Double Reading Of Screening Mammograms Research Articles

Optimizing the Pairs of Radiologists That Double Read Screening Mammograms.

Background Despite variation in performance characteristics among radiologists, the pairing of radiologists for the double reading of screening mammograms is performed randomly. It is unknown how to optimize pairing to improve screening performance. Purpose To investigate whether radiologist performance characteristics can be used to determine the optimal set of pairs of radiologists to double read screening mammograms for improved accuracy. Materials and Methods This retrospective study was performed with reading outcomes from breast cancer screening programs in Sweden (2008-2015), England (2012-2014), and Norway (2004-2018). Cancer detection rates (CDRs) and abnormal interpretation rates (AIRs) were calculated, with AIR defined as either reader flagging an examination as abnormal. Individual readers were divided into performance categories based on their high and low CDR and AIR. The performance of individuals determined the classification of pairs. Random pair performance, for which any type of pair was equally represented, was compared with the performance of specific pairing strategies, which consisted of pairs of readers who were either opposite or similar in AIR and/or CDR. Results Based on a minimum number of examinations per reader and per pair, the final study sample consisted of 3 592 414 examinations (Sweden, n = 965 263; England, n = 837 048; Norway, n = 1 790 103). The overall AIRs and CDRs for all specific pairing strategies (Sweden AIR range, 45.5-56.9 per 1000 examinations and CDR range, 3.1-3.6 per 1000; England AIR range, 68.2-70.5 per 1000 and CDR range, 8.9-9.4 per 1000; Norway AIR range, 81.6-88.1 per 1000 and CDR range, 6.1-6.8 per 1000) were not significantly different from the random pairing strategy (Sweden AIR, 54.1 per 1000 examinations and CDR, 3.3 per 1000; England AIR, 69.3 per 1000 and CDR, 9.1 per 1000; Norway AIR, 84.1 per 1000 and CDR, 6.3 per 1000). Conclusion Pairing a set of readers based on different pairing strategies did not show a significant difference in screening performance when compared with random pairing. © RSNA, 2023.

Screening outcome for interpretation by the first and second reader in a population-based mammographic screening program with independent double reading.

Double reading of screening mammograms is associated with a higher rate of screen-detected cancer than single reading, but different strategies exist regarding reader pairing and blinding. Knowledge about these aspects is important when considering strategies for future use of artificial intelligence in mammographic screening. To investigate screening outcome, histopathological tumor characteristics, and mammographic features stratified by the first and the second reader in a population based screening program for breast cancer. The study sample consisted of data from 3,499,048 screening examinations from 834,691 women performed during 1996-2018 in BreastScreen Norway. All examinations were interpreted independently by two radiologists, 272 in total. We analyzed interpretation score, recall, and cancer detection, as well as histopathological tumor characteristics and mammographic features of the cancers, stratified by the first and second readers. For Reader 1, the rate of positive interpretations was 4.8%, recall 2.3%, and cancer detection 0.5%. The corresponding percentages for Reader 2 were 4.9%, 2.5%, and 0.5% (P < 0.05 compared with Reader 1). No statistical difference was observed for histopathological tumor characteristics or mammographic features when stratified by Readers 1 and 2. Recall and cancer detection were statistically higher and histopathological tumor characteristics less favorable for cases detected after concordant positive compared with discordant interpretations. Despite reaching statistical significance, mainly due to the large study sample, we consider the differences in interpretation scores, recall, and cancer detection between the first and second readers to be clinically negligible. For practical and clinical purposes, double reading in BreastScreen Norway is independent.

Characteristics of screen-detected cancers following concordant or discordant recalls at blinded double reading in biennial digital screening mammography.

To analyse which mammographic and tumour characteristics led to concordant versus discordant recalls at blinded double reading to further optimise our breast cancer screening programme. We included a consecutive series of 99,013 screening mammograms obtained between July 2013 and January 2015. All mammograms were double read in a blinded fashion. Discordant readings were routinely recalled without consensus or arbitration. During the 2-year follow-up, relevant data of the recalled women were collected. We compared mammographic characteristics, screening outcome and tumour characteristics between concordant and discordant recalls. There were 2,543 concordant recalls (71.4%) and 997 discordant recalls (28.0%). The positive predictive value of a concordant recall was significantly higher (23.5% vs. 10.0%, p < 0.001). The proportion of BI-RADS 0 was significantly higher in the discordant recall group (75.7% vs. 56.3%, p < 0.001). Discordant recalls were more often an asymmetry or architectural distortion (21.8% vs. 13.2% and 9.3% vs. 6.5%, respectively, p < 0.001). There were no differences in the distribution of DCIS and invasive cancers and tumour characteristics were comparable for the two groups, except for a more favourable tumour grade in the discordant recall group (54.7% vs. 39.9% grade I tumours, p = 0.022). Screen-detected cancers detected by a discordant reading show a more favourable tumour grade than cancers diagnosed after a concordant recall. The higher proportion of asymmetries and architectural distortions in this group provide a possible target for improving screening programmes by additional training of screening radiologists and the implementation of digital breast tomosynthesis. • With blinded double reading of screening mammograms, screen-detected cancers detected by a discordant reading show a more favourable tumour grade than cancers diagnosed after a concordant recall. • The proportions of asymmetries and architectural distortions are higher in case of a discordant reading. • Possible improvement strategies could target additional training of screening radiologists and the implementation of digital breast tomosynthesis in breast cancer screening programmes.

Arbitration of discrepant BI-RADS 0 recalls by a third reader at screening mammography lowers recall rate but not the cancer detection rate and sensitivity at blinded and non-blinded double reading

PurposeTo evaluate the characteristics of low suspicion lesions (BI-RADS 0) at blinded and non-blinded double reading of screening mammograms and to determine the potential effect of arbitration of discrepant BI-RADS 0 recalls by a third reader on screening outcome. MethodsWe included a series of 84,927 consecutive digital screening mammograms, double read in a blinded (43,184 screens) or non-blinded (41,743 screens) fashion, between July 2009 and July 2011. Discrepant readings were routinely recalled for further evaluation. During 2 years of follow-up, radiology, surgical and pathology reports were collected of all recalled women. Arbitration of discrepant BI-RADS 0 recalls (only one radiologist assigning a BI-RADS 0 score) was retrospectively performed by a third screening radiologist. ResultsAt blinded and non-blinded double reading, 32.0% and 32.5% of recalls were assigned BI-RADS 0 with a positive predictive value (PPV) of 7.2% and 6.8%, respectively. Compared to non-blinded double reading, BI-RADS 0 recalls at blinded double reading showed a higher discrepancy rate (9.0 versus 4.3 per 1000 screens, p < 0.001) and false positive recall rate (10.1 versus 8.4 per 1000 screens, p = 0.012).Arbitration of discrepant BI-RADS 0 recalls would have significantly lowered recall rate (from 3.4% to 2.8% at blinded double reading, p < 0.001, and from 2.8% to 2.5% at non-blinded double reading, p = 0.008), without a decrease in cancer detection rate (from 7.5‰ to 7.3‰, p = 0.751, and from 6.6‰ to 6.5‰, p = 0.832, respectively) and program sensitivity (from 83.2% to 81.2%, p = 0.453, and from 76.0% to 74.6%, p = 0.667, respectively). Arbitration would have significantly increased the PPV at blinded double reading (from 22.3% to 26.3%, p = 0.015). ConclusionWe advise arbitration of discrepant BI-RADS 0 recalls, at (non-)blinded double reading of screening mammograms, to reduce recall rates and improve the PPV of recall at blinded double reading.

Variations in screening outcome among pairs of screening radiologists at non-blinded double reading of screening mammograms: a population-based study

Substantial inter-observer variability in screening mammography interpretation has been reported at single reading. However, screening results of pairs of screening radiologists have not yet been published. We determined variations in screening performances among pairs of screening radiologists at non-blinded double reading. We included pairs of screening radiologists with at least 7,500 screening examinations per pair, obtained between 1997 and 2011. During 2-year follow-up, breast imaging reports, surgical reports and pathology results were collected of all referred women and interval cancers. Referral rate, cancer detection rate, positive predictive value and sensitivity were calculated for each pair. A total of 310,906 screening mammograms, read by 26 pairs of screening radiologists, were included for analysis. The referral rate ranged from 1.0 % (95 % CI 0.8 %-1.2 %) to 1.5 % (95 % CI 1.3 %-1.8 %), the cancer detection rate from 4.0 (95 % CI 2.8-5.2) to 6.3 (95 % CI 4.5-8.0) per 1,000 screens. The programme sensitivity and positive predictive value of referral ranged from 55.1 % (95 % CI 45.1 %-65.1 %) to 81.5 % (95 % CI 73.4 %-89.6 %) and from 28.7 % (95 % CI 20.8 %-36.6 %) to 49.5 % (95 % CI 39.7 %-59.3 %), respectively. We found significant variations in screening outcomes among pairs of screening radiologists at non-blinded double reading. This stresses the importance of monitoring screening results on a local scale. • Substantial inter-observer variability in screening mammography interpretation is known at single reading • Population-based study showed significant variations in outcomes among pairs of screening radiologists • Local monitoring and regular feedback are important to optimise screening outcome.

Benefits of double reading of screening mammograms: retrospective study on a consecutive series

The purpose of this study was to assess the performance of delayed second reading of screening mammograms when added to real-time reading plus immediate assessment. The study setting was the mammography screening programme of an Italian Local Health Unit. Recall rate and cancer detection rate at first reading or informed second reading only were assessed in a cohort of 23,629 women aged 50-69 years screened during 2007-2008. Incremental recall rate, incremental cancer detection rate and incremental cost of second reading were determined. Recall rate was 13.0% at first and 2.7% at second reading (incremental recall rate +21.1%). Overall, recalls were more frequent in the younger decade and in the presence of denser breasts. Cancer detection rate was 7.06‰ (n=167) at first and 0.93‰ (n=22) at second reading (incremental cancer detection rate +13.1%). Compared with first reading, second reading detected more cancers depicted as isolated microcalcifications and distortions (40.9% vs. 16.2%, p=0.02) and at a lower stage (stage 0-I 81.8% vs. 69.5%, p=0.34). The cost of adding delayed second reading was + <euro> 3.65 per screened individual or <euro> 3,926.61 per incremental cancer detected. The study confirms the efficacy of second reading, even as an adjunct to real-time single reading plus immediate assessment. Incremental recall rate is acceptable in view of the incremental cancer detection rate, and both figures are within the range of literature reports on double-reading performance.

Breast cancers missed by the arbitration process

In our institution double reading of screening mammo-grams with arbitration by consensus was introduced in January 2000. Discordant double reading opinions are discussed at a weekly arbitration meeting that all screen readers attend and a decision is made as to whether the patient should be recalled to assessment or returned to routine screening. To date, there is little known about the outcome of arbitrated cases that are returned to routine screening. We aim to identify the number of breast cancer cases that subsequently developed at the arbitrated site that were returned to routine screening at the arbitration meeting and to determine the mammographic and pathological features of these cancers.

Consensus approach to discrepancies arising from independent double reading of screening mammograms

Consensus approach to discrepancies arising from independent double reading of screening mammograms

Comparison of Computer-Aided Detection to Double Reading of Screening Mammograms: Review of 231,221 Mammograms

This study compares the efficacy of single reading with computer-aided detection (CAD) to double reading and also to the first reader (without CAD) in a double-reading program. A review was performed of 231,221 screening mammograms interpreted by experienced mammographers from 2001 through 2005 in a community-based mammography program. In 112,413 (48.6%), mammographers performed the first of two readings. In 118,808 (51.4%), they performed a single reading aided by CAD. For double reading, the first reader's recall rate was 10.2%; sensitivity, 81.4%; positive predictive value, as a percentage of positive screening mammograms resulting in a tissue diagnosis of cancer within 1 year (PPV(1)), 4.1%; and cancer detection rate, 4.12 per 1,000. After the double-reading process, the final recall rate was 11.9%; sensitivity, 88.0%; PPV(1), 3.7%; and cancer detection rate, 4.46 per 1,000. For single-reading with CAD, the recall rate was 10.6%; sensitivity, 90.4%; PPV(1), 3.9%; and cancer detection rate, 4.20 per 1,000. Statistically significant results included a lower recall rate with CAD compared with double reading (10.6% vs 11.9%, respectively; p < 0.0001); increased sensitivity with CAD compared with the first reader (90.4% vs 81.4%, p < 0.0001); and increased recall rate with CAD compared with the first reader (10.6% vs 10.2%, p < 0.0001). Double reading increased sensitivity with a modest increase in the recall rate compared with single reading. Single reading with CAD, compared with double reading, resulted in a small, but not statistically significant, increase in sensitivity with a lower recall rate. Our results indicate that CAD enhances performance of a single reader, yielding increased sensitivity with only a small increase in recall rate.

Additional Double Reading of Screening Mammograms by Radiologic Technologists: Impact on Screening Performance Parameters

Studies have shown that having mammography technologists review screening mammograms in addition to radiologist review may increase the number of breast cancers that are detected at screening mammography. We prospectively examined the effects on screening performance of adding independent double reading of screening mammograms by technologists to standard double reading by radiologists. Twenty-one screening mammography technologists and eight certified screening radiologists participated in this study. From January 1, 2003, to January 1, 2005, all 61,251 screening mammograms obtained at two mammography screening units in The Netherlands were independently read (although the second reader was not blinded to the first reader's interpretation) by two technologists and, in turn, by two radiologists. Radiologists were blinded to the referral opinion of the technologists. During a 2-year follow-up period, we collected clinical data, breast imaging reports, biopsy results, and breast surgery reports of all women with a positive screening result (i.e., those that required additional imaging) from any reader. The distributions of categorical variables between subgroups were compared using chi-square or Fisher's exact tests. Differences in referral and detection by radiologists and technologists were assessed using McNemar's test. All statistical tests were two-sided. The radiologists referred 905 women (referral rate = 1.48%, 95% confidence interval [CI] = 1.38% to 1.57%), of whom 323 had breast cancer, corresponding to a positive predictive value of referral (PPV) of 35.7% (95% CI = 32.6% to 38.8%). Review of 446 additional technologist-positive readings led to another 80 referrals, resulting in the detection of 22 additional cancers. These extra referrals increased the initial referral rate from 1.48% to 1.61% (difference = 0.13%; 95% CI = 0.10% to 0.16%) and the cancer detection rate (CDR) from 5.27 to 5.63 cancers per 1000 women screened (difference = 0.36 cancers per 1000 women screened; 95% CI = 0.24 to 0.55). With technologist double reading only, 829 women would have been referred (referral rate = 1.35%, 95% CI = 1.26% to 1.45%); among these women, 286 cancers were diagnosed (PPV = 34.5%, 95% CI = 31.3% to 37.7%; CDR = 4.67 cancers per 1000 women screened, 95% CI = 4.13 to 5.21). Referral of all 1351 radiologist- and/or technologist-positive readings would have led to 362 cancers found at screening. The cancer detection rate for radiologist double reading would have increased from 5.27 to 5.91 cancers per 1000 women screened (relative increase = 12.1%, 95% CI = 8.8% to 16.5%; difference = 0.64 cancers per 1000 women screened, 95% CI = 0.47 to 0.87). A referral strategy that includes all technologist-positive readings, which would have increased the cancer detection rate while maintaining a low referral rate, should be considered.

Single Reading with Computer-aided Detection and Double Reading of Screening Mammograms in the United Kingdom National Breast Screening Program

To retrospectively determine if the use of a computer-aided detection (CAD) system can improve the performance of single reading of screening mammograms to match that of double reading in the United Kingdom. Local research ethics committee approval was obtained; informed consent was not required. This study included a sample of 10 267 mammograms obtained in women aged 50 years or older who underwent routine screening at one of two breast screening centers in 1996. Mammograms that were double read in 1996 were randomly allocated to be re-read by eight different radiologists using CAD. The cancer detection and recall rates from double reading and single reading with CAD were compared. Statistical significance and confidence intervals were calculated with the McNemar test to account for the matched nature of the data. Single reading with CAD led to a cancer detection rate that was significantly (P = .02) higher than that achieved with double reading: 6.5% more cancers were detected by means of single reading with CAD than by means of double reading. However, the recall rate was higher for single reading with CAD than for double reading (8.6% vs 6.5%, respectively; P < .001). This was equivalent to relative increases of 15% and 32% in the cancer detection and recall rates, respectively. Single reading with CAD leads to an improved cancer detection rate and an increased recall rate.

The role of arbitration of discordant reports at double reading of screening mammograms

To asses the effectiveness of arbitration of discordant double readings in mammography screening. A retrospective study of 1217 consecutive arbitrations. A subset of discordant double readings from the Florence screening programme underwent arbitration by a third reader. Positive arbitration of 1217 discordant double readings prompted assessment in 476 cases (39.2%), detecting 30 cancers (6.3%). Of 741 negative arbitrations (60.8%), 311 have been followed up thus far, and two cancers (0.64%) occurred in the site previously suspected at one of the two independent readings. Arbitration had a sensitivity of 86.3% and a negative predictive value of 99.3%. Arbitration reduced the overall referral rates from 3.82% to 2.59% (relative decrease 32.1%). Due to false-negative arbitration, cancers detected per 1000 women screened would decrease from 4.58 to 4.50 (relative decrease 1.7%). For every cancer missed due to false-negative arbitration, 151 unnecessary recalls and 21,248 euro would have been saved, whereas the saved cost per screened woman due to arbitration was 1.72 euro. Arbitration of discordant double reading would substantially reduce referral rates with a limited reduction in cancer detection rate, and may be recommended as a routine procedure. Greater benefit from arbitration might be expected in the presence of high referral rates at independent double reading, a common scenario in a newly implemented service screening.

Can computer-aided detection with double reading of screening mammograms help decrease the false-negative rate? Initial experience.

To retrospectively evaluate the role of computer-aided detection (CAD) in reducing the rate of false-negative (FN) findings on screening mammograms considered normal at initial double reading. At the authors' institution, independent prospective double readings in which the second reader is not blinded to results of the first reading are performed routinely for all mammograms. When cancer is diagnosed, prior mammograms also are reviewed with double reading to determine cancer visibility. Findings are categorized as (a) no evidence of cancer on any prior screening mammogram and patient presents more than 1 year after prior screening, (b) no evidence of cancer on any prior screening mammogram and patient presents with symptoms within 1 year after prior screening (year-interval occult false-negative), or (c) cancer visible. The clinical director separately evaluates each case in the same way. In 2000, 519 histologically proved breast cancers were diagnosed, including 132 for which patients sought a second opinion and FN findings were not tracked. Prior screening mammograms were available in 318 of the other 387 cases. Five radiologists in two reading sessions independently reviewed current and prior mammograms to categorize visible cancers as either threshold or actionable FN findings. Visible cancers deemed actionable by at least three of five readers were analyzed with a commercially available CAD system. FN rates were calculated prior to and after CAD analysis. Twenty-seven occult and 71 visible cancers were found (total FN findings, 98). Three of five readers considered 52 (73%) of 71 visible cancers actionable. The CAD system correctly marked 37 (71%) of these 52 on prior screening mammograms (19 [65%] of 29 masses, seven [88%] of eight microcalcifications, seven [78%] of nine architectural distortions, and four [67%] of six masses with microcalcifications). The FN rate was 98 (31%) of 318 before CAD and 61 (19%) of 318 after CAD. In this retrospective review of this small subset of cancers, it appears that CAD has the potential to decrease the FN rate at double reading by more than one-third (from 31% to 19%). The CAD system correctly marked 37 (71%) of 52 actionable findings read as negative in previous screening years.

Independent double reading of screening mammograms in The Netherlands: effect of arbitration following reader disagreements.

To prospectively determine the value of arbitration by a panel of radiologists when two radiologists performing independent readings of screening mammograms do not reach a consensus about referral. The study population consisted of women who participated in the Dutch Nationwide Breast Cancer Screening Program, in which biennial screening is offered to women aged 50-75 years. An arbitration panel of three radiologists assessed those screening mammograms for which two screening radiologists did not reach a consensus about referral necessity. Women were referred for further analysis if at least one arbitration panel radiologist considered referral to be necessary. The two screening radiologists agreed on the recommendation for referral of 498 (0.8%) of 65,779 screened women and on the recommendation for no referral of 64,949 (98.7%) women. They initially disagreed about the referral in 332 (0.5%) cases. After a mutual consultation, disagreement persisted regarding 183 (0.3%) mammograms. The arbitration panel referred 89 of these cases for further analysis, which revealed cancer in 20 (22%) cases. In three (3%) of the 94 cases that were not referred by the panel, breast cancer was detected at the site of previously discrepant mammographic findings seen at subsequent screening performed 2 years later. If all 183 discrepant cases had been referred, the referral rate would have increased from 0.8% to 0.9% at subsequent (incident) screenings and from 1.5% to 1.7% at initial screenings. In addition, at subsequent screenings, the number of cancers detected per 1,000 women screened would have increased from 4.4 to 4.5. Mammograms with discrepant findings constitute a very important subset of screening mammograms. All lesions that are subsequently proved to be malignant may not be detected with panel arbitration.

Double reading of screening mammograms: the use of a third reader to arbitrate on disagreements

Abstract Screening mammograms from a County wide breast screening programme were double read by any 2 of a team of 3 radiologists. There were 398 cases where the 2 readers disagreed. The third radiologist was asked to arbitrate on these. The third readers recalled 49% and returned the other 51% to normal screening. From the 196 recalled for assessment, 9 underwent histological examination and from these, 4 malignant lesions were found. From the 202 sent to normal screen, one true interval cancer has occurred during a 3 year follow-up period. Double reading of mammograms increases the sensitivity of screening. If this is combined with arbitration in disagreements the number of unnecessary recalls is reduced.

Cost-effectiveness of Consensus Double Reading of Screening Mammograms

HomeRadiologyVol. 209, No. 2 PreviousNext LettersCost-effectiveness of Consensus Double Reading of Screening MammogramsPieter Van Der Valk, David Beijerinck, Jan J. M. DeurenbergPieter Van Der Valk, David Beijerinck, Jan J. M. DeurenbergAuthor AffiliationsUniversity Hospital Utrecht, Room E01.132, Heidelberglaan 100, 3584 CX Utrecht, NetherlandsPieter Van Der ValkDavid BeijerinckJan J. M. DeurenbergPublished Online:Nov 1 1998https://doi.org/10.1148/radiology.209.2.588-bMoreSectionsPDF ToolsImage ViewerAdd to favoritesCiteTrack CitationsPermissionsReprints ShareShare onFacebookTwitterLinked In Article HistoryPublished in print: Nov 1998 FiguresReferencesRelatedDetailsCited ByEuropean Journal of Cancer, Vol. 44, No. 6Short-term Follow-up Results in 795 Nonpalpable Probably Benign Lesions Detected at Screening Mammography1Isidro Vizcaíno, , Leopoldo Gadea, , Luis Andreo, , Dolores Salas, , Francisco Ruiz-Perales, , Dolores Cuevas, , Constantino Herranz, , Francisco Bueno, , and 1 May 2001 | Radiology, Vol. 219, No. 2Recommended Articles RSNA Education Exhibits RSNA Case Collection Vol. 209, No. 2 Metrics Altmetric Score PDF download

Double reading of screening mammograms will have resource implications

EDITOR,—Jackie Brown and colleagues show that in their breast screening unit double reading of screening mammograms with recall by consensus between film readers is cost effective.1 Unfortunately, these results cannot be directly extrapolated to the NHS breast screening programme as the study was carried out during the first (prevalence) screening round, when the overall rate of detection of cancer is higher and the number of additional cancers detected by double reading is likely to be greater than in …

Mammography screening: an incremental cost effectiveness analysis of double versus single reading of mammograms

To compare mammography reading by one radiologist with independent reading by two radiologists. An observational non-randomised trial at St Margaret's Hospital, Epping. 33 734 consecutive attenders for breast screening in the main trial and a sample of 132 attenders for assessment who provided data on private costs. Three reporting policies were compared: single reading, consensus double reading, and non-consensus double reading. Numbers of cancers detected, recall rates, screening and assessment costs, and cost effectiveness ratios. A policy of double reading followed by consensus detected an additional nine cancers per 10 000 women screened (95% confidence interval 5 to 13) compared with single reading. A non-consensus double reading policy detected an additional 10 cancers per 10 000 women screened (95% confidence interval 6 to 14). The difference in numbers of cancers detected between the consensus and non-consensus double reading policies was not significant (95% confidence interval -0.2 to 2.2). The proportion of women recalled for assessment after consensus double reading was significantly lower than after single reading (difference 2.7%; 95% confidence interval 2.4% to 3.0%). The recall rate with the non-consensus policy was significantly higher than with single reading (difference 3.0%; 2.5% to 3.5%). Consensus double reading cost less than single reading (saving 4853 pounds per 10 000 women screened). Non-consensus double reading cost more than single reading (difference 19 259 pounds per 10 000 women screened). In the screening unit studied a consensus double reading policy was more effective and less costly than a single reading policy.

Independent double reading of screening mammograms.

To evaluate the cost effectiveness of independent double reading of screening mammograms. Prospective study of 18,817 women undergoing first or repeat screening in a population based programme in the Florence district. Mammograms were independently double read by experienced radiologists. Subjects with mammographic abnormalities reported by at least one reader were recalled for diagnostic assessment. The mean increase in recall rate, cancer detection rate, and screening costs attributable to double reading was calculated. Eleven of 125 cancers were detected by only one reader. The mean increase in cancer detection rate attributable to double reading compared with single reading was 4.6% (95% confidence interval (CI) 1.1 to 8.9). From a total of 748 cases referred for diagnostic assessment, 196 subjects were referred by one reader only. The mean increase in referral rate attributable to double reading compared with single reading was 15.1% (CI 12.3 to 17.8). Double reading caused a marked increase in the cost for each woman screened -8.5% at the first screening and 6.2% at repeat screening and a more limited increase in the cost for each cancer detected -3.5% at the first screening and 2.7% at repeat screening. Cancers detected by only one screener were at an earlier stage than those detected by both screeners (P = 0.6, not significant). Independent double reading results in only a modest increase in the detection of cancers and therefore may not be cost effective.