Since the publication of the National Lung Screening Trial’s (NLST) results in 2011, low-dose CT screening has been the focus of substantial research, guidelines, and general discussion.
Dr. Denise R. Aberle
The NLST and Beyond
“Most of the evidence that drove the decisions to approve screening came from the National Lung Screening Trial,” said Denise R. Aberle, MD, of the David Geffen School of Medicine at the University of California, Los Angeles, who chaired the Education Session. It is the only screening trial that has published a mortality benefit, and Dr. Aberle added that “it is certainly the largest dedicated lung cancer screening trial in history.”
The NLST included 53,454 heavy smokers, aged 55 to 74, who were randomly selected to either low-dose CT or chest x-ray once annually for 3 years. After a follow-up of 6.5 years, there was a 20% relative lung cancer mortality reduction, and a 6.7% all-cause mortality reduction. The number of people needed to screen in order to prevent one death was 320, a number that Dr. Aberle said compares favorably to other screening tests, such as mammography.
“The major limitation was false positivity,” Dr. Aberle said. Any nodule of 4 mm or larger was considered a positive screen, and the vast majority of positive findings were false positives. Still, there were few complications among the false positives (< 0.1%), and as a result of the NLST, the U.S. Preventive Services Task Force recommended low-dose CT screening in 2013. And by 2015, the Centers for Medicare and Medicaid Services (CMS) also determined that screening should be a covered benefit.
Still, some of the details of screening remain up for debate. “Age and smoking are themselves insufficient,” Dr. Aberle said. “We need a better way to identify the at-risk population.”
Among the disagreements on screening criteria is the upper age limit. The NLST included people up to age 74, but the U.S. Preventive Services Task Force recommendation expanded that to age 80. CMS set the upper limit at age 77.
The National Comprehensive Cancer Network devised two separate categories for screening eligibility, one based on the NLST criteria and a second category that adjusted some of the details. That second category included anyone aged 50 and older and lowered the smoking history threshold from 30 pack-years to 20 with no requirement for smoking cessation date but required at least one additional risk factor. Those factors could include a family history of lung cancer or personal cancer history, other smoking-related lung diseases such as emphysema, or exposure to radon or other lung carcinogens.
Dr. Aberle said that there are little comparative data as of yet between the various screening eligibility criteria, but in February 2015, the Lahey Hospital and Medical Center in Burlington, Massachusetts, published its experience comparing the National Comprehensive Cancer Network categories in more than 1,700 patients. They found that the annualized rates of lung cancer were 1.6% in patients considered category 1 (NLST criteria) and 1.8% in patients considered category 2. This suggests that expanding the criteria beyond the NLST could capture more patients at risk without significantly changing rates of positivity.
Another approach has arisen from the Prostate, Lung, Colorectal, and Ovarian (PLCO) Screening Trial. The PLCO2012 model used logistic regression involving a number of variables to create a risk-assessment tool, and a study published in 2013 compared that with NLST criteria.
Using that model, the NNT dropped from 320 people with the NLST down to 255 people. There were 8.8% fewer people screened using PLCO2012, yet 12.4% more cancers were found.
Another potentially powerful way to reduce false positives is to change the criteria regarding nodule size, Dr. Aberle said. In the NLST, nodules between 4 and 6 mm accounted for 51.0% of all positive screens but less than 1% of cancers found. “We can safely say that the minimum threshold for positive screens can be increased without significantly degrading the performance of CT screening,” Dr. Aberle said.
The Lung CT Screening Reporting and Data System (Lung-RADS) approach has now emerged as a rule-based alternative to assessing CT screens. It includes nodule size and consistency, grades nodules 1to 4, and includes specific management guidelines for each screening result. Other methods, such as the Pan-Canadian Lung Cancer Screening Initiative, also known as the Brock Model, have also been proposed.
“The NLST has informed a practice, and it does provide a baseline from which I think we can only improve,” Dr. Aberle said. “We’ll know that we’ve gotten there once we begin to see these reproducible algorithms at the point of care.”
Implementing What We Have Learned
Dr. Abbie Begnaud
Since her health system began a screening program several years ago, Dr. Begnaud said, “It has been a very exciting rollercoaster of regulations, guidelines, complications, and updates. There has never been a dull moment.”
Among the complications is the existence of a wide variety of guidelines and recommendations from professional societies, including ASCO, the American College of Chest Physicians, the American Cancer Society, and others. Although these differ in some of the details, Dr. Begnaud said there are some common threads to follow when implementing a program. The basic eligibility criteria components of age and smoking history are consistent, as are guidelines regarding imaging technique and interpretation. Many guidelines also include recommendations regarding smoking cessation, shared decision making between patient and caregivers, as well as multidisciplinary team approaches and oversight of the screening program.
With those in mind, beginning a program has a few basic steps: market analysis; assembling a multidisciplinary team of both clinicians and non-clinical staff; establishing workflow for ordering exams, image interpretation, and follow-up; and a thorough review of recent regulations.
Medicare has implemented several requirements that Dr. Begnaud said have added to the challenge of starting a screening program. For example, it is required that all patients who are screened participate in a registry, which she said is extremely useful and important but does add steps to the process.
Ensuring eligibility has also been a challenge, Dr. Begnaud said, for a variety of reasons. “There are patients who hear about screening, and they want to be screened,” she said. “It’s hard to say ‘no’ to that patient in your clinic, especially if you’re not armed with all the data.”
Shared decision-making is also now required for patients receiving Medicare, and at the University of Minnesota program, it is now recommended for all patients. Dr. Begnaud said that before August 2015, only 15% of screens involved shared decision making; after implementing changes encouraging the practice, that improved first to 39% from August to November 2015, and then to 62% for the following three months.
Still, challenges remain. An analysis of Dr. Begnaud’s program showed that for a large proportion of patients (41%), the electronic medical records had inadequate information to actually determine the patient’s screening eligibility. She said that was largely due to incomplete smoking histories.
“We still have a lot of work to do to either complete our charting or figure out the right people who should be getting screening,” Dr. Begnaud said.
Rethinking the Tobacco Component
Smoking history plays an enormous role in the assessment of risk for lung cancer. Graham W. Warren, MD, PhD, of the Medical University of South Carolina in Charleston, spoke during the Education Session and said that there must be a new focus on how smoking changes cancer outcomes and treatment responses.
“Smoking causes adverse outcomes in [patients with] cancer,” he said. Smoking increases the risk of overall mortality in patients with cancer by 51% and of cancer mortality by 61%, according to a 2014 Surgeon General’s Report to which Dr. Warren contributed. And importantly, this is not just a lung cancer issue, with significant negative outcomes associated with tobacco use in breast cancer, hematologic malignancies, prostate cancer, and others. “This isn’t really a lung and head and neck cancer problem; this is a cancer problem,” Dr. Warren said.
There is an important gap in addressing tobacco in clinical trials, given those important associations with outcomes. A 2012 analysis of cooperative group trials found that 29% had any assessment of tobacco, and 71% had no such assessment at all. Surveys of oncologists suggest that most do ask patients about smoking, but fewer than half discuss cessation medications or actively treat patients with regard to cessation. Dr. Warren pointed out that when you survey patients rather than oncologists, the reported rates of such discussions and treatment are even lower.
Incorporating cessation efforts into both clinical trials and clinical practice represents an enormous opportunity to improve outcomes, Dr. Warren said. He described a system of automatic referral to a cessation-support program for patients with cancer; 81.3% of patients were able to be contacted regarding cessation, and remarkably, fewer than 3% of those refused participation in a cessation program. In contrast, sending mailings to patients about cessation had a dismal response rate, meaning an opt-out program is likely the preferable method. Identifying appropriate patients can also be quite simple: asking only three questions identified 98.8% of patients in Dr. Warren’s program.
A particularly important question is, “When did you last have a cigarette?” This can capture patients who have recently decided to quit, but, unfortunately, are at very high risk of relapse. “Anyone who has used tobacco in the last 30 days needs cessation support,” Dr. Warren said.
Smoking cessation had a powerful effect on lung cancer mortality in a recent study, Dr. Warren said. “Smoking cessation was as good as [LDCT] screening with 7 years of cessation, and it was better after that,” he said.
There is also a large gap in research regarding the effects of tobacco use on treatment responses. Some research has showed a dramatic reduction in efficacy of radiation therapy in smokers, for example. There is also an opportunity to develop biomarkers that could guide specific cessation methods.
Other questions also loom large. “The effects of smoking [on treatment response] seem to be reversible, but not on the same day,” Dr. Warren said. “What if we just delay treatment for a month? Can we get a better outcome?”
He called the lack of attention on tobacco use a failure in personal cancer treatment, and an enormous opportunity given the now ubiquitous discussions of value in cancer care. “With the cost of cancer drugs, addressing tobacco may be an extremely valuable method to improve our cancer treatment outcomes.”