© The Author 2007. Published by Oxford University Press.
CORRESPONDENCE |
Response: Re: Extended Lung Cancer Incidence Follow-up in the Mayo Lung Project and Overdiagnosis
Affiliations of authors: Biometry Research Group, Division of Cancer Prevention, National Cancer Institute, Bethesda, MD (PMM); Survey Research Center, Mayo Clinic, Rochester, MN (AH, KPO)
Correspondence to: Pamela M. Marcus, PhD, Biometry Research Group, Division of Cancer Prevention, National Cancer Institute, 6130 Executive Blvd, Ste 3131, Bethesda, MD 20895-7354 (e-mail: pm145q{at}nih.gov).
We thank Morabia and Markowitz for their comments concerning our extended lung cancer incidence follow-up of the Mayo Lung Project cohort. They posit that oversurveillance, rather than overdiagnosis, explains the excess of lung cancer cases diagnosed among intervention arm participants.
Because the correspondence does not explicitly define the term oversurveillance, we interpret it to mean that participants in the intervention arm received more medical attention than participants in the usual-care arm. The reasons for such visits are varied: evaluation or treatment of a chronic medical condition, or preventive visits, including annual physical and cancer screening exams. We mentioned in our article (1) that continuation of screening by intervention arm participants could produce an excess of cases because catch-up (i.e., after screening ends, the usual-care arm counterparts of the screen-detected cases in the intervention arm are diagnosed from symptomatic presentation) would not occur. Data in table 4 of our article, however, indicated otherwise. Patterns of imaging from 1999 through 2001 and frequency of imaging since 1983 were similar across arms. There was a suggestion, although not statistically significant, that persons in the usual-care arm actually were more likely to have frequent imagining.
We question whether the method of Morabia and Markowitz, which involved the source of lung cancer information, is more useful than reported imaging information in assessing oversurveillance. Furthermore, we are uncertain as to why their measure of oversurveillance was based solely on Mayo Clinic records and death certificates. Neither source provides direct evidence of oversurveillance. Mayo Clinic records may reflect choice of a major medical facility for treatment of cancer diagnosed elsewhere among persons familiar with that facility as a result of their trial experiences. Death certificates are not a reliable source for cancer diagnoses. We did not collect death certificates on either a complete or a random sample of participants but used death certificates in a final attempt to obtain evidence of a lung cancer diagnosis for participants known only to be dead. We disagree with the authors’ use of the statements from Swensen et al. (2) to support the use of death certificates as a measure of oversurveillance. Swensen et al. state that "some lung cancer deaths may have been misattributed to other causes" because of "temporal improvements in the detection and diagnosis of cancer." Temporal improvements should affect the arms equally. Swensen et al. did not show misattribution; they merely hypothesized its existence for some lung cancers.
It would be naive to believe that no oversurveillance, particularly in the form of screening, existed in this cohort after 1983 or that overdiagnosis did not exist before 1983. To fully discredit the existence of overdiagnosis, one would need to explain away 85, or 15%, of lung cancers diagnosed in the intervention arm. The estimate of 12.5% from Morabia and Markowitz, which was based only on reports apparently supportive of their argument, does not suffice. An estimate that incorporates all sources (10.3%, 39 excess cases in the intervention arm of a total of 379 cases in the intervention arm), thus eliminating selectivity that could influence conclusions, further discounts the authors’ argument against overdiagnosis.
NOTES
The opinions expressed in this correspondence represent the views of the authors and do not necessarily represent those of the United States Department of Health and Human Services or the United States Federal Government.
REFERENCES
(1) Marcus PM, Bergstralh EJ, Zweig MH, Harris A, Offord KP, Fontana RS. Extended lung cancer incidence follow-up in the Mayo Lung Project and overdiagnosis. J Natl Cancer Inst (2006) 98:748–56.
(2) Swensen SJ, Jett JR, Hartman TE, Midthun DE, Mandrekar SJ, Hillman SL, et al. CT screening for lung cancer: five-year prospective experience. Radiology (2005) 235:259–65.
Related Correspondence
CiteULike 
Connotea 
Del.icio.us What's this?
J Natl Cancer Inst 2007 99: 898-899.
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||