|Home | About | Journals | Submit | Contact Us | Français|
Recent cancer research efforts have focused on better understanding tumor biology to facilitate the development of targeted therapies. Improved outcomes for various cancers, including HER2-positive breast cancer,1 chronic myeloid leukemia,2 and neuroblastoma3 have resulted directly from tumor biology studies followed by targeted drug development. However, progress is limited when sufficient tumor tissue is not available for study. Here we review the medical, scientific, and social benefits offered by autopsy that includes tumor tissue harvesting after the death of cancer patients (Table 1). Although largely overlooked as a source of research data, the autopsy has become increasingly important to the advancement of cancer biology studies and the development of targeted therapies. Our hope is to promote a strong initiative among practicing oncologists to seek autopsy consent for both diagnostic purposes and research tissue acquisition. Optimizing collection of tumor specimens would facilitate the development of novel therapies for many types of cancer.
Tumor tissue obtained from living patients is often suboptimal for research needs. Some tumors, particularly small, clinically benign-appearing lesions, may be excised in settings where they are not preserved in a manner appropriate for research. Even at centers with cancer expertise, tumor specimens taken at the time of initial diagnosis may be extremely small, especially when obtained by needle core biopsy or via a small incision. Advancing technologies in interventional radiology and video-assisted endoscopy have also reduced the size of biopsy specimens, commonly providing only enough tissue to establish the diagnosis and leaving none for special study. Further, these specimens may not include the most biologically aggressive region of the tumor. The fact that tumors can be genetically heterogeneous has been increasingly appreciated. For example, it is now well known that NMYC amplification can be restricted to certain areas of neuroblastoma.4 Intratumor karyotypic heterogeneity has been documented in various cancers, including carcinomas and soft tissue sarcomas.5, 6 When tumors recur, the small size of the biopsy sample (if obtained) again often precludes research use or limits the research-related assays that can be performed.
In contrast, patients often have a very large tumor burden at the time of death. A larger quantity of tumor tissue can be collected after death and preserved by different methods to support a variety of analyses. Examples include collection of fresh tissue in cell culture medium for cytogenetics, generation of cell lines/primary cell cultures, and xenograft tumor models, frozen tissue for molecular genetics studies using tumor-derived DNA and RNA, and paraffin-embedded tissue for routine light microscopy, immunohistochemistry, and fluorescence in situ hybridization to detect tumor-specific translocations. Autopsy also allows collection of a broad range of matched diseased and normal tissues from various organs, and tumor tissue may be collected from both the primary and metastatic sites, including tumor deposits in multiple organs, This tissue can be used to monitor the effects of novel therapies and to correlate histological effects with clinical ones. What may be unique about the autopsy of a patient with widely metastatic cancer is that the evolution of molecular abnormalities at different metastatic sites may be investigated in a way not otherwise possible. Review of pre-mortem imaging studies may be helpful to guide the autopsy procedure for identifying obscure sites of disease. Limited experience with post-mortem magnetic resonance imaging has shown a good correlation with pathologic findings, suggesting that post-mortem imaging may be helpful in rapidly identifying sites of disease that warrant attention during the autopsy.7
Tumor specimens obtained at autopsy might be particularly valuable for identifying both biomarkers of disease and targets for therapeutic intervention, because the genes and proteins expressed at the time of relapse or progression are more likely to reveal factors driving tumor growth.8, 9 Molecular changes observed at the time of treatment failure can provide clues to drug resistance or may be valuable as biomarkers of aggressive disease or “drivers” of tumor progression. Ideally, such biomarkers would provide sufficient sensitivity and specificity to predict treatment success or failure earlier than diagnostic imaging, which detects tumor recurrence only at a relatively late stage. Identification of such changes provides an opportunity to develop molecularly targeted therapies (i.e., therapies designed to target a specific gene or protein) to inhibit tumor progression. The advent of molecularly targeted therapies underlines the need to identify specific molecular defects in high-risk cancers.
A recent innovative application is the “molecular genetic autopsy,” which allows comparisons between the primary tumor and various metastases to detect the frequency and nature of secondary mutations associated with cancer progression, metastasis to specific anatomic sites, or the development of chemotherapy resistance.10–12 These studies have shown that genetic findings in primary tumor specimens may differ from those in metastases; this difference may explain the failure of targeted agents selected on the basis of findings in the diagnostic tumor biopsy.13, 14 Clonal evolution of the tumor may produce mutations that confer drug resistance or other phenotypes fostering tumor progression.15–17 However, the molecular genetic autopsy may also confirm that certain tumor features are not lost during disease progression. In non-small cell lung cancer, for example, an autopsy series demonstrated that epidermal growth factor receptor (EGFR) expression in most metastases was consistent with its expression in the primary tumor, suggesting that loss of EGFR expression is rare during disease progression and that this receptor may be effectively targeted even late in the disease course.18 As molecularly targeted therapy becomes increasingly available, understanding the expression pattern of these targets in patients whose disease has progressed after initial therapy will be critical. The sampling of all sites of disease may also yield other important biologic insights with therapeutic implications. For example, an autopsy series found that lethal metastatic prostate cancer generally has a monoclonal origin despite the presence of multifocal primary tumors.19 When disease progression has occurred after frontline therapy, tumor specimens taken at autopsy may be a unique source of information about markers of aggressive disease and potential therapeutic targets.
Beyond the broader tumor sampling that may provide insight into therapeutics, material from autopsy provides an opportunity to study apparently normal tissue to learn more about the root causes of cancer. As one example, mathematical modeling predicts that sporadic human cancer is likely caused by multiple “hits” during embryonic development. This prediction implies that neonates would display the different genetic aberrations as a “somatic mosaic” pattern in a range of non-cancerous tissues and that cancer would then arise after subsequent accumulation of additional genetic or epigenetic events.20 Direct proof of this concept would largely change how cancer susceptibility is viewed. Biopsy of the wide range of tissue samples needed to investigate this possibility is not currently feasible, but the availability of autopsy material could make it possible.
Studies of rare tumors are particularly limited by the inherent paucity of biological material. As an example, rhabdomyosarcoma is a rare tumor overall, despite being the most common pediatric soft tissue sarcoma. The estimated five-year survival of children with rhabdomyosarcoma in the U.S. improved from 45% to 63% between 1973 and 2002.21 These gains can be attributed largely to clinical trials conducted by the Intergroup Rhabdomyosarcoma Study Group and later the Children's Oncology Group (COG), which have improved systemic therapy, local tumor control, and supportive care.22 In recent years, however, survival has reached a plateau despite the use of novel cytotoxic therapies and refinement of radiotherapy. In fact, for patients with metastatic disease, the 3-year overall survival rate remains only about 35%.23 Similar observations apply to other rare cancers and to subsets of more common cancers, in which the limits of cytotoxic chemotherapy have been reached and targeted therapies have yet to be developed.24–26
A recent COG Soft Tissue Sarcoma Committee biology symposium emphasized the potential contributions of autopsy to the advancement of knowledge about pediatric soft tissue sarcomas. Tissue procurement in COG protocols has traditionally focused on primary tumor material obtained at the time of initial diagnosis and/or disease recurrence. For example, the Fall 2010 inventory report from the Biopathology Center (a biospecimen repository operated under the aegis of the Cooperative Human Tissue Network in Columbus, OH) noted the availability of samples from 1023 cases of pediatric soft tissue sarcoma in which snap-frozen primary tumor was obtained at diagnosis. However, only 58 snap-frozen samples had been obtained from a metastatic site, only fourteen cases included samples of recurrent tumor, and postmortem samples were even scarcer. It is important to raise awareness that postmortem samples can in many ways be of greater value than antemortem material, as wider sampling and larger volumes of material hold the potential of revealing more clinically and biologically relevant information, as noted above.
Postmortem tumor recovery not only provides ample tissue for basic tumor research but also advances the science of medicine by providing data useful to clinicians and to epidemiologists who classify and report disease incidence and causes of death. An unbiased and thorough evaluation of the body and its organs may determine the cause of death, the nature of any medical conditions present, and the effects of any treatment given. These findings may confirm or refute clinical impressions, uncover medical errors, and identify unexpected underlying conditions or complications of treatment. Engaging the clinicians who cared for the patient provides opportunity for interpreting the findings of the autopsy and may assist in continuous performance improvement. Autopsy is an important quality assurance/quality improvement tool in both adult and pediatric hospitals27–31 and an important element of medical education.32, 33 Comparison of clinical diagnoses with autopsy findings may identify medical conditions that were unrecognized before the patient's death, thereby benefiting physicians of all specialties who cared for the patient, including oncologists, surgeons, radiologists, and pathologists.28, 31, 34–36 A recent review of adult and pediatric autopsy series showed that the post-mortem examination continues to identify major diagnostic errors in 8% to 24% of cases; in a substantial subset of these, the errors are considered sufficiently serious to affect patient outcome.34 The rate of unrecognized diagnoses is high even in cancer patients. In a series of 86 autopsies of cancer patients who died in a medical-surgical ICU, 26% revealed major diagnoses that were missed, and about half of these diagnoses had a potential adverse impact on survival and would have changed management.37 Thus, an autopsy may provide valuable learning experiences for individual physicians and improve their future clinical care. Autopsy of oncology patients can usually better define the extent and distribution of cancer, more thoroughly identify and characterize the effects of medications, and identify secondary complications such as infections.38–41
Medical knowledge is also advanced when autopsy findings are reported in the peer-reviewed literature. Case reports of individual autopsies whose findings are sufficiently novel to warrant publication remain common. In fact, new cancer-related syndromes and findings that are clinically relevant to cancer patients are still being reported on the basis of autopsy material.42–45 Data from multiple autopsies that are collated into published series provide an opportunity to study radiographic and clinicopathologic correlations, side effects of new therapies, and tumor heterogeneity.46, 47 These observations can also offer valuable clinical insights.
One limitation of postmortem tumor tissue collection is that autopsy may be performed hours to days after a patient has died, in part due to scheduling constraints. A delay in refrigeration of a corpulent body habitus that retains heat can accelerate autolysis (decomposition). Delay in performing the autopsy may degrade the quality of nucleic acids recovered or alter the integrity and phosphorylation state of proteins.48, 49 The histologic morphology and antigenicity may be compromised if fixation is delayed, but viable tumor cells adapted to a hypoxic environment may be isolated for culture and functional studies in vitro if autopsy is performed expeditiously. Limited data are available about the acceptable time lapse between death and tissue collection for research purposes. A study of autopsy samples of diffuse pontine glioma found that minimally degraded DNA could be isolated in nearly 90% of cases (although RNA recovery was only 70%) when autopsy was performed within 5 hours of death.50 Another study documented a 60% success rate in xenografting primary and metastatic pancreatic tumors recovered by autopsy within 6 hours of death.51 This time range need not contraindicate fresh tissue collection after even more delay; in the experience of one of our authors (CK), primary cell cultures can be established from autopsy specimens collected even 18 hours after death.
Importantly, the results of autopsy may be beneficial to the family of the deceased. The autopsy may help explain why the patient died, improve the family's understanding of the circumstances that led to death, clarify the extent of disease, and provide reassurance that the cancer was incurable. It may grant peace of mind by demonstrating that nothing more could have been done. On occasion, an autopsy may disclose an unrecognized genetic disorder that may have clinical implications for family members.
Families report that improving medical knowledge and facilitating research is a motivating factor for consenting to an autopsy.52 For this reason, patients and their family members may find meaning in authorizing an autopsy that includes collection of tumor tissue. Among families of individuals who underwent autopsy for medico-legal reasons, 96% authorized retention of tissue samples for research,53 suggesting that collection of tumor tissue for research purposes may be readily accepted by families who authorize an autopsy for other reasons. To our knowledge, only one study has examined the willingness of family members to authorize autopsy solely for the purpose of tumor harvesting; in that study, 40% of families were willing to consent to an autopsy for this reason.50
When family members consent to autopsy, they most commonly report being motivated by the desire to understand the cause of death more clearly, to help others, and to assist medical research.54 However, a number of barriers must be overcome to increase the proportion of autopsies of patients who die of cancer (Table 2). Failure to obtain permission for autopsy may reflect attributes of the requesting clinician, the patient's family, or both. There are times when a clinician may not seek permission for autopsy.55 If permission is sought, a clinician who conveys a low level of interest may negatively influence the family's decision.56 A clinician who is well educated about the many benefits of autopsy is likely to convey these benefits more clearly to the family, thereby facilitating agreement. In a study of deaths in a public teaching hospital, an educational intervention for the physicians responsible for patient care and completion of death certificates more than doubled the autopsy rate, from 8% to 17%.55 Of the cases in which no autopsy was done, permission was not sought by the physician in about one third and permission was denied by the family in about two thirds. Physicians reported not seeking autopsy permission because the patient had a “do not resuscitate” order or because they knew the cause of death. A survey of resident physicians suggests that inadequate knowledge about the procedure and unfamiliarity with the process of obtaining informed consent from family members are barriers.57 Some clinicians may hesitate to seek consent for autopsy for fear that a previously unrecognized therapy-associated complication may be discovered. This fear should be countered by the fact that complications must be recognized if they are to be prevented in the future. Further, autopsy findings can be powerful in refuting unjustified medicolegal claims.
The most common reasons that families provide for declining an autopsy are fear of mutilation of the body, desire to respect the wishes of the deceased, and concern about delaying funeral arrangements.58, 59 Other reported barriers to family consent include lack of knowledge about the procedure, lack of rapport with the physician, feeling that the deceased has suffered enough, concern about potential costs, lack of consensus among family members, stress at the time of the death, and religious or cultural practices.59, 60 The increasing use of home hospice care may also make coordination of an autopsy difficult.
Many of the barriers to obtaining permission for autopsy can be overcome. As noted above, simply educating physicians about the potential benefits of autopsy and providing better training in how to seek informed consent may increase the proportion of physicians who request consent for autopsy.61, 62 Explaining the procedure to families in clear, unambiguous language (e.g., “removing the tumor to study it to help find ways to cure similar tumors”) and providing supportive educational materials are key components of the informed consent process. Conveying the potential benefits to the family, the medical team, and to society at large is also important. Key points that should be covered include the fact that the procedure does not disfigure the body, that there is usually no cost to the family, and that autopsy need not delay funeral arrangements. Offering options to restrict the autopsy to limited anatomic sites or small samples, such as multiple core needle biopsies, limits the information that can be gleaned from autopsy but may also facilitate greater willingness to consent to the procedure. For physicians, a good understanding of various cultural and religious perspectives on autopsy may also be helpful in addressing the family's concerns.61, 63–65
Through the connectivity of the internet, patients and their families are increasingly driving the trend toward post-mortem tissue retrieval for research. A striking example is the parent-to-parent plea of one mother after losing her son to metastatic medulloblastoma (see http://www.kidsvcancer.org/tissue-donation/). After arranging for autopsy tissue from her son to be distributed to multiple laboratories for the specific purpose of studying the biology of leptomeningeal metastasis, she created a website to encourage other parents to do the same. While this example is unusual, patients and their families may be a unique source of suggestions for guidelines on how physicians and caregivers should seek permission for post-mortem tumor tissue collection for research purposes.
Addressing systemic barriers will also be necessary to maximize the benefits of autopsy in patients dying of cancer. Communication and collaboration between the treating physician, the pathologist, and, when relevant, tissue banking personnel is important to facilitate a timely and appropriately focused autopsy. Support of hospital autopsy services by institutional leadership is critical, as is funding to support the research effort since autopsy costs are not reimbursed by public or private insurers. For patients who die in the hospital, autopsy costs are generally covered through existing budgets, although times of service may be restricted due to personnel requirements and resources. Patients who die outside the hospital may have access to autopsy but fees for the service and transportation to the hospital facility may be incurred for which a defined funding mechanism may not exist. A robust autopsy program that includes collection of tissue for research purposes could be supported by endowments, philanthropic organizations, advocacy groups, and federal and non-federal research grants. Although collection of biologic specimens for research purposes occurs at virtually all academic medical centers, research that includes the vulnerable population of dying patients requires special ethical consideration. Various institutional mechanisms exist to consider these ethical issues and oversee the protection of research subjects.66, 67 Of particular note, deceased patients are not afforded the same legal protection as living ones; for example, the Health Information Protection and Portability Act (HIPPA) laws do not cover deceased patients in the same manner as living ones (see Federal Register http://www.hhs.gov/ocr/privacy/hipaa/administrative/privacyrule/1999nprm.pdf)
The Institute of Medicine report on improving care at the end of life recommended an integrated, multidisciplinary approach with a focus on identifying and addressing the dying patient's preferences and goals.68 In this context, discussion about autopsy with the patient and/or family, as appropriate, may be be approached before death occurs and when the patient can be a full participant in the discussion. Some patients may be comforted by the knowledge that an autopsy may address unanswered questions for their loved ones and, by permitting them to donate tumor tissue for research, may allow them to make an important contribution to the care of future patients with the same condition. Family members, particularly the individual designated as the next-of-kin, may also benefit by having an opportunity to discuss the potential benefits of a post-mortem examination with the patient and develop a more informed understanding of the patient's wishes. The informed consent discussion should review the nature, goals, and benefits of the autopsy and should offer patients and families the opportunity to ask open-ended questions. It should also clarify whether the autopsy will include collection of research specimens and the extent of the procedure (complete, body-only [excluding the brain], regional [chest only, abdomen only, etc.], or organ-focused [lung only, brain only, etc.]). When a decision to perform an autopsy is made before the time of death, informed consent documents for the procedure may be signed by the patient, by the parent or legal guardian of a minor child, or by the authorized health care representative of an adult who lacks decisional capacity. After death occurs, informed consent for autopsy must be obtained from the next-of-kin.
Optimally, informed consent discussion regarding autopsy should take place within the larger context of comprehensive end-of-life care in which there is an ongoing dialogue between the patient, family, and healthcare team about the goals of care. The autopsy should not be viewed as the final interaction with the patient and family, but rather an opportunity for exchange of valuable information and a first step in the bereavement process. Surviving family members may take great comfort when the autopsy confirms their understanding of their loved one's condition or answers important questions, such as whether more could have been done for the patient. When the autopsy includes donation of tumor tissue for research, bereaved family members may also experience the healing knowledge that their loved one left behind a valuable legacy for future patients. It may be helpful to schedule a time, weeks to months after the death, for the family to return for a discussion of the autopsy findings and any other lingering questions.
Autopsy is an underutilized procedure with significant benefits for bereaved family members, for the clinicians who cared for the decedent, and for research aimed at improving cancer survival. Autopsy provides a precious opportunity to learn about the disease process and the effects of treatment. Collection of tumor tissue at autopsy can provide valuable tumor specimens to advance the development of targeted therapies for many types of cancer. Emerging studies are already demonstrating the feasibility of tumor tissue harvesting at the time of autopsy for the purpose of establishing live cell cultures for basic and translational research. We propose that all patients who die of cancer should undergo autopsy when possible. Oncologists should be educated about the multiple potential benefits of autopsy for clinicians, families, and future cancer patients, so that a strong and informed awareness of its value will be conveyed when consent is sought.
The authors gratefully acknowledge ongoing technical and administrative assistance by Natalie Beeler, Julie Moore, Laura Monovich, and others at the Biopathology Center who support the COG Soft Tissue Sarcoma Biology and Tissue Banking Study. We also thank Sharon Naron, MPA, ELS, for editing the manuscript.
Research funding: This research was supported in part by Cancer Center Support (CORE) grant P30 CA 21765 from the National Cancer Institute (SLS), by the American Lebanese Syrian Associated Charities (ALSAC) (SLS), and by grant 3R01CA133229-04S1 from the National Cancer Institute (H.R. 1553: Caroline Pryce Walker Conquer Childhood Cancer Act of 2008) (CK).
Financial disclosures: None