One of the most extensive and ambitious medical-screening trials in US has been stopped. It’s because of good news: Low-dose helical CT (LDCT) exams can cut lung-cancer deaths by as much as 20%, compared with chest radiography alone, according to early results from the National Lung Screening Trial (NLST). The NLST was halted in early November, 2010, after researchers determined that a 20.3% reduction in mortality rates occurred among patients examined using LDCT, compared with mortality rates among patients who underwent conventional chest radiography.
The trial lasted eight years and included nearly 53,500 participants between the ages of 55 and 74 who were asymptomatic current or former smokers, each with at least 30 pack-years of cigarette consumption prior to enrolling in the study.
The NLST participants were separated into two groups randomly, with one group receiving LDCT scans and the other, chest radiographs. The study confirmed that LDCT was better at finding small cancers in the smoking cohort than chest radiography was. The results further showed that significant numbers of current or former smokers can be saved from lung-cancer death if their tumors are discovered and treated early enough.
Denise R. Aberle, MD, a principal investigator in the study, notes, “I would say the results are very exciting. This is the single most significant medical intervention that has been done in patients to reduce lung-cancer mortality in several years.” Aberle adds, however, “CT is not a panacea. You can’t suggest it will save a life if there is high risk of lung cancer. That needs to be understood. This screening reduces mortality, but it by no means guarantees survival.”
Aberle is a professor of radiology and vice chair of research in the thoracic-imaging section at the University of California–Los Angeles (UCLA). She is also the NLST’s principal investigator for the ACR® Imaging Network (ACRIN).
The NLST was initiated jointly by the National Cancer Institute (NCI) and ACRIN, with NCI funding. The recommendation to end the study was taken by the trial’s data and safety monitoring board after results confirmed the life-saving benefits of LDCT over x-ray. It was a decision Aberle agreed with. She says the 20% reduction in mortality in the CT patient group was statistically significant and crossed a “+predetermined boundary” under which participants would be notified of screening benefits in one arm or the other. “We sent letters to all 54,000 participants,” she added.
The Biggest KillerAccording to a recent report¹ on the NLST, lung cancer kills more than 150,000 people annually in the United States. The smoking and lung-cancer impact on health care isn’t going away any time soon, but the results of the NLST suggest that thousands of deaths among current and former smokers could be averted with LDCT screening and treatment. The problem, which won’t be solved overnight, is how to implement such a program. Who will pay for it is a major consideration.
CT scans, as Aberle notes, are expensive. Even for the NLST participants, she says, each LDCT cost about $300. On the open market, costs can be significantly higher. Costs are high enough that even the NLST participants in the radiography arm are on their own if they want to be rescreened using CT, Aberle says; the National Cancer Institute (NCI) doesn’t have the funding to screen them. She says, “We advised the radiography patients to consider CT, in conjunction with their physicians.”
Aberle urged caution for patients and physicians who look at the NLST results with an eye toward ordering an immediate CT screening exam. There are currently no national screening guidelines, she says. Moreover, reimbursement might be a problem, whether from private insurers or from Medicare.
Don’t RushAberle also says that she’s concerned that imaging providers might be too quick to offer CT screening based on the NLST results. “There is an opportunity for capitalizing on this in ways that are not in the interest of the public health,” she says. “There are financial incentives to offer screening.”
Aberle cautions patients and physicians to seek qualified imaging providers and health networks before ordering screening. Provider decisions should not be made without forethought, she says. Imagers should have chest-CT expertise, and provider networks should be capable of dealing with abnormal findings and should meet standards/guidelines for practice following positive screenings, she says.
Reaching broad implementation of CT screening for lung cancer will require a string of decisions about who will pay and how the screening will be carried out, Aberle says. “We will need to establish reasonable boundaries for price. To pay $3,000 per scan is not appropriate,” she says. “As important, any institution offering screening is going to have to invoke a multidisciplinary effort.”
This means coupling CT lung-cancer screening with pulmonary clinics, smoking-cessation programs, and follow-up medical treatment. Aberle says, “We are going to have to be able to standardize, as much as possible, how to manage someone with a positive screening.”
Based on the NLST experience, about 25% of CT-screened patients will have some indeterminate finding concerning a nodule and will have to be re-examined using the most judicious radiation dose possible, Aberle adds. She says that the NLST ACRIN data still have to be analyzed and modeled to determine the cost effectiveness of CT screening, based on quality-adjusted life years and other factors, before insurers and Medicare can get a takeaway message on broad implementation.
“My sense is that there will be a series of recommendations by a variety of medical professional groups—the ACR, the Society of Thoracic Radiology, and the US Preventive Services Task Force. All of them will make recommendations for screening. Once that happens, that will motivate insurers and Medicare to address the issue,” Aberle says. “I would say that will happen over the next year.”
Aberle is hesitant to say which patient groups should be identified for CT lung-cancer screening, beyond current and former heavy smokers like those in the NLST. “People between the ages of 55 and 74 achieved a 20% mortality benefit; that’s all I know, scientifically,” she says.
She suggests, though, that other likely candidates for screening might be people with pulmonary fibrosis or chronic obstructive pulmonary disease, those with occupational exposure to lung carcinogens, and those with a family history of lung cancer. “At this point, I wouldn’t want to presuppose who might be appropriate,” she says. “That will be better informed by modeling.”
NLST ProtocolsPatients in the NLST received either a chest radiograph or a CT scan upon enrolling and were then rescreened annually for the next two years. They were followed for five years thereafter. All deaths were documented. According to the NCI, as of October 20, 2010, the CT patient group had produced 354 lung-cancer deaths, while 442 in the radiography group had died from the disease.
Aberle says that 14 different types of CT scanners were used across the trial, with 18 different image-acquisition parameters standardized across the trial as well. Radiation-exposure estimates for LDCT lung scanning hover near 1.5 mSv, but in the NLST, exposures varied based on scanner type, Aberle says. “They all got essentially the same scan,” she says. She says that the high-contrast visibility of lung-cancer lesions allowed LDCT to be used at exposure levels that were about 20% to 25% of those associated with conventional diagnostic CT.
Aberle says that the number of detectors on the scanner was not as important as the use of low dosage and capturing the image in a single breath hold. The scanners used in the NLST had a minimum of four detectors, but many had 16, 24, or more, Aberle says.
Prior to its use in the trial, each CT scanner was reviewed by a physicist using a phantom to calibrate radiation dose, Aberle says, and similar reviews were conducted yearly on all scanners. Scanners were also calibrated weekly by radiology staff. Aberle says, “There was fairly rigorous oversight.”
Patients with newly discovered lesions or suspected lesions—or, for that matter, any observed abdominal or thoracic abnormalities—were contacted by radiologists. The radiologists were empowered to handle referrals to clinicians using their best judgment, Aberle says. “We’re going to look at those now and find out what happened,” Aberle says.
Exciting DataOne of the most interesting aspects of the NLST, in fact, is that there is a huge amount of data; analyzing and modeling it will take years. The major endpoint of the NLST was lung-cancer mortality, which was shown to be reduced by LDCT scanning. Analysis must still be performed, though, to determine possible negative aspects of the screening, including cumulative radiation exposure and the complications of lung-cancer false positives and of additional diagnostic workups for noncancer findings.
Aberle says that researchers must still tease out from the data how many deaths were attributable strictly to lung cancer. The all-cause mortality in the trial was four to five times higher than just the lung-cancer deaths, Aberle says. This means that as many as 4,000 enrollees died during the trial, with only 796 deaths, so far, attributed definitively to lung cancer. The radiography patients had an all-cause death rate that was 7% higher than that of the CT patients, however. Aberle says that researchers think that at least half those excess deaths were also attributable to lung cancer.
Blood, sputum, and urine samples were collected from patients, and these still have to be analyzed. Who will do that testing is to be determined. The samples are preserved, waiting for researchers with funding to analyze them, Aberle says.
The same goes for thousands of resected tumors in paraffin blocks that are preserved at UCLA. Researchers with specific biomarkers of interest should contact the ACRIN biostatistical center to apply to use these materials, Aberle says.
“Because these specimens were acquired in a screening scenario, they’re more enriched for early lung-cancer diagnosis,” Aberle says. “That will potentially be hugely helpful in finding molecular markers, or panels of molecular markers, that may predict lung cancer. I’m pretty confident that we will identify them. It’s likely we will come up with a panel of biomarkers that have reasonable sensitivity and can be translated to clinical practice.”
In the meantime, there is one final message to be taken from the NLST: “The message is stop smoking,” Aberle says. “Couple screening with smoking-cessation programs.”
George Wiley is a contributing writer for ImagingBiz.com.