Is Breast transillumination a viable option for breast cancer screening in limited resource settings?
Authors: Elobu EA1 M.Med, Galukande M1 M M.Med, MSc, FCS, Namuguzi D1 M.Med, Muyinda Z2 M.Med. Affiliations:
1-Department of Surgery, College of Health Sciences, Makerere University 2-Department of Radiology, Mulago National Referral Hospital, Uganda Correspondence: Moses Galukande, Department of Surgery, College of Health Sciences, Makerere University P. O. Box 7072 Kampala, Uganda Email:
Background: Mammography is an established screening tool for breast cancer in high-income countries but may not be feasible for most re-source poor nations. Alternative modalities are needed to mitigate the impact of the increasing incidence and mortality due to breast cancer. This may require the development of new technologies or reevaluation of old technologies applicable to resource limited settings.
Objective: To determine the sensitivity and specificity of breast transil-lumination as compared to mammography and to describe features of benign and malignant breast lesions as seen with breast transil-lumination.
Methods: A single group descriptive analytical study was conducted over a six month period (2011) in the breast unit of Mulago National hospital. Eligible participants were consecutively sampled. Participants underwent Clinical Breast Examination (CBE), breast transillumination (BT), mammography (MG) and histopathological analysis of identified breast lumps. Sensitivity, specificity and predictive values were calculated. Features of the masses detected by transillumination were then described.
Results: The number of participants recruited was 201 (mean age 42 years, range 30-80 years). The average palpable lump size was 3.8 cm (range 0.5 to 10 cm). BT had a sensitivity of 63.2% (PPV 86.8%) and a specificity of 89.5% (NPV 61.2%) with mammography as the reference standard. Also, 73.3% of breast lumps with irregular margins and 88.5% with dense opacity at transillumination turn out to be malignant at histopathology examination.
Conclusion: The Breast transillumination technique had a moderate sensitivity of 63.2%. This warrants a large scale population-based evaluation of BT as a screening tool. This technique may not substitute mammography but to be considered an option where mammography access is limited.
Breast cancer accounts for about one third of all cancers diagnosed globally with both incidence and mortality re-ported to be increasing in sub-Saharan Africa. In Uganda breast cancer is the third most commonly diagnosed can-cer after Kaposi’s sarcoma and cancer of the cervix (1). The majority (77%) of women present in late stages (III & IV) with low 5-year survival (of 39%). In the past, resource allocation for health was skewed towards infectious dis-eases such as HIV/AIDS and malaria. Currently however, a rapidly growing burden of non-communicable diseases (NCDs) is demanding similar attention.
Screening reduces mortality due to breast cancer. Mam-mography has been shown to reduce breast cancer mor-tality by up to 25-30 % in women over 50 years old (2-4). Access to mammography is limited in developing countries. There are currently only a handful of mam-mography machines in Uganda which are not equitably distributed and some charge a fee thus precluding access for most women (5). Women below the age of 30 years are ineligible for mammography yet a sizeable portion of them are at risk and therefore need some form of screen-ing. Alternative screening modalities should therefore be explored either by re-evaluation of old techniques or in-vention of new ones (6,7). Culter first described the use of breast transillumination some 80 years ago (8). Advent of x-rays and mammogra-phy led to its abandon. There is justification for its revalu-ation with a view to determining its utility in areas where there is no mammography. The Breast light is a commer-cially available modification of earlier prototypes used for breast transillumination. It is comparatively much cheaper than mammography, easy to use and has mini-mal running costs.
The primary aim of this study therefore was to determine the sensitivity, specificity and predictive values of breast transillumination compared to mammography in a hos-pital setting. Is Breast transillumination a viable option for breast cancer screening in limited resource settings? Elobu EA, Galukande M, Namuguzi D, Muyinda Z
A descriptive study was conducted from January to June 2011 in the breast unit of Mulago Hospital, one of three national referral hospitals but the only public hospital with comprehensive breast care services in the country. Mulago has a bed capacity of 1500 and is the teaching Hospital for Makerere University College of Health Sci-ences. The outpatient Breast Clinic runs once a week and in the past three years over 600 incident cases of breast cancer were registered there.
All women aged 30 years or more who attended the breast clinic were prospectively included. Pregnant or lactating women and those with ulcerated breast lesions were excluded because of radiation risks in pregnancy and high breast density during lactation which inter-feres with mammography interpretation respectively. The sample size was determined using the formula for sample size for a descriptive study of a dichotomous variable (6, 7), alpha was 0.05 for 95% confidence interval, a sensitivity of BT of 67% was used. Consecutive sampling was done.
CBE was performed to determine the presence or absence of a breast lump prior to transillumination. At transil-lumination, the number, size, site, density of opacity and regularity of margins of detected lumps were noted. Opacity was measured against a three point scale devised for this study as translucent, opaque or densely opaque. A lump with more than three quarters of its margin being even was considered regular, less than ¾ was considered irregular.
Translucent – most light comes through (Red/pink)
Opaque – some light comes through (Gray)
Densely opaque – no light comes through (Black)
In a dark room, the Breast light (model BL 801 manufac-tured by PWB health, G82 3PW, UK) was used by a single observer to examine the woman’s breasts in seated posi-tion with the arms raised to hold the back of the head. A water based gel lubricant was applied to both breasts to ease the examination for the user and make it more com-fortable for the participant. The light was switched on and applied firmly to the skin on the inferior aspect of the breast beginning with the right, thereby giving a cranio-caudal view. The Breast light was moved to either side up to the edge of the breast. The light was then moved for-ward and upwards over the breast to ensure that all parts of the breast were transilluminated. The Breast light was finally moved superolaterally along the milk line towards the axilla to examine the axillary tail. At all times the in-vestigator observed the transmission of light through the breast at a point directly opposite the light bulb of the Breast light. The same was repeated on the left breast. The participant was cleaned and allowed to dress.
Mammography was performed on all participants and read as normal (BI-RADS 1) or abnormal (BI-RADS 2-5) by a total of four consultant radiologists at different times. Participants who had a clinically palpable abnor-mality and consented underwent a core needle biopsy. Histopathological analysis was performed by an experi-enced team of pathologists. Findings were recorded in a structured data collection tool.
The investigators performing the breast transillumina-tion, mammography and histological examination were blinded to findings of any of the other tests.
Data management and analysis
Raw data were entered using Epidata Version 3.1 software, and then exported to Epidata Analysis v.126.96.36.199 for analysis. Sensitivity, specificity and predictive values were calculated for transillumination technique. Features of malignant and benign lumps at transillumination were described.
Ethical clearance was secured from the Ethics and re-search Committee. Informed consent was obtained from all participants.
A total of 201 participants were recruited, 10 (5% ) of whom were asymptomatic. The average palpable lump size was 3.8 cm (range 0.5 to 10 cm). The smallest lump detected by CBE, BT and MG was 0.5 cm. Participants were aged 30- 80 years with an average of 42 years.
Overall, mammography detected more abnormalities than the other two techniques (Fig 1-2). Mammography performed best for lumps less than 2 cm where it detected 7.5 times as many lumps as BT but for lumps 2-5 cm, MG detected only 1.08 times more lumps than transillumina-tion. BT picked up more than CBE for all lumps < 5cm.
Transillumination was negative in 14 (14.4%) of partici-pants with abnormal CBE and 46 (36.8%) with abnor-mal mammography. The average lump size was 2.1 cm for CBE and 1.5 cm for mammography (Table 1). 50% of these were located in the deep in breast tissue close to the chest wall. Non-lump lesions included mammographic abnormalities not described as lumps e.g. focal area of in-creased density, calcifications, or tramline calcifications. These accounted for 52.4% of the mammograms.
Click to view figure 1
Click to view figure 2
Click to view table 1
Click to view table 2
Histology was done on a total of 60 lumps, 61.7% of which were malignant, the rest were benign. Of the lumps with irregular margins, 73.3% were histologically confirmed to be malignant as were 88.5% of those that appeared densely opaque (Table 2).
The challenges that face cancer control efforts in sub Saharan Africa are numerous and complex; the major-ity of women who need screening are young (therefore not eligible for the standard mammography screening), national screening programs are non existent and public awareness is low. Resource allocation is skewed to infec-tious disease such as HIV/AIDS and malaria. Exploring the possibility of using BT as a screening modality is an attempt to contribute to Breast cancer control initiatives in sub Saharan Africa. Breast transillumination is not meant to be a substitute to x-ray screening but a stop gap measure in the absence of x-ray screening or other more appropriate screening tools to vulnerable populations. We also anticipate that in popularizing its use, awareness might be raised and access to screening enhanced.
Sensitivity and Specificity
This study found that BT in comparison to mammogra-phy had a moderate sensitivity of 63.2%, a specificity of 89.5%, and positive predictive value of 90.8%. BT detect-ed 85.6% of palpable lumps and was able to detect lumps as small as 0.5 cm in widest diameter. Irregular margins and dense opacity at transillumination were found to be highly suggestive of malignancy. Although there were concerns by the investigators as to the applicability of transillumination in populations with the black skin as previous studies were conducted in predominantly white populations of the USA and UK (7-12), it worked well in this study. Mammography proved superior to BT in several ways: higher overall proportion of abnormalities in the study population, (62.2% vs. 45.3%), higher true positive test of abnormalities among CBE positive par-ticipants, (91.8% vs. 85.6%) and greater abnormalities in source settings? Elobu EA, Galukande M, Namuguzi D, Muyinda Z
CBE negative participants, (34.6% vs. 7.7%). However, BT showed some strength as well. We were able to de-tect lumps as small as 0.5cm with breast transillumina-tion and most of the lumps detected with CBE negative participants were less than 2 cm in their widest diameter possibly demonstrating its potential for early detection of breast lumps. These findings demonstrate that breast transillumination is a promising tool, further evaluation of which is justifiable. This is particularly important giv-en the high rate of presentation with advanced disease in our environment. The most important feature of a screen-ing tool in a resource limited setting is sensitivity (13) to detect as many truly positive cases as possible since few opportunities exist for patients to undergo screening. The specificity of 89.5% means that the majority of nor-mal patients will have negative transillumination (i.e. few false positives). This is important since false alarms are costly and may reduce confidence of the end users.
Coupled with a NPV of 59.6%, this minimizes the chanc-es of a participant with disease having a negative test. Currently, transillumination detects benign and breast cancer alike. Approaches to such a distinction are evalu-ated in this paper. Whereas it is common practice in the developed world to monitor mammographically be-nign disease, in our setting a definitive diagnosis is often sought and appropriate treatment offered. This is because of high loss to follow up of many patients. Therefore, de-tection of benign disease should not raise similar con-cerns as in the developed world.
Transillumination had sensitivities of 67-95% in various other studies with histology as the reference standard (7). Although it may not be easy to directly compare sensi-tivities of transillumination and mammography because they were each calculated using different reference stan-dards, it may still be possible to make some meaningful inferences. The reported sensitivity of mammography is 63%-95% (5,14). The fact that the sensitivity of transil-lumination at 63.2% fits within the reported sensitivity range of mammography makes it comparable in its use as a technique for detection of breast lumps.
An in depth analysis was done of the missed lumps that accounted for 36.8% of the false negative cases rate. The average size was 1.5 cm (range 0.5- 4cm) and 61.2 % of them were in breasts of density BI-RADS 2, but because lumps of similar sizes were found in other breast of simi- lar density and size, size and breast density are unlikely explanations Nearly half of these lumps were located deep in the breast tissue and close to but not attached to the chest wall. Previous studies (8,15) showed that lumps deep in breast tissue, close to the chest wall do not trans-illuminate satisfactorily as was seen in this study. Many of these lumps were small and were in thin/flat (small volume) breasts. Such lumps did not transilluminate well in our study as was the case for Cutler (5). Transillumi-nated light may diffract around these lumps giving false negative appearances, but such a phenomenon may not occur with the much higher energy x-rays used in mam-mography. Another possible explanation for false nega-tives is that some mammographic diagnostic criteria such as calcifications are not detectable by transillumination. Although calcifications should not transilluminate, their small sizes possibly allow diffraction of light around them and therefore are not visualized. As was reported by Sickles (9), this remains a limiting factor for the use of breast transillumination, and represents a challenge for an improved transillumination device.
This study also set out to evaluate the diagnostic potential of breast transillumination. The inability to distinguish benign from malignant disease was earlier brought out as a major weakness of the breast transillumination tech-nique (9,15). In a previous study, tumour vascularity was investigated as a possible distinguishing feature between benign and malignant tumours but there was no signifi-cant difference. We once again attempted to determine with different tumour characteristics if a benign – malig-nant distinction could be made. We based on regularity of margins and degree of opacity measured against a pre-determined standard. It was found that 73.3 % of lumps with irregular margins and 88.5 % of those with dense opacity were malignant. This correlates well with Ohls-son’s findings in which malignant tumors were described as dark with irregular margins (10). Reproducibility of these distinctive characteristics may be need further vali-dation.
Low cost screening
With a high burden of breast cancer disease and the ma-jority of patients presenting with advanced disease (1,16-17), it is essential that the population at risk accesses screening tools that not only exhibit high sensitivity and specificity but are affordable and can be used for all age groups as well as being culturally acceptable . In addition it should be easy to use for both the health worker and the patient.
Screening mammography remains the standard of care and its benefits in improving survival through early detec-tion are well documented (2-4). However such benefits may not be achievable in these resource poor settings due to the inhibitory resource requirements related to mam-mography and the ineligibility of nearly half of women who are in dire need of such screening services. One may therefore wish to consider breast Ultrasonography but it is also not without limitations (5,18).
In the absence of these two modalities what can be of-fered to the woman concerned about her breast health is the premise for re-evaluation of breast transillumination, a once promising imaging technique that was overtaken by advances in investigative science.
A Breast light costs about US$ 150 and has minimal maintenance costs. It runs on dry cells easily available in rural areas where most of our patients stay. The transil-lumination devices could be individually or communally owned thus further lowering costs and increasing accessi-bility. Indeed even individual women can be trained and empowered to examine their own breasts.
Limitations of the study included the fact that it was conducted in a hospital setting and was not restricted to asymptomatic patients which may limit its external validity as a screening tool. However, none of the previ-ous studies (7-12) on transillumination were restricted to asymptomatic participants either. To our knowledge, this is the first study of its kind in a black African popula-tion, and the cost of a population based study limited this initial study to the hospital setting. In addition, the greater number of patients with positive findings in the hospital made it suitable for the first study. And now that this study has given us insight into the capabilities of BT in comparison to mammography, a larger study in an as-ymptomatic population may be justified.
BT was conducted by a single observer without significant prior experience in its use. This may have an effect on ac-curacy of interpretation of the findings, but perhaps the use of BT to determine presence or absence of a lump should not require specialized training and prior experi-ence. This suggests that BT can be performed by various cadres of health care providers. Although no measure of variability was done, the use of more than one experi-enced mammography interpreter possibly reduced such variability. A total of four consultant level radiologists who routinely read mammography did the reporting on mammography in this study.
Also, since no measure of intra or inter observer variabil-ity was done, no strong inferences can be made on the reliability of transillumination technique.
Breast transillumination has moderate sensitivity and high specificity for detecting breast lumps. Breast transil-lumination should be considered for a large scale popu-lation based evaluation as a screening tool for breast can-cer in resource limited settings, though it is not intended to substitute x-ray screening where it is available.
We thank the participants and the staff at the breast clinic and the department of surgery.
Gakwaya A, Galukande M, Luwaga A et al. Breast cancer guidelines for Uganda (2nd Edition 2008). Afr Health Sci, 2008;8(2):126-32.
Prasad S N, Houserkova D. A comparison of mammogra-phy and ultrasonography in the evaluation of breast masses. Biomed Pap Med Fac Univ Palacky Olomouc Czech Repub, 2007;151(2):315-22.
Tabár L , Duffy S W, Vitak B, Chen H. H , Prevost T C. The natural history of breast carcinoma. Cancer, 1999; 86(3): 449-462.
Barton, Mary B.,Harris, Russell,Fletcher, Suzanne W. Does This Patient Have Breast Cancer?: The Screening Clinical Breast Examination: Should It Be Done? How? JAMA, 1999; 282(13): 1270-1280.
Cutler M. Transillumination of the Breast. Ann Surg, 1931; 93(1): 223-34.
Galukande M, Kiguli-Malwadde E. Rethinking breast cancer screening strategies in resource-limited settings. Afr Health Sci, 2010; 10(1): 89-92.
Hulley S B , Cummings S R , Browner W S et al. Designing Clinical Research. second ed. 2001: Lippincott Williams and Wilkins.
Iwuchukwu O, Keaney N, Dordea M. Analysis of Breastlight findings in patients withbiopsies. , in European Institute of Oncology’ s 12th Milan Breast Cancer Conference. 2010:Mi-lan.
Is Breast transillumination a viable option for breast cancer screening in limited resource settings? Elobu EA, Galukande M, Namuguzi D, Muyinda Z
Sickles E A. Breast cancer detection with transillumination and mammography. AJR Am J Roentgenol, 1984; 142(4):841-4.
Ohlsson B, Gundersen J , Nilsson DM. Diaphanography: A method for evaluation of the female breast. World Journal of Surgery, 1980; 4: 701-705.
Gisvold J J, Brown L R, Swee R G, Raygor D J, et al. Com-parison of mammography and transillumination light scan-ning in the detection of breast lesions. AJR Am J Roentgenol, 1986;147(1): 191-4.
Bartrum R J, Jr.Crow H C. Transillumination lightscanning to diagnose breast cancer: a feasibility study. AJR Am J Roent-genol, 1984;142(2): 409-14.
Wilson JM, Jungner G. Principles and practice of screening for disease. Geneva: World Health Organization, 1968. 34th ed: 1–163.
Mande N, Gakwaya A M, Byanyima R K, et al. The triple assessment in the pre-operative evaluation of patients with breast cancer in Mulago Hospital, Kampala - Uganda. East and Central Afr J of Surg, 2004; Vol. 9(2): 111- 116.
Geslien G E , Fisher J R ,DeLaney C. Transillumination in breast cancer detection: screening failures and potential. AJR Am J Roentgenol, 1985; 144(3): 619-22.
Gakwaya A, Kigula-Mugambe J. B, Kavuma A, et al. Cancer of the breast: 5-year survival in a tertiary hospital in Uganda. Br J Cancer, 2008; 99(1): 63-7.
Parkin D M , Nambooze S ,Wabwire-Mangen F, et al. Changing cancer incidence in Kampala, Uganda, 1991- 2006. Int J Cancer, 2010;. 126(5): 1187-95.
Anderson B O, Shyyan R, Eniu A, et al. Breast Cancer in Lim-ited-Resource Countries: An Overview of the Breast Health Global Initiative 2005 Guidelines. The Breast Journal, 2006;