Though there is an emphasis in modern conventional medicine on hard evidence to support our clinical management of patients, and this focus on “evidence-based medicine” is probably especially strong in oncology (and with GRACE), every oncologist falls somewhere on a spectrum between being rigidly focused on following exactly what the evidence supports vs. being completely guided by a combination of a good rationale and some intuition.   Patients and families also have very different expectations about whether their doctor/medical team is very driven by “current best practices” or more inclined toward “outside the box thinking”.  Complementary and alternative medicines are not often accompanied by hard evidence, but their exceptional popularity speaks to the fact that many people don’t feel that they require heaps of evidence to act.  Meanwhile, it typically takes years for the cutting edge research concepts of today to become the standard of care later: carefully conducted studies take a lot of time, and in the meantime, we all struggle with which ideas to adopt and which ones to wait on until we learn more. 

    Neither approach in itself is always the right one.  We’ve seen cases where giving an EGFR inhibitor after chemo and radiation for locally advanced NSCLC can be harmful for many patients, despite this being an unexpected finding at the time (and still not something we can explain).  We’ve seen that integration of Avastin (bevacizumab), a targeted agent that improves survival in many patients with advanced lung cancer, into chemo/radiation treatment for stage III NSCLC or limited disease SCLC can lead to dangerous or even fatal side effects that weren’t anticipated.  But there’s no doubt that some patients today could benefit from what will ultimately prove to be the standard therapies of tomorrow (or at least of 2014).  We just aren’t terrific at predicting.

    Earlier this week, I presented my general framework for using or not using molecular markers to guide my clinical recommendations, and this represents a setting in which new data are exciting but precede more conclusive results by years.  I explained that my personal approach is to not use markers like ERCC1, RRM1, and thymidylate synthase levels to shape my treatment decisions for patients, because they haven’t been proven to be helpful in large prospective trials.  Instead, they’ve mostly looked like they could be helpful based on retrospective assessments in which we pick through the data after the study has been completed, which isn’t as helpful as planning to use a variable and then using it to guide treatment prospectively.

    Yet, less than 24 hours after I presented this view, I found myself offering a second opinion in which I recommended that we send tumor tissue from a 30 year-old man with NSCLC for molecular marker testing to help us consider what chemo to pursue.  Why be such a hypocrite?  In his particular case, he was diagnosed in the late fall with a stage IIIA N2 node-positive NSCLC and had received cisplatin-based chemo and concurrent radiation from another oncologist, followed by surgery.  Unfortunately, he had a small amount of residual viable cancer in a few mediastinal lymph nodes, a situation that suggests a less favorable prognosis than if the mediastinal nodes were all completely negative for cancer (mediastinal sterilization). 

   This was a case in which we simply have no good data (and by that I really mean any data), to tell us how we should proceed.  The trials that have looked at this situation have generally tried to give a couple of additional post-operative cycles of the same chemo that was given pre-operatively.  But there has never been a trial that compared continuing the same treatment vs. trying a to kill the cancer by shooting from a different angle and using different chemotherapy.  And in this situation, without any good reference point for guidance, I’d be happy to use imperfect data to help inform my decision.  If the molecular marker studies suggest that gemcitabine or more or alimta or more platinum is a very good or poor choice, I’ll take that hint, because it’s not overriding a standard of care here.  Similarly, I feel that it’s pretty reasonable to use ERCC1 status to help make a decision about the value of adjuvant chemo for resected NSCLC if a patient is otherwise on the fence about whether it’s worth doing or not, but I’m not comfortable using that data to steer someone away from it if all other indicators suggest that it’s the appropriate thing to do.  In other words, I think the threshold to use tentative, imperfect data can be lower for a situation in which there is no right answer than in a situation in which you’d potentially be thumbing your nose at a well-established, evidence-based standard of care.

   This is really also the way I feel about alternative medicine approaches.  I would hate to see someone dismiss surgery for early stage lung cancer or an EGFR inhibitor if they have an EGFR mutation because they would rather pursue alternative medicines.  On the other hand, it’s certainly very appropriate for people to pursue whatever avenue for treatment concepts they’d like to follow if conventional oncology approaches can’t offer much promise.

   To sum it up: “In the land of the blind, the one-eyed man is king.”   Though it’s ideal to have a solid path you know is right, reality isn’t always that convenient, and it makes sense to just do the best you can under the circumstances in an ambiguous situation.   And my own general recommendations are still adapted to the individual person in front of me, as it should probably always be. 

   As always, I welcome your thoughts.

No related posts.

Related posts brought to you by Yet Another Related Posts Plugin.

As always, you’ll find below the audio and video versions of the podcast, along with the transcript and figures.

dr-manning-pitfalls-in-cancer-imaging-audio-podcast

dr-manning-pitfalls-in-cancer-imaging-transcript

dr-manning-pitfalls-in-cancer-imaging-figures

It’s hard to believe that with all of our advances in imaging we still can’t reliably determine the cause of the lesions we see, but even the improvements in the field leave us making educated guesses and relying on follow-up over time, along with judicious use of biopsies, to help clarify what’s going on.  I hope this is helpful.

No related posts.

Related posts brought to you by Yet Another Related Posts Plugin.

I receive and have no plans to receive payment of any kind, including salary, honorarium or research funding from Veridex.

What are circulating tumor cells (CTCs)?  Leukemias as a case in point
Circulating tumor cells, or CTCs are just what they sound like: cancer cells that circulate in the bloodstream alongside normal cells such as white blood cells, red blood cells, and platelets. The classic examples of circulating tumor cells are the leukemias, or blood cancers.

As every criminal starts his or her life as a baby then crosses the line into criminality somewhere along the way, every cancer cell starts its life as a normal cell.  Leukemia cells start as bone marrow cells.  The normal job of bone marrow cells is to provide blood cells, so it’s no surprise that leukemia, by its very nature, circulates in the blood.  This gives leukemia a unique property—it can be biopsied every day with extremely minimal risk to the patient, because the biopsy entails just a simple blood draw.  For this reason, leukemias were the first cancer to have detectable CTCs — the technique is called flow cytometry.

Flow cytometry works by separating the cells in the blood by size, the scatter of light (which in turn measures the complexity of the nucleus) and cell surface markers (they’re mostly named CD-something).  You get plots of each cell seen and if they don’t look normal, you’ve not only identified the cancer, but you’ve learned what surface makers it bears.  This test is a routine part of sub-typing leukemia; an example of flow results from a patient with M1 leukemia is shown below.  Leukemias, like lung cancers, have molecular tests that help determine treatment (for example, BCR-Abl for CML, JAK2 for the myeloproliferative disorders, and FLT-3 for AML) and they are routinely tested on blood because of the large quantity of CTCs found there.

(click on image to enlarge)

 The theory of CTCs in solid (non-leukemia / non-lymphoma) cancers
We’ve believed that CTCs are sometimes present in other cancers for a long time.  In the process of going from a local cancer to a metastatic one, cancer spreads to lymph nodes then to distant sites.  We have long believed that these cells travel through blood to get to distant sites.  Therefore, the idea that we might one day be able to detect these cells is not new.  There are two main reasons to care about detecting these cells: to count them, and to characterize them.  Achieving both of these aims could potentially improve patient care and speed research.  If CTCs are detected early in chemotherapy for metastatic disease, and the chemo is successful, then we would expect to find fewer (or no) CTCs in the blood.  Perhaps through this method, CTCs could allow earlier detection of progression.  Earlier detection of progression might allow for greater safety in chemotherapy breaks.  When the sad day of progression happens, you’d rather catch it early as an increase in CTCs than pain in your belly because of liver mets.  Maybe CTCs can even decrease the number of CT scans needed, decreasing radiation exposure, increasing patient convenience, and decreasing the cost of care.

Perhaps the presence or absence of CTCs can help predict who will benefit from adjuvant chemotherapy.  Maybe patients with CTCs detected before surgery have a higher risk of micrometastatic disease and perhaps they benefit more from adjuvant chemotherapy.  Finally, if we can study the CTCs, perhaps patients can have full molecular characterization of their cancer without the risk and discomfort of a biopsy.  If the latter goal were achieved, a key barrier in lung cancer research would also be lifted — we would always have “biopsies” for molecular correlate studies to better understand how novel therapies are affecting cancer cells and to target which patients benefit most from which therapies.

Technical issues in CTC detection for solid malignancies
Most of the cells circulating in the blood in a patient with a “solid” cancer such as lung cancer are normal blood cells, particularly white blood cells.  Because of this issue, flow cytometry has trouble detecting the relatively small numbers of CTCs against this large background. That’s why the technology for seeing the CTCs is new, even though the idea that they were there is older. The basic idea is to exploit what is unique about CTCs as compared to the larger numbers of white blood cells surrounding them. Non-cancerous epithelial cells don’t circulate in the blood; in contrast, cancers such as lung, prostate, breast, and colon cancer that are derived from them do. These cancer cells tend to bear the same cytokeratins that their non-cancerous ancestors do, they bear a nucleus (as do the large number of surrounding white bloods cell) and (unlike white blood cells) do not bear CD45.  The basic idea is depicted graphically below:

The state of CTCs in other cancers
The cellsearch technology is FDA approved in breast, prostate, and colon cancer.  The most promising data for the value of counting CTCs are from an article in the Journal of Clinical Oncology from November, 2009.  The authors observed 68 patients being treated for breast cancer.  They collected blood from these women every 3-4 weeks and re-imaged them every 9-12 weeks.  They then looked to see how well CTC assessment correlated with progression-free survival, with a “positive” CTC result defined as > 5 cells detected and a “negative” result as <5 cells detected.  The test had much better sensitivity than specificity:

The CTC test had a strong discriminatory power for progression-free survival:

The authors continue to observe these patients for survival, intending to present data on the correlation between CTC findings and overall survival once the data are sufficiently mature.

CTCs have also been studied in colon and prostate cancer, where the technology is also FDA approved. I chose to focus on the breast results because I think that they are the most advanced and best show what we may be able to achieve in lung cancer. We’ll turn to CTCs in lung cancer next.

No related posts.

Related posts brought to you by Yet Another Related Posts Plugin.

To answer this question themselves, the investigators send out questionnaires that they received back from over 4,000 physicians, non-cancer physicians, medical oncologists,  surgeons, and  radiation oncologists.  Specifically, they asked physicians when they would bring up the issues of prognosis, DNR preferences, hospice, and preferred location of death for a hypothetical patient with an estimated survival of 4-6 months but who felt well now.  Options ranged from the time of the initial visit (”now”) to a point in the future when the patient was symptomatic, or there were no further feasible treatments, or when a patient was hospitalized, or not until the patient and/or family brought up these issues.  The investigators also asked about physicians’ experience with managing terminally ill patients and how confident they were in their knowledge about managing terminal care issues for patients.

Overall, 65% of physicians reported that they would discuss prognosis from the beginning, with 15% on the other extreme of not discussing prognosis until or unless the patient brings it up.   For discussion of DNR status,  44% favored discussing DNR preferences at the initial visit, with the remainder scattered across the milestones.  Discussion of hospice was most commonly reported as something to defer until there were no further potentially effective palliative treatments available.  Least likely to be discussed “now” was preferred location of death, which 21% of physicians saying they’d bring up from the initial visit, and 24% not discussing this until the patient brings up the topic.

There were also some differences among the physicians surveyed.  Younger physicians were more likely to be proactive about initiating these discussions earlier.  Medical oncologists and surgeons were significantly more likely to bring up prognosis than other physicians.   Not surprisingly, the physicians who reported more experience and greater confidence about end-of-life care brought the issues up earlier than others.

The results are interesting and overall show a lot of personal differences, in keeping with the evidence as well as what I think most of us see in the real world.  Some physicians feel that this is part of appropriate care for terminally ill patients, since studies show that physicians are not especially good at assuming correctly the preferences of a patient for issues like DNR status.  Other physicians are concerned that discussing end-of-life care dispels hope and sets a negative tone and evade these issues as long as possible.

I personally feel that it’s important for patients to have some real understanding of their prognosis if they are really going to be equipped to make an informed decision about treatment.  Someone who is agreeing to start a potentially challenging regimen of two or three agents for advanced lung cancer should know the limitations of what treatment can offer, and specifically that we aren’t giving it with an intent or realistic expectation that it will be curative treatment.  On the other hand, I’ve also become more nuanced and try to read my patients in terms of what they’re receptive to discussing, compared with being more universally proactive in my earlier years in practice.  And even though I’m definitely among the oncologists at my institution least likely to avoid these discussions by the time they become pressing issues, I’m certainly not someone who routinely discusses hospice and a patient’s preference for where they would prefer to have their death occur on an initial visit.

But is that wrong?  My hospital now requires that we specify a preference for “Full code” or DNR for every patient admission, but I know that we physicians don’t always have these discussions with patients in whom they don’t envision this to be an urgent issue (and they often still don’t when it really will be an immediate issue).  The guidelines suggest that doctors initiate these discussions from the very beginning if a patient has a prognosis of less than a year, even if they feel well.  I personally don’t feel that I should, based merely on guidelines from an expert panel that are based on no particularly strong evidence that this is more helpful than waiting until you’ve established a relationship with a patient and end-of-life issues may be relevant in the near future.  Some of the physicians who responded that they don’t pursue these discussions in an early visit probably feel the same way, though I’m sure other physicians just don’t want to have a discussion that can be awkward and, in some patients, may alienate or offend them.

Over time, I’m coming to feel that our approach to these issues should be individualized, just like we individualize our medical approach based on patient health, performance status, molecular markers, preferences about aggressiveness, and many other factors.  Some patients may take longer than others to reach a point when they can discuss these issues, and I think that’s OK and even quite appropriate. To me, using guidelines to direct the timeline for having these discussions is too “color-by-numbers” in a time when people need individualized management.

I think it’s very important to hear from people on the other side about what patients and caregivers would be receptive to, or whether they’d prefer to bring these issues up and not have these topics initiated by their physicians.

No related posts.

Related posts brought to you by Yet Another Related Posts Plugin.

Increasingly, though, links between clotting and cancer are emerging. As I result, I find myself going back to something that I found a lot less interesting during my training than I do now: the clotting cascade.

(click on image to enlarge)

Efficient clotting has been essential to our survival: without it, we would risk fatal bleeding with every scrape. Defects in this system lead to the disease hemophilia (Greek for loves to bleed), a disorder characterized by prolonged bleeding after minor trauma. Advances in the treatment of this disease have been a direct result of knowing which clotting factor (denoted by the roman numerals in the figure) is missing and then replacing it.

In cancer, we are usually dealing with excess clotting rather than bleeding; unfortunately, the reasons for this link aren’t as simple as an excess of a single factor. We have known for a long-time that blood clots are associated with cancer. In fact, patients with an unexplained blood clot are up to ten times more likely to harbor a cancer than those with no clots or someone with a blood clot and a known cause for it. Increasingly, we are realizing that this is more than just a casual association: it appears that the clotting cascade may facilitate the growth of cancer even as cancer fuels clotting.

One of the end-products of the cascade is a protein called fibrin, which links together to form a tough mesh. Very useful for clot formation and for cancer survival, it turns out. This mesh may provide a nest that protects cancer cells as they grow and invade. Cancer researchers are often amazed that cancer cells are able to detach and make the trip through the bloodstream, constantly in danger of destruction by our body’s immune cells. Then, they have to arrive in a hostile, new tissue and somehow establish a blood supply. Fibrin may help by surrounding and protecting these fragile migrating cancer cells, preventing immune detection and attack. Fibrin may also draw new blood vessels to cancer cells and provide a platform on which the endothelial cells that form blood vessels can grow. Interestingly, cancers implanted into mice deficient in fibrin were unable to form metastases. Key players in the clotting cascade, such as tissue factor (TF) may also directly stimulate pathways involved in cancer growth and survival.

So, if the clotting process drives cancer growth, can we help patients with cancer by inhibiting the clotting process? Currently, we use anticoagulants to treat cancer patients who have developed a blood clot. Low molecular weight heparins such as Lovenox or Fragmin (both administered by subcutaneous shots) are preferred over Coumadin (oral) because of increased effectiveness at preventing recurrent clots in cancer patients. The first question is, given the high risk of blood clots, should we be giving prophylactic anticoagulation to all cancer patients? The recently completed PROTECHT trial randomized patients with solid tumors who were undergoing chemotherapy to either placebo or the injectable anticoagulant nadroparin. The study found that patients on the anticoagulant were half as likely to develop a blood clot. But the numbers who developed clots were small, only 4% in the placebo group compared to 2% in the nadroparin group.

The results for the lung cancer patients in the trial were more striking, with a 9% risk of developing a clot in the placebo group compared to 3.5% in the treatment group. Serious side effects were no more common with nadroparin than placebo. There was no difference in survival between the two groups though the authors note that patients remained on nadroparin for a relatively short time (median 4 months) which might make it difficult to see an effect on survival.

Other investigators have suggested that, independent of the reduction in blood clots, anti-coagulants may have a direct anti-cancer effect. Some of the most intriguing results have come from studies in small cell lung cancer (SCLC), a disease where we can definitely use some new ideas. In one small trial, patients with SCLC were randomized to chemotherapy or chemotherapy plus the anticoagulant dalteparin (Fragmin). Those treated with anticoagulation and chemo had a median survival of 13 months compared to 8 months for those on chemo alone. Even though only 84 patients were included, the results were statistically significant (p = .01).

I was thinking about all of this when I happened to hear a fascinating presentation at a recent meeting in Boston. Dr. Caroline Dive was presenting her work on circulating tumor cells and she discussed some interesting findings in patients with SCLC. She described discovering large numbers of tumor microemboli in a patient who had a particularly aggressive tumor. I wondered if this may be a key to the early dissemination of SCLC and the difficulty of curing this disease. If SCLCs are particularly good at making fibrin, this may explain both their propensity for dissemination and why anti-coagulants might have an effect on survival in this disease.

Part of making progress in cancer involves improving our supportive care. Certainly, we have been able to help patients by using drugs to reduce the risk of neutropenia or severe nausea with chemotherapy. Chemotherapy may actually contribute to the pro-clotting environment in cancer patients. Given the results of the PROTECHT study for lung cancer, I’m wondering if we should add prophylactic anticoagulation to our arsenal of supportive care. At the very least, I think that the issue of anticoagulation in lung cancer patients deserves further study and attention, even at the risk of having to learn the clotting cascade all over again.

No related posts.

Related posts brought to you by Yet Another Related Posts Plugin.

Most oncologists, including those offering insights here, probably seem very conservative in their responses when patients and caregivers come into the office or write in with breathless excitement about the latest report in the mass media about the newest treatment  (see Dr. Weiss’s post on “How to Vet a Treatment Idea” for a balanced perspective).  Yes, the person on 60 Minutes (or the 11 o’clock news, or Oprah, or whatever) said it was incredibly promising, but the regular reports of miracle treatments have been far more likely to be mirages than revolutionary treatments.

I give credit to the authors of a recent article in the Journal of the National Cancer Institute for highlighting the problem of how the media handle reporting about cancer.   It describes some of the reasons why far too much new information is described in extremely sensational terms.  This is a real pet peeve of mine, because I hate the idea that the mass media willfully generate more viewership or readership by whipping up a frenzy of danger or false hope (and I’ve thought that GRACE should announce the  ”winner” of an annual GRACIE award for the most socially irresponsible and sensational report of the year).

The authors note that willful misrepresentation is part of the issue, but part of the issue is also the fact that many journalists don’t have the knowledge needed to put the work into the proper context.  Many of our reports only include reference to relative risk or benefit (30% higher risk of developing cancer, or 20% improvement in survival) without noting the effect in absolute terms.  When journalists just echo what is put out in a press release that is deliberately ambiguous, it’s the fault of the investigators or people putting out the press release to highlight a “20% improvement” without noting the limitation that this may amount to an improvement of only 3 weeks in absolute terms, or that a doubling of risk of developing a rare cancer may mean that the risk escalates from 0.1% to 0.2% over a person’s lifetime.

Another common practice in the media is to obtain a quote or two, often from the person who  conducted the research.  When you think about it, it would be hard to find anyone more likely to be biased about the importance of the work than the person who led it.  I consider it particularly dubious when the leader of very early research, still years away from being vetted in appropriate clinical trials, declare that their research represents a likely breakthrough for cancer patients.  This may just be a case of an investigator who can’t be objective because it’s their research (most of us think our own children are beautiful), or it may be motivated by their desire to leverage the attention into more research funding and an opportunity to be perceived as more of a leader in the field.  Regardless, a reflexive request by the mass media to have investigators weigh the value of their own work is as pointless as accepting a director’s assessment of his own movie (although I’m pretty sure that the people involved in Ishtar or just about any Paulie Shore movie must not have been able to hide their shame behind pretty words).

One of the other recommendations that is also included in the JNCI paper is that journals demand that investigators describe the limitations of their own work, and/or that the journal offer some context.  The Annals of Internal Medicine has recently pursued such a policy, but they’re ahead of the curve in this respect.

There are certainly some journalists (Gina Kolata of the New York Times, for instance, and there are others) who regularly cover the gray areas of a topic and highlight the limitations of early research.  We should strive to have more journalists work to provide more than a one-sided story, while medical researchers need to provide more context and less grandstanding.  And the general public needs to recognize that there they should be suspicious of easy answers being spoon-fed to them, whether they are fear-mongering or touting a new miracle treatment.

No related posts.

Related posts brought to you by Yet Another Related Posts Plugin.

Some new ideas succeed and, after careful evaluation, truly advance care. Others fail. As a clinical trialist, I regularly evaluate ideas to consider a clinical trial.   Discussion with cards7up led me to realize that it might be helpful to share this thought process with this group—on behalf of yourselves and your loved ones, you vet ideas from the popular media, friends, family and the world wide web. I hope that these thoughts can help you to do so with more confidence.

This slide is from a talk that I gave a few weeks ago to a patient group in Philadelphia. It shows the four basic sources of ideas for a trial:

•  Laboratory data
•  Clinical observations
•  Alternative Medicine
•  Successes in other cancer

Some people might be surprised to see alternative medicine as an idea source.   It is a common misperception that Western doctors are not open minded to alternative medicine.  Taxol started as the bark of the Pacific Yew tree, ginger is gaining acceptance as a nausea treatment, and acupuncture has found a variety of roles in mainstream medicine.  What I am closed-minded to is treating patients routinely with something whose efficacy and safety are unproven.  So, new ideas are good, as long as they’re tested properly.   We first take a new idea to the lab, and learn everything we can about how it might work. We test it on cancer cells in a dish.   If it’s the next cure for cancer in a dish, we take it to animal testing where we test for efficacy, safety and dose.   If it looks like the next magic bullet for rodent cancer (again, I can’t tell you how many times we’ve cured cancer in various rodent models) then the idea can start the clinical trials vetting process.

The vetting of a trial starts with a funding source.   The investigator writes a proposal to the NIH, a pharma company, or a foundation laying out all of the data leading to the proposed trial and explains the basic structure of the proposed trial.   If the funding source finds the idea promising and chooses to fund it, the investigator will next write a protocol.  The protocol is a detailed guide for how a trial will be run.   Once completed, it is submitted to a scientific review board.   The job of this committee is to evaluate if the idea has sufficient scientific merit and promise and to ensure that the clinical rationale is sound.   If they sign off, it next goes to the institutional review board (IRB) who ensures that the study is ethical and protects human subjects.   If they approve the study, it may open, with either a data safety monitoring board (for large trials) or a medical monitor (for small trials) monitoring the trail to ensure subjects’ protection.   Finally, at completion of a study, the results, positive or negative, should be presented and published.

Okay, so that’s how an investigator thinks.   What about a practicing oncologist faced with an onslaught of new ideas published in various journals, presented at various conferences, and promoted in the media?   How does a doctor, working a gazillion hours a week seeing patients choose what to read and what to believe? How does he or she evaluate a paper?

First, it is key to understand the idea of levels of evidence.   Different ideas about how to treat patients have different levels of proof behind them.   I’ve summarized these in the trial ladder below.

My next major point also answers the question of why your oncologist has never heard of the great new cancer therapy that you read about on the internet, or even the news.   Given finite time, we want to read (and you should value most highly) the best, most proven papers.   How do we know which ones they are?   There’s a hierarchy of prestige levels in the cancer journal world, and authors try to get their papers accepted to the best possible journals.  While I’m sure that oncologists will argue about which ones are best, , here’s my opinionated, very rough guide for clinical trials in lung cancer:

•  Top tier: Journal of Clinical Oncology (JCO), New England Journal of Medicine (NEJM), Lancet, International Journal of Radiation Oncology Biophysics (”Red journal”)

•  Second tier: Journal of Thoracic Oncology (JTO), Chest, Lung Cancer, Journal of the National Cancer Institute (JNCI), Clinical Lung Cancer, Journal of the American Medical Association (JAMA)

Often worth of reading: Other peer-reviewed, medline-listed medical journals

There’s also a hierarchy of conference presentations, with the American Society of Clinical Oncology (ASCO), the European Society of Medical Oncology (ESMO) and the World Lung Conference at the top.   Note that findings usually first come out at conferences, then are published in journals — it should raise a red flag if a study is presented but never published.

So, to summarize, it might be worth reviewing brief criteria to know if an idea is worthy of routine implementation.  We read high-quality journals.   We then evaluate the findings presented and mentally place it somewhere on the trial ladder shown above.   Finally, if we’ve accepted the idea that the data is true, we then start the more comprehensible process of asking if the difference is big enough and if it’s worth the toxicity.

In the end, properly vetting an idea is challenging even for trained oncologists.  We meet in journal clubs to discuss new findings and we discuss them with colleagues.   Those more pressed for time can turn to review articles and conferences. In these media, trusted experts summarize the data and give their interpretation.  Time consuming? Yes.  Difficult? Yes.   But the stakes are high—failure to move a promising idea forward denies better therapy to patients in need. Treating patients with untested therapy denies them more promising therapy, risks toxicity, and denies the opportunity to move properly-vetted concepts forward to advance care.

No related posts.

Related posts brought to you by Yet Another Related Posts Plugin.

    I treat cancer for two and only two reasons.  I want my patients to live longer and with a higher quality of life.  All other measures are surrogates, and that includes cancer growth or shrinkage.  For years, oncologists focused on survival as the primary outcome of trials. We assumed that any therapy that could knock down cancer would also make patients feel better.  In fairness to the older trials, this assumption often was right.  When cancer grows, it can cause such terrible symptoms that even a therapy with bad side effects can lead to a better quality of life than if the cancer was left untreated.  However, as our drugs have improved, we have become pickier and now demand that new drugs not only knock down cancer better, but also have fewer side effects.  Most new trials incorporate robust measures of both side effects and quality of life. Some trials even seek improvements in quality of life as their primary endpoint.

   Every patient has different priorities and values, but I keep hearing the theme of “quality over quantity” from a sizable portion of my incurable patients.  I think that most of my colleagues, including trialists, now acknowledge the importance of quality of life in palliative therapy.  However, this recognition has lagged in potentially curable disease, whether the cure is attempted via chemoradiation or surgery plus adjuvant chemotherapy.  With this issue in mind, two recent trials in the Journal of Clinical Oncology (JCO) are heartening.

   The first study is derived from RTOG 9801.  This trial randomized patients receiving chemoradiotheray with curative intention to receive amifostine or placebo.  The goal of the trial was for amifostine to decrease the toxicity of radiation.  Unfortunately, the trial was negative, with no difference between the amifostine or placebo group in survival, toxicity, or any other measure.  However, the trial did teach us an unexpected but very important lesson.

   Since the trial was negative, the amifostine and placebo groups are equivalent and all patients in the trial can be evaluated for other prognostic factors without regard to assignment to receive amifostine or placebo.  When this is done, the most important prognostic factor predicting for survival is patient-reported quality of life at the beginning of the trial.  The yellow curve below represents patients with an above average quality of life at study initiation and the blue curve patients with below average quality of life.  You could drive a Mack truck between the survival curves with the yellow curve representing survival of patients with a score above the median, and the blue curve representing patients with a less than average score. 

(Click on image to enlarge)

   The quality of life score was so predictive of survival that even Karnofsky Performance Status, a classic practitioner-measured instrument for performance status, fell out as a prognostic factor in a model for overall survival.  The implications of this finding are not just philosophical, they are also practical.  In trials, we often stratify patients by variables that we think are important for survival. For example, if a trial allows patients with varying performance status, we make sure that an equal number of patients with poor or great performance status are assigned to each of the two arms of the study.  In this way, any differences between the groups can be attributed to the intervention (usually a drug) and not to chance imbalances in some important factor.  These results suggest that we can make better trials by stratifying by patient-reported measures of quality of life than by practitioner-measured performance status.

   The second trial involves an impressive use of complex mathematics to prove an incredibly human point: that adjuvant chemotherapy after surgery improves quality of life. At a time when we did not yet know that chemotherapy added to survival, the JBR.10 study randomized patients to either chemotherapy or observation alone after surgery. Patients who had stage IB or stage II disease were randomized to receive cisplatin and vinorelbine (in my opinion, the first of the “modern” regimens) or observation.  Overall, the study showed a 15% improvement in absolute survival at 5 years with patients living an average of 73 months without chemotherapy and 94 months with it.

   However, these results left many doctors and patients wondering: is 15% worth the toxicity of therapy?  Our Candadian colleagues took a stab at this question using a measure called Q-TWiST, which originally was used to evaluate adjuvant chemotherapy in breast cancer.  Q-TWiST stands for quality-adjusted time without symptoms and toxicity.  To get to Q-TWiST, you start with TWiST.  TWiST subtracts “time with toxicity” from “disease-free survival” (time alive and without cancer) to get the amount of time that the patient was alive, cancer-free, and feeling well. Q-TWiST then adjusts this score based on the relative values that a patient places on each of these states.  Ideally, these valuations would come from the patients themselves.  Since the authors did not ask them, they instead derived them from various (reasonable) sources.  No matter how they calculated it, chemo came out on top.

(Click to enlarge — it’s not visible otherwise)

   In order to make chemo lose, the patient would have to value time with toxicity less than half that of time without symptoms.  The authors summarize their findings well, “as long as living with chemotherapy-related adverse effects, such as fatigue, nausea, and vomiting seen in the JBR.10 is worth half as much as being healthy, then chemotherapy will pay off.”

   Both of these trials should arm your oncologist to have a better conversation with you about quality of life in the context of curative-intention therapy. Both suggest routes to better clinical trials.  Intelligent discussion of both requires some pretty heavy statistical terms. I have spent time looking at GRACE posts from the past few months, and I am impressed by the descriptions of really complex cancer findings in fairly simple English.  Dr. West posted a great article on cancer trials terminology in 2007 to which I refer the confused.  I am also considering posts on the following topics:

• *How do researchers get ideas for trials
  *Clinical trial vetting process and pt safety protection
  *What are the different kinds of trials
  *Why you should consider trials (in case it doesn’t shine through I’m a huge trials believer)
  *Trials open nationwide
  *How to “trial shop”

   Since this site is here to serve you, please let us know what topics would be most helpful!

No related posts.

Related posts brought to you by Yet Another Related Posts Plugin.

His efforts have been rewarded because there is now a growing body of literature showing that TH, while not actually causing cancer, may act in a permissive manner on existing cancers.  It seems that many cancers have on their outer cell membrane a great number of receptors for TH.  When TH is present in the system, it attaches to this receptor.  What are called growth signals are then sent from the receptor through the cytoplasm to the nucleus, allowing the cancer cell to multiply by dividing.  Such receptors are, in addition, found in abundance on the blood vessels which feed the cancer.  This, too, leads to more cancer because more blood vessels feeding the cancer is the result, leading to progression of that cancer.  Normal cells also have such receptors but in far lesser numbers.

Now, in this country, many patients, more often women than men, are given TH for a number of reasons, including fatigue, weight reduction, or a clinically insignificant degree of low thyroid function.  This is usually discovered by a routine blood test called TSH [thyroid-stimulating hormone].  Any level more than 5 is considered abnormal. When the level is 5-10, many doctors feel that it is necessary to give TH in order to get the TSH level back below 5.  The American Thyroid Association, however, does not feel this is necessary because almost no one has any symptoms at a level of 5-10.  In fact, many individuals feel just fine with TSH levels much higher than that.  When they are then given TH, they feel no different.

This is of potential importance to us because a number of studies, some epidemiological (focusing on populations and not really individuals), have shown those patients with a history of excess TH or who received TH do worse with their cancers than those without such a history. Likewise, those with low thyroid [hypothyroidism] or a history of such seem to do better.
It is now my policy to be sure that those patients with either an active cancer or a history of such really need TH.  I do this by checking their TSH levels and adjusting the TH dose so that the TSH level is 5-10, and not less.  In fact, some are now trying to get the TSH level even higher than 10 in those patients with an active cancer and have it rise to a level where patients may have minimal or no symptoms of hypothyroidism.

In fact, there is one study of patients with recurrent brain cancer who were given medications to make patients hypothyroid.  That group actually did better than the group who didn’t become hypothyroid.  Such drugs are now being given to patients with other types of cancer.  In that regard, there is some evidence that a drug which is quite effective against kidney cancer, Sutent (sunitinib), may exert some of its effect by inducing low thyroid function, which we have routinely viewed as a negative side effect.  Indeed, drugs are being developed which can block the TH receptor mentioned above.

Thus, for patients with active cancer or who have had the disease, it seems to be important to make sure that their dose of TH is appropriate, or even necessary.  And manipulating thyroid function may emerge as a relevant way to modify cancer outcomes, along with affecting the many other body systems affected by thyroid function.

No related posts.

Related posts brought to you by Yet Another Related Posts Plugin.

Clinical trials designed to test new cancer therapies are the only way that potentially effective drugs can make if from research labs into cancer patients in dire need of help. In April I talked about a couple of recent papers (see link above) that documented the complicated process of opening a clinical trial, involving hundreds of distinct processes in the bureaucratic morass of review, and taking multiple years from submission to trial opening.

In an early release publication in the Journal of Clinical Oncology, a group led by Dr. Razelle Kurzrock at the M.D. Anderson Cancer Center in Houston describes a new streamlined process for getting a phase I trial approved and the first cancer patient enrolled in what must be a record time (aptly called “Project Zero Delay”). This paper begins with a very interesting review of the myriad problems with the current system. For example, the authors note an average of 4-5 years for a company to take a new drug from the lab through enough initial safety studies in animals to get to clinical trials. After that, there is another 7.5 years before a promising drug makes it to FDA approval! All of this while over half a million people per year are dying of cancer in the USA, and of course millions elsewhere. The group then describes many of the same details I mentioned in my initial post about the delays in getting a specific trial opened.

Project Zero Delay focuses just on this aspect of the problem, namely getting a single new trial open. MD Anderson took it upon itself to also look at the exorbitant delay in opening trials, and designed a plan to fix it. The first step MD Anderson took was to negotiate a master agreement with a particular pharmaceutical company that would cover all trials over a 3-5 period, designed to overcome the months of budget and contract negotiations that normally take place for a new trial.

Most of the remaining problems with this process can be related to sequential levels of bureaucracy. In other words, drug companies, academic institutions, and the federal government each have an enormous amount of steps necessary to produce what is known as an Investigational New Drug (IND) application and then to open the trial, and these operations are usually done sequentially (one after the other) rather than simultaneously to be efficient. Project Zero Delay was designed to have all regulatory processes happen in parallel rather than sequentially, so that when the final OK came from the FDA the trial could open immediately.

The protocol described in the paper was designed cooperatively with the company and the MD Anderson principle investigator (PI). As the PI of a number of trials myself, I can tell you that the way it normally works is I write a protocol and then send it back to the company for review. This can then start a back and forth review and re-review process that can literally take years. In this paper, the authors were able to streamline this by having the company and PI cooperate at the inception of the concept so that the protocol could be completed fast.

Another sticking point is IRB (Institutional Review Board) review. Normally this necessary safety step, done to insure the safety of the study patient population, does not happen until the FDA has reviewed and approved a new IND for the study, adding months of time. In Project Zero Delay, the MD Anderson IRB agreed to review and approve the protocol PRIOR to FDA approval, so that when the IND arrived the trial was ready to open. I won’t go into every step, but all of the other regulatory steps also happened in parallel so that they all finished at the same time, without eliminating any steps! Look at the figures below. The first shows the normal process and timeline, while the second shows how all of the processes can be done simultaneously.

 (click on image to enlarge)

The final result is that the trial took only 46 days from protocol submission to enrolling the first patient on trial, and only 2 days from IND approval from the FDA! A month and a half. Not years, not six months, but six weeks! This is so amazing to me that I am still in a bit of shock and am still having trouble envisioning this applying broadly to every trial everywhere. I doubt every trial at MD Anderson moves this smoothly, and the paper doesn’t mention what level of personnel and resources are needed to make this work so quickly. But it does illustrate that with recognition of the problem, and with concerted effort to identify the issues and correct them, that clinical trials can be opened in a very reasonable amount of time. I am taking this paper to my boss and asking him if possibly we could do this too!

Related posts:

1. Challenges with Clinical Research Thanks to

Related posts brought to you by Yet Another Related Posts Plugin.