By Author
  By Title
  By Keywords

September 2022, Volume 72, Issue 9

Research Article

Positive predictive value of Birads IV Lesions on mammogram in detection of breast carcinoma

Anila Rahim  ( Dow Institute of Radiology, Dow University of Health Sciences, Karachi, Pakistan; )
Amjad Sattar  ( Dow Institute of Radiology, Dow University of Health Sciences, Karachi, Pakistan )
Syed Omair Adil  ( School of Public Health, Dow University of Health Sciences, Karachi, Pakistan. )
Sohail Ahmed Khan  ( Dow Institute of Radiology, Dow University of Health Sciences, Karachi, Pakistan )


Objective: To determine the positive predictive value of Breast Imaging Reporting and Data System category IV lesions, and to evaluate interobserver agreement among radiologists in interpreting the subcategories.


Method: The cross-sectional study was conducted at the Dow Institute of Radiology, Dow University of Health Sciences, Karachi, from September 2018 to February 2019, and comprised adult females aged 30-60 years diagnosed with Breast Imaging Reporting and Data System IV. Already diagnosed breast carcinoma patients were excluded. The positive predictive value of Breast Imaging Reporting and Data System category IV was calculated using histopathology findings as the gold standard. Data was analysed using SPSS 22.


Results: There were 191 patients with a mean age of 49.01±11.79 years. Positive family history of breast cancer was found in 33(17.3%) subjects. Category IV-A was noted in 53(28.0%) subjects, IV-B in 45(23.6%) and IV-C in 93(48.7%). The overall positive predictive value of category IV was 151(79.1%), IV-A 27(50.9%), IV-B 35(77.7%) and IV-C 89(95.6%). Malignant disease was significantly associated with architectural distortion and suspicious microcalcifications (p<0.001). A good agreement was observed in the categorisation of Breast Imaging Reporting and Data System category IV between observers 1 and 2 (kappa 0.85, 95% confidence interval: 0.70-0.99), between observers 1 and 3 (kappa 0.77, 95% confidence interval:

0.61-0.94), and between observers 2 and 3 (kappa 0.63, 95% confidence interval: 0.43-0.83).


Conclusion: Breast Imaging Reporting and Data System BI-RADS IV was found to have a high positive predictive value in the diagnosis of breast cancer. Besides the interobserver agreement in terms of identifying BI-RADS IV subcategories was good.


Keywords: BIRADS IV, Interobserver agreement, Positive predicted value. (JPMA 72: 1750; 2022)






Carcinoma of the breast is the most common malignancy in women globally.1According to a recent World Health Organisation (WHO) estimate, 2.3 million women are diagnosed with breast cancer worldwide.2 In addition, 1 in every 9 women in Pakistan suffer from breast cancer, which is the uppermost incidence rate.3

It is reported that breast cancer has better prognosis if detected early.4 The imaging modalities for the evaluation of breast pathologies include conventional and digital mammography, ultrasound (US), strain and shear wave elastography, and magnetic resonance imaging (MRI).5,6

The American College of Radiology (ACR) has proposed the Breast Imaging Reporting and Data System (BI-RADS) to systematise reporting of mammograms as well as USs and MRIs.7,8 All categories raise the level of suspicion of malignancy and these categories of assessment finally confirmed by histopathology.9

The BI-RADS 5th edition, 2013, comprise mammography, US breast and MRI as well as further subcategorization of the BI-RADS category IV subtype into three further divisions; IV-A having low suspicion for malignancy, IV-B having moderate suspicion, and IV-C having high suspicion. The category IV overall is used for the evaluation of a large number of pathologies in which histopathological correlation is required to confirm the diagnosis.10

An international study portrayed that high breast biopsy rates were diagnostic overkill, drastic and valuable.11 A study found that most BI-RADS IV lesions had fibrocystic change on histopathological assessment, and the second most common was ductal carcinoma in situ (DCIS).12 However, in our clinical practice, most BI-RADS IV lesions on mammogram fall in IV-C subcategory and turn out to be malignant.

First-line imaging technique is usually mammography due to its cost-effectiveness, easy availability and high efficacy.13,14 Literature has reported a low positive predictive value (PPV) of BI-RADS IV, especially of IV-A and IV-B lesions. A local study reported high diagnostic accuracy of BI-RADS for breast cancer diagnosis with sensitivity 92.3% and specificity 88.8%.15 Although the study had a larger sample size, it did not stratify the accuracy according to each BI-RADS category which was its limitation.15 Also, it did not specify PPVs.

The current study was planned to determine the PPV of BI-RADS category IV and to evaluate interobserver agreement among radiologists in interpreting the subcategories.


Subjects and Methods


The cross-sectional, multicentre study was conducted at the Ojha, Latif Ebrahim Jamal (LEJ) and Nazimabad campuses of the Dow Institute of Radiology, Dow University of Health Sciences (DUHS), Karachi, from September 2018 to February 2019. After approval from the institutional ethics review board, the sample size was calculated using EpiInfo calculator,16 taking PPV of BI-RADS IV to be 98%,17 along with confidence interval (CI) 97% and precision 2.2%.

The sample was raised using consecutive sampling technique. Those included were adult female patients aged 30-60 years having post-unilateral mastectomy/post-lumpectomy status and who were labelled as BI-RADS IV on mammography. Patients who already had biopsy-proven breast carcinoma, residual disease following surgery, chemoradiotherapy, contraindication to mammogram, including pregnant or lactating patients, contraindications to trucut biopsy, and severe psychiatric disorders, which makes them impossible to collaborate, were excluded.

After taking written informed consent from all the subjects, a detailed history was obtained and clinical examination of the lesion was carried out. A senior female technologist with >6 years of experience performed mammography on GE Performa (GE Healthcare, United States) and Mammomat 1000 (Siemens). Mediolateral oblique (MLO) and craniocaudal (CC) views were obtained. If required, additional views, such as magnification, cone compression or axillary, were undertaken for lesion localisation. Female imaging specialist with >5 years of experience evaluated all the mammograms for the presence of mass, its density, peculiar characteristics, microcalcification, their pattern, any asymmetrical density, architectural distortion, and abnormally appearing axillary lymph nodes.

BI-RADS IV was diagnosed as per ACR 2013 guidelines.7 BI-RADS IV-A was labelled as positive based on the presence of characteristics, such as partially circumscribed mass, a palpable, single, complex cystic and solid cyst, along with a probable abscess. BI-RADS IV-B was labelled as positive based on the presence of characteristics, such as group of amorphous / fine pleomorphic or coarse heterogeneous calcification(s) and nondescript solid mass with indistinct margins. BI-RADS IV-C was labelled as positive based on the presence of characteristics, such as a new group of fine linear / fine linear branching calcification(s) and new indistinct, irregular single mass. The final diagnosis of BI-RADS IV lesion on mammography was compared with histopathology. In case the histopathology report was not traced due to loss to follow-up, such patients were excluded. Besides, interobserver agreement was noted among three radiologists.

Data was analysed using SPSS22. Mean ± standard deviation was computed for age, and duration of symptoms. Frequencies and percentages were computed for categorical variables, like marital status, family history of breast cancer, smoking, gravida, parity, lactation status, menstrual cycle, mammographic findings, subcategories of BI-RADS IV and final diagnosis on histopathology. PPV was calculated by dividing the number of patients truly diagnosed as positive (true positive; TP) with the total positive cases (true positive + false positive [FP]). PPV was calculated considering histopathology as the gold standard. Further stratification was done on the basis of health characteristics and suspicious microcalcification to observe the PPV of BI-RADS. Comparison was also done to see the association of malignant disease with architectural distortion and suspicious microcalcifications For which chi-square test was used. P<0.05 was taken as significant. For interobserver agreement, kappa (k) statistics were used.




There were 191 patients with a mean age of 49.01±11.79 years. The mean weight of the sample was 67.39±11.20 kg, mean height 5.12±0.25 m, and mean body mass index (BMI) 26.21±4.57 kg/m2. A positive family history of breast cancer was found in 33(17.3%) subjects. Category IV-A was noted in 53(28.0%) subjects, IV-B in 45(23.6%) and IV-C in 93(48.7%).

There were 58(31.7%) multigravida and 22(12%) multiparous patients. The history of lactation was observed in 160(87.9%) patients. In addition, most of the women presented with menopause 97(50.8%), followed by those with normal menstrual cycle 56(29.3%), and perimenopausal 38(19.9%).

Mammographic findings showed well-defined margins in 17(8.9%) patients; 19(9.9%) presented with amorphous microcalcifications; skin thickening was found in 32(16.8%); and architectural distortion in 141(73.8%), Invasive ductal carcinoma (IDC) was noted in 150(78.5%) patients

(Table 1).



Majority of the patients had upper outer quadrant involved 105(57.4%), and the most affected side was the left breast 113(59.2%).

The overall PPV of BI-RADS IV was 151(79.1%) (Table 2), while for subcategory IV-A it was 27(50.9%), IV-B 35(77.7%) and IV-C 89(95.6%).



Stratification was done to see the PPV of BI-RADS IV with respect to sociodemographic and clinical characteristics, and a higher PPV was observed in all cases (Table 3). Malignant disease was significantly associated with architectural distortion and suspicious microcalcifications (p<0.001).A good agreement was observed in the categorisation of BI-RADS IV between observers 1 and 2 (k: 0.85, 95% CI: 0.70-0.99), between observers 1 and 3 (k: 0.77, 95% CI: 0.61-0.94), and between observers 2 and 3 (k: 0.63, 95% CI: 0.43-0.83) (Table 4).







The current study showed that BI-RADS IV had a high PPV of 79% in the diagnosis of breast cancer. Various local studies have shown a high sensitivity and specificity of BI-RADS in diagnosing breast cancer,15,18 while some studies showed low PPV as well.19,20 According to a study conducted in Singapore, BI-RADS IV PPV was reported as 27.2%.20 The difference may be attributed to the fact that the two studies18,19 were old and from relatively developed Asian countries. Another reason that maybe that an older version of BI-RADS was utilised in the previous studies,19,20 whereas the current study utilised the updated BI-RADS version, 2013. According to another local study, the PPV of BI-RADS IV was 98%17 compared to 79% found by the current study.

In the current study, category IV-A had PPV 50.94%, IV-B 77.78% and IV-C 95.69%. An earlier study reported corresponding values of 9%, 21% and 57%.19

The current study showed BI-RADS IV PPV 78% in patients aged ≤50 years and 80.5% in patients aged >50 years. Conversely, the recent trend has shown an increased incidence of breast cancer among women aged <45 years.2 A study reported that an advanced stage breast cancer diagnosis in Pakistani women was at least a decade earlier than in the West.21 Though screening mammography is recommended at age 40, we get referrals by physicians and surgeons for women aged 30 years and above with palpable mass and strong family history. As such, those patients were also included in the current study. Also, recent literature has reported an increased incidence of breast cancer in young patients.2

The current study included unilateral mastectomy as well as patients with post-lumpectomy with breast conservation surgery. Moreover, the findings showed that BI-RADS IV PPV in underweight population was 83.3%, in normal weight 87.1%, in overweight 94.4% and in obese 91.30%. Increasing BMI increases the risk of breast cancer.22,23 This factor might have attributed to high PPV.

Most of the patients in our study had IDC grade II. A local study conducted on 150 patients also demonstrated that grade II was common compared to grades I or III.24 Similar results have been reported in various international studies as well.25 However, another local study has reported the frequency of IDC grade III to be the highest.26 Therefore, further studies with large sample size are needed to validate the histopathological patterns of breast cancer in Pakistan.

There was a good interobserver agreement in the current study, which is in contrast with a study that showed fair agreement for final assessment of BI-RADS IV.8 Another study showed fair agreement for final BI-RADS categories on mammogram and a good agreement for BI-RADS categories on US.27The current study showed that amorphous, fine pleomorphic and fine linear/linear branching calcification had the highest PPV in diagnosing breast cancer. Similar findings have been reported earlier,28 the PPV reported was 100% compared to 95.8% in the current study. The PPV of fine pleomorphic microcalcification was 63% whereas the current study reported it to be 85%. One study29 compared microcalcifications reported by polymerase chain reaction (RCR) 5-point breast imaging system and BI-RADS and reported that PPV fine linear/fine linear branching was 77.8%29 compared to 85% in the current study.

In the current study, there was a significant association of architectural distortion with malignant BI-RADS category IV, which was more pronounced in subcategories IV-B and IV-C. An international study showed that in screening mammograms, architectural distortion less likely represented malignancy compared to diagnostic mammograms, and the results showed a significant association of malignancy with architectural distortion on diagnostic mammography30 almost comparable to the findings of the current study.

The current study has certain limitations Firstly, the sample size was small. Secondly, it had a cross-sectional design in which the temporal relationship between different risk factors and outcome was not determined. Furthermore, the data regarding use of hormone replacement therapy (HRT) among postmenopausal women was not collected as the use of HRT can have effect on breast cancer risk.

Despite these limitations, however, the findings are based on data collected from one of the largest public-sector academic medical centres in which different modalities, such as X-ray, US, MRI or computed tomography (CT) scans were available under one roof. Moreover, the data was collected prospective, and the study was multicentric. Lastly, interobserver agreement was also calculated.

In the light of the findings, it is recommended that further multicentric studies based upon larger sample sizes shall be carried out. Moreover, studies by incorporating variables that were not collected in the current study, such as the status of oestrogen receptors and progesterone receptors positivity, use of HRT, cultural and ethnic status. Furthermore, studies evaluating temporal relationship between risk factors and outcome shall also be carried out.




BI-RADS IV was found to have a high PPV in the diagnosis of breast cancer. Further stratification on the basis of subcategories showed PPV of IV-C to be the highest, followed by IV-B and IV-A. As far as suspicious microcalcification is concerned, amorphous and fine linear/linear branching had better PPV. Most importantly, interobserver agreement in identifying BI-RADS IV subcategories was found to be good.


Disclaimer: The text is based on an academic thesis.


Conflict of Interest: None.


Source of Funding: None.




1.       Torre LA, Islami F, Siegel RL, Ward EM, Jemal A. Global Cancer in Women: Burden and Trends. Cancer Epidemiol Biomarkers Prev 2017;26:444-57. doi: 10.1158/1055-9965.EPI-16-0858.

2.       World Health Organization. Facts Sheet: Breast Cancer. News release. The WHO Media Center. [Online] 2021 [Cited 2021 December 08]. Available from URL: detail/breast-cancer.

3.       Sohail S, Alam SN. Breast cancer in Pakistan-awareness and early detection. J Coll Physicians Surg Pak 2007;17:711-2.

4.       Mariotto AB, Yabroff KR, Shao Y, Feuer EJ, Brown ML. Projections of the cost of cancer care in the United States: 2010-2020. J Natl Cancer Inst 2011;103:117-28. doi: 10.1093/jnci/djq495.

5.       Li DD, Xu HX, Guo LH, Bo XW, Li XL, Wu R, et al. Combination of two-dimensional shear wave elastography with ultrasound breast imaging reporting and data system in the diagnosis of breast lesions: a new method to increase the diagnostic performance. Eur Radiol 2016;26:3290-300. doi: 10.1007/s00330-015-4163-8.

6.       Barr RG, Nakashima K, Amy D, Cosgrove D, Farrokh A, Schafer F, et al. WFUMB guidelines and recommendations for clinical use of ultrasound elastography: Part 2: breast. Ultrasound Med Biol 2015;41:1148-60. doi: 10.1016/j.ultrasmedbio.2015.03.008.

7.       D’Orsi C, Sickles E, Mendelson E, Morris EJGS. ACR BI-RADS® atlas, breast imaging reporting and data system. Reston, VA: American College of Radiology; 2016.

8.       Lazarus E, Mainiero MB, Schepps B, Koelliker SL, Livingston LS. BI-RADS lexicon for US and mammography: interobserver variability and positive predictive value. Radiology 2006;239:385-91. doi: 10.1148/radiol.2392042127.

9.       Rao AA, Feneis J, Lalonde C, Ojeda-Fournier H. A Pictorial Review of Changes in the BI-RADS Fifth Edition. Radiographics 2016;36:623-39. doi: 10.1148/rg.2016150178.

10.     Burivong W, Amornvithayacharn O. Accuracy of subcategories A, B, C in BI-RADS 4 lesions by combined mammography and breast ultrasound findings. J Med Med Sci 2011;2:728-33.

11.     Yan XS, Stark A, Chitale D, Burke M, Zarbo R, Nathanson D, et al. PS1-12: Suspicious Mammogram (BIRADS 4) Outcome and Breast Biopsy: Preliminary Findings from a Cohort of 6,198 Women. Clin Med Res 2012;10:147. doi:10.3121/cmr.2012.1100.ps1-12

12.     Sanders MA, Roland L, Sahoo S. Clinical implications of subcategorizing BI-RADS 4 breast lesions associated with microcalcification: a radiology-pathology correlation study. Breast J 2010; 16:28-31. doi: 10.1111/j.1524-4741.2009.00863.x.

13.     Brown AL, Phillips J, Slanetz PJ, Fein-Zachary V, Venkataraman S, Dialani V, et al. Clinical Value of Mammography in the Evaluation of Palpable Breast Lumps in Women 30 Years Old and Older. AJR Am J Roentgenol 2017;209:935-42. doi: 10.2214/AJR.16.17088.

14.     Muradali D, Kennedy EB, Eisen A, Holloway CMB, Smith CR, Chiarelli AM. Breast screening for survivors of breast cancer: A systematic review. Prev Med 2017;103:70-5. doi: 10.1016/j.ypmed.2017.07.026.

15.     Shahid R, Rasheed B, Gulzar R. Diagnostic Accuracy of Breast Imaging Reporting and Data System (BI-RADS) in Determination of Breast Malignancy Taking Histopathology as Gold Standard. Int J Pathol 2018;16:106-11.

16.     Sullivan KM, Dean A, Soe MM. OpenEpi: a web-based epidemiologic and statistical calculator for public health. Public Health Rep 2009;124:471-4. doi: 10.1177/003335490912400320.

17.     Fatima N, uz Zaman M, Uddin Q, Ahsan Z. Accuracy of mammography and ultrasound for detecting breast cancer at a breast care clinic in Karachi, Pakistan. J Biomed Graph Comput 2011;1:44-50. DOI: 10.5430/jbgc.v1n1p44

18.     Majeed AI, Naz N, Arif M, Majeed A. Diagnostic Accuracy of Mammogrpahy in Detecting Breast Cancer Keeping Histopathology as Gold Standard. Ann Pak Inst Med Sci 2016;12:118-121.

19.     Wiratkapun C, Bunyapaiboonsri W, Wibulpolprasert B, Lertsithichai P. Biopsy rate and positive predictive value for breast cancer in BI-RADS category 4 breast lesions. J Med Assoc Thai 2010;93:830-7.

20.     Tan YY, Wee SB, Tan MP, Chong BK. Positive predictive value of BI-RADS categorization in an Asian population. Asian J Surg 2004;27:186-91. doi: 10.1016/S1015-9584(09)60030-0.

21.     Soomro R, Faridi S, Khurshaidi N, Zahid N, Mamshad I. Age and stage of breast cancer in Pakistan: An experience at a tertiary care center. J Pak Med Assoc 2018;68:1682-5.

22.     Lahmann PH, Hoffmann K, Allen N, van Gils CH, Khaw KT, Tehard B, et al. Body size and breast cancer risk: findings from the European Prospective Investigation into Cancer And Nutrition (EPIC). Int J Cancer 2004;111:762-71. doi: 10.1002/ijc.20315.

23.     Liu K, Zhang W, Dai Z, Wang M, Tian T, Liu X, et al. Association between body mass index and breast cancer risk: evidence based on a dose-response meta-analysis. Cancer Manag Res 2018;10:143 - 51. doi: 10.2147/CMAR.S144619.

24.     Hadi NI, Jamal Q. Comparison of clinicopathological characteristics of lymph node positive and lymph node negative breast cancer. Pak J Med Sci 2016;32:863-8. doi: 10.12669/pjms.324.10324

25.     He ZY, Wu SG, Yang Q, Sun JY, Li FY, Lin Q, et al. Breast Cancer Subtype is Associated With Axillary Lymph Node Metastasis: A Retrospective Cohort Study. Medicine (Baltimore) 2015;94:e2213. doi: 10.1097/MD. 0000000000002213.

26.     Khokher S, Qureshi MU, Mahmood S, Sadiq S. Determinants of advanced stage at initial diagnosis of breast cancer in Pakistan: adverse tumor biology vs delay in diagnosis. Asian Pac J Cancer Prev 2016;17:759-65. doi: 10.7314/apjcp.2016.17.2.759.

27.     Elverici E, Zengin B, Barca AN, Yilmaz PD, Alimli A, Araz L. Interobserver and intraobserver agreement of sonographic BIRADS lexicon in the assessment of breast masses. Iran J Radiol 2013;10:122- 7. doi: 10.5812/iranjradiol.10708.

28.     Kim SY, Kim HY, Kim EK, Kim MJ, Moon HJ, Yoon JH. Evaluation of malignancy risk stratification of microcalcifications detected on mammography: a study based on the 5th edition of BI-RADS. Ann Surg Oncol 2015;22:2895-901. doi: 10.1245/s10434-014-4362-6.

29.     Metaxa L, Healy NA, O'Keeffe SA. Breast microcalcifications: the UK RCR 5-point breast imaging system or BI-RADS; which is the better predictor of malignancy? Br J Radiol 2019;92:20190177. doi: 10.1259/bjr.20190177.

30. Bahl M, Baker JA, Kinsey EN, Ghate SV. Architectural Distortion on Mammography: Correlation With Pathologic Outcomes and Predictors of Malignancy. AJR Am J Roentgenol 2015;205:1339-45. doi: 10.2214/AJR.15.14628.

Journal of the Pakistan Medical Association has agreed to receive and publish manuscripts in accordance with the principles of the following committees: