Dual reading as a routine procedure in mass radiography

Danish Tuberculosis Index

Abstract

Independent dual reading of survey films of the chest has been used as a routine method in the nation-wide Danish antituberculosis mass campaign. An attempt to evaluate the usefulness of this procedure is made in the present paper, which gives the results of dual reading of 431,550 technically adequate 35-mm chest films taken on adults in 12 Danish counties.

The films were first read locally by the chief physician of the county and then subjected to an independent reading by a second (central) reader. All persons whose films were designated as suspicious (i.e., representing photofluorographic indication of pulmonary disease) by at least one of the two readers were referred to the county chest clinics for a large roentgenogram. Altogether 10,391 persons—or 2.4% of the total examined—were referred because of a suspicious chest film. The local and the central readers on an average agreed on only one-third of the referrals.

Of the persons subjected to roentgenographic follow-up, 291 were found to have active pulmonary tuberculosis requiring treatment: 243 of these cases had been detected by the local readers, the central readers adding 48, or 20%. A further 2,804 persons presented evidence of clinically significant pulmonary disease other than active tuberculosis requiring treatment or observation: 1,924 of these referrals had been made by the local readers, the central readers adding 880, or 45%. In the remaining 6,907 persons, the large roentgenograms revealed either nothing abnormal or, in a small number of cases, only signs of healed pulmonary disease.

Thus, of the 10,002 photofluorograms read as suspicious and checked by large roentgenograms, 6,907 must be classed as “false positives” and 3,095 as “true positives”. While the local reading alone resulted in 2.1 false positives to each true positive, dual reading produced 2.2 false positives to each true positive—only a slight increase in the relative extent of over-reading.

The reading of the photofluorographic material of the mass campaign was performed by 12 experienced chief tuberculosis physicians. All without exception overlooked clinically significant pulmonary changes in a considerable number of cases, but the individual variations were great. An independent second reading of the photofluorograms thus considerably increased the value, at relatively small expense, of the mass X-ray survey as a case-finding procedure.