Discover more from PulmCCM
Lung cancer screening with CT: Does it work in the real world?
The first real-world results from a population-based deployment of lung cancer screening are in, from a demonstration project at Veterans Affairs (VA) hospitals. Screening was generally effective at identifying early-stage lung cancer, but with far more effort per cancer detected than in the seminal National Lung Screening Trial. High false positive rates led to a large organizational burden of repeat CT scanning and greater than expected harm to patients from anxiety and unnecessary invasive procedures.
Investigators performed population outreach to enroll and screen 2,100 high-risk veterans meeting criteria for inclusion in the original NLST. The demonstration program followed them after at least one initial CT scan, and any follow up scans or interventions cascading from there, over about a year.
Only about 20 early-stage lung cancers were found (1% of patients). The short followup period was not designed to detect any survival benefit. Other key findings included:
Sixty percent of patients had an abnormal result on their single CT scan. This was more than double the 27% abnormality rate seen with three CT scans in the NLST. In subsequent years of screening, one would expect even an even higher proportion of abnormal CT scans at the VA.
About 99% of the abnormal VA scans were false positives (not lung cancer), vs. 96% in the NLST.
About 40 patients underwent bronchoscopy or thoracotomy as a direct result of screening; at least one quarter of these procedures were unnecessary.
The scans also generated other incidental findings (e.g., coronary calcifications) in 40% of patients, creating significant burden for follow-up and unquantified additional testing, treatments, with unknown consequences (harm or benefits).
By comparison, in the National Lung Screening Trial, 4% of participants were diagnosed with lung cancer (over about three years). Adjusted for person-years, the rate of detected cancer (i.e., screening efficacy) was lower in the VA study, but not drastically so.
Compared to NSLT subjects, veterans analyzed in the pilot study were older, often current smokers, and men -- making them more likely to benefit from screening, in theory. In almost 40% of patients initially invited for screening, their recorded smoking history or quit date was inaccurate and screening was not indicated.
If Lung Cancer Screening Isn't "Worth It," Neither Is Anything Else
While authors and editorialists fretted at the perceived low utility of low dose CT for lung cancer screening, it measured up quite well against other established cancer screening programs, such as for breast cancer, colon cancer, and (erstwhile) prostate cancer.
In mammography, for instance, only 162 breast cancers are detected per 100,000 women screened (0.16%), and only 30 of these (0.03%, or 1 in 3,300 women screened) are believed to be both aggressive and curable. The rest (80%) represent overdiagnosis of relatively harmless cancers whose detection and treatment does not save lives (if these data are correct).
Lung cancer screening, even at the reduced efficacy seen at the VA and with its own overdiagnosis rate ≥18%, looks like a resounding success by comparison.
Rather than the VA screening program’s efficacy, authors instead bristled against the huge administrative burden required to create and scale up a population based screening program. Extrapolating from the effort expended on their 2,100 patients, they noted glumly:
It is estimated that nearly 900 000 of a population of 6.7 million VHA patients met the criteria for LCS. Implementation of LCS in the VHA will likely lead to large numbers of patients eligible for LCS and will require substantial clinical effort for both patients and staff.”
On the other hand, applying the updated Fleischner guidelines for pulmonary nodules would have slashed the VA’s false-positive to 20%, and the NLST's to 12%, authors note. This would go far to reducing the burden of repeat CT scanning. The effort required, although substantial, is primarily administrative, requiring coordinators, registry clerks, IT/EMR customization, etc.
Lung Cancer Screening: Is Outreach an Obligation?
The VA study tends to confirm what was predicted by most observers: deployed in the real world, the benefits of lung cancer screening would be lower than seen in the NLST, and its costs burdensome to already-stretched health systems.
Given the cost and effort required by lung cancer screening relative to its benefits, many health systems may drag their bureaucratic feet at expending resources to build a robust screening program. Lung cancer screening must be offered, but some health systems and physicians will ‘go slow’ in creating and scaling their programs.
There are already more precise lung cancer risk calculators that can better identify the highest-risk patients than the NLST criteria. Wider use would reduce the false positives that will be generated by any lung cancer screening program referencing NLST. Such an approach would be helped if the USPSTF's modified its recommendation for lung cancer screening to endorse the use of other risk screening tools (for outreach, not eligibility).
The burden on primary physicians, the main initiators of screening, could be substantially reduced as well. A nurse can ask patients over 55, Have you smoked in the last 15 years? If so, give them a brochure that describes lung cancer screening and provides the phone number to call to arrange the shared decision making visit. That visit could be done in a group setting by an associate provider, paid by an insurance company to counsel patients from multiple practices in a community. Computer-literate patients can visit shouldiscreen.com to better quantify their own risk and guide their decision.
Ultimately, one's conclusions about lung cancer screening depends on one's values and opinions about screening generally. If you’re in medicine as a business, all those follow-up scans, the false positives and incidentalomas, and the frightened patients they belong to, apparently represent a fantastic revenue stream, especially if you do the right marketing -- oops, I meant “outreach.” A profit motive can save lives -- that's in the U.S. health system's DNA.
But for closed health systems like the VA, it makes more sense to make the limited benefits of lung cancer screening available to PCPs and patients who request it, without erecting edifices of administrative waste.
Source: JAMA Int Med