Autologous Cellular Immunotherapy Treatment of Metastatic Prostate Cancer

CMS NCA document | source_status=Closed | review_type=New | public_comment_open=False | document_id=CAG-00422N

The Centers for Medicare and Medicaid Services (CMS) has determined that the evidence is adequate to conclude that the use of autologous cellular immunotherapy treatment - sipuleucel-T; PROVENGE® improves health outcomes for Medicare beneficiaries with asymptomatic or minimally symptomatic metastatic castrate-resistant (hormone refractory) prostate cancer, and thus is reasonable and necessary for that indication under 1862(a)(1)(A) of the Social Security Act (the Act).

Decision Memo: To: Administrative File CAG-00422N From: Louis B. Jacques, MD Director, Coverage and Analysis Group Tamara Syrek Jensen, JD Deputy Director, Coverage and Analysis Group James Rollins, MD, MSHA, PhD. Director, Division of Items and Devices Leslye K. Fitterman, Ph.D. Lead Analyst and Epidemiologist, Division of Items and Devices Lori Paserchia, MD Medical Officer, Division of Medical and Surgical Services Subject: Internal Request for Autologous Cellular Immunotherapy Treatment of Metastatic Prostate Cancer Date: June 30, 2011 I. Final Decision The Centers for Medicare and Medicaid Services (CMS) has determined that the evidence is adequate to conclude that the use of autologous cellular immunotherapy treatment - sipuleucel-T; PROVENGE® improves health outcomes for Medicare beneficiaries with asymptomatic or minimally symptomatic metastatic castrate-resistant (hormone refractory) prostate cancer, and thus is reasonable and necessary for that indication under 1862(a)(1)(A) of the Social Security Act (the Act). II. Background Information Prostate cancer Prostate cancer is the most common non-cutaneous cancer in men in the United States (Jemal, 2009). In 2009 an estimated 192,280 new cases of prostate cancer were diagnosed and an estimated 27,360 deaths were reported. The National Cancer Institute (NCI) states that prostate cancer is predominantly a cancer of older men; the median age at diagnosis is 72 years. "More than 2 million men in the U.S. who have been diagnosed with prostate cancer at some point are still alive today" (Mark, 2010). The TNM classification is used to define the extent of the cancer where "T" stands for primary tumor; "N" stands for regional lymph nodes; and "M" stands for distant metastasis. The extent of the tumor is generally determined using a number of techniques including digital rectal examination, a surgical procedure/biopsy and imaging studies such as CT, ultrasound, radionuclide bone scan and perhaps magnetic resonance imaging (MRI). The location of the primary tumor can range from confined within the prostate gland to extension beyond the wall of the prostate gland to adjacent structures to spread throughout the body. The extent of regional nodal involvement as well as distant metastasis is stated as present or absent. To further characterize the tumor, the Gleason score, determined upon microscopic examination of a tissue sample, is used to report the histopathologic grade of the tumor. Based on the TNM designations as well as Gleason score, one of four stages of prostate cancer is declared. The TNM-based stage is used in clinical practice to determine the prognosis. The NCI (NCI Prostate cancer treatment [PDQ®]) notes that the "extent of the tumor is related to survival for patients with prostate cancer. When the cancer is confined to the prostate gland, median survival in excess of five years can be anticipated. Patients with locally advanced [i.e., the cancer is no longer confined to the prostate gland but has not spread to distant parts of the body] cancer are not usually curable, and a substantial fraction will eventually die of the tumor, though median survival may be as long as five years. If prostate cancer has spread to distant organs, current therapy will not cure it. Median survival is usually one to three years, and most such patients will die of prostate cancer. Even in this group of patients, however, indolent clinical courses lasting for many years may be observed." Patient’s age, the serum PSA level and comorbidities can also impact the prognosis. The TNM-based stage as well as the serum prostate specific antigen (PSA) level are also used in clinical practice to guide treatment and estimate the risk of recurrence after treatment. Hence, treatment recommendations will vary considerably; potential therapies include surgery, anticancer chemotherapy, radiation therapy and/or androgen deprivation therapy. Active surveillance (a.k.a., watchful waiting) may also play a role depending on the clinical circumstances and patient preference. As noted by Mark, et al., "After initial treatment, patients are monitored for recurrence by measuring PSA levels on a regular basis and imaging tests, if a distant metastasis is clinically suspected. If cancer recurs, and imaging workup indicates the presence or high suspicion of metastasis, androgen-deprivation therapy (ADT) is the standard therapy. ADT in the form of various medications or bilateral orchiectomy are equally effective. Effective ADT will produce a decrease in serum PSA levels, pain relief, and regression of soft tissue metastases. However, ADT does not permanently suppress the progression of cancer, and eventually most patients will experience a rise in PSA levels, followed by development and/or progression of metastases. Such a state of advanced cancer is called castrate-resistant, metastatic prostate cancer." This clinical state is also referred to as castration-resistant, metastatic prostate cancer. Once the patient has castration-resistant, metastatic prostate cancer, further treatment consists of some combination of anticancer chemotherapy with docetaxel and steroids, mitoxantrone and steroids, secondary ADT, palliative radiotherapy or radionucleide therapy and bisphosphonates (NCCN, 2011). Additional subsequent therapies according to the 2011 NCCN guideline include best supportive care, enrollment in a clinical trial and salvage chemotherapy with mitoxantrone and steroids or with cabazitaxel and steroids. Despite these treatments, the median survival for patients with castration-resistant, metastatic prostate cancer is generally less than two years (Mark, 2010). As of the date of this final decision memorandum (DM), docetaxel (Taxotere® et al.), cabazitaxel (Jevtana®), and abiraterone (Zytiga®) are the only anticancer treatments that have demonstrated a prolongation of survival in randomized clinical trials for patients with castration-resistant prostate cancer. The median survival for patients with castration-resistant, metastatic prostate cancer who received docetaxel was 18.9 months compared to 16.5 months in patients who received mitoxantrone in a Phase 3 randomized trial that served as the basis for Food and Drug Administration (FDA) approval (FDA SBRA document). The patients in the docetaxel arm of that trial also had a greater relief in pain and a greater increase in quality of life measurements compared to patients in the mitoxantrone arm (Mark, 2010). It should be noted that mitoxantrone has shown in clinical studies to have a palliative effect in patients with prostate cancer but not an effect on survival. The most common (frequency ≥30%) adverse events associated with docetaxel were hair loss, nausea and vomiting, fatigue, sensory neuropathy, and neutropenia (Mark, 2010). Cabazitaxel, which received FDA approval in June of 2010, is indicated for use in combination with prednisone for the treatment of patients with castration-resistant, metastatic prostate cancer who were previously treated with a docetaxel-containing regimen. In the randomized, open-label clinical trial the median survival was 15.1 months for patients who received cabazitaxel and 12.7 months for patients who received mitoxantrone. The most common (frequency ≥10%) adverse events in the cabazitaxel arm included neutropenia, anemia, thrombocytopenia, diarrhea, nausea, vomiting, hematuria, peripheral neuropathy and dyspnea (U.S. Food and Drug Administration; June 18, 2010). Abiraterone was approved by the FDA on April 28, 2011. It is indicated for use in combination with prednisone to treat patients with late-stage (metastatic) castration-resistant prostate cancer who have received prior chemotherapy containing docetaxel. According to the FDA news release on April 28, 2011, “patients who received the abiraterone and prednisone combination had a median overall survival of 14.8 months compared to 10.8 months for patients receiving the placebo and prednisone combination. The most commonly reported side effects in patients receiving abiraterone included joint swelling or discomfort, low levels of potassium in the blood, fluid retention (usually of the legs and feet), muscle discomfort, hot flashes, urinary tract infection, cough, high blood pressure, heartbeat disorders, urinary frequency, increased nighttime urination, upset stomach or indigestion and upper respiratory tract infection.” (U.S. Food and Drug Administration; April 28, 2011). Autologous Cellular Immunotherapy In 2010 the FDA approved sipuleucel-T (Provenge® ; APC8015), for patients with castration-resistant, metastatic prostate cancer. The posited mechanism of action, immunotherapy, is different from that of anticancer chemotherapy such as docetaxel. This is the first immunotherapy for prostate cancer to receive FDA approval. The goal of immunotherapy is to stimulate the body's natural defenses (such as the white blood cells called dendritic cells, T-lymphocytes and mononuclear cells) in a specific manner so that they attack and destroy, or at least prevent the proliferation of, cancer cells. Specificity is attained by intentionally exposing a patient's white blood cells to a particular protein (called an antigen) associated with the prostate cancer. This exposure "trains" the white blood cells to target and attack the prostate cancer cells. Clinically this is expected to result in a decrease in the size and/or number of cancer sites, an increase in the time to cancer progression, and/or an increase in survival of the patient. For the benefit of the lay reader, the context of sipuleucel-T use differs from traditional infused anti-cancer therapies. Most such anti-cancer therapies are manufactured in quantity, packaged for general use and sold by a biopharmaceutical company. Aside from the dosage amount, the therapy itself is not specific to any particular patient. In contrast, once the decision is made to treat with sipuleucel-T, the manufacturing process produces a patient-specific preparation of sipuleucel-T. Sipuleucel-T is made uniquely for each patient with his own white blood cells. The patient’s white blood cells are removed via a procedure called leukapheresis. In a laboratory the white blood cells are exposed to PA2024, which is a molecule created by linking prostatic acid phosphatase (PAP) with granulocyte/macrophage-colony stimulating factor (GM-CSF). PAP is an antigen specifically associated with prostate cancer cells; GM-CSF is a protein that targets a receptor on the surface of white blood cells. Hence, PAP serves to externally manipulate the immunological functioning of the patient's white blood cells while GM-CSF serves to stimulate the white blood cells into action. As noted in the FDA's clinical review, each dose of sipuleucel-T contains a minimum of 40 million treated white blood cells, however there is "high inherent variability" in yield of sipuleucel-T from leukapheresis to leukapheresis in the same patient as well as from patient to patient. The treated white blood cells are then infused back into the same patient. The FDA-approved dosing regimen is three doses with each dose administered two weeks apart. The total treatment period is four weeks. III. History of Medicare Coverage Before this decision Medicare had no National Coverage Determination (NCD) for autologous cellular immunotherapy treatment - sipuleucel-T; PROVENGE®. Local Medicare contractors had discretion to determine coverage in the absence of an NCD. A. Current Consideration CMS opened this review to determine whether or not autologous cellular immunotherapy is reasonable and necessary under sections 1862(a)(1)(A) and/or 1862(a)(1)(E) of the Act. B. Benefit Category Medicare is a defined benefit program. An item or service must fall within a benefit category as a prerequisite to Medicare coverage is that an item or services must meet one of the statutorily defined benefit categories in the Social Security Act and not otherwise be excluded. Autologous cellular immunotherapy treatment, specifically sipuleucel-T (PROVENGE®), may be considered to be within the benefit category of Social Security Act section §1861(s)(2)(A) when furnished in a physician's office; and §1861(s)(2)(B) when furnished in an outpatient hospital setting. This may not be an exhaustive list of all applicable Medicare benefit categories for this item or service. IV. Timeline of Recent Activities June 30, 2010 – Tracking sheet posted on CMS website. July 30, 2010 – End of comment period on the tracking sheet November 17, 2010 – MEDCAC Meeting March 30, 2011 – Proposed decision memorandum posted on the CMS website June 30, 2011 – Final decision memorandum posted on the CMS website V. FDA Status In 2010 PROVENGE® (sipuleucel-T) Suspension for Intravenous Infusion received U.S. FDA approval for one indication and usage. As stated in the label, “PROVENGE is an autologous cellular immunotherapy indicated for the treatment of asymptomatic or minimally symptomatic metastatic castrate resistant (hormone refractory) prostate cancer.” VI. General Methodological Principles When making NCDs under §1862(a)(1)(A), CMS generally evaluates relevant clinical evidence to determine whether or not the evidence is of sufficient quality to support a finding that an item or service falling within a benefit category is reasonable and necessary for the diagnosis or treatment of illness or injury or to improve the functioning of a malformed body member. The critical appraisal of the evidence enables us to determine to what degree we are confident that: 1) the specific assessment questions can be answered conclusively; and 2) the intervention will improve health outcomes for beneficiaries. An improved health outcome is one of several considerations in determining whether an item or service is reasonable and necessary. A detailed account of the methodological principles of study design that the Agency normally utilizes to assess the relevant literature on a therapeutic or diagnostic item or service for specific conditions can be found in Appendix A. Public commenters sometimes cite the published clinical evidence and provide CMS with useful information. Public comments that provide information based on unpublished evidence, such as the results of individual practitioners or patients, are less rigorous and, therefore, less useful for making a coverage determination. CMS uses the initial comment period to inform the public of its proposed decision. CMS responds in detail to the public comments that were received in response to the proposed decision when it issues the final decision memorandum. VII. Evidence A. Introduction This DM organizes the evidence according to FDA status: labeled indication (a.k.a, on label) or off label uses. Only evidence from Phase 3 studies published as full-text, peer-reviewed literature articles was assessed. Unlike Phase 1 and Phase 2 trial designs, Phase 3 trial designs by nature incorporate methods such as randomization, control and blinding to minimize the potential for bias and confounding that can negatively impact the internal validity of a study. Appendix A presents more information regarding methodological considerations. Furthermore, Phase 3 studies in general focus on an endpoint that addresses clinical utility such as survival (as opposed to a surrogate endpoint or a biomarker), which is consistent with CMS’ focus on health outcomes. Lastly, Phase 3 studies generally have larger sample sizes, which permit a robust analysis of the primary efficacy endpoint as well as the safety profile. Health outcomes of interest to CMS for patients with prostate cancer include survival, disease progression and quality of life. CMS prefers to use a holistic approach to assessing the benefit of a medical product for Medicare patients. For example, while a delay in tumor progression may be considered by some to be a good intermediate or surrogate outcome, a coincident onset or increase of toxicity due to treatment may contradict the usefulness of tumor progression as a predictor of improved overall survival or improved quality of life. Under such a circumstance, an alternative management program may be a better strategy for the patient with prostate cancer. B. Discussion of Evidence Reviewed 1. Questions: The development of an assessment in support of Medicare coverage decisions under §1862(a)(1)(A) is based on the same general question for most DMs: "Is the evidence sufficient to conclude that the application of the item or service under study will improve health outcomes for Medicare patients?" For this DM, the questions of interest are: Is the evidence sufficient to conclude that autologous cellular immunotherapy treatment of metastatic prostate cancer in men whose disease is castration-resistant and who are asymptomatic or minimally symptomatic improves health outcomes of Medicare beneficiaries? Is the evidence sufficient to conclude that autologous cellular immunotherapy treatment of prostate cancer in men that is not metastatic and/or whose disease is not castration-resistant and/or who are more than minimally symptomatic improves health outcomes of Medicare beneficiaries? 2. External Technology Assessments (TA) An Internet-based search of the Cochrane Library (http://www.thecochranelibrary.com/view/0/index.html), performed on October 19, 2010 using the search terms “Provenge® ,” “sipuleucel-T,” or “APC8015,” revealed two TA’s. One TA, dated December 2005 and titled “Vaccines for metastatic hormone-refractory prostate cancer,” was from the Canadian Coordinating Office for Health Technology Assessment (www.ccohta.ca). Two vaccines, Provenge® and GVAX, were assessed. The TA noted that Provenge® was not yet on the market in any country. The TA also commented that Provenge® is “safe and well-tolerated,” “demonstrates a statistically significant survival benefit” and that future “clinical trials should compare Provenge® with the current treatment (Taxotere plus prednisone).” The second TA, dated September 2007 and titled “Sipuleucel-T (Provenge® ): Active cellular immunotherapy for advanced prostate cancer,” was from the Canadian Agency for Drugs and Technologies in Health (http://www.cadth.ca/media/pdf/E0037_Sipuleucel-T_prostate_cancer_cetap_e.pdf). This TA, which focused only on sipuleucel-T, noted that the evidence so far indicated a survival benefit in men with androgen-independent prostate cancer but that these were preliminary results that need to be confirmed in future, larger clinical trials. In addition, the TA noted that future studies are necessary to evaluate the effectiveness and safety of sipuleucel-T in men with earlier stages of prostate cancer and in combination with other treatments for prostate cancer. Another TA, dated April 2010 and titled “Special report: vaccines for the treatment of prostate cancer” was conducted by the Blue Cross Blue Shield Association Technology Assessment Evaluation Center. This assessment was performed and the report released just prior to the FDA approval of Provenge®. The report noted that Provenge® is one of only two "vaccines" that have progressed to Phase 3 clinical investigation and, given the preliminary nature of the evidence, Provenge® has demonstrated the “curious result” of an increase in overall survival but no impact on cancer progression. In June 2010 CMS commissioned an external TA for this DM through the Agency for Healthcare Research and Quality (AHRQ) Evidence-based Practice Centers Program. The Blue Cross Blue Shield Association Technology Assessment Evaluation Center performed the work. Mark et al. prepared a final TA dated February 10, 2011 and titled “Outcomes of Sipuleucel-T Therapy.” The TA addresses three key questions posed by CMS: What is the evidence regarding the clinical outcomes of sipuleucel-T for its FDA-approved indication; asymptomatic or minimally symptomatic metastatic androgen-independent prostate cancer? 1a. What is the evidence regarding the relationship between baseline patient characteristics, measurable characteristics of treatment such as cell number or immune response characteristics of patients, post-treatment factors, and sipuleucel-T on outcomes of treatment? What is the level of evidence and summary of evidence for off label indications for sipuleucel-T? 2a. For off label indications, what is the evidence regarding the relationship between baseline patient characteristics, measurable characteristics of treatment such as cell number or immune response characteristics of patients, post-treatment factors, and sipuleucel-T on outcomes of treatment? What is the evidence regarding adverse events potentially attributable to the use of sipuleucel-T? The authors conducted a search of the Medline, Embase and Cochrane Controlled Trials Register databases using the search terms “Provenge,” “sipuleucel,” “Dendreon,” “APC-8015,” “prostate,” “prostatic” and “dendritic cell.” Publically available documents from the FDA, clinicaltrial.gov and conference abstract websites were also searched. Articles or abstracts were selected for further review if they presented the results from randomized, controlled trials or case series studies, focused on the use of sipuleucel-T in patients with prostate cancer, and included at least one clinically relevant outcome of interest. The quality of a comparative study was assessed using a process developed by the US Preventive Services Task Force. Studies were rated as good, fair or poor. The TA also rated the strength of the overall body of evidence using a system developed by AHRQ, which evaluates risk of bias, consistency, directness and precision. Evidence strength was given a grade of high, moderate, low or insufficient. For more detailed information about these two assessments please see the TA. A brief summary of the results of the TA are presented below by key question: 1. What is the evidence regarding the clinical outcomes of sipuleucel-T for its FDA-approved indication; asymptomatic or minimally symptomatic metastatic androgen-independent prostate cancer? Mark et al. evaluated evidence from three randomized, controlled trials that had been published in numerous source documents including FDA reviews and the peer-reviewed medical literature. They identified the trials as “IMPACT,” “D9901” and “D9902A.” They remarked that the numerous sources of documents introduced three concerns during the TA. The first concern was the potential for “discordant results seeming to arise out of what appears to be the same analysis. Such issues could arise from error, slight differences in the data used, such as different cutoff date, or minor selection criteria. We noticed several instances where calculations varied beyond the tenth place in a decimal calculation. We used the peer-reviewed publication value whenever there was discordance.” The second concern was “the presentation of alternative or “sensitivity” analyses, which are variations on a particular analysis usually meant to support the validity of the primary analysis. These are sometimes problematic, because they are usually performed post-hoc, often incompletely described and presented, and may be inappropriate. There are many such analyses in the published materials we reviewed for this assessment. Presentation of all these analyses may give the impression that there is more evidence than there really is. Analyses based on the same data set are obviously highly correlated; the unmeasured biases of that study exist across all analyses. In these materials, it is evident that some analyses performed by FDA statisticians on the clinical trial data were meant to critique the validity of the sensitivity analyses performed by the sponsor. Points of contention raised by competing statistical analyses raise issues that may not be solvable but point to an underlying uncertainty in the conclusions of a particular analysis.” Given these caveats, the authors opted to focus on the principal results of the studies and simply summarize the results of the numerous sensitivity analyses. In this DM, CMS will take an identical approach. The last concern was that “the studies were not fully independent investigations, although they will be presented as such. Decisions regarding outcome measures, selection criteria, and analysis were made based on the findings of the earlier studies. For example, because study D9901 did not attain statistical significance for its original end point of progression-free survival, enrollment for study D9902A was terminated; thus, its sample size is smaller than originally planned. Because of results obtained from analyses of D9901 and D9902A, the principal outcome of IMPACT was changed to overall survival, and selection criteria for the study were altered.” Mark et al. identified a number of issues with the study design, which was very similar for all three trials. The sipuleucel-T administered to patients originally in the placebo arm that experienced disease progression and were subsequently unblinded (the authors referred to this as frozen salvage product) was different from the sipuleucel-T administered to the patients randomized to the sipuleucel-T arm. As stated by the authors: “In addition to being prepared from cryopreserved cells, another difference between frozen salvage product and sipuleucel-T is that the frozen cells have never been exposed to circulating sipuleucel-T in the patient’s body. In the treatment groups of the clinical trial, the cells extracted at the second and third leukapheresis sessions are drawn from subjects who have had prior exposure to sipuleucel-T from the first dose.” This difference may have impacted the internal validity of the study. The lack of a truly inert placebo was another issue identified by the authors that may have impacted internal validity. Given the concerns and issues just described, the quality of each of the three trials was rated as good. The TA also rated the strength of the overall body of evidence as “moderate” using a system developed by AHRQ. “The principal reason for the moderate grade is the risk of bias due to the unequal provision of subsequent treatments. The trial design resulted in a systematic bias against the control group due to a delay induced by treatment with frozen salvage product. The statistical methods used to account for subsequent treatments are limited in that time-dependent confounding effects cannot be accounted for.” Appendix B contains tables from the TA that show the survival results from each of the trials. Duration of follow-up in each trial was sufficient to follow patient survival outcome until at least 67% had died. Median survival difference between each arm was 4.1 months for the IMPACT study, 4.5 months in D9901 and 3.3 months in D9902A. The authors noted however that “in D9902A, survival times were shorter in both sipuleucel-T and placebo groups, such that the median survival time in the sipuleucel-T group was shorter than the placebo groups from the other 2 trials. There does not appear to be any difference in patient characteristics in this trial to explain this difference in survival times. However, given the relatively small sample size of the study, the result could be due to chance. The hazard ratio for death calculated from a Cox proportional hazard model shows a reduction in mortality for sipuleucel-T treated groups of 0.77, 0.59, and 0.79, from the IMPACT, D9901, and D9902A studies, respectively. These differences were statistically significant for the IMPACT and D9901, but not significant for the D9902A.” Appendix C contains a table from the TA that shows the results for the disease progression outcomes and other secondary outcomes from each of the trials. None of the results for the disease progression outcomes achieved statistical significance. The TA considered the impact on health outcome of subsequent salvage treatment or chemotherapy after disease progression. In the IMPACT study 63.7% received salvage treatment while 75.6% and 66.7% received salvage treatment in D9901 and D9902A, respectively. In the IMPACT study the median survival for patients who received salvage treatment was 23.8 months compared to 11.6 months for those who did not. The authors noted that this “comparison should not be used to infer a potentially beneficial effect of frozen salvage product, because it is not randomized and is subject to survivor bias. Assignment to the frozen salvage product group is conditional on survival up to the point of receipt of that treatment, producing a survivor bias.” In each of the three studies some patients with disease progression from either the sipuleucel-T arm or the control arm subsequently received chemotherapy. The impact of docetaxel, which has shown a survival benefit in previous clinical trials, was assessed. For the IMPACT study, “a greater proportion of sipuleucel-T-treated patients received docetaxel chemotherapy (57.2 percent versus 50.3 percent), and they also received it earlier (median 7.2 months versus 9.6 months). In D9902A, slightly more sipuleucel-T treated-patients received docetaxel (38.6 percent versus 34.4 percent), but in D9901, more placebo-treated patients received docetaxel (47.6 percent versus 35.9 percent). The difference in median time to receipt of docetaxel in IMPACT might be partially explained by the use of frozen salvage product in the placebo group, which requires one month to administer.” Mark et al. noted the different types of analyses, and the results of these analyses, that were conducted to adjust for the potential confounding effects of the administration of docetaxel. “In one type of analysis, patients are removed (“censored”) from the study upon docetaxel initiation. Assuming that the patients censored are similar to the patients not censored, such an analysis intends to estimate the survival of patients who did not receive docetaxel. This analysis of the IMPACT data showed a treatment effect hazard ratio of 0.649, which was statistically significant (p=0.009). Another analysis of IMPACT data using a time-dependent variable indicating the time of docetaxel use showed a hazard ratio of 0.777 which was also statistically significant (p=0.034). An analysis with time-dependent variables assumes that patients who receive docetaxel are similar to patients who do not receive docetaxel, and that their estimated survival is altered by some fixed magnitude upon receiving docetaxel treatment. Analyses of D9901 and D9902A, which are likely to be time-dependent analyses, show similar magnitudes of treatment hazard ratio of sipuleucel-T to the other analyses, but do not meet the standard level of statistical significance. The authors expressed uncertainty that the analyses performed can fully account for all of the potential confounding effects of subsequent docetaxel treatment. Specifically, these analyses did not “account for potential differences between treatment regimens in terms of dose or length of treatment. The analyses require assumptions of the events which are not observable in the trial. The usual assumption of an analysis censoring subjects at the time of docetaxel use is that the censoring time provides no further information about the subjects’ likelihood of future survival. Survival curves will be biased unless those who were censored for docetaxel use have similar expected survival to those who were not censored for docetaxel use. This assumption is implausible, since docetaxel is a treatment for disease progression. However, the effect of this bias on the estimate of the relative effect of sipuleucel-T on survival would depend on the degree of this bias in each treatment arm. It could be possible that each treatment arm is similarly affected by this bias, producing an unbiased estimate of treatment effect. Time-dependent analyses also assume that the change of exposure and its timing are not related to the probability of future survival.” The TA noted that this analytic issue is called time-dependent confounding. To summarize their assessment for this key question the authors stated that “all three studies showed improved median and 36-month survival of sipuleucel-T-treated subjects compared to placebo-treated subjects. In two of the studies, the difference met traditional levels of statistical significance. The third smaller study did not meet statistical significance. The third study showed overall shorter survival times, but chance or other unmeasured difference in study participants could explain the finding. There was no difference in disease progression end points. Analyses undertaken to account for potential confounding effects of subsequent treatments did not change the magnitude or statistical significance of the findings, but such methods may be limited in the ability to fully account for such effects.” 1a. What is the evidence regarding the relationship between baseline patient characteristics, measurable characteristics of treatment such as cell number or immune response characteristics of patients, post-treatment factors, and sipuleucel-T on outcomes of treatment? In this question the authors were asked to assess the results of subgroup analyses. While acknowledging the potential benefits of conducting subgroup analyses such as the identification of treatment modifiers, the authors also stated the potential statistical limitations of such analyses including a lack of power to detect a real difference due to the small sample size of each subgroup. The authors noted that this problem was compounded by the “fairly small” size of the database from D9901 and from D9902A. “Thus it is unlikely that examination of subgroup effects in these trials would generate any definitive findings unless the underlying treatment interactions were extremely strong. Any suggestion of a subgroup effect in these analyses would require further research and confirmation.” Therefore, while the authors presented the results of the subgroup analyses for the pooled efficacy databases, their opinion was that “the broad confidence intervals encompassing each subgroup hazard ratio preclude any conclusions or signals of subgroup treatment effects. It cannot be determined whether any large difference between the two hazard ratios between any pair of subgroups is due to random variation (chance) or a real interaction.” Mark et al. expressed a similar judgment upon examining the results of subgroup analyses for just the IMPACT efficacy database with one exception: the effect of patient age (i.e., younger than 65 years v. 65 years of age or older). However, as noted above, in the analysis of the pooled efficacy database, this difference in outcome by age group was not found hence the authors conclude that it is “inconclusive whether there is a true treatment interaction” based on age group. In their assessment of cell product parameters and patient outcome the authors examined the association of survival with CD54 up-regulation ratio, total nucleated cell count and CD54 cell count. The only results available were from patients in the sipuleucel-T arm. Without the ability to compare these results to those obtained from patients in the control arm, it was not possible to differentiate “a treatment effect versus a characteristic associated with inherent survival.” The assessment of the association of patient immune response, as measured by antibody titer against T-cell proliferation to PA2024 or PAP, and patient outcome was based on results from only patients from the IMPACT study. The authors noted that there “were no associations between T-cell proliferation to PA2024 or PAP and survival. None of these analyses appear to have been adjusted for potential confounding variables.” Mark et al. concluded that “some analyses of product parameters and patient immune responses show an association between the characteristic and survival, but the clinical significance of these associations are unknown. Because the biologic mechanism of the therapeutic effect of sipuleucel-T is not fully understood, these analyses do not inform the question of the overall efficacy of sipuleucel-T. The quantity of data and the analyses performed so far are not sufficient to determine whether such product parameters or measures of patient immune response are clinically useful.” Finally, Mark et al. examined the efficacy results to “determine if there is an interaction of sipuleucel-T and post-treatment chemotherapy. That is, is there a differential effectiveness of sipuleucel-T depending on whether post-treatment chemotherapy is given or not?” The authors continue by stating that “Unfortunately, given the data and analysis available, this is difficult to determine. Examination of the survival curves of each initial and subsequent treatment group (sipuleucel-T/placebo, no docetaxel/docetaxel) may be biased by potential confounding and survival biases. The groups receiving post-progression docetaxel survive longer than the other groups because receiving such treatment was conditional on being alive to receive such treatment. If sipuleucel-T is effective, then it is effective in a context in which a substantial proportion of patients receive subsequent chemotherapy. Determination of the independent and/or interactive effects of sipuleucel-T and subsequent therapies would require further study using study designs where patients are randomized to subsequent treatments or data collection and analyses are performed which can account for time-dependent confounding variables.” 2. What is the level of evidence and summary of evidence for off label indications for sipuleucel-T? A formal rating of the quality of each study was not performed given the lack of comparative studies to rate. Dr. Mark et al. noted that the peer-reviewed and published studies that investigated the use of sipuleucel-T in patients without both metastatic and castration resistant prostate cancer “were early Phase I and II studies which did not have control groups. They were largely intended to assess potential biologic activity, immune response, and safety, and thus were not intended to provide definitive evidence for efficacy. They may have not been designed or conceived with strict treatment indications in mind. The shortcomings of those studies in determining efficacy should be viewed in this light. In addition, the dose and scheduling of treatment differed from the three RCTs previously reviewed for the on-label indication. The manufacturing process and quality control criteria may have differed from the currently available treatment.” One abstract was found of a randomized, controlled trial but there was minimal detail available about the patient population and results; this study has yet to be published in a peer-reviewed article. In conclusion, the authors stated that there “is insufficient evidence to evaluate the outcomes for off-label indications.” 2a. For off label indications, what is the evidence regarding the relationship between baseline patient characteristics, measurable characteristics of treatment such as cell number or immune response characteristics of patients, post-treatment factors, and sipuleucel-T on outcomes of treatment? The authors succinctly noted that “Since none of the studies provide evidence of efficacy of sipuleucel-T for off label indications, this question is moot.” 3. What is the evidence regarding adverse events potentially attributable to the use of sipuleucel-T? Marks et al. used the pooled safety analysis from the FDA Clinical Review to assess adverse events. Thus, the TA’s safety analysis was based on the 601 patients treated with sipuleucel-T and 303 placebo-treated patients. The authors stated that given the length of the safety analysis report in the FDA’s clinical review, the TA focuses on “some of the issues analyzed in the review we judged to be relevant: 1) deaths occurring proximate in time to treatment; 2) nonfatal serious adverse events; 3) cerebrovascular events; 4) infections; and 5) infusion-related adverse events.” In addition, “Using the questions proposed to evaluate the quality of reporting harms based on the McMaster Quality Assessment Scale for Harms, based on the description of adverse events reporting from the protocol document for the IMPACT trial and the FDA Clinical Review, we judged that all 6 questions could be answered affirmatively and thus the adverse event reporting in this document was of good quality.” The authors introduced a number of reasons that increased the complexity of assessing the adverse effects of sipuleucel-T. The first reason concerned the advanced age of the typical patient with metastatic, castration-resistant prostate cancer. Comorbidities are common in this age group. The authors noted that “As patients are followed for a survival end point, as disease progresses it would become increasingly difficult to attribute any particular event to the patients’ existing comorbidities, progressive cancer, sipuleucel-T, or other subsequent treatments. In all of the randomized clinical trials, after progression of prostate cancer (in IMPACT) or after 16 weeks (in D9901 and D9902A), adverse events, with the exception of cerebrovascular events (CVEs), were only collected if they were thought by the investigators to be related to sipuleucel-T treatment. Since the studies became unblinded at the time of disease progression, such a judgment of causality could be biased.” In the authors’ opinion the lack of a truly inert placebo was a second reason for increased complexity. Patients in the control arm of each study “were subjected to leukapheresis procedures and received an infusion of cultured but untreated cells. Thus any adverse effects caused by procedures in common between the treated and placebo groups might be balanced in the two groups. In the usual clinical trial with an inert placebo, an equal incidence of an adverse event in active and placebo groups implies that the event is due to inherent baseline risk, natural history, or psychological effects. This conclusion should not be made in these clinical trials, particularly for types of events that are suspected or known to be caused by infusions. If, for example, contaminated infusions cause an equal incidence of bacterial infection in both sipuleucel-T and placebo groups, it should not be concluded that sipuleucel-T does not cause bacterial infection.” The third reason for increased complexity: “after progression a large proportion of placebo-treated patients received frozen sipuleucel-T salvage treatment. We could not locate reports of the adverse events associated with frozen salvage treatment. It is unknown whether the potential risks of standard sipuleucel-T may occur with frozen sipuleucel-T salvage product. In addition, as reported previously, many patients in both groups received chemotherapy. Given these multiple confounding effects, it is very difficult to tell if events occurring distant in time to the initial treatment with sipuleucel-T or placebo can accurately be attributed to sipuleucel-T.” The authors concluded that “there are a few solid conclusions that can be reached.” However, they did identify two adverse events, infusion reactions and infections, which have an association with sipuleucel-T treatment. They stated that sipuleucel-T “can cause symptoms consistent with an infusion reaction” given the greater frequency of infusions reactions in the sipuleucel-T group compared to the placebo group as well as the temporal proximity of infusion reactions to the administration of sipuleucel-T. They also noted that infections during sipuleucel-T treatment are “probably in relation to leukapheresis and infusion procedures. Catheter-related infections are attributable to sipuleucel-T treatment. Some infections proximate in time to infusion are possibly related to sipuleucel-T treatment. Attribution is difficult because the control groups in the RCTs also underwent leukapheresis and infusion procedures. Contaminated infusion product has been documented.” Beyond these two types of serious adverse events, “it is unclear whether there is an association with sipuleucel-T treatment. CVEs were a particular focus of attention, and although rates were slightly higher, it is not possible with the data available to determine causality. No associations with product parameters or interactions with patient characteristics were identified.” Overall Conclusions In their overall conclusion for the TA the authors stated that “Three randomized clinical trials of sipuleucel-T are consistent with longer overall survival in patients meeting the FDA-labeled indication. This conclusion is tempered by consideration of a trial design with inherent potential for confounding due to systematic differences in post-progression treatment, making the estimate of the quantity of benefit less certain. This treatment effect occurs in the context of use of post-progression chemotherapy. There is insufficient evidence regarding potential interactions, associations with characteristics of the product, and interactions with other treatment. There is insufficient evidence for any off-label indication. Sipuleucel-T can cause infusion reactions and infections.” “Interpretation of the existing clinical trials of sipuleucel-T was hampered by a study design that had the original intended purpose of assessing progression-free survival in an objective manner. This dictated measures such as blinding and placebo in order to avoid bias in the assessment of outcome. The likely presence of time-varying subsequent treatment and confounding adds further complexities. Since it appears that sipuleucel-T has little or no effect in delaying measurable disease progression, it would be important for future trials to be robustly designed for a survival end point. Although it is not possible to dictate all possible treatments being employed in clinical trials, particularly as patients’ disease progresses, study designs should avoid the potential for systematic biases in the use of post-progression treatments and ensure an equal standard of care for patients in all treatment arms.” “Because the effect of sipuleucel-T is not apparent early in the course of disease after treatment and only in the context of a substantial amount of eventual chemotherapeutic treatment, it would be important to understand the existence of and nature of interactions between sipuleucel-T and subsequent treatments. The current existing analyses are insufficient to know to what degree sipuleucel-T is effective in the absence of chemotherapy or depends on chemotherapy to demonstrate improvement in survival. Such information is critical for decisions physicians and patients need to make as they plan how to treat the patient’s cancer. Future clinical trials with properly designed treatment arms and data collection may be able to determine these interactions.” 3. Internal technology assessment CMS performed a literature search using PubMed on October 15, 2010 with the search terms “Provenge,” or “sipuleucel-T,” or “APC8015.” The following limitations were applied: Humans, Clinical Trial, Meta-Analysis, Practice Guideline, Randomized Controlled Trial, Case Reports, Comparative Study, Controlled Clinical Trial, and Multicenter Study. Ten literature articles were identified due to this search. Eight articles addressed the on-label use of Provenge while the remaining two addressed off-label use. This same search performed May 23, 2011 did not identify any new articles. On label Use Eight articles were found that focus on on-label use. Three of the eight articles (Small, 2006; Higano, 2009; Kantoff, 2010) presented results from randomized, controlled trials. These trials comprised the majority of the evidence that FDA used as the basis for its decision. CMS’ presentation of these trials is located in the Evidence Summary below. The remaining five articles are not addressed in this DM because they either concerned an early (Phase 1 or 2), uncontrolled study that did not assess a health outcome as its primary endpoint (Small, 2000; Burch, 2000; Burch, 2004) or presented only preliminary results of a Phase 3 study (Lee, 2003; Schellhammer, 2005) that were presented completely in a subsequent article by Small et al., 2006. Off label Use No Phase 3 studies were found during the literature search. Evidence Summary On label Use Small EJ, et al. Placebo-controlled phase III trial of immunologic therapy with sipuleucel-T (APC8015) in patients with metastatic, asymptomatic hormone refractory prostate cancer. I 2006;24:3089-3094. Note: This article presents the results of the trial labeled as D9901 in the FDA clinical review. Small, et al. presented the results of a US-based, multicenter, double-blind, placebo-controlled, randomized (2 sipuleucel-T:1 placebo) Phase 3 trial in patients with metastatic, castration-resistant prostate cancer. Patients had to have radiologic evidence of metastases, a serum testosterone less than 50 ng/dl, and an ECOG performance status of 0 or 1. Concurrent bisphosphonates treatment was permitted but not systemic corticosteroids. Patients with cancer-related bone pain or who were taking opioid analgesics for cancer pain were excluded. Patients randomized to the placebo arm who subsequently were unblinded due to disease progression were eligible for treatment with sipuleucel-T as salvage therapy. Treatment with any type of anti-cancer therapy was prohibited until the primary efficacy endpoint was met. At that time, each patient was treated by his own physician. Sipuleucel-T or placebo was prepared after each of three leukaphereses. Sipuleucel-T was manufactured by the procedure presented in the Background section of this DM. Placebo consisted of the reinfusion of only one-third of the APCs isolated during leukapheresis without any exposure to PA2024. The remaining APCs were frozen to be used in the patient if he experienced disease prog