It was a privilege to present our clinical research, “Do the surgical results in the National Lung Screening Trial reflect modern thoracic surgical practice?” at the 98th American Association of Thoracic Surgery (AATS) Annual Meeting last week in San Diego. I was pleased that the AATS recognized the importance of our message by placing the presentation in one of the featured plenary sessions.
My main goal in presenting our study was to readdress the issue of low dose computed tomography (LDCT) screening for lung cancer. Despite coverage by private insurance and Medicare/Medicaid, it has been estimated that less than 10 percent of eligible patients are actually being screened with LDCT for lung cancer. Unfortunately, lung cancer screening came of age in an era in which there is a increased skepticism regarding cancer screening in general. That, combined with patient level barriers, provider concerns over the potential “harms” due to screening, and the systematic concerns over the cost of cancer screening have limited the adoption of LDCT screening.
Unfortunately, this is to the detriment of patients at risk for lung cancer. Microsimulation modeling based upon current and future rates of tobacco use, predict that by the year 2030 we could avoid 34,681 lung cancer deaths if screening adherence rates in eligible populations were 25 percent; 62,425 deaths if adherence rates were 45 percent; and 104,0442 deaths if adherence rates were 75 percent. Clearly the stakes are high and a large number of lives depend upon lung cancer screening.
In our study, we specifically sought to address the perceived “harms” of lung cancer screening, harms which in our opinion are often overstated. In the NLST Study Guide for Patients and Physicians published by the National Cancer Institute, the authors wrote that with LDCT screening, “2 in 1,000 more (patients) had a major complication from invasive procedures.” However, what is often overlooked is that the rate of major complications was heavily driven by patients who were actually found to have lung cancer. Furthermore, only 6 in 10,000 patients without lung cancer had major complications after invasive diagnostic procedures, compared to 11.2 percent of lung cancer patients.
Clearly, it is the cancer patients who carry most of the risk—risk that must be balanced with the need for treatment of lung cancer. We therefore thought it was important to examine those surgical risks more closely. Although the results of the NLST trial were originally published in 2011, the surgical results in patients with lung cancer in the trial have never been closely scrutinized. We retrieved the NLST data and evaluated the perioperative outcomes in 1,029 patients undergoing surgery for lung cancer.
Our study had several key findings:
Interestingly, we found that diagnostic and surgical choices affected complications. In the original NLST study, the management of screen-detected nodules was left to the discretion of treating physicians. As such, there existed a great deal of variability.
It is anticipated that the Lung-RADS reporting system developed by the American College of Radiology will standardize the workup of screen detected lung nodules. It has already been shown that such an approach will halve the number of “positive” screens. In the NLST study, it is our belief that too few patients (only 29 percent) received non-surgical confirmatory biopsies, such as CT guided or bronchoscopic biopsies, before proceeding with invasive surgical procedures. In a screening population in which the incidence of lung cancer in screen detected nodules in less than 5 percent, the burden is on surgeons to prove the diagnosis of lung cancer prior to exposing the patient to surgical risks.
We also found that surgical approach and extent of resection influenced complications. Minimally invasive surgery was not broadly utilized in NLST patients, with only 30 percent of surgical patients undergoing video assisted thoracic surgery (VATS) for resection of their cancers. VATS has previously been shown to have lower complications than open surgery and is now performed in almost two-thirds of lung cancer resections at specialty thoracic surgery centers. Lung cancer patients in the NLST undergoing VATS showed a strong trend towards less surgical complications and mortality (HR 0.76, CI 0.56–1.04, p=0.09).
Similarly, sublobar resection, either segmentectomy or wedge resection, has been associated with less complications than lobectomy in lung cancer patients. Although the oncologic efficacy of sublobar resection remains to be clarified by ongoing surgical trials (CALGB 140503, JCOG 0802), patients with screen detected lung cancer may be the optimal candidates for lung sparing surgery as their tumors are generally smaller and perhaps less biologically aggressive. Additionally, sublobar resection will likely preserve pulmonary function in this younger population of patients at risk for emphysema or second lung cancers. In the NLST, sublobar resection was performed in only 16 percent of lung cancer patients. However, its use was associated with a marked decrease in complications and mortality (HR 0.59, CI 0.38–0.94, p<0.02).
In conclusion, the rates of surgical mortality and major complications with long-term implications are very low in lung cancer patients in the NLST data set. Greater utilization of sublobar resection and minimally invasive surgery will likely further reduce morbidity in patients with screen-detected cancers. This combined with more standardized approaches to nodule evaluation and management will decrease the “harms” of lung cancer screening and hopefully lead to the screening of more eligible patients. The lives of lung cancer patients depend upon it.
Dr. Brendon M. Stiles is a board-certified thoracic surgeon who treats lung and esophageal cancer. He is a cancer researcher, patient advocate and an associate professor at the Weill Cornell Medicine New-York Presbyterian Hospital.