|Lung cancer trial results show mortality benefit with low-dose CT: Twenty percent fewer lung cancer deaths seen among those who were screened with low-dose spiral CT than with chest X-ray|
Brief Description: The National Cancer Institute (NCI) is today releasing initial results from a large-scale test of screening methods to reduce deaths from lung cancer by detecting cancers at relatively early stages.
Coordinator:†Welcome and thank you for standing by. Currently all participants are on listen-only for the presentation. At that time of the question and answer session, please press star then 1 to ask your question. Todayís call is being recorded. If anyone objects they may disconnect. Iíd like to turn the call over to Dr. Harold Varmus. You may begin.
Dr. Harold Varmus:†Thank you. Good morning ladies and gentlemen. Iím Harold Varmus, Director of the National Cancer Institute, a component of the National Institutes of Health, and Iím calling from Bethesda. Today weíre releasing the long-awaited results, initial results, of the National Lung Screening Trial, better known as the NLST. And for this announcement Iím joined this morning by Deputy Director of the National Cancer Institute, Douglas Lowy, spelled L-O-W-Y, and by Richard Fagerstrom. His name is spelled F-A-G-E-R-S-T-R-O-M, who is also on the staff of the NCI and one of the Lead Statisticians with the NLST. There are a number of other experts who are standing by to help answer your questions after we make these announcements. Let me begin by telling you something about the NLST. Beginning in 2002 the NLST enrolled more than 53,000 current and former heavy smokers then aged 55 to 74 to determine whether screening these older, high-risk individuals with low-dose helical computerized tomography, or CT, would reduce lung cancer mortality rates.
This study was performed by many investigators at 33 centers across the country. Itís supported by the National Cancer Institute with funding that has exceeded $250 million. The NLST is a randomized clinical trial, which is the most effective type of study to answer questions of this kind. Participants who met the qualifications of age and smoking history were randomly assigned to receive annual screening for three years, either with low-dose helical CT or with conventional X-rays. They were then followed for results from the screenings for outcomes of further tests, for any resulting diagnoses and for the ultimate indicator, death from any cause and from lung cancer. All deaths from lung cancer were carefully documented and were reviewed by an independent board. The primary result of these studies can be stated simply. The NLST found that low-dose helical CT screenings reduces the death rate from lung cancer in this high-risk population by 20%, and the study showed the benefit is statistically significant.
Let me address the question of why this study is being released now, prior to formal publication. As is commonly done in large clinical trials, the NLST was closely followed by a Data and Safety Monitoring Board, or DSMB. That group of independent experts has met regularly since 2003 to ensure the safety of participants and to determine if the primary scientific objective of these studies had been met. Late last month the DMSB - DSMB members determined that there was sufficient data to show that the studyís primary objective had indeed be met - been met. They decided that the study should now be halted and that the participants and the general public should be informed of the outcome. The DSMB conveyed these conclusions to me in a letter dated October 28, and as youíll see the press release for todayís announcement contains a link to that letter.
Let me comment on the importance of this study. Lung cancer as many of you know is the major cause of death from cancer in the United States and in the world. It is estimated that lung cancer will cause about 157,000 deaths in this year in the United States. Over 85% of those victims are present or former smokers. As a consequence this finding has important implications for public health, with the potential to save many lives among those at greatest risk for lung cancer. Let me remind you about smoking. I donít want you to come away from this announcement believing that itís now safe to continue to smoke or to start smoking. The NLST has demonstrated that low-dose helical CT screenings can reduce deaths from lung cancer among older heavy smokers by 20%, but this screening does not prevent lung cancer and it does not protect the large majority of subjects from death from lung cancer. Furthermore smoking has many other detrimental effects on health, including other cancers and cardiovascular disease. So not smoking and quitting smoking remain important public health goals and remain the best defenses against lung cancer.
Iíd like to turn the microphone over now to Dr. Lowy who will reinforce some other cautions about the screening process.
Dr. Douglas Lowy:†Yes, good morning. Despite the positive findings of the trial that you have already heard about, itís important to note that screening procedures such as helical CT may also have some disadvantages. One of them is cost. At this time most insurance carriers and Medicare do not provide reimbursement for screening CT scans. Some plans however do cover diagnostic scans, but those scans look for disease when there are already signs or symptoms. Another concern is false positives. Recent studies indicate that between 20% and 50% of screening CT scans of current and former smokers will show abnormalities. It was about 25% in the NLST study. Most of these abnormalities will not prove to be lung cancer but could be scars from smoking, areas of inflammation or other non-cancerous conditions. Another issue involves subsequent medical procedures to evaluate the suspicious findings. These procedures include lung biopsy and thoracic surgery, which are potentially risky procedures that can cause a host of complications. Radiation exposure from even low-dose helical CT scanning is significantly greater than exposure from conventional X-rays, and patients often undergo additional scans to evaluate the abnormalities detected by the screening procedure. However it remains to be determined how or if the radiation doses from the screening CT scans used in the NLST may have increased the risk for cancer over the remainder of the participantís lifetime. All of these considerations have implications for recommendations. These factors must be taken into consideration when recommendations are eventually made regarding the use of low-dose helical CT for lung cancer screening. The NLST findings should not be interpreted to mean the general public should now get regular CT scans. This trial answered specific questions about a well-defined high-risk population. Further analysis and modeling of NLST data will be needed before specific recommendations can be made about the possible broader use of this screening method. I would now like to turn the microphone over to Dr. Fagerstrom, who is going to discuss some issues that still remain open and uncertain.
Dr. Richard Fagerstrom:†Thank you and good morning. Until further analysis is carried out we remain unable to answer a number of intriguing and important questions. For example we do not know whether low-dose helical CT screening has different levels of benefit in subpopulations, that is groups defined by gender, age, ethnicity or amount of smoking or in groups that were not included in the trial, such as people younger than 55, non-smokers or light smokers. Analysis of the very large data sets from the NLST will continue to examine other objectives of this study. In fact the study team has already noted one further intriguing finding; a modest 7% decline in all caused mortality, meaning deaths due to any factor, and this effect however has not yet been explained and will be investigated into its mechanism. After more analysis, within the next few months results will be promptly published in a peer review journal with full and immediate access to the public. I would now like to turn the microphone over to Dr. Varmus.
Dr. Harold Varmus:†Thank you Richard. So to recap and conclude, todayís release of initial results from the National Lung Screening Trial is an important moment for public health.
The NLST has rigorously defined the ability of helical CT screenings to reduce death rates from lung cancer by 20% in a high-risk population. This finding will be an important factor in subsequent efforts to protect the tens of millions of former and current smokers in this country against the lethality of lung cancer. In concluding, let me mention that we owe a great debt to the many investigators who conducted this remarkable trial, and especially to the men and women who so generously participated in the study and made the results possible. At this point we and - the three of us and our colleagues in the room are happy to take your questions. The operator will announce your name and affiliation.
Coordinator:†Thank you. At this time if you would like to ask a question, please press star then 1. To withdraw a question, please press star then 2. Once again to ask a question, please press star then 1. One moment for questions. At this time we do have several questions. Our first question is from Rob Stein, the Washington Post. Your line is open.
Rob Stein:†Yes, hi. Thanks very much for taking my question. I just wanted to clarify what if any recommendation you might be making based on these results. I mean, would you recommend that the people that meet the criteria of the folks in the study population get screenings on a regular basis?
Dr. Harold Varmus:†Weíre making no recommendations at this point. Subsequently, once the paper has been published with the data, the recommendations will be made by any of several bodies that have commonly been called upon to make recommendations for the medical health services community and for patients. Those might include the U.S. Preventive Services Task Force, the American Cancer Society, the Radiological Society of North America, the American College of Radiologists and several others. Right now weíre attempting to communicate to the public that an endpoint has been reaped in this important study to give an indication of what helical CT scanning can do to reduce deaths from lung cancer and to prepare the healthcare community to respond to these new findings. There will be further analysis as Dr. Fagerstrom has indicated required before a formal paper is written. We hope to have that paper finished in the next couple of months and submitted for publication in the peer-reviewed journal, at which point the paper and findings will be distributed to many interested parties for evaluation and formulation of recommendations. Is that helpful Rob? I...
Rob Stein:†Yes, thatís helpful. I guess - if I could just follow up quickly. I just - so what would you say should be the take home message for the average person from this study at this point?
Dr. Harold Varmus:†At this point the conclusion to be drawn is that at least in this high risk population, people aged 55 to 74, an older age group who have a history of heavy smoking, those are the ones who were included here. To be included in the study required 30 pack-years of smoking; that is an average of a pack a day for 30 years, and you had to either be a current smoker or to have quit smoking within the last 15 years. The data we have applies to those individuals. The study was set up to be able to detect a 20% reduction in mortality with a 90% level of competence, and we reached that level very recently. And the point is that this procedure can significantly reduce the number of deaths, the ultimate standard here for effectiveness of a screening procedure in that high-risk group. The application of these methods to a more general population may be difficult to decipher. We donít know that as yet and we donít yet have a recommendation for the schedule of screening or how - for how many years it should be used. That will have to be left to others who formulate the recommendations once the full data set is evaluated. This method is effective at reducing deaths from lung cancer by a significant amount, but we want to emphasize in any brief message that this is no reason to feel that smoking is not just as dangerous as it always was. Do you want to add to it Douglas?
Dr. Douglas Lowy:†I just would like to emphasize that this is really the first clear demonstration that a screening procedure can be effective in reducing mortality from lung cancer.
Dr. Harold Varmus:†Richard?
Dr. Richard Fagerstrom:†I would reiterate that. In fact what has been published up to this point has shown very little. Itís not as if it were almost significant and just missed. There has been really no good evidence up to this point of effectiveness, so this is of clearly strong health importance for the public.
Dr. Harold Varmus:†Next question.
Coordinator:†Our next question is from Lauran Neergaard, Associated Press. Your line is open. You may ask your question.
Lauran Neergaard:†Hi, thank you. Iíd like to have a little more discussion of the risk versus the benefit. You mentioned the 25% false alarm. What about false negatives? Were these...?
Dr. Harold Varmus:†Do you want to speak Richard to the question of how many lung cancers may have been missed in deaths who had false negatives?
Dr. Richard Fagerstrom:†Of false negatives - we have not actually evaluated those data extensively yet, and I think itís a bit premature at this time to report about it. We - I can say generally speaking that there werenít a tremendous number of them. Obviously it is a much smaller number in the spiral CT arm than in the chest X-ray because of the very much higher sensitivity of the test.
Dr. Harold Varmus: And we do know that the patients who received the chest - the conventional chest X-ray frequently were not alerted to lung cancer because the death rate from lung cancer was higher than the detection rate.
Dr. Richard Fagerstrom:†Yes.
Dr. Harold Varmus:†Secondly, to get to your point about the false positives let me just say a little bit more about that. Most of those individuals were then subjected to additional, more conventional so-called diagnostic helical CT scans. And in the majority of those cases they were found to have abnormalities that were very unlikely to be cancer. But in a significant number of patients who were subjected to as Dr. Lowy indicated earlier, either to biopsy or in some cases to chest surgery, so-called thoracotomy. And that obviously is one of the risks of this procedure as in most screening procedures for cancer. It sometimes requires obtaining tissue to ascertain whether or not a cancer diagnosis is accurate. Next question.
Coordinator:†Our next question from Katherine Hobson.
Coordinator:†We have a question from Katherine Hobson from the Wall Street Journal. Your line is open.
Katherine Hobson:†Hi, thanks for taking my questions. I have two quick ones. One is can you give us a - do you have an estimate for the number needed to treat in this? And then also can you say what the cost of screening this population would be per year or if not that, how much does one of these screens cost? Thanks.
Dr. Harold Varmus:†Iím not sure what you mean by the number to treat, but the respective cost we can say - first of all as youíve heard, the screening procedure is generally not reimbursed at this time by insurance companies or by Medicare.†The cost of a diagnostic as opposed to a screening helical CT scan is approximately $300, so you can estimate that the cost at the moment of a helical CT scan is likely to be in the range of a few hundred dollars. We canít say what that would be and itís always going to vary at different places in the country, depending on the type of CT machine thatís used for the screen.
Dr. Richard Fagerstrom:†Katherine, I have one comment that may be relative to your number needed to treat. Iím not certain exactly what was meant there, but I can say one thing and that is that in this trial, for every 300 NLST participants screened with helical CT, we estimate that one life has been extended.
Dr. Harold Varmus:†I know that Constantine Gatsonis, who is the Lead Statistician for a major participant in this study, the American College of Radiology Imaging Network, is also on the call. Heís at Brown University. Constantine, do you want to comment further on that question?
Constantine Gatsonis: Yes. On the question of cost, we have collected cost data from the study and we will be publishing a formal cost effectiveness analysis in the coming months.
Dr. Harold Varmus:†And I might also say that weíve had a comment from Dr. Don Berwick at the Center for Medicare and Medicaid Services, who said, ďCMMS looks forward to evaluating the results of the NLST as soon as they are published, and to reviewing low cost spiral CT as a screening procedure of potential value for Medicare and Medicaid beneficiaries.Ē Next question.
Coordinator:†Our next question is from Richard Knox, National Public Radio. Your line is open.
Richard Knox:†Hi, thanks very much. I appreciate the chance. This large...
Dr. Harold Varmus:†Hello. We canít hear you Richard.
Coordinator:†One moment. Richard, your line is open.
Richard Knox:†Yes, Iím talking. Can you hear me?
Dr. Harold Varmus:†No Richard, we havenít been able to hear you. Why donít you start again? Now I can hear you.
Richard Knox:†Okay, thanks. Iím sorry. This large trial was powered to detect a 20% difference and thatís all it found. Is that a disappointment? Had you hoped for a higher effect? And also Iím not clear - if this had been allowed to continue might the effects have become bigger than what you ended up finding at this point?
Dr. Harold Varmus:†I will let Richard answer that but of course our perspective here was not what we hoped to find. We hoped to get a definite answer to the question. We werenít rooting for one side or the other.
Dr. Richard Fagerstrom:†Yes, we werenít rooting either way.
Dr. Harold Varmus:†Itís better to have something that works and this clearly at least partially worked.
Dr. Richard Fagerstrom:†And one thing I want to emphasize is that 90% is a really big chance, so obviously we had power somewhat lesser but nonetheless substantial. Many trials are designed with only 80% power and we could have predict - we could have detected somewhat smaller differences with 80% power. And what was the other half of your question Richard?
Richard Knox:†Well I was wondering if there - just from a statistical point of view, if this had been allowed to continue is there any reason to expect that that might have - the effect size might have gotten bigger?
Dr. Richard Fagerstrom:†No, I do not believe so because this is what is known as a stop screening trial, in that we only gave three screens and three screens - if screening is not maintained itís ability to pick up cancers diminishes. So the effect will actually be diluted as we continue the trial if we donít continue to screen. So the answer is no, we would not have expected a larger...
Dr. Harold Varmus: So just to amplify on that slightly, the screen was done once a year for three years at the beginning of the study. So screenings stopped in 1985 or í86 and - sorry, 2005 or í06. Iím sorry. Iím slightly disoriented this morning. And as Richard points out if anything the differences might have decreased with continued time. But this raises an important point that we - thatís as yet not resolved and needs to be brought to the attention of the public, namely that we donít know the ideal way as yet to do this screening, whether it should be annually or for how long. So those are the questions that will be hopefully illuminated by closer inspection of the data, by continued modeling of the effects of the screening process and possibly by additional work by us or others.
Richard Knox:†Thank you. By the way, could I ask you to identify yourselves as you, you know, switch off because itís hard to know whose speaking.
Dr. Harold Varmus:†That was Harold Varmus who just finished and Richard Fagerstrom before that.
Richard Knox:† Thank you.
Dr. Harold Varmus:†Next please.
Coordinator:†Our next question is from Gardiner Harris from the New York Times. Your line is open.
Gardiner Harris:†Most of my questions have been answered except for one. Could you guys give - compare the radiation dose from the standard chest X-ray to the low-dose CT screen, and also compare it to what a woman would get in a standard mammography if you donít mind? Like what are the relative shares of radiation?
Dr. Harold Varmus:†Right. Well Iím going to use this question to introduce Denise Aberle, who is the Principal Investigator from the Akron arm of the NLST study. Sheís a Professor of Thoracic Radiology at the David Geffen School at UCLA. So Denise, could you address this point?
Denise Aberle:†Sure. If you look at the published literature, the effective radiation dose in the literature for a two view chest X-ray, which is a frontal and a lateral view, ranges anywhere from .05 to .24 millisieverts with an average of about .1. We actually did only a single view which was the frontal radiograph, so it would be lower than that. The effective radiation dose for a typical diagnostic CT in the United States varies greatly. An average number would be 7 millisieverts with a range of 4 to 18. We used whatís called low-dose helical CT. Thereís not a formal definition for what we mean by low-dose, but generally we consider a CT scan to be low-dose if itís anything less than what you might receive from non-medical background radiation sources per year, which is about 3 millisieverts. Our study using low-dose helical CT probably used about 20% of the dose from a conventional diagnostic CT scan, about 1.4 millisieverts.
Gardiner Harris:†About 1.4. And then how does that compare to a standard mammography if you donít mind?
Denise Aberle:††I would have to guess on that. Chris, do you have an answer for that?
Dr. Harold Varmus:†Identify?
Christine Berg:†In general - oh, excuse me. Iím Christine Berg, Principal Investigator from the Lung Screening Study with the National Cancer Institute. By - a two-view mammography to the breast in general is in the range of 2.5 to 3.5 millisieverts, so itís similar.
Gardiner Harris:†So theyíre about equivalent then it sounds like.
Gardiner Harris:†Okay. Thanks so much.
Dr. Harold Varmus:†Okay, thank you Gardiner. Next question please.
Coordinator:†Our next question is from Maggie Fox, Reuters. Your line is open.
Maggie Fox:†Thanks, and thanks for identifying yourselves each time you speak. That really helps. Iím just assuming that these guys live longer because the CT scan imaged a tumor better and it was removed earlier. Is that correct?
Dr. Harold Varmus:†Thatís the assumption. Also detected smaller tumors, yes. Tumors found earlier and treated successfully in at least some cases to produce a 20% reduction. Further question?
Maggie Fox: I didnít hear who was speaking and I didnít - youíre just presuming that? We havenít - you havenít established thatís what happened?
Dr. Harold Varmus:†Sorry, say it again.
Maggie Fox:†Youíre just presuming that the tumorís removed earlier? That hasnít been established in the study?
Dr. Harold Varmus:†Well the way the procedure works is that one identifies the tumors through this process and they are, I mean, if a cancer is found it is surgically removed and the assumption, which will be subjected to further analysis as the data is more closely examined, is that a larger number of early cancers that would have been lethal were removed in patients who had undergone the helical CT scanning, and that accounts - that very likely accounts for the reduced mortality from lung cancer.
Maggie Fox:†And Iím sorry. I didnít hear who that was who was speaking.
Dr. Harold Varmus:Thatís Dr. Howard Varmus again - same person who answered the first time. Next question please.
Coordinator:†Our next question is from Peggy Eastman from Oncology Times Newspaper. Your line is open.
Peggy Eastman:†Thank you. My question was similar to Maggieís and it had to do with the treatment of lung cancer, which is notoriously not very good if the lung cancerís...
Dr. Harold Varmus:†Yes, weíre not hearing the entire question. Iím not sure what the question was.
Dr. Harold Varmus:†It was cut off in the middle of the question.
Coordinator:†It looks like somebody is moving them into the queue and when they do that it does close down the other personís line. Itíll take me just a moment...
Dr. Harold Varmus:†Well itís not happening at this end.
Coordinator:†Okay, itíll take just a moment and Iíll open the line. Okay Peggy, your line is open.
Peggy Eastman:†Okay, thank you. My question was similar to Maggieís and it has to do with the implications for treatment. Treatment of lung cancer at later stages is notoriously not very good so is it pretty - a pretty good assumption do you think, I mean, can we say with some confidence that the benefit of this low-dose helical CT catches lung cancers early when they can be surgically removed successfully and have not metastasized?
Dr. Harold Varmus:†That is the assumption on which this study is built, but I want to emphasize that simply identifying a lung cancer at an early stage with helical CT scanning and removing it is not evidence that helical CT scanning is an effective modality, because some of those small tumors could in principal never go on to cause lethality. The gold standard here in showing that this screening procedure has the intended public health benefit is to carry out a large, long study like this one that shows that carrying out the screening process affects the ultimate endpoint, which is death from lung cancer. And thatís the strength of this study and itís important to emphasize that. Itís been shown before that helical CT scanning can identify small tumors that appeared to be early in the process of growth, but itís never been shown before conclusively that that procedure and the attendant surgery and the risks that are involved have an effect upon the ultimate outcome, namely lung cancer mortality.
Peggy Eastman: Right, but is it accurate to say Dr. Varmus that it is important to pick up these early small lung cancer tumors because you...?
Dr. Harold Varmus:†What we can say at this point is what Iíve said already.
Dr. Douglas Lowy:†This is Doug Lowy. The problem is that the analysis has not been carried out sufficiently at this point to determine which group of patients with lung cancer have been protected as a result of the screen. And so itís plausible to believe that it was predominantly people who had lung cancer detected at an early stage, but it remains to be demonstrated and that is part of the ongoing analysis which will be available in a few months in conjunction with the fuller analysis of these studies.
Dr. Harold Varmus:†Just to follow up on that, this is Dr. Varmus again. It is perfectly possible, even likely, that many patients, some patients who went through helical CT scanning and had small lung cancers removed would not have died of those cancers during the study or ever. So the - again the important endpoint here is reduction of mortality from lung cancer and we believe itís likely we can attribute that to having removed some lung cancers early in their development.
Peggy Eastman:†Thank you.
Dr. Harold Varmus: Next please.
Coordinator:††At this time we do have Eliot Marshall, Science Magazine. Your line is open.
Eliot Marshall:†Thank you. Iím curious about the follow up on this because you said intriguing 7%...
Dr. Harold Varmus: Weíve been cut off Eliot. Can you hear me? This is Harold.
Eliot Marshall:†I can hear you. Can you hear me?
Dr. Harold Varmus:†I didnít hear the question. We heard that you were interested in this intriguing 7%...
Eliot Marshall:†Yes, 7% all-cause mortality benefit. What kinds of possible explanations will you be looking into? I mean, what are the potential...?
Dr. Richard Fagerstrom:†Potentials would be other - this is Richard, excuse me, Richard Fagerstrom. Potentials would be other cancers than lungs; also a variety of different respiratory illnesses. Those would probably be the primary ones. Another possibility would be cardiovascular illnesses.
Dr. Harold Varmus:†But the point - this is Harold Varmus. The point Eliot is that at this stage we have made an observation of reduction in all-cause mortality. The actual explanation for that reduction remains to be determined. Other questions?
Coordinator:†We do have a question from John Farber from Milwaukee Journal Sentinel. Your line is open.
John Farber: Yes, thanks for taking the question. Actually I have two questions. I was surprised that the study was set up so that it was - there were annual screenings. So I thought that it was lung cancer and CT scanning there - the idea would be that there would be something like colonoscopy where you get a one-time screening or once every five or ten years. And the fact that you did it every - you did it three times suggests that a cancer may not have been discovered the first time, but was discovered the second or third time. Thatís my first question. The other question is what does this tell people who didnít - who smoked, you know, for letís say ten years a pack a day and then quit? Are they likely to get less mortality benefit from this procedure than someone who smoked for 30 years?
Dr. Richard Fagerstrom:†Well John, your second question canít be answered at this stage. Perhaps some predictions could be made when the analysis is more complete, but those patients simply werenít studied so we donít know patients with 10 pack-year history. Iíd like to turn the first question over to Chris Berg from the NCI whom youíve heard from before.
Christine Berg:†Yes, the - there are many, many differences between colon cancer and lung cancer. A colonoscopy actually can remove adenomas and therefore prevent the development of invasive malignancy.And you can then repeat the colonoscopy much later because you are removing the preneoplastic lesions. That is not what weíre achieving with lung cancer screening. We think - we donít know yet. Lung cancer represents a number of different histologic types: adenocarcinomas which are preferentially in the outside of the lung area and more easily detected by lung cancer screening, and then very aggressive tumors such as small cell and large cell carcinomas, and then central tumors which are squamous cell carcinomas.† These are very aggressive malignancies with rapid growth rates so it was felt that that more regular screening would be much more important. The other issue is that helical CT screening as mentioned earlier by Dr. Varmus picks up a lot of small little nodules. Many of these nodules are indiscriminate. We cannot tell exactly what they are and even things like PET scanning doesnít help us because theyíre so small.You need to do further tests, either a diagnostic CT or repeat the low-dose CT six months or a year later to assess stability of the nodule to make sure that you donít need to go in with an invasive procedure. So we felt that annual screening achieved the best balance between frequency and the aggressiveness of the malignancy. In terms of the duration of the study, we thought that we - it was our obligation to try to answer this very important question, and three screens in this size population was the most effective design to answer the specific question as rapidly as possible.
Dr. Harold Varmus:†John, this is Harold Varmus again. Let me say one more general - make one more general point about your question. Chris has explained to you why the study was designed as it was. This does not mean that when recommendations are written for how the public should be screened in the future, that the annual screens will be the recommended frequency. And we hope that some of the statistical analysis that we carry out on the very large data sets that are now available may help us determine whether screening is more appropriate on an annual or bi-annual or tri-annual basis. Next question.
Coordinator:†We have a question from Amy Burkholder with CBS News. Your line is open.
Amy Burkholder:†Yes, hi. Thank you very much. I want to go back to your point please when you said there has been no good evidence up to this point on the effectiveness of CT. This is an area thatís sort of been dogged with conflict of interest allegations and how does this very large study now sort of speak to all the evidence out there, including the JAMA report that it was not effective?
Dr. Harold Varmus:†Well Iím going to let - Dr. Lowy made that point which we all agree with, but Iím going to let him expand on it.
Dr. Douglas Lowy:†Prior trials have not been randomized trials and therefore inferences about protection were difficult to establish unequivocally. In addition the endpoints of prior trials were not decreased in mortality from lung cancer. The current trial overcomes both of those deficiencies by being a randomized trial and also using as the endpoint decrease in mortality from lung cancer.
Dr. Harold Varmus:†Let me just amplify -- this is Harold Varmus -- one simple point here. There are many biases being produced by a study in which one asks about survival after diagnosis, as opposed to what was done in this study to ask about the mortality rate amongst all of those who were screened. The second is the gold standard for determining whether or not there is an effect on what we care about most, namely death from lung cancer.
Amy Burkholder:†Thank you.
Dr. Harold Varmus:†Next question please.
Coordinator:†We do have a question from Liz Szabo from US TODAY. Liz, your line is open.
Liz Szabo:†Hi. Dr. Varmus just alluded to this. I was just wanting to know if thereís any problem with lead time bias in a study like this or if not...?
Dr. Harold Varmus: Not in this study Liz, but in the - in studies that are conducted without the randomization and using lung cancer mortality as the endpoint, yes, lead time bias can be a very significant attribute and one of the ways in which conclusions can be inaccurately drawn.
Liz Szabo:†And just to double check and I think you said the false positive rate in this study was about 25%?
Dr. Harold Varmus:†Thatís roughly 25% in the helical CT group. There were abnormal findings that required initially a next step which was in general, not always but in general, a diagnostic helical CT scan. We called them positives. That doesnít mean that there had been a conclusion that cancer was diagnosed. It means that there was an abnormality of some kind that required follow up. And this will be an important consideration in drawing up recommendations for how this technology should be used even in this high risk group, because this is - there are economic, medical and psychological consequences of these abnormalities, findings of abnormalities, and there will have to be careful consideration of their frequency and what is done to pursue them in the drafting of guidelines, and in the response that patients and their physicians add to those guidelines.
Liz Szabo:†Thank you.
Coordinator:†At this time Iím showing two more questions. We have a question from Gardiner Harris, New York Times. Your line is open.
Gardiner Harris:†Hi, thanks so much for taking this second question. I just had a question about chest X-rays as the comparator. You know, I think there have been some studies that suggested that chest X-rays might lead to slightly higher lung cancer mortality rates compared to no screening. Are you guys going to release in the results comparisons not only with chest X-rays but also with no screening, and is there any possible bias given that the present results are just related to chest X-ray use?
Dr. Harold Varmus:†Iíll ask Chris Berg to respond to that.
Dr. Richard Fagerstrom:†This is Richard Fagerstrom. I can indicate that there is a very large cancer screening trial of which you may be aware called the PLCO Cancer Screening Trial.
Gardiner Harris: Right.
Dr. Richard Fagerstrom:†It does look at the issue of chest X-ray versus no screening control, and there is in fact a subgroup of the participants in that trial who meet the eligibility criteria of NLST. And the investigators in that trial are looking at this issue so that we can have comparability across the two trials with regard to what the message is with respect to whether chest X-ray is better than not screening at all. And we also have shown in this trial that lung cancer mortality is reduced in spiral CT versus screening with chest X-ray. So we do have that particular angle covered in the PLCO Trial.
Gardiner Harris:†And do you know when those data might be released, the comparison basically?
Dr. Harold Varmus: This is Chris Berg.
Christine Berg:†Yes, this is Chris Berg. The PLCO DSMB has provided authorization for me to state that in the 30,000 individuals who have comparable high risk criteria to the NLST group, when you compare the chest X-ray groups across both studies they have comparable mortality. If you look at chest X-ray compared to community care in the areas, which is not a lot of chest X-ray screening these days, you see no benefit of chest X-ray. But importantly you do not see harm of chest X-ray. We are not seeing harms of chest X-ray that had been postulated in some of the early studies. Those studies were small, underpowered and seriously flawed which is why the PLCO was undertaken.
Gardiner Harris: Thatís what I needed. Thank you so much.
Dr. Harold Varmus:†Good. Anyone else?
Coordinator:†At this time Iím showing two more questions. We have Zach Jump with American Lung Association. Your line is open.
Dr. Harold Varmus:†If your line is open weíre not hearing you.
Coordinator:†Mr. Jump, your line is open. Please check your mute button. Your line is open if you would like to ask a question at this time.
Zach Jump:†Can you hear me now?
Dr. Harold Varmus: Yes.
Zach Jump:†All right, let me try again. Thank you for taking my question. My first one had to do with cutting the trial short, the decision by the DSMB. Was this an a priori decision? Did you have a pre-decided follow up date that you were going to measure through, or was it just decided that there was a large enough body of evidence?
Dr. Harold Varmus:†Richard.
Dr. Richard Fagerstrom:†This is Richard Fagerstrom. It is largely that there is a large enough body of evidence. Let me just give a little detail about how the DSMB for this particular trial works. It met twice yearly. Once a year it would look at the accumulating mortality data that we had to see whether there was enough evidence to be able to report an effect, either as a significantly beneficial effect or no effect. And once there is sufficient evidence then the DSMB decides that we have clear reason to report the results of this trial and go public, and thatís what has happened. This is - this analysis process of looking at the data and making this decision has been happening since 2006. And every year weíve looked at it. Weíve decided we donít have enough evidence yet. At the past meeting on October 20 we received the word from the DSMB there is enough evidence. We should report the results. Itís significant. There is a reduction in lung cancer mortality.
Zach Jump:†I guess Iím wondering why the decision wasnít made to continue through another year too and see if that significant difference still existed, or if it had just been a temporary measurement at that time.
Dr. Richard Fagerstrom:†Those issues are considered in the statistical analysis plan. This is what is known as a sequential monitoring design and we - if you cross one of these boundaries itís extremely strong evidence that you donít have a spurious effect.
Zach Jump: Thank you. I had one other question. Do you know why the decision was made - youíve alluded to this earlier that they said - the DSMB announced you should let people know the 20% difference. But considering this thereís nothing really else to say about it. Why was - do you know what their thinking was with not waiting until the full study was published?
Dr. Harold Varmus:†Well youíre welcome to look at the letter that was written by the Chairman of the DSMB to me. This is Harold Varmus. The - first of all when - the study being concluded they wanted to reveal the results as quickly as possible to participants. Those results would then quickly be disseminated without the adequate explanations we are trying to provide today to the general public to explain what the next steps are and to say what we can say at this juncture. In general all of us would rather release something that had been subjected to peer review when itís published in its entirety, but given the public health import of this announcement we believe that itís more appropriate to be transparent about what we do know and we donít know, and to ask for the public patience until we can provide a full set of data, until many who will be thinking about recommendations to the health services community can do their work.
Zach Jump: Thank you very much.
Dr. Harold Varmus:†And the last question? Weíll take that.
Zach Jump:†Thank you very much.
Coordinator:†We do have our last question from John Farber, the Milwaukee Journal Sentinel. Your line is open.
John Farber:†Hi, thanks for taking my additional question. I think you said before that you didnít think the mortality benefit would be any different had the study gone on longer. I - but I wasnít quite sure why you came to that conclusion, especially given that it appears that the subjects in the trial were followed for up to five years. And Iím just, you know, thinking if somebodyís cancer was detected in the - after the third scan, you know, their - it just seems like the possibility of additional mortality benefit would have been there. Could you explain that I guess in terms that I can understand?
Dr. Harold Varmus:†Richard, do you want to do that or do - there are - I think thereís several ways to approach this but go ahead.
Dr. Richard Fagerstrom:†Well, one thing I want to say -- this is Richard Fagerstrom -- is that while the screening period of this trial extended until 2007, the vast majority of the screens were done up to a year before that. So we actually had substantial follow up from that last screen on just about everyone in the trial. And as I mentioned before with respect to the benefit of screening, especially with a cancer like lung, itís something youíre going to see because these are generally not huge lead-time. Youíre going to see this benefit fairly quickly. If you start following longer without continuing to be screened, then this benefit will be attenuated because youíre not picking up these newer developing, somewhat more aggressive tumors and they will be missed and it dilutes the effect. So I think that we do have follow up for most of these individuals for three years or close to three years, and with a disease like this I think that it is sufficient.
John Farber:†But arenít there a significant number of people who die from lung cancer, who die beyond three years, who survive for three, four, five, six, seven years and then die? And wouldnít they be - they would not be accounted for in this study, correct?
Dr. Richard Fagerstrom:†Yes, the extremely what you might call somewhat indolent lung cancers would not necessarily be accounted for. Itís entirely possible. Now youíre talking - this is getting to a little bit of a round of what they call overdiagnosis. If somethingís sufficiently indolent it may not make that much of a difference whether itís detected or at what time it is detected. So I donít think our effects would get any bigger.
John Farber: Thank you.
Dr. Harold Varmus:†With that I think weíll draw a conclusion to this. I thank many of you for taking the time to attend this important announcement, and I want to thank my colleagues today for participating in this useful presentation of new data from the National Cancer Institute. And again I want to thank particularly those who participated in this trial and made it possible for us to come to these clear conclusions. Thank you all.
Coordinator:†Todayís call has concluded. All parties may disconnect at this time. Thank you.