NEW YORKTwo important developments are bringing about a
reappraisal of major institutional policy recommendations against
lung cancer screening: (1) The amassing of compelling evidence that
resection of early lung cancer has a major impact on survival and (2)
the emergence of new techniques allowing earlier disease detection.
The reappraisal was clear from presentations at the First
International Conference on Screening for Lung Cancer, convened among
experts in the field, and was manifested in the formation of a new
group called ICScreen (International Collaboration to Screen for Lung Cancer).
Noting that lung cancer kills more individuals than breast, colon,
cervix, and prostate cancer combined, the consensus statement issued
postconference provided a rationale for a change of strategy.
Recent scientific advances create an extraordinary potential to
develop a lung cancer screening program that would prevent untimely
deaths of vast numbers of current and former smokers who remain at
high risk despite smoking cessation, the statement said. An
estimated 160,000 lung cancer deaths will occur this year in the
The criteria for an acceptable screening test are that it be simple,
inexpensive, noninvasive, potentially widely available, and accurate.
The consensus statement from the conference suggests that, for lung
cancer screening, these criteria are met by the combined use of
spiral computed tomography (CT) and automated airway cell marker analysis.
While CT scanning detects peripheral tumors, airway cell marker
analysis detects positive cells in the sputum and has good
sensitivity for picking up squamous-cell and small-cell cancers.
Confidence in these modalities rests on international experience
involving more than 20,000 spiral CT scans and 9,000 airway cell
marker analyses. Research from Japan, where experience with spiral CT
scan screening for lung cancer is most extensive, suggests that the
technique can aid in early detection and save lives.
Tomotaka Sobue, MD, MPH, National Cancer Center Research Institute,
Tokyo, Japan, compared data from lung cancer screening collected at
two distinct time points in a cohort consisting primarily of smokers
over the age of 50. Screening from 1975 to 1993 was by chest x-ray
plus sputum cytology twice yearly, and screening from 1993 to 1998
was with low-dose spiral CT scan and sputum cytology twice yearly.
Nearly 26,000 exams were conducted in the first period among 2,529
subjects; 43 lung cancers were detected, Dr. Sobue said. In the
second period, 10,000 exams among 1,678 subjects detected 36 lung
cancer cases. The crude detection rates, Dr. Sobue noted,
were 0.16% in the first period and 0.37% in the second.
While cautioning that the follow-up period is insufficient, Dr. Sobue
reported that 5-year survival is higher among individuals whose
cancers were detected through spiral CT scanning (82.6% vs 48.8% for
the first period). CT alone detected 90% of the cancers found in the
second period, and survival was nearly 90% in that group. These
figures indicate that CT-detected lung cancer cases are highly
curable, Dr. Sobue said.
Dr. Sobue acknowledged that biases leading to overdiagnosis are a
possibility, but noted that analysis of current retrospective data
cannot clearly separate out such biases. Importantly, he estimates
that the length of the preclinical detectable phase of lung cancer
for CT screening is more than 5 years.
Screening is of value, the conference attendees agreed, only when it
is linked with appropriate diagnostic interventions and treatment.
They concluded that further evaluations of lung cancer screening
should be conducted within the framework of an overall research
program leading to a standardization of diagnostic evaluations and
treatments, to minimize unnecessary diagnostics, invasive procedures,
Further research is needed to determine more precisely which
populations should be screened and how often, and to identify
noninvasive diagnostic algorithms for abnormal screens, as well as