Online First/Online Only
Articles/items published ahead of print or only online.
Sunday, December 21, 2014
BY LOLA BUTCHER
Payment reform in oncology got a major boost with the announcement that the University of Texas MD Anderson Cancer Center and UnitedHealthcare are launching a pilot to test bundled payments for head and neck cancers.
Unlike more limited bundled payment or episode-of-care pilots, this initiative calls for UnitedHealthcare to make a single payment to MD Anderson for all services provided to a patient in a year.
“We will send one bill, and UnitedHealthcare will send us one payment,” explained Thomas Feeley, MD, Head of Anesthesiology and Critical Care and the Institute for Cancer Care Innovation at MD Anderson. “The course of the bundle is a year, and absolutely anything the patient needs is covered in that price.”
That includes all inpatient and outpatient care, surgery, chemotherapy, emergency care, and ancillary services provided at MD Anderson. Dental restorations are not included in the bundle, nor are cancer-related services that are rendered by providers outside the MD Anderson system.
The announcement underscores the emerging consensus that the bundled payment concept may work for cancer care. Also in the works: The federal government’s Center for Medicare & Medicaid Innovation (CMMI) is developing the Oncology Care Model (OCM) as part of its Bundled Payment for Care Improvement Initiative. And a consortium of payers, physicians, and other stakeholders convened by the Center for American Progress (CAP) is working on a multi-payer demonstration of oncology bundled payments. And, like UnitedHealthcare, private insurers are conducting pilots with individual provider organizations.
Everything is an experiment at this point. Speakers at the “Bundled Payments in Oncology” webinar presented by the National Institute for Health Care Management Foundation and the Center for American Progress earlier this year said the only thing certain is that fee-for-service payment for cancer care is on its way out.
“We are in the early stages of a movement to build a better health care system, one not bound by historical and political considerations, but driven by science, teamwork, and compassion,” said Glenn Pomerantz, MD, Vice President and Chief Medical Officer at Horizon Blue Cross Blue Shield of New Jersey. “We are at a tipping point of monumental change.”
Bundles in Context
The terms “bundled payments” and “episode-of-care payments” are often used synonymously, although some people distinguish between the two. Both terms refer to a payment system in which providers—working alone or with others—receive a single payment for a pre-defined set of services. In some cases, the “bundle” is all-inclusive—for example, all inpatient, outpatient, and ancillary services from 30 days before a knee replacement to 120 days after. In others, the “episode” is much narrower, such as all outpatient physician services needed to treat a patient with a specific condition during a 90-day period.
In some cases, “bundles of care” or “episodes of care” are paid for upfront; in others, insurers pay providers on the standard fee-for-service schedule for a certain period of time, then make a “retrospective” adjustment that may reward the providers for keeping costs low or punish them for failing to do so.
The proving ground for the bundled/episode payment concept was the Medicare Acute Care Episode Demonstration that started in 2009. That initiative showed that the use of a bundled payment for both hospital and physician services for certain inpatient procedures—for example, hip and knee replacements and cardiac pacemaker implantation—prompted providers to work together in new ways. The result: quality of care improved; overall costs were lower; and physicians and hospitals benefitted financially.
Since then, bundled payment has become the most popular payment innovation in the country. More than 6,000 hospitals, physician groups and skilled nursing facilities are participating in the CMMI Bundled Payments for Care Improvement Initiative, which includes treatment of heart attacks, cellulitis, and dozens of other conditions as well as procedures.
In oncology, the biggest bundled-payment experiment heretofore was UnitedHealthCare’s pilot with five medical oncology groups (OT, 11/25/11 issue). That three-year pilot, which included 810 patients with breast, colon, and lung cancers, found that a new payment model resulted in a 34 percent reduction in total medical costs (Newcomer et al: Changing Physician Incentives for Affordable, Quality Cancer Care: Results of an Episode Payment Model. Journal of Oncology Practice doi: 10.1200/JOP.2014.001488).
In that case, the “bundle” was limited to services provided by medical oncologists. Chemotherapy medications were reimbursed based on average sales price, so an oncologist’s income was not tied to drug sales.
New Pilot Different in Almost Every Way
Despite the financial success of that pilot, the new UnitedHealthCare pilot with MD Anderson is different in almost every way. For one thing, the scope of the bundle is all-inclusive because MD Anderson offers a full range of cancer care services. For another, the pilot is limited to cancers of the salivary glands, oral cavity (including the mouth, lips, tongue, and soft palate), throat, and larynx.
Those diagnoses account for a much smaller cohort of patients than those for more common cancers, which makes it easier for MD Anderson to assume financial risk for those patients.
“Also, our patients aren’t as expensive to care of [compared with patients with some other cancer types] so there is much less risk starting out in our program,” said Randal Weber MD, Chair of MD Anderson’s Department of Head and Neck Surgery.
Oncologists will not know whether an individual patient is assigned to the pilot or not. However, the awareness that MD Anderson is at financial risk for some patients is expected to make all members of the multidisciplinary team more cost-conscious, he said.
“We wanted to not have our opinions and judgments clouded by whether there is a financial issue here. We want to take care of patients the way we always have,” Weber said. “Globally, this is raising our awareness of ‘is that PET scan necessary and important?’”
As the MD Anderson-UnitedHealthCare pilot gets under way, a group of oncologists, payers, patient groups and policymakers are developing a multi-payer demonstration for bundled payments. The group, convened by CAP, has been working together for nearly two years.
“Much of the recent literature about the success and failure of bundled payments and other payment changes is that you need a very large signal from payers, and in cancer, that large signal has to include both CMS and private payers because of the age spectrum of cancer diagnoses,” said Ezekiel Emanuel, MD, PhD, Chair of the Department of Medical Ethics and Health Policy at the University of Pennsylvania, and a Senior Fellow at CAP.
In the webinar, he described the decisions to date: The initiative will begin with high-prevalence cancers: metastatic non-small cell lung cancer, adjuvant colon cancer, and metastatic colon cancer. The episode payment will cover the total cost of a patient’s care. While several physicians might provide some elements of care, the medical oncologist will be the “accountable provider,” serving as quarterback for the entire care plan.
At the beginning, the oncologist would be eligible for shared savings—that is, to share the savings with the payer if total costs of care are lower than a pre-determined benchmark—without downside risk. Eventually, however, the oncologist would be subject to two-sided risk, meaning that he or she would receive shared savings if costs are low but would risk losing money on patient care if costs are higher than expected.
A standardized set of quality measures would be used by all payers.
“One of the widespread agreements going into this was that the existing set of quality measures for oncology were inadequate—not meaningful in terms of the continuum of care and not meaningful, either to the patients or to providers, in terms of measuring really what we care about in terms of high-quality care,” Emanuel continued.
Consortium members have developed a relatively small set of quality measures, most of which apply to all diseases and a few of which are disease-specific. Measures that apply to both adjuvant and metastatic disease include the percentage of patients with an inpatient admission associated with treatment-related complications and discussion of patient’s goals before initiation of any new line of therapy.
The consortium intends to publish its recommendations, encourage broad adoption of the model by both payers and providers, and track to see how the bundled payment concept works.
“We all know we’re not going to get things perfect out of the box,” he said. “We will participate in refining the model and working to ensure that best practices and efforts to actually transform the care get widely disseminated so that we can shorten the adoption time and the learning curve.”
Meanwhile, the government is expected to propose its Oncology Care Model very soon.
“We’ve heard from many specialists that they want to move towards alternative payment models and away from fee-for-service, but they need to see the pathway to move in that direction,” Patrick Conway, Deputy Administrator for Innovation and Quality and CMS Chief Medical Officer, said in the webinar.
Erin Smith, Director of CMS’s Division of Technical Model Support, identified some of the broad parameters of what the Oncology Care Model (OCM) will entail:
· All cancer types will be eligible for the payment model;
· An episode will be defined as the six months after chemotherapy is initiated, and if a patient is continuing treatment after that, a second episode will begin;
· All inpatient and outpatient services and drugs that a patient receives during the episode period will be included in the episode;
· Physician practices that participate in the OCM must agree to engage in “practice transformation” activities including the use of treatments that comply with nationally recognized clinical guidelines, electronic health record technology, data that support quality improvement, patient navigators, and documented care plans for every patient; and
· Participating physicians would receive a per-beneficiary-per-month payment to help support the enhanced services, and they would receive a monthly care management payment for all OCM patients during their six-month episodes. Additionally, physicians would be eligible for financial incentives if they lowered the total cost of care and achieved certain quality-of-care thresholds.
The government plans to recruit other payers, such as commercial insurers and Medicaid managed care organizations, to participate in the OCM by supporting practices and offering financial incentives that are aligned with the government’s pay model. Although Medicare is the single largest payer for cancer care, it is responsible for only about 50 percent of cancer patients; CMS officials agree with Emanuel that physician practices cannot afford to change the way they practice for just one segment of their patient population.
“In order to create practice transformation at a higher level and not just for Medicare beneficiaries, we think it’s critical to include other payers in this model,” Smith said. “Being able to leverage additional payers to cover more of a practice’s patient population would make that transition easier and provide a better value proposition for practices to be able to do this.”
Episode-of-Care Payments, New Jersey-style
And in yet another payment reform development, Horizon Blue Cross Blue Shield of New Jersey and Regional Cancer Care Associates (RCCA), one of the nation’s largest oncology physician groups, launched an episode-of-care program for breast cancer patients this fall.
In that pilot, Horizon makes a single payment—including outpatient and inpatient clinical services as well as diagnostics and drugs—for a discrete diagnosis from the onset of symptoms until treatment is complete.
RCCA—with more than 100 physicians working in 24 locations—collaborated with Horizon to develop the new payment system because its leaders believe it will improve the quality and outcomes of cancer care.
“The core objective is: You must maintain or improve clinical outcomes while you’re controlling cost,” said RCCA Co-founder and President Andrew Pecora, MD.
To start the program, Horizon and RCCA agreed to a fixed price for a bundle of services for a given diagnosis for a set period of time. They also agreed on quality and safety measures pertinent to that medical situation.
As medical services are provided, Horizon pays RCCA oncologists and all other providers on a fee-for-service basis. The oncologist assumes responsibility for safety and quality outcomes, including overall survival, disease-free progression, appropriate treatment, and patient’s functional status.
“If the safety and quality thresholds are met and there are savings, we share those savings with the oncologists,” Pomerantz said.
After oncologists gain experience managing the quality and safety of patient care, they will assume risk for managing the cost of care as well. At that point, Horizon will pay the agreed-upon price for services for a given diagnosis and RCCA will make or lose money based on its ability to control costs while still meeting the quality and safety targets.
The program, which began in October, is enabled by a data platform that uses the molecular subtype of a patient’s cancer to identify individualized treatment. The technology pulls in electronic health record data to provide real-time tracking of survival, toxicity, dose and dose intensity, cost, and other outcomes. It also alerts physicians about possible pivot points—for example, if outcomes are veering away from expectations—at which they may need to take action.
Sunday, December 21, 2014
Robert C. Young, MD
On January 13, 2013, CBS News Online reported that eating 150 grams of processed meat daily (the equivalent of six strips of bacon) increases your risk of developing pancreatic cancer by 57 percent. This and many other odd and commonplace statistical risks are explored in this book, a delightful romp through the world of statistics and psychology of risk perception.
The authors, two Brits--Michael Blastland a journalist and David Spiegelhalter, a Cambridge University expert in risk perception—use a lighthearted, humorous style for a serious exploration of statistical risk and our human responses to those risks.
To provide us with a way to compare different risks, the authors employ an ingenious little tool called a MicroMort (MM). Since roughly one in a million Brits die of accidents each year, this one in a million defines 1 MM. Actually life in the United States is riskier (1.6 MMs). Using this tool one can equate the risk of anesthesia (5 MMs) with 20 miles of motorcycle travel: climbing Mount Everest (43,000 MMs), one scuba dive (7 MMs)--you get the idea.
MicroMorts measure acute risks. For chronic risks like smoking or obesity, the authors use MicroLives (MLs). One ML is about 30 minutes of life, selected because the average adult has about one million half hours of life remaining. Smoking costs you about 10 MLs a day, obesity 3MLs a day, being male 4MLs a day. Since MMs are acute risks, they disappear after the event, but MLs accumulate.
The book includes several large charts at the end comparing many common and unusual risks measured in MMs and MLs.
The authors skillfully explore, in uncomplicated terms, many arcane and at times misleading statistics. Take the bacon example mentioned earlier: First, it is unlikely that anyone would actually eat that quantity every day of their lives. And even if they did, about 1.5 percent of Americans develop pancreatic cancer in their lifetime. That’s 5 in 400, and the risk in bacon addicts rises to only 6 in 400. Relative risks behave like a magnifying glass: If we consider the entire at-risk population, 395 of the 400 people avoid pancreatic cancer. In those who gorged on bacon, 394 of 400 are still fine.
The authors examine a variety of intriguing statistics about the normal risks of life, and while we tend to focus on external risks, natural risks abound: Being a newborn is risky—6,100 MMs in the first year. Childbirth is 210 MMs, or the equivalent of an 800 mile trip on a motorcycle. We accept the natural risks better because we lack any meaningful control over them and because we so greatly value the outcome.
But Blastland and Spiegelhalter argue that most people are not statisticians and don’t see risks in the same way. To probe the psychology of risk perception, the book utilizes three characters: Norm, who seeks risk balance; Prudence, who is risk averse; and Kelvin, who enjoys risk. Statisticians see risk in terms of numbers, but people assess risk through stories, anecdotes, and personal experience. Throughout the text, these three folks are used to illustrate how different people respond to a similar level of risk.
Crime elicits different responses depending on the nature of the crime. The more gruesome the crime, the more vividly it is remembered--and as a result, more often overestimated in frequency. Reality is often different. Children are 20 times more likely to be injured by their parents than by strangers (0.4% of abductions) or by sex offenders (1.5% of abductions). Of the approximately 260,000 abductions that take place every year, 200,000 are by relatives.
Fear of flying is a dramatic example of a disconnect between perception and reality: Three to five percent of people won’t fly, 17 percent are afraid, and 30 to 40 percent have moderate anxiety when flying. Yet in reality, commercial flying is extremely safe--measured in MicroMorts, the risk is 0.02MMs. The authors remind us that you would have to take 50 million flights before you are likely to have a fatality. One flight a day would take you 120,000 years. Comforting thought.
‘Illusory Superiority,’ ‘Denominator Neglect’
But why do we have this misperception? First because it’s a catastrophic event and understandably frightening, and second because it’s completely beyond our control. We humans always have the belief that we can produce better outcomes if we are in control--what the authors call “illusory superiority.”
One of the most provocative chapters in the book is entitled “Nothing,” which examines the concept of “denominator neglect”—i.e., that it’s easier for humans to imagine events than to conceptualize non-events. The media contributes to the difficulty. For example, you never see headlines proclaiming “No children killed on the way to school today,” but rare, exotic events (“Mad Cow disease”) are covered widely. Comparatively common events (smoking-related deaths) are seldom covered.
The authors finish the book by attempting to reconcile the two faces of risk: The orderly view of population risk seen in the numbers, contrasted with the jumbled human response based on widely varying personal perceptions of individual risk-benefit.
Blastland and Spiegelhalter conclude that “Norm, Prudence, and Kelvin are not irrational when they choose to ignore the numbers and go their own way. The measure of what people perceive as risk is a matter of personal value and personal framing.”
Indeed, individual conclusions about risk are highly dependent on how the risk data are framed when presented. “Evidence-Based Risk Communication” was explored in a recent Annals of Internal Medicine article (2014;161:270-280). Those authors tested various ways of communicating risk, concluding that visual aids and absolute risk formats improve patient’s understanding, whereas numbers needed to treat can lessen understanding. Presenting benefits before harms improved accuracy, but decreased acceptance of treatments.
The Norm Chronicles is a spirited and entertaining expedition through the complex world of risk perception. It is heavily weighted on British statistics but contains enough American data to secure its relevance to domestic readers.
The upbeat treatment of the contrasting aspects of risk makes the book valuable for oncologists and other medical folks as well as for the general public.
2014, Basic Books, ISBN 0465085709; available in paperback and Kindle editions.
Tuesday, December 16, 2014
BY Thomas E. StinchCombe, MD
Co-director, Multidisciplinary Thoracic Oncology Program and Clinical Associate Professor in Clinical Research, Lineberger Comprehensive Cancer Center, University of North Carolina.
The development of novel therapies has improved the care for patients with non-small cell lung cancer (NSCLC). The identification of activating epidermal growth factor receptor (EGFR) mutation (defined as an exon 19 deletion or an exon 21 L858R point mutation), and development of EGFR tyrosine kinase inhibitors (TKI’s) and the identification of anaplastic lymphoma kinase (ALK) rearrangements and development of ALK inhibitors has dramatically changed the treatment paradigm for patients with advanced NSCLC. Unfortunately, progress for patients with NSCLC without these specific molecular alterations and SCLC has been slower. Several important clinical trials were published in peer-reviewed journals in 2014 that will impact clinical research and care over the next several years.
Erlotinib, gefitinib, and afatinib are first-line therapies for patients with a known EGFR mutation, and the progression-free survival (PFS) observed with these therapies is approximately 10-13 months. There is a clinical need to improve first-line therapy. A randomized phase II trial investigated erlotinib alone and with bevacizumab in patients with an activating EGFR mutation.1 The primary end-point was PFS by independent radiological review (IRR). Patients assigned to erlotinib and bevacizumab compared with erlotinib alone experienced a statistically significant improvement in PFS (hazard ratio [HR] of 0.54, p=0.0015; median of 16.0 months and 9.7 months, respectively). The most common grade ≥ 3 adverse events in the erlotinib and bevacizumab and erlotinib arms were rash (25% vs. 19%), hypertension (60% vs. 10%), and proteinuria (8% vs. 0%).
The optimal treatment for patients with EGFR mutant NSCLC who have progressed on EGFR TKI is an area of active investigation. A single-arm phase Ib trial investigated the activity of afatinib and cetuximab in patients who had progressed after erlotinib or gefitinib.2 In the intent-to-treat (ITT) patient population the objective response rate (ORR) was 29% and the median progression-free survival was 4.7 months. The ORR observed in patients with T790M positive and negative NSCLC was 32% and 25%, respectively (p=0.341). The most common grade 3 toxicities observed were rash (20%) and diarrhea (6%), and grade 4 toxicities observed were fatigue, pneumonitis, and lung infiltration, which occurred in two patients.
Crizotinib is approved by the Food and Drug Administration for patients with advanced NSCLC with an ALK rearrangement without regard to line of therapy. Phase III data demonstrating the superiority of crizotinib compared with second-line chemotherapy was available, but phase III data demonstrating the superiority of crizotinib compared with platinum-based chemotherapy was not available until this year.3 Patients with ALK rearranged NSCLC were randomized to crizotinib or platinum-pemetrexed, and the primary end-point was PFS by IRR. Patients assigned to crizotinib compared with chemotherapy experienced a statistically significant longer PFS (HR of 0.45; 95% CI, 0.35 to 0.60; p=0.001; median PFS of 10.9 and 7.0 months, respectively) and higher ORR (74% vs. 45%, p<0.001). Patients assigned to crizotinib experienced a greater reduction in lung cancer symptoms and greater improvement in quality of life. This trial supports the already common practice of using crizotinib in the first-line setting.
Until recently patients who progressed on crizotinib did not have an ALK-directed therapy, but ceritinib, a more potent ALK inhibitor, was approved by the FDA for patients who have progressed on or are intolerant to crizotinib based on a phase I trial with an expansion cohort.4 The maximum tolerated dose was 750 mg daily, and the dose-limiting toxicities (DLTs) observed were diarrhea, vomiting, dehydration, elevated aminotransferase levels, and hypophosphatemia. Among the 114 patients who received ceritinib 400 mg daily, the ORR was 58% (95% CI, 48 to 67), and the median PFS was 7.0 months (95% CI, 5.6 to 9.5). Among patients who had or had not previously received crizotinib, the ORR was 56% and 62%, respectively, and the PFS was 6.9 and 10.4 months, respectively.
At the 750 mg dose level, 62% required at least one dose reduction, and a significant proportion of patients required this dose reduction after the third cycle. Alecitinib was investigated in a phase I/II trial in patients with ALK rearranged NSCLC who had progressed on or were intolerant of crizotinib. The recommended dose for phase II trials was 600 mg twice a day, and the DLT’s were grade 3 neutropenia and headache. The most common grade 3 or 4 adverse events were adverse liver function tests, neutropenia, and hypophosphatemia. The ORR as assessed by the investigator was 55%, and objective responses were seen for central nervous system metastases.
ROS1 rearrangements occur in approximately 1% of patients with NSCLC, and ROS1 rearranged NSCLC represents a unique subset of NSCLC. Crizotinib is an inhibitor of the ROS1 tyrosine kinase. Fifty patients tested positive for ROS1 and the majority of patients on the trial had the ROS1 rearrangement detected using a break-apart fluorescence in situ hybridization (FISH) test. Patients were enrolled in an expansion cohort of a crizotinib phase 1 trial. The ORR observed was 72% (95% CI, 58 to 84), and the median PFS was 19.2 months (95% CI, 14.4 to not reach).5 No unexpected toxicities were observed.
This study demonstrates the value of testing for ROS1 rearrangements and crizotinib as a therapeutic agent.
A multivariate serum protein test (Veristrat®) can classify patients based on treatment outcome with EGFR TKI; patients are categorized as “poor” or “good.” A phase III trial prospectively tested second-line patients in a blinded manner, and patients were stratified based on test classification and randomized to erlotinib or second-line chemotherapy (pemetrexed or docetaxel).6 The primary end-point was OS, and the primary hypothesis was the existence of interaction between protein test classification and treatment. The OS was not significantly different in the two treatment arms in the ITT patient population, and a statistically significant interaction between proteomic classification and treatment was observed (p=0.031).
Patients with proteomic classification of poor assigned to erlotinib compared with chemotherapy experienced a worse OS (HR of 1.72, 95% CI, 1.08 to 2.64; p=0.022; median of 3.0 and 6.4 months, respectively). Patients with proteomic classification of good patients assigned to erlotinib compared with chemotherapy experienced a similar OS (HR of 1.06, 95% CI, 0.77 to 1.46; p=0.714; median 10.9 and 11.0 months, respectively).
This test should be used in the second-line setting in EGFR wild-type patients, and patients with a proteomic status of poor should not receive second-line erlotinib.
One challenge in performing molecular profiling of NSCLC is the need for sufficient tumor tissue, and patients may undergo repeat biopsy to determine EGFR mutation and ALK rearrangement status at time of diagnosis or at the time of disease progression to detect resistance mutations (e.g., EGFR exon 20 T790M). There has been increased interested in developing “liquid” biopsies using peripheral blood samples.7,8 The optimal technique for liquid biopsies and clinical situation is yet to be determined. There is also increased use of immunohistochemistry (IHC) for the detection of ALK rearrangements as an alternative to FISH testing.9
These techniques of assessing the molecular characteristics of NSCLC may become widely available in the near future.
EGFR Mutation and ALK Rearrangement-negative NSCLC
A phase III trial compared docetaxel with ramucirumab, a monoclonal antibody against vascular endothelial growth factor receptor 2, or placebo in patients who had progressed after first-line platinum-based therapy.10 Patients were not restricted based on histology, and the primary end-point was OS. Patients assigned to the ramucirumab compared with placebo arm experienced a statistically significant longer OS (HR of 0.86, 95% CI, 0.75 to 0.98; p=0.023; median 10.5 and 9.9 months, respectively), PFS (HR of 076, 95% CI, 0.68 to 0.86; p< 0.0001; median 4.5 and 3.0 months, respectively) and ORR (23% vs. 14%, P<0.0001).
Common grade ≥ 3 adverse events in the ramucirumab and placebo arms were neutropenia (49% vs. 40%), febrile neutropenia (16% vs. 10%), fatigue (14% vs. 10%), and hypertension (6% vs. 2%). The rate of grade 3 pulmonary hemorrhage was similar (1% in both arms). The global quality-of-life analysis showed that the time to deterioration did not differ between the treatment arms. Approximately a quarter of patients enrolled had squamous histology and the rates of gastrointestinal tract bleeding and pulmonary hemorrhage did not differ according to histology.
A phase III trial compared docetaxel with nintedanib or placebo in patients who had experienced disease progression after first-line platinum-based therapy. The primary end-point was PFS by IRR, and OS was a secondary end-point.11 Patients assigned to the nintedanib compared with placebo arm experienced a statistically significant longer PFS (HR of 0.79, 95% CI, 0.68-.02; p=0.0019; median 3.4 and 2.7 months, respectively), and similar OS (HR of 0.94, 95% CI, 0.83 to 1.05; p=0.272; median 10.1 and 9.1 months, respectively). A statistically significant longer OS was observed in patients with adenocarcinoma histology (HR of 0.83; 95% CI, 0.70 to 0.99; p=0.0359; median 12.6 and 10.3 months, respectively).
The future development of nintedanib will be in patients with adenocarcinoma histology.
Ramucirumab or nintedanib are currently not approved by the FDA, a biomarker to select patients likely to benefit is not available for either agent, and the OS benefit observed is modest.
Small Cell Lung Cancer
First-line therapy for extensive stage SCLC is platinum-etoposide and prophylactic cranial irradiation, but patients frequently have persistent intra-thoracic disease or experience intra-thoracic disease progression. A phase III trial investigated the role of thoracic radiation therapy (TRT) of 30 Gy in 10 fractions) compared with no TRT.12 Patients were required to have responded to chemotherapy and have a performance status of 0 to 2. The primary end-point was the one-year OS rate. The one-year OS rate was not significantly different between the TRT and no-TRT groups (HR for OS of 0.84, 95% CI, 0.69 to 1.01; p=0.066 and 1-year OS rate of 33% and 28%, respectively).
A secondary analysis revealed an improvement in the two-year OS rate in TRT compared with the no-TRT group (13% vs. 3%, p=0.004). The most common grade 3 adverse events in TRT were fatigue 4.5%, esophagitis (1.6%), dyspnea (1.2%), insomnia (1.2%), and headache (1.2%). This trial suggests that the benefit of consolidation TRT after first-line therapy is limited.
Looking Ahead to 2015
While it is difficult to anticipate the future developments for lung cancer, especially given the increasing pace or target identification and drug development, additional data on several novel agents are highly anticipated in 2015. Preliminary data demonstrating the activity of “third generation” EGFR TKI in patients who had progressed on EGFR TKI and had an exon T790M resistance mutation were presented in 2014. These agents spare the EGFR wild-type receptor and have a low rate of typical EGFR-associated toxicities of rash and diarrhea.
Several checkpoint inhibitors targeting the PD-1/PD-L1 interaction have revealed promising results in heavily treated patients with NSCLC. Several checkpoint inhibitors are being studied in different disease settings, and these agents may have a role in a broad population of patients with NSCLC. These two classes of agents have the potential to immediately impact the care of patients with advanced NSCLC.
1. Seto T, Kato T, Nishio M, et al. Erlotinib alone or with bevacizumab as first-line therapy in patients with advanced non-squamous non-small-cell lung cancer harbouring EGFR mutations (JO25567): an open-label, randomised, multicentre, phase 2 study. Lancet Oncol. Oct 2014;15(11):1236-1244.
2. Janjigian YY, Smit EF, Groen HJ, et al. Dual inhibition of EGFR with afatinib and cetuximab in kinase inhibitor-resistant EGFR-mutant lung cancer with and without T790M mutations. Cancer discovery. Sep 2014;4(9):1036-1045.
3. Solomon BJ, Mok T, Kim DW, et al. First-Line Crizotinib versus Chemotherapy in ALK-Positive Lung Cancer. N Engl J Med. Dec 4 2014;371(23):2167-2177.
4. Shaw AT, Kim DW, Mehra R, et al. Ceritinib in ALK-rearranged non-small-cell lung cancer. N Engl J Med. Mar 27 2014;370(13):1189-1197.
5. Shaw AT, Ou SH, Bang YJ, et al. Crizotinib in ROS1-rearranged non-small-cell lung cancer. N Engl J Med. Nov 20 2014;371(21):1963-1971.
6. Gregorc V, Novello S, Lazzari C, et al. Predictive value of a proteomic signature in patients with non-small-cell lung cancer treated with second-line erlotinib or chemotherapy (PROSE): a biomarker-stratified, randomised phase 3 trial. Lancet Oncol. Jun 2014;15(7):713-721.
7. Newman AM, Bratman SV, To J, et al. An ultrasensitive method for quantitating circulating tumor DNA with broad patient coverage. Nat Med. May 2014;20(5):548-554.
8. Oxnard GR, Paweletz CP, Kuang Y, et al. Noninvasive detection of response and resistance in EGFR-mutant lung cancer using quantitative next-generation genotyping of cell-free plasma DNA. Clin Cancer Res. Mar 15 2014;20(6):1698-1705.
9. Wynes MW, Sholl LM, Dietel M, et al. An international interpretation study using the ALK IHC antibody D5F3 and a sensitive detection kit demonstrates high concordance between ALK IHC and ALK FISH and between evaluators. J Thorac Oncol. May 2014;9(5):631-638.
10. Garon EB, Ciuleanu TE, Arrieta O, et al. Ramucirumab plus docetaxel versus placebo plus docetaxel for second-line treatment of stage IV non-small-cell lung cancer after disease progression on platinum-based therapy (REVEL): a multicentre, double-blind, randomised phase 3 trial. Lancet. Aug 23 2014;384(9944):665-673.
11. Reck M, Kaiser R, Mellemgaard A, et al. Docetaxel plus nintedanib versus docetaxel plus placebo in patients with previously treated non-small-cell lung cancer (LUME-Lung 1): a phase 3, double-blind, randomised controlled trial. Lancet Oncol. Feb 2014;15(2):143-155.
12. Slotman BJ, van Tinteren H, Praag JO, et al. Use of thoracic radiotherapy for extensive stage small-cell lung cancer: a phase 3 randomised controlled trial. Lancet. Sep 14 2014.
Tuesday, December 16, 2014
BY MEERI KIM, PHD
Routine surveillance imaging typically adds little to the detection of relapse for patients with diffuse large B-cell lymphoma (DLBCL), a new analysis has found (JCO 2014;32:3506-3512).
As reported by the team led by Carrie A. Thompson, MD, Consultant in the Division of Hematology and Assistant Professor of Medicine at the Mayo Clinic, for a cohort of patients from the Mayo Clinic/University of Iowa, the majority of relapses instead were identified ahead of their scheduled follow-up visit. In total, only 1.6 percent of patients observed after therapy had asymptomatic DLBCL relapse detected by surveillance imaging before clinical manifestations.
The analysis also revealed no differences in overall survival between individuals who had their relapse detected outside of a scheduled follow-up visit versus those whose relapse was found at a planned visit. A second, independent cohort of patients from France showed similar results.
These results go against the recommendation of current guidelines, and calls the utility of such routine imaging into question, the authors noted. Over-imaging has risks that should be minimized, including radiation exposure, cost, patient anxiety, and false positives that can lead to more testing.
Because up to a quarter of DLBCL patients in remission will relapse during observation, careful observation is necessary after treatment. But the frequency of imaging administered by clinicians can vary widely.
“The recommendations used to be very frequent follow-up including scans,” Thompson said in an interview. “The guidelines have changed over time, but some people are still practicing the older guidelines, and some of us have been confused about what is the best way to follow these patients.”
The current National Comprehensive Cancer Network guidelines suggest computed tomography scans no more than every six months, and then only as clinically indicated. However, a 2012 study of DLBCL patients at seven comprehensive cancer centers discovered that the median number of imaging studies after remission was actually 2.5 per year Published in Leukemia and Lymphoma (Abel et al: Leukemia & Lymphoma 2012;53:1113-1116).
“We’ve realized that this is indeed the case--that we scan too much in this country,” said Radhakrishnan Ramchandren, MD, Assistant Professor in the Department of Oncology at the Barbara Ann Karmanos Cancer Center, who was asked his perspective. “We do PET scans when they’re not indicated, and we do CT scans much more than necessary.”
He stresses that clinicians at Karmanos do not make a habit of using follow-up scans for DLBCL and other lymphoma subtypes. Instead, Ramchandren and his colleagues use imaging only on a case-by-case basis.
Two Large, Independent Cohorts
Previous studies had also demonstrated that most DLBCL relapses occur outside of the timeframe of a scheduled visit, and thus imaging may be of limited utility for preclinical relapse detection. Thompson and her colleagues wanted to build on them with their own investigation of two large, independent cohorts.
Between 2002 to 2009, a total of 680 DLBCL patients from the Molecular Epidemiology Resource of the University of Iowa/Mayo Clinic Lymphoma Specialized Program of Research Excellence (SPORE) were prospectively enrolled in one study cohort. That group was then narrowed down to those who achieved remission (552 out of 680 patients, 81%), and then later suffered a relapse (112 of 552 patients, 20%).
“Our goal was to see how the relapse was detected,” Thompson explained: “Did they have a scan at the time? Were symptoms reported in the chart? Was it a routine scheduled follow-up visit or in between follow-up visits?”
The majority of DLBCL relapses were detected outside of a scheduled visit (67 out of 104 patients with available records, 64%). This means that most relapses are caught in response to patient-reported symptoms, which then lead to abnormal physical exams or laboratory studies.
In total, only nine of 552 patients (1.6%) had their DLBCL relapse first detected by surveillance imaging. For a separate cohort from the Leon Berard Cancer Center in France, the rate was 4 out of 222 patients (1.8%).
Thompson said she believes that these results should be taken into consideration by those who make guidelines for how to follow DLBCL patients after remission: “Given that the majority of patients presented with a relapse outside of the follow-up visit, patient education is very important for them to understand signs and symptoms that they should be watching for,” she said.
The study was supported by funding from the Lymphoma SPORE, the Predolin Foundation, Mayo Clinic Robert D. and Patricia E. Kern Center for the Science of Health Care Delivery, and the Arnold and Kit Palmer Benefactor Award.
Signs to be Aware of
Signs to be aware of for DLBCL patients include enlarged lymph nodes, drenching night sweats, unexplained weight loss, pain (especially if the original disease presented with pain), and unexplained fevers.
Ramchandran also emphasizes vigilance when assessing patients for relapse. He said he sees his patients every three months for the first two years after remission--each time performing a thorough physical examination, discussing clinical signs and symptoms, and doing blood work.
Caveats from Andrew Zelenetz
Also asked his opinion on the study, lymphoma specialist Andrew D. Zelenetz, MD, PhD, Vice Chair of Medical Informatics at Memorial Sloan Kettering Cancer Center, highlighted the limitations of retrospective analysis and some methodological issues. For instance, he felt the authors did not give the imaging scans enough credit.
“Yes, they had both symptoms and the scan, but the scan showed the disease,” he said. “It may not have been with significant lead time before the symptoms, but these scans were positive.”
“It may be that those doctors assessing patients with a scan suggestive of relapse were more likely to either elicit or document symptoms,” commented Gregory A. Abel, MD, MPH, Assistant Professor of Medicine at Dana-Farber Cancer Institute.
Abel added that surveillance imaging is unlikely to go away completely, or at least until a new technology replaces it beyond the assessment of symptoms, abnormal physical exam, and blood testing. That new method of detecting relapses could be immunoglobulin gene sequencing. A multicenter, prospective trial is currently underway to investigate this technique in patients after remission versus CT, although results won't be available for another two to three years, he noted.
As the NCCN Guidelines Panel Chair for Non-Hodgkin’s Lymphomas, Zelenetz confirmed that the data will be reviewed at the next panel meeting this summer--but he said he doesn't foresee a change in the imaging recommendation. Also, in terms of wide variation of imaging frequency, he puts the responsibility on the clinician: “We work very hard to update these guidelines. Maybe not everyone is following the guidelines, but that’s not the fault of the guidelines--that’s the fault of the practitioner.”
Tuesday, December 16, 2014
By Kurt Samson
When used with current risk parameters, a targeted transcriptome panel assay could help identify postcystectomy at-risk muscle-invasive and/or node-positive bladder cancer patients, according to new data.
In the collaborative study, published in the Journal of the National Cancer Institute (doi: 10.1093/jnci/dju290) and led by investigators at the University of Southern California, the University of British Columbia, Stanford University, and GenomeDx Biosciences, transcriptome-wide expression profiles on archived tumors from 225 bladder cancer patients were examined, linking 15 specific transcript profiles to increased cancer recurrence risk and associated mortality.
“With this panel it might be possible to spare many patients who may not be at imminent risk for cancer recurrence following cystectomy from the toxicity associated with chemotherapy, while identifying others who need more aggressive management,” said the lead author, Anirban P. Mitra, MD, PhD, Assistant Professor of Medical and Molecular Genetics in the Department of Pathology and Center for Personalized Medicine at the University of Southern California.
“Compared with any of the other standard clinical metrics currently used to calculate risk, we found this does a better job,” he said in an interview. “It provides a different perspective, and when combined with these other prognostic tests it significantly increases accuracy.”
Genomic and clinical classifiers (GC and CC) for predicting postcystectomy recurrence were initially developed on a set of 133 patients. The median age of patients was 68.5, and there was a median follow-up of 9.3 years. In all, 51 percent of the patients had recurrence of their cancer within that time and almost 60 percent had died.
Performances of GC, CC, International Bladder Cancer Nomogram Consortium (IBCNC) data sets, and genomic and clinicopathologic classifiers (G-CC, G-IBCNC) were assessed in discovery and independent validation sets, while GC was further validated in four external datasets. Discrimination and prognostic abilities of these classifiers were then compared using area under receiver-operating characteristic curves (AUCs).
The team developed a15-feature genomic classifier on the discovery set, with an AUC of 0.77 in the validation set, which was higher than any individual clinical variable, and was comparable to CC (AUC = 0.78). As the researchers reported, accuracy was improved by combining GC with clinical outcome prediction nomograms, with the G-CC model having the highest predictive accuracy. GC also performed the best in validation compared with seven prior signatures and was more prognostic across four independent datasets.
Following normalization and feature selection, 15 markers were identified corresponding to RNAs from protein-coding and noncoding regions of the genome that were differentially expressed based on recurrence.
GC had an AUC of 0.88 in the discovery set; and as GC was optimized in the discovery set, it could not be compared with individual clinicopathologic variables. “We could, however, assess the prognostic value of adding genomic information to clinical risk-prediction models,” Mitra said.
When added to G-IBCNC, the AUC increased from 0.73 to 0.89, and when added to G-CC, it increased from 0.81 to 0.93. “We found that these genomic-based classifiers outperformed clinical models for predicting postcystectomy bladder cancer recurrence,” he said.
Clinical parameters are currently used to select candidates for adjuvant therapy, but they are of limited predictive power, he noted. “To our knowledge this represents the largest effort to date to discover and validate a prognostic genomic signature for clinically high-risk urothelial cancer of the bladder. This may be due to our use of a large cohort, clinically relevant endpoints, and high-density transcriptome-wide expression profiling.”
Standardized patient management and profiling using formalin-fixed paraffin-embedded tissues also make the signature more clinically applicable, although further studies are needed to better characterize the prognostic potential of the markers identified.
“Our results suggest that the combination of these markers and clinicopathologic parameters may be prognostic for clinically high-risk bladder cancer,” Mitra said.
He noted that clinically high-risk patients in prior studies have not always received consistent surgical management, and prognostic signatures have often been generated toward non-disease-specific outcomes such as overall survival. Previous studies have also been hampered by scarce frozen tumor specimens and typically used low-density platforms that would profile only select protein-coding transcripts.
“This panel is very specific for patients with muscle-invasive tumors, while other researchers have looked at tumors across all stages,” Mitra said. “What we found is that with each dataset, the panel proved to be extremely prognostic, so we have five validation sets that comprise over 400 patients to date.”
He added that the chip used is the widely available Affymetrix array, which costs about $500, while the test runs about $850.
‘One of Most Expensive Cancers to Treat’
Mitra called bladder cancer one of the most expensive cancers to treat, with outcomes of patients with invasive disease often poor. In fact, postcystectomy recurrence of bladder cancer is now fatal in approximately 85 to 95 percent of patients.
Regarding limitations, he said, the study was not able to further evaluate genomic classifiers due to the lack of public clinically annotated Human Exon array-profiled bladder cancer datasets, but the markers that were found could be mapped using other profiling platforms for assessment of performance. “We expect that our findings will provide a better understanding of the molecular alterations involved in aggressive bladder cancer and how best to manage patients based on their genomic profiles,” he said.
Asked for his opinion for this article, Dan Theodorescu, MD, PhD, the Paul A. Bunn Professor of Cancer Research and Director of the University of Colorado Cancer Center, said: “This is a very important step forward because of the large number of patients and the fact that analysis was conducted on formalin-fixed [specimens], which is what we use. It is also highly clinical and personal, and the data was benchmarked against two nomograms on postcystectomy recurrence risk from the International Bladder Cancer Nomogram Consortium. Existing genomic data is helpful, but not as much data in the nomogram.”
The technique could potentially be used to differentiate at-risk patients from high-risk individuals, although this needs further analysis and needs to be validated in a prospective clinical trial, Theodorescu added.
The cost of conducting such a test is unknown, but would be higher than just the price of the array chip. “That’s a loaded question, but while the cost of the chip might not be that high, when added to labor costs it could be much higher--you need experienced readers to interpret the results, for instance.”
Theodorescu pointed to a randomized study that he and his colleagues conducted, published in Lancet Oncology (2011;12:137-143) on the first molecular test to predict which bladder cancer patients may have cancer involvement in their lymph nodes at the time of surgery.
The test, which analyzes 20 genes on tumor biopsies, could help oncologists flag which patients would be good candidates for pre-surgical, or neo-adjuvant, chemotherapy. That prospective randomized trial evaluated tumor samples from the United States, Canada, and Germany and was funded by the National Cancer Institute.
“We validated the test’s ability to predict lymph node spread in a large sample of patients, and the predictive ability held up,” Theodorescu said. “If confirmed in additional clinical trials, the test could be used to help patients with positive nodes live longer while keeping node-negative individuals from overtreatment.”