for mismatch repair gene mutations are being used to identify Lynch syndrome in individuals who are newly diag- nosed with colorectal cancer.1 The intent is that family members of those with Lynch syndrome would then be screened, so that those identified as also having Lynch syndrome could take preventative measures in hopes of reducing their morbidity and mortality from colorectal cancer. COLORECTAL CANCER AND LyNCH SyNDROME SCREENING Colorectal cancer is the second leading cause of cancer death in the United States, accounting for about 50,000 deaths per year and affecting almost 150,000 people each year. One in every 19 people will be diagnosed with colorectal cancer in their lifetimes, and one person dies from the dis- ease every nine minutes. About 5 to 10 percent of colorectal cancer cases are familial, and it is estimated that about 1 to 5 percent of cases are due to mutations in highly penetrant single genes. A subset of these mutations cause Lynch syndrome, sometimes referred to as hereditary non-polyposis colorectal cancer (HNPCC) because individuals with the syndrome tend to have a relatively small numbers of polyps. Lynch syndrome may account for as much as 2 to 4 percent of all colorectal cancers, and it also increases an 1 The complete scenario provided to workshop participants is available in Appendix C. 

OCR for page 5
 ThE VALuE Of GENETIC AND GENOMIC TEChNOLOGIES individual’s risk of cancers of the stomach, small intestine, liver, gallblad- der ducts, upper urinary tract, brain, skin, and prostate as well as, among women, endometrial and ovarian cancer. Diagnosing Lynch Syndrome Lynch syndrome is associated with mutations in four major mismatch repair (MMR) genes (MLh, MSh, MSh, and PMS), and individuals who have one of these mutations have a 20 to 65 percent lifetime risk of colorectal cancer, compared with a lifetime risk among the general popula- tion of approximately 5 percent. Because inheritance of these mutations is autosomal dominant, close biological relatives are also at high risk. Classic signs that suggest an individual may have Lynch syndrome include: a family history of colorectal and associated cancers, specific pathologic characteristics of the tumors, a young age of onset, and syn- chronous or metachronous colorectal cancer. However, these signs, either alone or in combination, are not sufficiently sensitive to identify more than about 50 percent of patients with Lynch syndrome. A new strategy has been proposed to screen the tumors of patients presenting with colorectal cancer, using techniques designed specifically to identify Lynch syndrome (e.g., immunohistochemical (IHC) staining for the protein products of the four MMR genes or an assessment of the tumor for microsatellite instabil- ity (MSI)). If the tumor screening is positive, a mutation analysis is done to confirm a diagnosis of Lynch syndrome. Who Is Lynch Syndrome Screening for? Screening for Lynch syndrome in a patient with colorectal cancer combines initial testing of tumors with mutational analysis to definitively diagnose the presence of MMR gene mutations.2 A diagnosis of Lynch syndrome may have some effect on the patient’s treatment as well as on monitoring for colorectal cancer recurrence and for other cancers associated with Lynch syndrome (e.g., increasing the frequency of colonoscopies or considering a prophylactic surgery, such as a hysterectomy and salpingo- oophorectomy, which are used to reduce the risks of endometrial cancer and ovarian cancer in women). However, the real impetus for testing is the potential effect of the diagnosis on close relatives. First-degree relatives have a 50 percent chance of having inherited the MMR gene mutation, and, on 2 Due to the cost of MMR gene sequencing, preliminary tests, including microsatellite in- stability (MSI) testing and immunohistochemistry (IHC), are often conducted first to identify those who should be offered DNA sequencing. Further details regarding the screening process for Lynch syndrome are provided in the case scenario in Appendix C.

OCR for page 5
 TuMOR-BASED SCREENING fOR LYNCh SYNDROME average, there are three affected family members for each proband (i.e., for each of the first subjects in a study). In the United States, there are 142,000 newly diagnosed cases of colorectal cancer annually and, assuming a 3 percent prevalence rate, about 4,250 of those individuals have Lynch syndrome. This means that around 8,500 to 12,750 relatives would also be carrying one of these mutations and also have Lynch syndrome. Based on evidence reviews, the Evaluation of Genomic Applications in Practice and Prevention (EGAPP) Working Group has reported that the overall analytic validity of the preliminary and diagnostic genetics tests for Lynch syndrome is high, there is adequate evidence of clinical validity for the preliminary and diagnostic tests, and there is adequate evidence to sup- port the use of genetic testing strategies to reduce morbidity and mortality in relatives with Lynch syndrome (i.e., high rates of relatives consent to testing and adhere to recommended cancer surveillance recommendations, and there are limited harms compared to benefits) (EGAPP Working Group, 2009). Testing of family members is much less expensive than primary diag- nostic testing, since only the specific familial mutation is tested for (i.e., full gene sequencing is not necessary). For those related individuals who are identified as having Lynch syndrome, endometrial screening in female carriers or prophylactic surgery, or both, may also be appropriate, as well as more frequent colonoscopies, starting at an earlier age. There are also effective interventions if precancerous polyps are detected. Because of these facts, EGAPP has recommended offering screening followed by confirmatory genetic testing for Lynch syndrome in individuals newly diag- nosed with colorectal cancer to reduce morbidity and mortality in relatives (EGAPP Working Group, 2009). Costs and Benefits of Implementation Using published data, Mvundura and colleagues conducted a cost- effectiveness analysis and found that, from a U.S. healthcare system per- spective, tumor-based screening for Lynch syndrome is cost effective and that an IHC-first approach is superior (Mvundura et al., 2010). Offering the perspective of an integrated healthcare delivery system, Williams explained that Intermountain Healthcare has modeled various screening scenarios and has also concluded that a strategy using IHC as the preliminary test appeared to be the most efficient. Following a decision pathway based on screening results makes it possible to sequence only one or two specific genes, as opposed to sequencing all four. Overall, this strategy has improved the quality and consistency of care at Intermountain Healthcare. Because the screening is done on the tumor, the hospital pays for the cost of screen-

OCR for page 5
 ThE VALuE Of GENETIC AND GENOMIC TEChNOLOGIES ing out of the diagnosis-related group (DRG) reimbursement it receives, which reduces profit margins. However, there is the potential to increase revenue for providers, hospitals, and outpatient surgery centers, Williams said, because of the increased periodicity of screening and also because of the potential for family members to enter the system as new patients. There will be small increases in costs for health plans because they would need to cover confirmatory mutation testing as well, but there is also the potential for savings by avoiding significant costs as a result of prevention of cancer, particularly in relatives that may also be covered. From a patient perspective, screening provides information on sur- veillance, it may affect treatment, and it may offer a better prognosis. Privacy issues remain a concern, Williams said, but they may be somewhat mitigated as patients already have expressed colorectal cancer. For family members, identification of high-risk individuals provides opportunities for primary prevention and, in roughly one-half of at risk relatives, reassurance for family members who do not carry the Lynch syndrome mutation. In implementing the Lynch syndrome screening system, Intermoun- tain Healthcare provides patients with an information sheet, and it offers full counseling and consent for confirmatory mutation testing of both the patient and family members. However, it was decided after significant ethi- cal consultation that informed consent was not needed for the tumor-based screening because it was not considered genetic testing but rather screening for susceptibility. PANEL REACTION Patient Advocate Perspective From the perspective of the Colon Cancer Alliance, the oldest and larg- est national patient advocacy organization in the United States dedicated to colorectal cancer, there is great value in Lynch syndrome screening. The prospect that screening could help prevent some of the 50,000 colorectal cancer deaths that occur each year and, more importantly, prevent some people from ever experiencing this cancer, is very exciting, said Andrew Spiegel of the Alliance. Spiegel highlighted several issues for further consideration. While Williams noted in his introduction that privacy is generally less of an issue for Lynch syndrome testing because close associates are already aware that the individual has colorectal cancer, Spiegel countered that patients with Lynch syndrome are also at a higher risk for ovarian, endometrial, and other cancers. An insurance company or potential employer who learns that a person carries the mutations that can cause Lynch syndrome now knows that the person is susceptible not only to colon cancer but also to other

OCR for page 5
 TuMOR-BASED SCREENING fOR LYNCh SYNDROME cancers as well. Privacy is a major issue not only for the primary individual with colon cancer, Spiegel stressed, but also for the family members who are also identified as having Lynch syndrome. What assurances are there, he asked, that the test results will remain private? Spiegel also raised concerns about informed consent, noting that tumor screening, while perhaps not a “genetic test,” does suggest the presence of a specific genetic mutation, and the result is the same from the patient perspective—the patient either does or does not have Lynch syndrome. A number of questions surrounding such screening must be addressed, Spiegel said: Should informed consent be required for tumor screening to determine whether or not a patient is at risk for having Lynch syndrome? If the per- son does have Lynch syndrome, what kind of counseling will be provided, and should that counseling be mandatory? What is the responsibility to inform relatives of the proband, and what if the relatives do not want to know whether or not they carry Lynch syndrome mutations? Who bears the responsibility to notify family members that a parent or sibling has Lynch syndrome? If it becomes the responsibility of the patient, what happens if the patient fails to tell family members for certain personal reasons? Should a doctor refuse to treat a patient who refused to notify family members? Should it instead be the responsibility of the doctor to tell family members that they may carry the Lynch syndrome gene and that they should be tested? Or perhaps the State Department of Health? Under what authority can notification be forced? Spiegel also noted that criteria are needed for how to inform and for which family members will be told. Would such cri- teria be uniform across the country? Policies and procedures in a major city hospital may be very different from that in a rural setting. What follow-up will there be with family members who may be at risk for Lynch syndrome? Whose responsibility will it be to ensure those family members are screened and to make screening readily available and affordable? Diagnostic Pathology Perspective Mark Boguski of the Center for Biomedical Informatics at Harvard Medical School (HMS) and the pathology department at Beth Israel Dea- coness Medical Center, a teaching hospital of HMS, offered the perspec- tive of diagnostic pathologists. The problem with any single genetic test, Boguski said, is there are 24,000 human genes and single tests are not scal- able. Laboratories already perform many hundreds of tests on tens of thou- sands of specimens daily. What will be the operational role of pathology, he asked, and what are the economic and efficiency implications of doing single gene tests in the age of whole genome sequencing and personalized medicine? Boguski suggested that in the not too distant future a patient’s whole genome will be part of his or her existing electronic medical record.

OCR for page 5
0 ThE VALuE Of GENETIC AND GENOMIC TEChNOLOGIES Karyotyping or MSI testing, for example, will no longer be done and will instead have been replaced by whole genome or whole transcriptome (transcribed RNA) sequencing, or both, and carrying out different genetic screens will be a matter of applying different software filters. To prepare for this data-rich future, Boguski has begun a training program for HMS clinical pathology house staff to bring them up to date on genomics and personalized medicine, specifically the state of the field and the technology being used. Boguski concluded by noting that while the discussion at hand is about Lynch syndrome, the larger issue to be addressed is the ability to carry out individual genetic tests for 24,000 genes, the movement toward genome sequencing, and the software and decision support systems that will be the primary diagnostic modality in the near future. Insurance Provider Perspective Roy Gandolfi, a practicing internist and associate medical director of Select Health, offered perspective on the role of insurers and coverage deci- sions. As the insurance arm of Intermountain Healthcare, Select Health is a bit different from most other insurance plans. It is not for profit, it does not offer coverage for Medicare, and the median age of the 500,000 com- mercial lives it covers is 27 years old. The system is integrated: Select Health is owned by the parent company, Intermountain Healthcare, which also owns many hospitals and employs a significant number of the physicians who work for the system. So, while Select Health is an independent business entity, it is integrated within the larger health care system. Intermountain has very strong fetal-maternal medicine programs at the university as well as at Intermountain hospitals. While these are invaluable resources, they also mean that Select Health is faced with more genetic decisions than other health plans with the same population. As a health plan, Select Health looks not only at guidelines but at how those guidelines were derived and their level of evidence. Medical technology is carefully reviewed, with a focus on looking at what the specific genetic test is try- ing to achieve, conducting evidence-based literature reviews, and seeking local input from providers regarding coverage recommendations. Alterna- tive technologies are considered, and economics are assessed relative to all stakeholders: the hospitals, the plan as a payer, the plan members, and their employers, who pay for the insurance. There are many genetic tests that have been proposed both commer- cially and academically, and, unfortunately, Select Health does not have the resources at the present time to evaluate all tests. How then, can the plan make coverage decisions if it is not able to conduct a medical technology review? Clinics are a valuable resource for gaining perspectives on clinical

OCR for page 5
 TuMOR-BASED SCREENING fOR LYNCh SYNDROME utility and economics, Gandolfi said, as are experts at the Clinical Genetics Institute as well as local providers. Select Health feels a responsibility to the employers who are paying the insurer to manage the care being provided, both for good health outcomes and for fiscal responsibility. For Lynch syndrome, for example, while the focus is on identifying affected relatives, it is also very important from the health plan perspective to look at the proband, because the treatment course for that patient will change if he or she has the genetic disorder. As such, Select Health feels it is important that coverage be applied to the Lynch syndrome test. Private Practice Perspective Dennis Salisbury, a family physician at Rocky Mountain Clinic in southwest Montana, said that he sees one or two colon cancer patients a year in his practice, out of several thousand patients total. This makes the process of deciding who gets what information—and making sure they have enough information without overwhelming them—rather challenging. He has come to rely on information gained elsewhere. Of key importance in the decision-making process for recommending a test is the clarity of the association between a condition and an outcome along with the validity and predictive value of the testing. Incidence and prevalence of the condi- tion are both important, as is the severity of the impact of the condition and the significance of the potential benefit of the test. Costs must also be considered, relative to whether that patient is insured. Costs that could be avoided if testing is done are also a factor. Patient benefit, family benefit, and, to some degree, public benefit are all part of his decision-making pro- cess, Salisbury said. The biggest issue is educating the patient to facilitate his or her autonomous decision. Since many in private practice do not have either the time or the exper- tise to assess the validity and predictive value of the testing, they rely on other sources, such as EGAPP recommendations. In this case, Salisbury said, he is firmly in favor of screening for Lynch syndrome in a patient with newly diagnosed colon cancer. The question then becomes how to present information about the test to patients so that they can make an informed decision. It is important to have good information about what the test can mean for a patient’s future (e.g., how it can improve life, extend life, make life more complicated or more difficult, whether there will be complica- tions associated with the testing, and if testing can help family members). Salisbury also noted that how a provider presents the information can sway a patient’s decision. Salisbury said he would be happy to be the one to tell family members of the test results, but he questioned what should be done if the proband

OCR for page 5
 ThE VALuE Of GENETIC AND GENOMIC TEChNOLOGIES does not want a family member informed. If the family member is also a patient in the practice, this becomes a very difficult scenario, legally and ethically. Public Health Perspective Providing a government perspective, Don Lyman, chief of the Chronic Disease and Injury Control Division of the California Department of Public Health, said that as far as most official government agencies are concerned, there has not been any conclusion as to whether screening for Lynch syn- drome is worthwhile or not. Historically, profound issues tend to be addressed most productively at the local level first. The federal government may see the problem, but it often waits for several “strategic loci” (localities or states) to begin to address the issue locally. Once several jurisdictions have been through the process, the federal government looks for models that can then be applied nationally. The great advances in sanitation in the 1800s, for example, were first employed in the cities of Boston, Charleston, and Philadelphia. Envi- ronmental air quality got its start in California. Once several jurisdictions found a way to bring these problems under control, the federal govern- ment recognized the commonalities in the various successful approaches, passed national legislation, set policy, and provided national funding. In this regard, the activities occurring now at state, local, and private levels regarding genetic testing are very attractive, Lyman said. The key questions concerning genetic testing are whether the ends justify the means and whether the means justify the ends. In a “utilitarian” approach, the end point is set, and then one does what can be justified to get to that endpoint. The whole genome project is a utilitarian approach, he said, and the question is how useful is it? The new paradigm for physi- cians in the 21st century is not to extend the length of life but rather to improve the quality of life. The genetic methodology under discussion is directly applicable to that goal, Lyman said. But from the point of view of a public official, it looks more like a solution in search of a problem; there is a process, it is very attractive, and a lot of money has been spent develop- ing it. Now how can it be applied? What a public official wants to see is an application that brings measurable results for quality of life on a popula- tion basis and not for just a few people here and there. Drawing lessons from other public health initiatives that have measurably reduced illness and death rates, such as AIDS, drunk driving prevention, and the tobacco control program in California, could help identify useful approaches to the application of genomic science, Lyman said. Cost savings is a tricky issue, Lyman said, and it is one that he does not take into account. There are cases of measurably productive, spectacular

OCR for page 5
 TuMOR-BASED SCREENING fOR LYNCh SYNDROME interventions that have brought no cost savings at all. There is a discon- nect between what the medical professional community does and what the insurance industry does. In the early 1950s for example, penicillin, the polio vaccine, and sterile surgical techniques were introduced; there were measurable changes in morbidity and mortality patterns; and still the cost of the industry went up. It is the government’s responsibility to take what looks credible and move with it. As a public health official, Lyman said, there is a point in the academic exercise when one has sufficient evidence to be able to reach a decision. But no one knows exactly where that point is. It takes meetings such as IOM workshops, published literature, and expert groups to help determine when enough is enough and when a disease control application can be moved along in order to induce measurable improvements in qual- ity of life. OPEN DISCUSSION Privacy, Informed Consent, and Information Sharing Privacy-related issues received a great deal of attention during the open discussion. Williams pointed out that while the Genetic Information Nondiscrimination Act of 2008 (GINA) prohibits discrimination based on genetic information, it does not prevent discrimination based on mani- fest disease; that is, predispositional testing and the information obtained thereof are protected, but people who already express a disease are not pro- tected under GINA for the purposes of insurance and employment. Spiegel added that a patient’s insurance company knows if that patient has colon cancer because the insurance company is paying the medical bills, and the patient’s employer may also know because, for instance, the patient needed time off from work for surgery. But the employer may not know that the patient has Lynch syndrome, which puts the patient at a higher risk of developing other cancers such as endometrial or ovarian cancer. Boguski pointed out that companies like Microsoft and Google Health are not governed by HIPAA. When someone types “Lynch syndrome” into the Google search engine, both Google and the Internet service provider know where that search came from. Boguski mentioned an anecdote about Google being able to predict H1N1 flu trends ahead of the CDC, just by tracking people’s search patterns. It is naïve, he said, to think that in such a world privacy can still be regulated in the traditional way. Lyman mentioned California’s cancer registry, one of the largest in the world. That database is primarily for research purposes, and patients are identified for studies or interventions based on the information in the cancer registry. There are strict protocols which have addressed most of the

OCR for page 5
4 ThE VALuE Of GENETIC AND GENOMIC TEChNOLOGIES privacy issues, but some still have concerns. The Veterans Administration, for example, has reluctantly agreed to report its cancer cases to the state cancer registries in California and elsewhere, but it will not allow names to be entered, which results in a large proportion of reported cancer cases that have been effectively removed from potential study. Application of Genomic Testing Data Gathering for Decision Making A participant asked Williams for more information about the sources of the data and the resources used for Intermountain’s modeling of various Lynch syndrome testing scenarios. Williams responded that the data were primarily from published studies but that Intermountain also contacted authors and asked if there were any additional data (i.e., “gray literature”). One university group had accrued a large set of IHC, MSI, and mutation testing data that they were in the process of preparing for publication and that they were willing to share with Intermountain Healthcare under an agreement. However, even this type of arrangement raises the larger issue of where to find resources to support the generation and analysis of this evidence. In this case, the IHC, MSI, and gene testing were done by a uni- versity under grant funding, and Intermountain Healthcare used available resources to support staff time to analyze the data and populate the models. However, many times those discretionary funds are not available. An important question, Boguski said, is who is going to pay for the development and analysis of testing in the future. Would a genetic test be a commercially viable diagnostic developed by a biotechnology or diagnostics company? Would there be enough evidence for such a test—and enough of a market for it—that it would be developed and reimbursable? Beyond the basic research funding environment, how will these tests be developed? There is a provision in the Food and Drug Administration Amendment Acts (FDAAA) of 2007 that is designed to trigger the collection of data from Medicare, Medicaid, and a number of private payers and ultimately to lead to the accrual of data about drug use in 145 million people in the United States, Williams said. If that effort is successful, it will provide numerous opportunities for analysis. Perhaps a similar approach could be applied to data collection for the results of genomic testing and screening programs, Williams said. Another approach to data collection is the “coverage with evidence devel- opment” method that the Centers for Medicare and Medicaid Services (CMS) has put forward. As an insurer, Gandolfi said he must consider whether to pay for research and for the statistical analysis of that research from an unbiased source. The Blue Cross and Blue Shield Association established a

OCR for page 5
 TuMOR-BASED SCREENING fOR LYNCh SYNDROME Technology Evaluation Center (TEC), and the independent Blue Cross and Blue Shield plans pay TEC to review and assess medical technologies for them. Select Health pays for a subscription to Hayes, an independent external technology assessment firm, because it does not have the necessary internal resources. Insurers need good-quality evidence in order to make coverage decisions. However, many of the studies that are available are underpowered, having not enrolled enough participants to support sound statistical analysis. Instead of funding such inefficient studies, Gandolfi said, resources should be directed toward studies that will produce useful evidence. Burke observed that Lynch syndrome testing is an example of an emerg- ing paradigm that falls somewhere between the individual patient care model of medicine and a public health model. The major benefit of tumor testing in one individual appears when other family members are tested and measures are taken to prevent cancers in those who are positively identi- fied as having Lynch syndrome. There are no data yet regarding whether family members will, in fact, appear for testing when notified or what the outcomes of the screening will be. While screening is intuitively a good idea, evidence is needed, including how to effectively reach family members. If the evidence were to support the better health outcomes envisioned here, would this new model of healthcare be a good idea, Burke asked? It would be a useful model for Lynch syndrome, Boguski said, however it is difficult to generalize it to other diseases. To achieve economies of scale, different approaches will be needed, he said. Beyond Lynch syndrome, it is not clear what other genetic conditions this approach might be applicable to and what kinds of technologies, processes, and payment systems will be needed to address those conditions. An alternative approach is needed. The fragmentation of the healthcare system adds to the difficulty of making the economic cost case, Salisbury said, since family members are often covered under different health plans. Even so, the case could be made—and there are articles that support—the use of sequencing to test for the presence of such conditions as Lynch syndrome. Lyman reminded the participants that the notion of diagnosing one per- son and then tracking down family members for treatment or prevention is not new. It has been done for many years, often with limited success. Fifty years ago, for example, it was known that there were risks associated with high blood pressure and that high blood pressure runs in families. Efforts to do blood pressure testing in family members failed, he said, but at least every doctor’s office now has a blood pressure cuff. Nutritional counseling for patients with diabetes and their families is another example. In part, the doctor–patient, one-on-one medical model is standing in the way of reaching out beyond the individual patient. A participant asked whether there was any reason, other than practi- cality, to focus Lynch syndrome screening exclusively on those with newly

OCR for page 5
 ThE VALuE Of GENETIC AND GENOMIC TEChNOLOGIES diagnosed colon cancer? Is there any reason, for example, not to look back at cases of colon cancer that were diagnosed in the recent past? Williams responded that Intermountain has considered doing that and would have the capability of pulling those cases. Within its system, using individual patient data and publicly available genealogical data, the company also has the ability to construct linkages and identify families that would be con- sidered at higher risk. Before expanding testing, the ultimate question that must be asked is, If people are presented with this information, will they act on it? Williams raised the issue of opportunity cost: If resources are invested in this approach and then it turns out it does not work, could that money have been better spent on some other program that would have resulted in better outcomes? What Intermountain Health is trying to do, he said, is to put out a number of different testable hypotheses. If it can be determined relatively rapidly that something is not going to work, then resources can be dedicated to something else. In this regard, Williams said, Intermountain Healthcare made a deci- sion, in conjunction with its pathologists, to not conduct Lynch syndrome screening tests in house. While it was agreed that conducting the testing in house would bring money into the system, it was also determined that Intermountain did not have as high a level of expertise as others, and send- ing the tests out would provide better quality of care. The reality is that performance will vary from laboratory to laboratory, and that can influence where resources should be invested. Salisbury said that the clarity of the gene–disease association and the validity and predictive value of the test are significant issues that affect his decisions and how he counsels patients. Gandolfi noted that internists sim- ply do not have the resources or the time to keep up on certain niches of medicine and need to rely on other resources that help them make decisions, including genetic counselors, geneticists, and oncologists. Clinical utility A participant from the Office of the Air Force Surgeon General cited an article in which the authors assessed a cohort of about 1,500 colorectal can- cer cases and identified 153 that were positive for Lynch syndrome. From all the examined cases, only one family member was ever referred based on genetic family history (Hampel et al., 2008). The participant noted that the same thing occurs at the Department of Defense: Families at risk are not identified based on family history. The question is, Does the screening test have utility beyond family history, and can it help fill the gap? Williams concurred, noting that in the vast majority of cases, family history is not applied well, and actually, in practice, sensitivity is going to be much lower than would be predicted.

OCR for page 5
 TuMOR-BASED SCREENING fOR LYNCh SYNDROME Boguski suggested a hypothetical situation in which a large entity (e.g. Kaiser, the Department of Public Health, the Air Force) would sequence everyone it had access to and use that information to estimate the poten- tial genetic disease burden in that population for disease that might not express itself for 20 or 30 years. He wondered if that approach would not be more cost-effective over the long term, compared to individual, specialized screening tests based on criteria that vary from one disease to another. A participant responded that what Boguski was suggesting was basically high-throughput next-generation sequencing. The error rate and reproducibility of such an approach is not yet known, the participant said. What is known is that there are neurological and other diseases that are not adequately diagnosed by these sequencing technologies and which may represent an important, complex disease burden. The participant cautioned that before potentially costly genome sequencing is broadly considered, it will be critical to make sure that the type of information obtained is vali- dated. “Sequencing is highly overrated,” he said. Boguski pointed out that once the sequencing approach has been validated and is reproducible, the economics will be clear. If, for example, a karyotype and fluorescence in situ hybridization (FISH) analysis costs $800 and a whole genome sequence costs $1,000, it would make economic sense to pursue an automated system that can provide readouts on the genetic burden for numerous diseases, rather than conducting one specific diagnostic test. Williams agreed that the advances in sequencing will transform the future, but those in practice can- not just wait for change. They need to have something they can do today.

OCR for page 5