Recall Rate of Opportunistic Screening Mammography in a University Referral Breast Center in Iran


Afsaneh Alikhassi 1 , * , Maryam Rahmani 2 , Nasrin Ahmadinejad 3 , Sona Akbari 4 , Farzin Roozafzai 4 , Zahra Alikhassy Habibabadi 5

1 Department of Radiology, Cancer Institute of Iran, Tehran University of Medical Sciences, Tehran, Iran

2 Department of Radiology, Vali-Asr Hospital, Tehran University of Medical Sciences, Tehran, Iran

3 Department of Radiology, Medical Imaging Center, Tehran University of Medical Sciences, Tehran, Iran

4 Department of Radiology, Tehran University of Medical Sciences, Tehran, Iran

5 Department of Surgery, John Hopkins School of Medicine, Baltimore, Maryland

How to Cite: Alikhassi A, Rahmani M, Ahmadinejad N, Akbari S, Roozafzai F, et al. Recall Rate of Opportunistic Screening Mammography in a University Referral Breast Center in Iran, Iran Red Crescent Med J. 2017 ; 19(10):e58569. doi: 10.5812/ircmj.58569.


Iranian Red Crescent Medical Journal: 19 (10); e58569
Published Online: October 10, 2017
Article Type: Research Article
Received: November 21, 2016
Revised: January 6, 2017
Accepted: February 25, 2017




Background: No information has been published on the effectiveness of digital non-diagnostic opportunistic screening mammography in Iran that is measured by recall rate as one of its indices.

Objectives: In this longitudinal study, we measured recall rate of non-diagnostic mammography at a tertiary referral university hospital and made a comparison with reported international data.

Methods: We examined 9395 digital mammograms performed in 2014 - 2015 from which, 2930 were the first-time and 6465 were subsequent mammography. The patients were referred to the university hospital by their clinicians during annual check-ups while none of them had any chief complaint. The mean age was 49 years. We calculated recall rate, sensitivity, specificity, and cancer detection rate.

Results: Breast cancer was diagnosed in 80 patients. Recall rates were 29% for the first-time and 22% for subsequent mammography, and the overall rate of cancer incidence was 8.5 per 1000 mammograms (80/9395) with specificity of 75.9%, sensitivity of 97.5%, PPV of 3.4%, and NPV of 99%.

Conclusions: The recall rate was much higher in this setting than the acceptable range reported in literature. However, the sensitivity and detection rate were higher; thus, the higher recall rate could be due to some differences in the patient population such as being at younger ages and higher risks.


Recall Rate Mammography Breast Cancer Screening Detection Rate

Copyright © 2017, Iranian Red Crescent Medical Journal. This is an open-access article distributed under the terms of the Creative Commons Attribution-NonCommercial 4.0 International License ( which permits copy and redistribute the material just in noncommercial usages, provided the original work is properly cited.

1. Background

Mammography can be done for diagnosis after a chief complaint from the patient or just for check-up screening. The recall rate has been defined differently in previous studies. Some studies restrict the recall rate to the need for further imaging to make a final decision, while other studies include clinical examinations, further imaging, both or just biopsies. Obviously, the different definitions affect the recall rate. The recall rate in this study was defined as the frequency with which a radiologist needs additional imaging or repetition of imaging before a final recommendation or recalling the patient for biopsy.

The recall rate is commonly used as a measure in screening mammography practice in some countries (1). A low recall rate can be associated with decreased sensitivity and increased number of false-negative results, while a rate that is too high increases false-positive results, increased costs, patient anxiety, and overload job on staff (2-7). Reported recall rates range from less than 1% to about 15% (7). The recommended recall rate according to the American college of radiology and the U.S. Agency for health care policy and research is less than 10% (8, 9). European guidelines recommend a target recall rate of 5% (with an acceptable rate of less than 7% for the first screenings and less than 5% for subsequent screenings) (10, 11).

Multiple factors seem to have an impact on the recall rate, including patient population, radiologist, employed techniques, and systemic factors. Factors related to patient population are age (12), breast density (13), use of hormone replacement therapy (12), time interval since obtaining the previous mammogram (14, 15), family history (16), and previous benign biopsy results (16). Radiologist factors that have been proposed to affect recall rate include sex (17), fellowship training in mammography (18), years of work experience (17, 19), and affiliation to an academic medical center (20). Factors related to technique and types of mammography machine include the use of digital or analogue device, tomosynthesis (21, 22), 3D images (23), computer-aided detection rate and even, skill of the technologists (24). Systemic factors shown to affect recall rate include reading volume (25), double versus single reading (26), and computer-aided reading (27).

In this study, we determined the recall rate at a tertiary referral university hospital and compared the data with international data. Patients came to the hospital to do self-paid screening mammography after receiving information from media meant to inform the population or following referral by clinicians for mammography check-ups.

2. Methods

2.1. Design

In Iran, there is no governmental plan for mammography screening. Mass media are trying to teach people about the value of breast cancer screening, and general physicians, gynecologists, and surgeons refer women who are usually above 40 to imaging centers for screening mammography. We conducted a longitudinal study in an academic tertiary referral hospital to measure the recall rate of this type of non-diagnostic mammography. We included mammographies from patients who did not have any chief complaint or positive clinical examination.

Before doing the mammography, patients signed a written informed consent form. Mammography was performed using dedicated, fully digital Selenia Dimensions mammography system (Hologic Inc., Marlborough, USA) in the breast clinic of the cancer institute of Imam Khomeini hospital, where breast surgeons were also working. Craniocaudal and mediolateral oblique views of each breast were recorded. The data used in this study were collected from mammograms read by three breast imaging radiologists with 6 to 20 years of breast imaging experience between Jan 1, 2014, and Jan 1, 2015. Computer-assisted detection software was not used.

Mammogram reports were categorized using the fifth edition of the breast imaging reporting and data system (BI-RADS) from the American college of radiology. All examinations with BI-RADS scores of 0 (additional imaging required), 4 (suspicious finding), and 5 (highly suspicious finding) were regarded as recall patients.

For 12 months, the patients were followed up and in the case of biopsies, an expert pathologist examined breast tissue specimens of core needle biopsies from patients with BI-RADS 4 or 5.

2.2. Subject

After exclusion of 11 cases with missing data, we included 9395 screening mammograms from women without any chief complaint or positive clinical breast examination, representing 3135 mammograms per radiologist.

2.3. Data Acquisition

Using MS Office Excel (Microsoft, Redmond, USA) during medical records review, we gathered data pertaining age, mammography session (first or subsequent), mammographic findings (breast mass, asymmetry, distortion, and micro-calcification), histopathological type of breast lesion (benign, atypia, in situ, and invasive), BI-RADS score, and “recall” mammograms.

2.4. Data Analysis

Dividing the number of “recall” mammograms by the number of screening mammograms gave us the recall rate. We assessed the sensitivity, specificity, negative predictive value (NPV), and positive predictive value (PPV) of recall mammograms for detection of in situ or invasive breast carcinoma within the follow-up period. Histopathological diagnosis (biopsy) was the gold standard. We defined cancer detection rate as the number of cancers (in situ or invasive) with positive initial interpretation (true positive recall mammogram) per 1000 screening mammograms. The cancer detection rate was also defined as the number of cancers with positive initial interpretation (recall) among 1000 screening mammograms (TP/TP + FN + FP + TN). We used descriptive statistics, including frequency distribution, mean, and standard deviation, to report the findings. Kappa measurement was calculated to assess the agreement between three radiologists on reporting mammograms (recall rates). ANOVA and independent t test was used to compare age factor between the categories. To compare the recall rate in categorical variables, we used chi square and Fisher’s exact tests. Type I error was considered 0.05. All analyses were conducted using SPSS v.22 (IBM Corp., Armonk, USA).

2.5. Ethical Considerations

All identity revealing information is preserved. Researchers imposed no harms on the patients. Researchers in this project are committed to the principles of the declaration of Helsinki and declare no conflicts of interests.

3. Results

The age ranged 25 to 78 years with the mean (± standard deviation) of 49.84 (± 9.19) years. Table 1 represents the frequency distribution and age of patients in BI-RADS categories. Patients in BI-RADS category 4 (suspicious findings, n = 180) were the youngest (44.06 ± 6.40 years), and the mean age significantly differed among categories (P value < 0.001). The recall rate in total, in the first mammograms (n = 2930), and in subsequent mammograms (n = 6465) were 24.7%, 29%, and 22%, respectively. “Recall” patients (n = 2320) were younger than patients not recalled (n = 7075) (48.00 ± 8.27 versus 50.44 ± 9.39 years, independent t test statistic = 11.901, P value < 0.001), and the recall rate significantly decreased per age decade (P value < 0.001, Table 2).

Table 1. Age of Women Based on BI-RADS Categories of Screening Mammography
BI-RADS CategoryFrequency, No. (%)Age, Mean ± SDANOVA
F statisticP value
0a2060 (21.9)48.24 ± 8.1978.249< 0.001
11570 (16.7)47.78 ± 8.94
24885 (52.0)51.36 ± 9.49
3590 (6.3)49.05 ± 7.38
4a180 (1.9)44.06 ± 6.40
5a80 (0.9)50.63 ± 10.95
630 (0.3)66.33 ± 9.97
Total9395 (100)49.84 ± 9.19--

Abbreviations: ANOVA, Analysis of Variance; BI-RADS, Breast Imaging Reporting And Data System; SD, Standard Deviation.

a”Recall” patients.

Table 2. Recall Rates in Different Age Groups of Women Undergoing Screening Mammography
Age Group, yCount, No. (%)RecallChi-Square
FrequencyRecall rate, %StatisticP value
< 40880 (9.4)27030.7137.710< 0.001
40 - 494090 (43.5)118028.9
50-592955 (31.5)64021.7
60 - 691140 (12.1)19016.7
≥ 70330 (3.5)4012.1
Total9395 (100)232024.7--
Table 3. Frequency Distributions of Different Types of Breast Lesions in 230 Biopsies from 9395 Women Undergoing Screening Mammography
Type of Breast LesionFrequencyOverall Percentage
In situ BC100.1
Invasive BC700.7
Total biopsies2302.4

Abbreviation: BC, Breast Cancer.

The agreement between radiologists (A, B, and C) in classifying “recall” mammograms was strong (A and B: kappa = 0.7, A and C: kappa = 0.7, B and C: kappa = 0.8).

Total 80 patients (34.8% of 230 biopsies) had breast cancer (10 in situ and 70 invasive breast cancer, table3), and cancer detection rate was 8.5 per 1000 mammograms (80/9395) with specificity: 75.9%, sensitivity: 97.5%, and PPV: 3.4%. Table 4 represents the contingencies.

Table 4. Contingency for for Recall Mammograms and Histopathological Diagnosis of Breast Cancer (230 Biopsies) in 9395 Women Undergoing Screening Mammographya
VariablesCancerbNo CancerTotal
No recall270737075

aSpecificity, 75.9%; Sensitivity, 97.5%; PPV, 3.4%; NPV, 99%.

bConsidering in situ and invasive.

As represented in Table 5, micro-calcification was 48.5% prevalent and came with lower recall rate (21.1% versus 28.1%, P value < 0.001). Mass was detected in 16.0% of mammograms. Recall rate was higher in the presence of mass (49.3% versus 20.0%, P value < 0.001). Distortion was 2.4% prevalent in mammograms and recall rate was higher in patients with distorted lesions (34.8% versus 24.4%, P value < 0.001). Asymmetric lesions were detected in 32.9% of mammograms. Recall rate for asymmetric lesions was higher (48.1% versus 13.2%, P value < 0.001).

Table 5. Description and Analysis of “Recall Rates” in Different Screening Mammographic Findings
Mammographic FindingsCategoryFrequency No. %RecallChi-Square
FrequencyRecall rate, %StatisticP value
Micro-calcificationYes4555 (48.5)96021.162.247< 0.001
No4840 (51.5)136028.1
MassYes1500 (16.0)74049.3582.738< 0.001
No7895 (84.0)158020.0
DistortionYes230 (2.4)8034.812.904< 0.001
No9165 (97.6)224024.4
AsymmetryYes3095 (32.9)149048.11364.628< 0.001
No6300 (67.1)83013.2
Total9395 (100)232024.7--

Of significance (all P values < 0.001), micro-calcification was weakly correlated with age (Pearson’s correlation coefficient = 0.139), asymmetry (-0.229), mass (-0.179), and BI-RADS score (0.196). Table 6 shows multivariate logistic model for prediction of “recall” (Nagelkerke R2 = 0.631, Chi-square = 5196.457, dF = 6, P value < 0.001). Mammographic detection of mass was the strongest predictor of “recall” (OR = 11.467, 95% CI: 9.464 - 13.894). Notably and opposing to univariate analysis (Table 7), micro-calcification showed an increased probability of “recall” (OR = 2.347, 95% CI: 2.018 - 2.731), adjusted for age, distortion, asymmetry, mass, and BI-RADS score in multivariate logistic regression.

Table 6. Multivariate Logistic Regressions with “Recall” Mammogram as Outcome Variable
PredictorsBetaOR (95% CI)P Value
Age-0.0200.980 (0.972 - 0.988)< 0.001
BIRADS-1.7100.181 (0.167 - 0.196)< 0.001
Mass2.44011.467 (9.464 - 13.894)< 0.001
Asymmetry2.0868.052 (6.912 - 9.380)< 0.001
Distortion2.0818.016 (5.292 - 12.142)< 0.001
Micro-calcification0.8532.347 (2.018 - 2.731)< 0.001
Intercept-7.2170.001< 0.001
Table 7. Univariate Logistic Regressions with “Recall” Mammogram as Outcome Variable
PredictorBetaOR (95% CI)P Value
Asymmetry1.8116.118 (5.528 - 6.772)< 0.001
Mass1.3593.892 (3.468 - 4.367)< 0.001
Distortion0.5001.649 (1.252 - 2.172)< 0.001
BIRADS score-1.8670.155 (0.144 - 0.166)< 0.001
Micro-calcification-0.3810.683 (0.621 - 0.751)< 0.001
Age (years)-0.0300.970 (0.965 - 0.975)< 0.001
Age decade-0.3070.736 (0.698 - 0.776)< 0.001

4. Discussion

In this study, mass was detected in 16.0% of mammograms, and recall rate was higher in the presence of mass (49.3% versus 20.0%, P value < 0.001). Mammographic detection of mass was the strongest predictor of “recall” (OR = 11.467, 95% CI: 9.464 - 13.894). One of our recalled patients with mass is depicted in Figure 1.

Figure 1. A 44 Year Old Lady Without Any Chief Complaint Who Referred for Mammary Screening
A 44 Year Old Lady Without Any Chief Complaint Who Referred for Mammary Screening

A, full digital mammography in MLO view shows a small dense mass in lower part of left breast; B, the same mass in CC view is in the inner part; C, in focal compression magnification view, it has speculated border; D, in target sonography, the mass was found which made biopsy and wire localization possible; it was proven to be an invasive ductal carcinoma.

The recall rates for non-diagnostic check-up mammograms were 29% for the first-time mammograms and 22% for subsequent mammograms, which are high in comparison with the range of 5% - 20% reported in the United States (28-34) but more similar to an Asian study (35). The sensitivity was also higher in this study than most previous studies (2, 35). Various optimum recall rates have been suggested in literature. Yankaskas et al. (2) proposed a recall rate between 4.9% and 5.5% and showed a plateau in the association between sensitivity and recall rate above this range. In contrast, Gur et al. (30) found that increases in recall rates beyond 10% still increased the detection rate. Schell et al. (36) performed a cost-benefit analysis to determine an optimal recall rate and recommended average recall rates of 10.0% and 6.7% for the first mammograms and subsequent mammograms, respectively. Recently, Grabler suggested that a recall rate less than 10% may be too low (37).

Possible reasons for the increased recall rate in this center include higher percentage of patients with prior surgery or biopsy at the hospital site, complicated mammograms that are difficult to interpret, and higher-risk patients. Unfortunately, we did not have access to this information, but these factors are likely for this type of center (as a tertiary referral university hospital). Future local research could be helpful to confirm this issue. Carney et al. showed that the recall rate increases among higher-risk patients (16). In addition, the mean age of patients was significantly lower compared to other reports (2). This factor may also have contributed to the higher recall rates, as younger age has been associated with higher recall rates (2, 36). It is possible that a higher recall rate is mandatory for some populations to maintain an appropriate cancer detection rate, and the local patient population has to be considered in the measurement.

Our study had several limitations. First of all, the number of patients included in the study was not large enough, and it would be better to include more patients in future studies. Due to the retrospective nature and limitations of the information gathering systems used in this study, no data were available for race, family, or personal history of breast cancer, and prior surgery or biopsy. Another important limitation is that we analyzed recall rates from only three radiologists, all of whom were university professors experienced in breast imaging and working in the same center. The majority of mammograms in Iran are interpreted by general radiologists as a small percentage of their overall workload. These points make it difficult to generalize the conclusions of this study to the general practice in Iran. Some centers in other parts of the world decreased their recall rate by double readings of mammography reports (38, 39), tomography, and three-dimensional mammography (21), or by using computer-aided detection software (40). These are potential future options for improving the national mammography report system.

In conclusion, this study showed a high recall rate in our center, which could be due to different patient population. Improving the mammography reading system may decrease this rate in the future.



  • 1.

    Miglioretti DL, Ichikawa L, Smith RA, Bassett LW, Feig SA, Monsees B, et al. Criteria for identifying radiologists with acceptable screening mammography interpretive performance on basis of multiple performance measures. AJR Am J Roentgenol. 2015; 204(4) -91 [DOI][PubMed]

  • 2.

    Yankaskas BC, Cleveland RJ, Schell MJ, Kozar R. Association of recall rates with sensitivity and positive predictive values of screening mammography. AJR Am J Roentgenol. 2001; 177(3) : 543 -9 [DOI][PubMed]

  • 3.

    Kopans DB. The positive predictive value of mammography. AJR Am J Roentgenol. 1992; 158(3) : 521 -6 [DOI][PubMed]

  • 4.

    Chubak J, Boudreau DM, Fishman PA, Elmore JG. Cost of breast-related care in the year following false positive screening mammograms. Med Care. 2010; 48(9) : 815 -20 [DOI][PubMed]

  • 5.

    Lidbrink E, Elfving J, Frisell J, Jonsson E. Neglected aspects of false positive findings of mammography in breast cancer screening: analysis of false positive cases from the Stockholm trial. BMJ. 1996; 312(7026) : 273 -6 [DOI][PubMed]

  • 6.

    Gilbert FJ, Cordiner CM, Affleck IR, Hood DB, Mathieson D, Walker LG. Breast screening: the psychological sequelae of false-positive recall in women with and without a family history of breast cancer. Eur J Cancer. 1998; 34(13) : 2010 -4 [DOI][PubMed]

  • 7.

    Elmore JG, Nakano CY, Koepsell TD, Desnick LM, D'Orsi CJ, Ransohoff DF. International variation in screening mammography interpretations in community-based programs. J Natl Cancer Inst. 2003; 95(18) : 1384 -93 [DOI][PubMed]

  • 8.

    D'Orsi CJ, Getty DJ, Pickett RM, Sechopoulos I, Newell MS, Gundry KR, et al. Stereoscopic digital mammography: improved specificity and reduced rate of recall in a prospective clinical trial. Radiology. 2013; 266(1) : 81 -8 [DOI][PubMed]

  • 9.

    Bassett LW. Determinants of quality in mammography. Surg Oncol Clin N Am. 1997; 6(2) : 213 -32 [PubMed]

  • 10.

    Advisory Committee on Cancer Prevention . Recommendations on cancer screening in the European Union. Eur J Cancer. 2000; 36(12) : 1473 -8 [DOI]

  • 11.

    Perry NM. Breast cancer screening--the European experience. Int J Fertil Womens Med. 2004; 49(5) : 228 -30 [PubMed]

  • 12.

    Henderson LM, O'Meara ES, Braithwaite D, Onega T, Breast Cancer Surveillance C. Performance of digital screening mammography among older women in the United States. Cancer. 2015; 121(9) : 1379 -86 [DOI][PubMed]

  • 13.

    Carney PA, Miglioretti DL, Yankaskas BC, Kerlikowske K, Rosenberg R, Rutter CM, et al. Individual and combined effects of age, breast density, and hormone replacement therapy use on the accuracy of screening mammography. Ann Intern Med. 2003; 138(3) : 168 -75 [DOI][PubMed]

  • 14.

    Yankaskas BC, Taplin SH, Ichikawa L, Geller BM, Rosenberg RD, Carney PA, et al. Association between mammography timing and measures of screening performance in the United States. Radiology. 2005; 234(2) : 363 -73 [DOI][PubMed]

  • 15.

    Yu-Mei L, Hsueh-Hua Y. Demographic factors influencing consensus opinion on the recall for women screened by mobile mammography unit in taiwan. Iran J Radiol. 2013; 10(3) : 116 -21 [DOI][PubMed]

  • 16.

    Carney PA, Cook AJ, Miglioretti DL, Feig SA, Bowles EA, Geller BM, et al. Use of clinical history affects accuracy of interpretive performance of screening mammography. J Clin Epidemiol. 2012; 65(2) : 219 -30 [DOI][PubMed]

  • 17.

    Elmore JG, Jackson SL, Abraham L, Miglioretti DL, Carney PA, Geller BM, et al. Variability in interpretive performance at screening mammography and radiologists' characteristics associated with accuracy. Radiology. 2009; 253(3) : 641 -51 [DOI][PubMed]

  • 18.

    Hawley JR, Taylor CR, Cubbison AM, Erdal BS, Yildiz VO, Carkaci S. Influences of Radiology Trainees on Screening Mammography Interpretation. J Am Coll Radiol. 2016; 13(5) : 554 -61 [DOI][PubMed]

  • 19.

    Barlow WE, Chi C, Carney PA, Taplin SH, D'Orsi C, Cutter G, et al. Accuracy of screening mammography interpretation by characteristics of radiologists. J Natl Cancer Inst. 2004; 96(24) : 1840 -50 [DOI][PubMed]

  • 20.

    Rothschild J, Lourenco AP, Mainiero MB. Screening mammography recall rate: does practice site matter? Radiology. 2013; 269(2) : 348 -53 [DOI][PubMed]

  • 21.

    McCarthy AM, Kontos D, Synnestvedt M, Tan KS, Heitjan DF, Schnall M, et al. Screening outcomes following implementation of digital breast tomosynthesis in a general-population screening program. J Natl Cancer Inst. 2014; 106(11)[DOI][PubMed]

  • 22.

    Friedewald SM, Rafferty EA, Conant EF. Breast cancer screening with tomosynthesis and digital mammography-reply. JAMA. 2014; 312(16) : 1695 -6 [DOI][PubMed]

  • 23.

    Greenberg JS, Javitt MC, Katzen J, Michael S, Holland AE. Clinical performance metrics of 3D digital breast tomosynthesis compared with 2D digital mammography for breast cancer screening in community practice. AJR Am J Roentgenol. 2014; 203(3) : 687 -93 [DOI][PubMed]

  • 24.

    Henderson LM, Benefield T, Marsh MW, Schroeder BF, Durham DD, Yankaskas BC, et al. The influence of mammographic technologists on radiologists' ability to interpret screening mammograms in community practice. Acad Radiol. 2015; 22(3) : 278 -89 [DOI][PubMed]

  • 25.

    Buist DS, Anderson ML, Smith RA, Carney PA, Miglioretti DL, Monsees BS, et al. Effect of radiologists' diagnostic work-up volume on interpretive performance. Radiology. 2014; 273(2) : 351 -64 [DOI][PubMed]

  • 26.

    Harvey SC, Geller B, Oppenheimer RG, Pinet M, Riddell L, Garra B. Increase in cancer detection and recall rates with independent double interpretation of screening mammography. AJR Am J Roentgenol. 2003; 180(5) : 1461 -7 [DOI][PubMed]

  • 27.

    Taylor P, Potts HW. Computer aids and human second reading as interventions in screening mammography: two systematic reviews to compare effects on cancer detection and recall rate. Eur J Cancer. 2008; 44(6) : 798 -807 [DOI][PubMed]

  • 28.

    Burnside ES, Sickles EA, Bassett LW, Rubin DL, Lee CH, Ikeda DM, et al. The ACR BI-RADS experience: learning from history. J Am Coll Radiol. 2009; 6(12) : 851 -60 [DOI][PubMed]

  • 29.

    Ghate SV, Soo MS, Baker JA, Walsh R, Gimenez EI, Rosen EL. Comparison of recall and cancer detection rates for immediate versus batch interpretation of screening mammograms. Radiology. 2005; 235(1) : 31 -5 [DOI][PubMed]

  • 30.

    Gur D, Sumkin JH, Hardesty LA, Clearfield RJ, Cohen CS, Ganott MA, et al. Recall and detection rates in screening mammography. Cancer. 2004; 100(8) : 1590 -4 [DOI][PubMed]

  • 31.

    Smith-Bindman R, Chu PW, Miglioretti DL, Sickles EA, Blanks R, Ballard-Barbash R, et al. Comparison of screening mammography in the United States and the United kingdom. JAMA. 2003; 290(16) : 2129 -37 [DOI][PubMed]

  • 32.

    Ganott MA, Sumkin JH, King JL, Klym AH, Catullo VJ, Cohen CS, et al. Screening mammography: do women prefer a higher recall rate given the possibility of earlier detection of cancer? Radiology. 2006; 238(3) : 793 -800 [DOI][PubMed]

  • 33.

    Gur D, Sumkin JH, Rockette HE, Ganott M, Hakim C, Hardesty L, et al. Changes in breast cancer detection and mammography recall rates after the introduction of a computer-aided detection system. J Natl Cancer Inst. 2004; 96(3) : 185 -90 [DOI][PubMed]

  • 34.

    Yankaskas BC, May RC, Matuszewski J, Bowling JM, Jarman MP, Schroeder BF. Effect of observing change from comparison mammograms on performance of screening mammography in a large community-based population. Radiology. 2011; 261(3) : 762 -70 [DOI][PubMed]

  • 35.

    Lee EH, Kim KW, Kim YJ, Shin DR, Park YM, Lim HS, et al. Performance of Screening Mammography: A Report of the Alliance for Breast Cancer Screening in Korea. Korean J Radiol. 2016; 17(4) : 489 -96 [DOI][PubMed]

  • 36.

    Schell MJ, Yankaskas BC, Ballard-Barbash R, Qaqish BF, Barlow WE, Rosenberg RD, et al. Evidence-based target recall rates for screening mammography. Radiology. 2007; 243(3) : 681 -9 [DOI][PubMed]

  • 37.

    Grabler P, Sighoko D, Wang L, Allgood K, Ansell D. Recall and Cancer Detection Rates for Screening Mammography: Finding the Sweet Spot. AJR Am J Roentgenol. 2017; 208(1) : 208 -13 [DOI][PubMed]

  • 38.

    Ciatto S, Ambrogetti D, Bonardi R, Catarzi S, Risso G, Rosselli Del Turco M, et al. Second reading of screening mammograms increases cancer detection and recall rates. Results in the Florence screening programme. J Med Screen. 2005; 12(2) : 103 -6 [DOI][PubMed]

  • 39.

    Dinnes J, Moss S, Melia J, Blanks R, Song F, Kleijnen J. Effectiveness and cost-effectiveness of double reading of mammograms in breast cancer screening: findings of a systematic review. Breast. 2001; 10(6) : 455 -63 [DOI][PubMed]

  • 40.

    Gromet M. Comparison of computer-aided detection to double reading of screening mammograms: review of 231,221 mammograms. AJR Am J Roentgenol. 2008; 190(4) : 854 -9 [DOI][PubMed]