Report from the 12th International Conference on Screening for Lung Cancer
Author: Fred GrannisThe 12th International Conference on Screening for Lung Cancer was held in Nara New Hall, Nara, Japan April 8-10, 2005 under the sponsorship of the Screening Committee of the Japan Lung Cancer Society and the International Association for the study of Lung cancer, with Professor Kengi Eguchi presiding.
Professor Tomotaka Sobue reported upon the history of lung cancer screening in Japan. In 1987, the Japanese government considered cessation of chest radiographic (CXR) population screening initially performed for early detection of tuberculosis. Despite a lack of data indicating a benefit in population mortality, CXR screening for lung cancer became national policy. As many as 7.8 million Japanese over the age of 40 have received CXR screening annually. Although there is no national data to demonstrate reduction in population mortality, a series of case-control studies show evidence of improved survival, and in the prefectures of Miyagi, Gunma, Niigata and Okayama show evidence of possible reduction in lung cancer population mortality.
Starting in the late 1980s, Japanese investigators began to study low-dose CT scan as a more sensitive modality for screening. Hironobu Ohmatsu of the National Cancer Hospital East reported the results of the Anti-Lung Cancer Association (ALCA) study comparing the results of these two modalities in screening 45,000 individuals. The percentage of patients detected by CXR vs. CT in stage IA was 42% vs. 74%, mean tumor size 30 vs 17 mm., and 5-year survival 49% vs. 80% in the two groups.
Torio Nakayama reported interim results of the Japan Lung Cancer Study Group, a multi-institutional, non-concurrent cohort study of 137,000 screened individuals CXR vs. CT with a case control arm. 84% of CT cases were detected in stage I with a 5 yr. Survival of 92%. Sub-centimeter cancers had 100% 5-yr. survival, statistically significantly better than that of NSCLC >30 mm (p=0.05). Lung cancer mortality rate appears to be dropping with increasing years of annual CT scan follow-up.
James Jett reported the results of a Mayo Clinic study on 1520 patients. 75% of study subjects had a nodule on one or more CT scans, 61% <4mm; 34% 4-7 mm and only 6% 8mm >. 61% of lung cancers were in stage I on initial scan, and 59% on annual repeat scans. There were 12 operations for non-malignant nodules and five year survival was 80%. Lung cancer mortality was 1.6 per 1000 person-years. Statistical evidence of a stage shift or improvement in survival were not achieved in this study. Although there were standard recommendations, the study does not report compliance with study recommendations beyond annual CT scan.
Claudia Henschke, the principal investigator of the International Early Lung Cancer Action Program (I-ELCAP), updated the ongoing experience with low dose, non-contrast, spiral computerized tomograms (CT scans) in more than 28,000 patients at high risk for lung cancer screened by I-ELCAP, including more than 20,000 repeat scans, as part of an prospective, international, multi-institutional study. She emphasized the importance of the management algorithm used by I-ELCAP. The I-ELCAP experience indicates that solid nodules smaller than 5 mm and non-solid nodules smaller than 8 mm in diameter require only an annual repeat CT scan in one year. Only those individuals with nodules above that cutoff level are considered to be positive tests. Accordingly, 15 % on initial prevalence scan and 6% on annual repeat require additional workup or intervention. “Lung cancer screening is a process, not a single test.”
The results of the I-ELCAP study suggest that although nodules larger than 15mm, that do not diminish in size with antibiotic treatment should be biopsied, smaller nodules do not require biopsy unless they show growth on 3 month interim CT scans or have a positive result on a PET scan. When the I-ELCAP protocol
<
is followed, more than 80% of cancers are detected in stage I, compared to approximately 20% in unscreened individuals in the U.S. currently. Henschke was joined by Javier Zulueta of the University of Navarre in Pamplona, Spain, in recommending that all PET negative nodules require close follow up, since approximately half of small PET negative lesions are later shown to be malignant. 90% of I-ELCAP screen-detected cancers less than 15 mm in diameter were stage I. Of 28 small cell lung cancers detected, 2/3 of lesions smaller than 25 mm were in limited stage.
Nasser Altorki of Cornell U. reported on the results of resection in 364 I-ELCAP cases. Post-operative mortality was 0.6%. 5.5% were N=1; 8% N=2 and 86.5 N=0. At a median follow up of 33 months, there were 3.9% lung cancer deaths in the group of 333 patients who had surgical resection alone.
Growing evidence from Japanese investigators suggests that adenocarcinoma of the lung frequently originates as atypical adenomatous hyperplasia and progresses through what is called bronchioloalveolar carcinoma in the west and as A, B and C adenocarcinomas in Japan in the pathologic classification system presented by Masayuki Noguchi. Non-solid, non-invasive carcinomas (A, B) then collapse into part-solid and solid, invasive Noguchi D, E and F classifications. All non-solid nodules diagnosed as cancers in the I-ELCAP experience were Stage I. Numerous Japanese investigators, including Kenji Suzuki of the Japanese Oncology Group, Ken Kodama of Osaka and Hiroaki Kuwahara of Hokkaido reported that non-solid nodules as well as part solid nodules with a % solid component less than 50% can often be safely treated by limited surgical resection, thus sparing lung tissue. Risk of local recurrence in this circumstance is small and long-term, disease-free survival approaches 100%.
Numerous investigators discussed experimental non-surgical alternatives to resection of small, screen-detected cancers, including radiation therapy, cryotherapy, radiofrequency ablation, microwave ablation and photodynamic treatment.
Because opponents of lung cancer screening have suggested that there is no difference in survival by size, and that there is a strong influence of overdiagnosis bias secondary to a purportedly high incidence of non-lethal cancers and mortality due to co-morbid disease, these topics were prominently discussed. Henschke and Yankelevitch described Cornell U., I-ELCAP and SEER databases to indicate that there is a statistically significant stage shift toward Stage IA, as well as increased survival, with diminishing diameter in both registry and screen-detected lung cancers. Among more than 44,000 registry cases in the SEER database, 5 year survival among stage I cancers <15 mm was 54%, falling to 46%, 34%, 25% and 14% respectively, as size increased to > 16, 26, 36 and 46 mm in diameter. Yankelevitch emphasized that registry data dilute size-stage correlation compared to asymptomatic screen-detected cancers.
Analysis of I-ELCAP cases with delay in diagnosis allowed calculation of doubling time. More than 90% had doubling times less than 400 days and 41% progressed in stage, indicating that the incidence of putative “non-lethal” cancers would have to be very small. The Pamplona group reported that there are no differences in gene amplification, chromosomal abnormalities or gene mutations between screen-detected and invasive adeno CA, including metastatic lesions.
Rowena Yip examined the survival of I-ELCAP patients with co-morbid disease. Although mortality increased by pack-year of smoking and age, risk of death from co-morbid disease was relatively small compared to the risk of lung cancer death. 97% of study subjects without lung cancer were alive at 5 yr., and 93% at 10 yrs. Lisa Guirguis reported on 62 patients with non-small cell lung cancer treated by limited resection at City of Hope in Duarte, California, because of severe comorbidity. Most deaths were secondary to lung cancer rather than comorbid disease.
Who should be screened? I-ELCAP data show that individuals 40-49 yrs. With 60+ pack-years of smoking history have a 1.1% lung cancer detection rate, rising to 3.7% above age 75.
A number of collateral benefits to CT screening in the form of detection of comorbid diseases, including coronary artery disease (coronary calcification scores), aortic aneurysm, emphysema and other disease were reported. Discussion with patients and demonstration by the radiologist of such co-morbid disease (e.g. emphysema in 8% of study subjects) was also described as fostering smoking cessation. Improved rates of smoking cessation (14-23%) in conjunction with CT screening were reported by the Mayo Clinic, Depisan (France) and I-ELCAP groups.
Keiichi Nagao of Chiba University and Ralph Aye of Swedish Cancer Institute in Seattle Washington emphasize the importance of high quality CT technology, skilled radiologic reading, defined algorithms, careful communication between radiologists, patients, and primary physicians, careful follow-up and appropriate radiologic and surgical intervention in the context of lung cancer screening programs. Screening prices, now $60-130 in Japan, will have to be subsidized by government to allow access to all citizens. Radiologists must be specially trained in the nuances of lung cancer screening, and assisted by computer-assisted detection technology. Cancer centers will need to continue to optimize diagnostic and treatment protocols.
The importance of integration of lung cancer screening protocols into medical education curricula, and the availability of web-based lung cancer screening teaching materials available at www.ielcap.org <
Frederic W. Grannis Jr. MD Department of Thoracic Surgery City of Hope National Medical Center Duarte CA fgrannis@coh.org