Degree of Malignancy of T-Cell Acute Lymphocytic Leukemia Related to Autofluorescence in an EL4-Based Model

T cell acute lymphocytic leukemia (T-ALL) is an aggressive hematologic malignancy in terms of its pathology and populations. It is important to understand the phenomenon of autofluorescence in living cells because normal and cancer cells can be distinguished by this feature. However, the autofluorescence link to T-ALL is poorly understood. We provide direct evidence that EL4-based T-ALL model expresses both high autofluorescence (FL1+) and low autofluorescence (FL1-) in the current study. The FL1+ phenotype is not as stable as the FL1-phenotype. Interestingly, CD25+ cells of FL1+ can transform to CD25+ cells when injected into mice. Autofluorescence associated with differential expression of CD25 is related to the malignant degree of disease, which is presented by the tissue infiltration and growth speed of EL4 cells. Of note, the phenotypes of autofluorescence and CD25 expression of EL4 cells are related to the sensitivity to NK cells killing and the survival time of EL4-treated leukemic mice in vivo. Thus, autofluorescence monitoring might provide as new insight for investigating the malignant degree and prognosis index of T-ALL.

Keywords: Autofluorescence; T-ALL; CD25; Survival Term

T-cell acute lymphoblastic leukemia (T-ALL) accounts for ~20% of acute lymphoma leukemias [1]. Patients usually exhibit hepatomegaly, splenomegaly and lymphadenectasis because of leukemic cell infiltration. In a number of non-hematopoietic and non-lymphatic tissues, such as endothelial cancers, including lung, breast, esophageal, skin, cervical and colorectal cancers, autofluorescence is abnormal in these neoplastic tissues; its imaging has also been used for the detection of precancerous lesions [2-5]. It is well known that autofluorescence is produced by living tissues or cells. As a safe, noninvasive, and rapid diagnostic strategy, autofluorescence detection is accepted by some institutions to research the function and phenotype of living cells [6]. Therefore, the purpose of our study was to research the characteristics of autofluorescence in T-ALL to improve strategies of management for this disease. Autofluorescence can distinguish between normal or malignant tissues; the same malignant tissue may differ according to its macroscopic type [7]. Fluorescence spectroscopy can detect the morphology and biochemistry of living cells, which are related to the metabolic activity, cellular proliferation and death of cells undergoing malignant changes. In neoplastic tissues, several factors result in reduced autofluorescence, such as decreased cytoplasmic ratio, loss of the submucosal collagen and increased hemoglobin concentration during neovascularization [8]. As a single-cell level, tumor endothelial cells present differing fluorescence compared with normal cells [9]. Impaired mitochondrial metabolism can change autofluorescence by mitochondrial NADH and FAD in cancer cells [10]. Excitation of mitochondrial NADH and FAD can result in cytoplasmic autofluorescence [11,12]. Dysplastic changes of cells can increase cytoplasmic autofluorescence by changing NADH and FAD levels [13]. Changes in cell physiology and pathology can alter the concentration of NADH [14]; NADH has been regarded as a biomarker for mitochondrial anomalies associated with cancer [15,16]. Relatively few studies on autofluorescence imaging and its mechanisms have been carried out; however, analyses of malignant hematopoietic and lymphatic tissues have been carried out more frequently than with endothelial cancers. It has been reported that leukocytes present autofluorescence [17]. In a lymphoma cell line, Chiaretti S et al. used autofluorescence to distinguish between normal and malignant cells [2]. Thus, it is necessary to further elucidate whether autofluorescence exists in T cell acute lymphocytic leukemia and lymphoma. We used EL4 cells, a T lymphoma leukemia cell line of mice, which can perfectly mimic the process of T-ALL pathogenesis in vivo and represents a valuable cell culture model to study the biological characteristics of T-ALL. Firstly, we found that EL4 cells express FL1+ and FL1- when injected into mice or cultured in vitro. Also, the FL1+ phenotype is not as stable as that of FL1- cells. The expression of interleukin (IL)-2 receptor α chain (CD25) is strongly induced at the transcriptional level after T cell activation [18]. CD25 is a marker of activation of T cells; as EL4 cells are a T lymphoma leukemia cell line, it is needed to explore CD25 expression within such cells. Interestingly, CD25+FL1+ cells cultured in vitro transform to CD25- cells after the inoculation of mice; although FL1+ and FL1- cells cannot transform into each other. We demonstrated that the differing growth speeds among different EL4 cell groups may be related to their expression of autofluorescence and CD25 (interleukin-2 receptor α; IL-2Rα). Furthermore, the EL4 groups with rapid growth speeds in vitro result in the short-term survival of mice when injected in vivo. These results may provide a cell biology-based insight into autofluorescence and its associated characteristics of T-ALL.

EL4 cell line (C57BL/6 origin) was obtained from the lab of the Translational Medicine Institution in Jilin University and was thawed from a frozen stock. Then was cultured on standard tissue culture plastic (12-well round-bottom plates) at an original concentration of 5×10³ cells/mL and 1 mL/well in RPMI 1640 medium (Thermo Fisher Scientific) supplemented with 100 U/mL penicillin (Thermo Fisher Scientific), 10% fetal bovine serum (Thermo Fisher Scientific), 100 μg/mL streptomycin(Thermo Fisher Scientific), and 50 μmol/L 2-mercaptoethanol (Thermo Fisher Scientific), at 37 °C with 5% CO2. 1×10⁶ cells/mL EL4 cells were cultured with or without 100 μl/mL IL-2 in 12-well-plates. We harvested EL4 cells and renovated complete medium every three days. On day 12, EL4 cells in each well were removed to culture on standard tissue culture plastic (6-well round-bottom plates) and 3mL/well in the same complete medium. EL4 cells were harvested and counted on day 14. EL4 cells were cultured with NK cells in a 4-hour chromium release assay as reported previously [19].

Wild-type (WT) C57BL/6 (B6) or C57BL/6(LY 5.2) mice were bred in pathogen-free conditions of the animal center affiliated to the First Bethune Hospital.

Subsets of EL4 cells were stained with rat anti-mouse CD25-Allophycocyanin (APC). Propidium iodine (PI) staining was used to exclude dead cells. Rat anti-mouse FcγR mAb 2.4G2 was used to block non-specific FcγR binding. Antibodies were from BD Biosciences (San Diego, CA). Aliquots of 1×10⁶ EL4 cells labeled with rat anti-mouse CD25-APC antibody were placed on ice for half an hour. After washing, we tested CD25 expression on FACS Canto™ II (BD Biosciences). In some experiments, FL1^-, FL1⁺25⁺, FL1⁺CD25^-, CD25^- EL4 cells were prepared by using the MACS system according to the manufacturer’s instructions (Miltenyi Biotec) and cell sorting prior to injection into mice or culture in vitro. The purify (>99%) was tested by FACS.

This research was performed according to the Care and Use of Laboratory Animals of the National Institutes of Health. The Ethical Review Boards of First Bethune Hospital of Jilin University approved this study (Protocol Number: 2017-004).

Survival rate were analyzed using a Kaplan-Meier curve+logrank test with GraphPad Prism 5 (GraphPad Software, Inc. La Jolla, CA, USA). One-way ANOVA for multiple comparisons and unpaired student’s t-test for two independent comparisons were used. Flow cytometric analyses were analyzed using FlowJo 10.0.7 software (BD FACSDivaTM software, New Jersey, USA). The statistical analysis was conducted using JMP13.0. Values indicate mean±standard deviation (SD). The data are presented as mean±SD. P<0.05 statistically significant.

To identify the biological characteristics of EL4 cells, including the distribution and their phenotype in vivo, original EL4 cells were recovered from liquid nitrogen and cultured for 7 days in vitro. WT B6 mice received EL4 cells and were then sacrificed when moribund. Vβ12 is the marker of EL4 cells and 100% EL4 cells were determined as Vβ12 positive (Vβ12⁺) cells. As shown in Figure 1A, all EL4 cells expressing Vβ12⁺ were distinguished from normal cells (Vβ12^-) by flow cytometry [20]. A total of 1x10⁶ EL4 cells were injected into naïve B6 mice-induced T-ALL mice. Vβ12⁺ EL4 cells infiltrated every tissue, such as the spleen, bone marrow (BM), lymph nodes (LN), peripheral blood (PBMC), thymus, ovary, liver, lung (21.45±13.98%), and the kidney. The percentage of EL4 cells in every tissue was significantly different as determined by one-way ANOVA for multiple comparisons (*P<0.05). Especially in the liver, almost 57% EL4 cells were detected due to its percentage of nucleated cells. Thus, EL4 cells harvested from the liver were used in the following experiments as the source of cells. We compared the phenotype of EL4 cells harvested in vivo and cultured in vitro. The same phenotype of CD3 on the surfaces of EL4 cells was seen in vivo and in vitro. As shown in Figure 1B, FL1⁺ and FL1^-EL4 cells were detected; EL4 cells expressed CD3 both in vivo and in vitro. Interestingly, there was a different phenotype on the surfaces of EL4 cells. Although the expression of CD25 could be detected within FL1+EL4 cells in vitro, no FL1⁺CD25⁺ cells were noted in vivo (Figure 1C). It is well known that CD25 expression can change during T cell activation. EL4 cells, as a T cell acute lymphoma leukemia cell line, were expected to show altered CD25 expression due to some unknown mechanisms. Thus, it was necessary to explore how CD25⁺ cells change when injected in vivo. From the results above, we concluded that some phenotypes of FL1+ EL4 cells were not as stable as that of FL1^- EL4 cells, particularly CD25 expression. To determine the associated property of autofluorescence on EL4 cells, CD25 was employed as an important factor for investigation.

To elucidate the phenotypic changes of CD25⁺ cells in vivo, original EL4 cells were recovered from liquid nitrogen and cultured for 7 days in vitro. Cells were then separated into FL1⁺CD25⁺, FL1⁺CD25^-, and FL1^- cells by a MACS separation system. Then A20⁺ C57BL/6(LY 5.2) mice received 1x10⁶ A20^-FL1⁺CD25⁺, FL1⁺CD25^- and FL1^- EL4 subsets respectively. Two weeks later, mice were sacrificed when moribund. As normal cells were FL1^- in B6 mice, we harvested A20^- cells to observe changes in EL4 cells. The process of phenotype change was investigated when injected in vivo (Figure 2). FL1⁺CD25^- EL4 cells harvested from the livers of mice may originate from FL1⁺CD25⁺ cells in vitro as there were no FL1⁺CD25^- cells in vivo (Figure 2A). In addition, no FL1⁺CD25^- cells in vitro transformed into FL1^-CD25^- cells in vivo (Figure 2B). Furthermore, FL1^- cells in vitro could not transform into FL1⁺ cells; CD25^- cells in vitro could not transform into CD25⁺ cells in vivo (Figure 2C). These results indicated that FL1⁺ and FL1^- EL4 cells cannot transform into each other; however, CD25⁺ cells were found to transform into CD25^- cells. Therefore, no CD25⁺ EL4 cells were seen in vivo. It was suggested that the expression of CD25 is in proportion to the FL1⁺ phenotype of EL4 cells. We also aimed to study the factors that affect the expression of CD25. IL^-2 is known to induce CD25 expression on activated T cells. In our study, however, no changes in CD25 expression were seen when EL4 cells were cultured with or without IL^-2 at day 7 or 14 (Figure 2D). Therefore, it is possible that there are other factors that regulate the expression of CD25 on EL4 cells. Although mechanisms remain unclear, our study shows that changes in CD25 expression may independent of IL^-2.

To further test whether there were differences in the biological characteristics of CD25⁺ or CD25^- EL4 cells, WT B6 mice were injected the same number of CD25⁺ or CD25^- EL4 cells (1x10⁶), and were sacrificed 2 weeks later. We found that there were differences in the sizes of liver infiltrated by EL4 cells. Hepatomegaly was markedly severe in mice injected with CD25^- EL4 cells than those injected with CD25⁺ EL4 cells (Figure 3A). The average liver weight between CD25⁺ and CD25^- EL4^-treated mice is significantly different (1.763±0.072 g vs. 1.458±0.076 g) (n=5 per group, mean ± SD, P<0.05). Therefore, it is necessary to further demonstrate whether there is a disparity of survival term between these two groups of mice due to the notable degree of tissue infiltration of EL4 cells. As shown in Figure 3B, WT B6 mice received the same number of CD25⁺ or CD25^- EL4 cells (5x10); the survival term was significantly reduced in mice injected with CD25^- cells than CD25⁺ cells (P<0.01). These results demonstrated that CD25^- EL4 cells induced notably severe hepatomegaly and short survival term compared with CD25⁺ EL4 cells. Therefore, it is likely that CD25^- EL4 cells are more malignant, leading to rapid disease progression. Furthermore, an NK cell cytotoxicity assay was performed with CD25⁺ and CD25^- EL4 cells. The result showed that CD25⁺ EL4 cells were more sensitive to NK cells (Figure 3C). Therefore, it is possible that the decreased malignancy of CD25⁺ EL4 cells may be attributed to NK cell killing. Another possibility is that the transformation of CD25⁺ EL4 cells may affect the degree of malignancy. As there were differing degrees of disease progression induced by EL4 cells according to their expression of CD25, we investigated whether there were differences in growth speed between CD25⁺ and CD25^- EL4 cells. Since no expression of CD25 in FL1^-EL4 cells was detected, EL4 cells were separated into FL1+CD25+, FL1+CD25^- and FL1^- populations by a MACS separation system. We found that FL1+CD25⁺ EL4 cells grew significantly slower than FL1⁺CD25^- EL4 (P=0.004) and FL1^- EL4 cells (P=0.013) (Figure 4A). Similar growth speeds were seen between FL1⁺CD25^- and FL1^- cells. Then, we continued to test the hypothesis that the mortality of mice treated with various EL4 cell groups differs according to their phenotype in vivo. So, 5x104 FL1^-, FL1⁺CD25⁺ and FL1⁺CD25^- EL4 cells were respectively injected into naïve B6 mice. The survival time of FL1⁺CD25⁺ EL4-treated mice was longer than that of FL1⁺CD25^- EL4- and FL1^- EL4-treated mice. As shown in Figure 4B, the difference in mortality of mice was statistically significant for the FL1⁺CD25⁺ EL4-treated mice versus the FL1^- EL4-treated mice (P=0.04), and the FL1⁺CD25⁺ EL4-treated mice versus FL1⁺CD25^- EL4-treated mice (P=0.037). Taken together, these results demonstrated that FL1⁺CD25⁺ EL4 cells induced slow progression of disease, which was presented as slow growth speed in vitro and long survival term in vivo.

The illumination of unstained cells that naturally emit light is termed as autofluorescence. It is important to understand phenomenon of autofluorescence in living cells as this distinguishes normal and cancer cells. For many endothelial tissues, dysplastic changes result in increased contents of mitochondrial NADH and FAD, which leads to changes in autofluorescence in cancer [13]. For diseases of malignant hematopoietic and lymphatic tissues, a previous report showed that autofluorescence can distinguish between a type of malignant B cell and normal lymphocytes [2]; however, the characteristics of autofluorescence in T cells lymphoma leukemia remains unknown. T cell acute lymphoma leukemia is characterized as a heterogeneous disease with poor prognosis [21]. By using an EL4-based T-ALL model, to the best of our knowledge, we first reported that these cells express FL1⁺ and FL1^- both in vivo and in vitro. CD25⁺ EL4 cells only persisted in FL1⁺ populations in vitro as they are observed to transform to CD25^- cells following injection into mice. It is possible that the transformation may require time as CD25⁺ cells induce notably longer terms of survival compared with CD25^-cells in vivo. Another possibility is that CD25⁺ EL4 cells are more sensitive to NK cell cytotoxicity. Furthermore, notable severe tissue infiltration, such as hepatomegaly, appears in mice with CD25^- cells when the same numbers of CD25⁺ or CD25^- EL4 cells are injected at the same time. These results indicate that the expression of CD25 in FL1⁺ might be related to the degree of malignancy of T-ALL in an EL4 model. This was further determined by separating EL4 cells into FL1⁺CD25+, FL1⁺CD25^- and FL1^- populations. We observed that FL1⁺CD25⁺ cells grow more slowly than FL1⁺CD25^- and FL1^- cells, which was statistically significant as determined with growth curves. Additionally, FL1⁺CD25⁺ cells result in statistically significant long-term survival compared with FL1⁺CD25^- and FL1^- cells. The survival term of FL1⁺CD25^- and FL1^- cells is similar without statistical significance. Autofluorescence was associated with expression of CD25 and may therefore be related to tissue infiltration, speed of growth and survival term in the EL4-based T-ALL model in the current study.

Over the past few decades, autofluorescence analysis is regarded as a promising way to distinguish between normal and cancer cells. The diagnostic applications of this method has notably increased [22]. The characteristics of autofluorescence in tissues are determined by the morphological and biochemical states of cells [23]. Malignant tissues and the corresponding normal tissues can be distinguished by autofluorescence [24]. Regarding the autofluorescence of malignant lymphatic tissues, a study reported that the fluorescence of lymphoma is significantly lower than that of lymphoid hyperplasia and normal tissues. It has also been shown that the accuracy is 91.5% by the visual classification of autofluorescence imaging (AFI). Thus, AFI may be regarded as a useful tool to indicate the disease stage and appropriate therapy in the certain types of lymphoma [25]. Furthermore, autofluorescence has been observed in leukocytes, such as macrophages, eosinophils and neutrophils [26]. Our data indicates that autofluorescence appears in an EL4-based T-ALL model; investigation of its characteristics might have clinical value for patients with T-ALL. CD25 is termed as the interleukin-2 receptor α chain, which is expressed on the surfaces of activated T cells and a type of T cell lymphoma leukemia, known as adult T-cell leukemia [27]. In our study, we reported that EL4 cells, as a model of T-ALL, partially express CD25 in vitro. We also demonstrated that CD25⁺ cells are not detected in vivo because they transform into CD25^- cells. Since both the expression and transformation of CD25 appear with high autofluorescence of EL4 cells, these data raise a possibility that high autofluorescence might not be as stable as low autofluorescence. Furthermore, different phenotypes associated with autofluorescence result in varying extents of tissues infiltration, growth speed and survival term. This may be consistent with the findings of the current study in which cell autofluorescence was related to the degree of cell differentiation. It has been identified that autofluorescence decreased when cells differentiate [28]. Until recently, the biological knowledge of T-ALL was limited [29]. Thus, autofluorescence monitoring might provide new insight into the processes of differentiation, which may further contribute to investigations into the malignant degree and prognosis index of T-ALL. However, a limitation of the present study is that we did not analyze clinical samples of T-ALL patients; we lacked clinical data to further identify our results. In the future, we will collect peripheral blood samples of T-ALL patients to investigate the expression CD25 and in relation to autofluorescence. In conclusion this new finding that autofluorescence exists in EL4-based T-ALL cells provides an insight into hematologic malignancies and lymphoma disease. Although cells with different autofluorescence intensities cannot transform into each other from in vitro to in vivo, its stability is not coincident such as different expression of CD3, CD25, and TCRαβ only existed in FL1⁺ EL4 cells. The expression of CD25 can transform in FL1⁺ cells as seen in vitro and in vivo, and is related to growth speed and survival term. Although no exact mechanisms have been clarified, our study shows that T-ALL, at least in the EL4 model, may present varying intensities of autofluorescence. This application of this approach showed that the expression of CD25 in FL1⁺ cells is related to the degree of malignancy of T-ALL, which may facilitate future mechanistic studies. Therefore, autofluorescence combined with CD25 expression may be considered as a useful adjunctive diagnostic tool and prognosis index in EL4-based T-ALL. Future studies of autofluorescence and the expression of CD25 should be broadened to research the underlying mechanism of the changes, regulations and biological characteristics of CD25 expression, in which we can collect extensive information to address the clinical application of potential treatments for T-ALL.

This work was supported by National Natural Science Foundation of China (81770149), National Natural Science Foundation of China (81100350).

Zhonghua Du: designed and performed experiments, analyzed data, and drafted the manuscript; Lixia Wang: performed experiments and analyzed data; Yongguang Yang: contributed to the development of the project and edited the manuscript; Yanping Yang: conceived the research project, designed experiments, analyzed data, and wrote the paper.

We piloted this project in the community of Rehri Goth, a peri-urban slum area in the neighborhood of Bin Qasim, Karachi. The population of Rehri Goth is estimated to be 75,000 and there are approximately 12,000 houses in the community [10]. The town of Rehri Goth is divided into eighteen smaller clusters called ‘Para’. We included all ‘Paras’ for the purpose of screening of breast cancer. The Aga Khan University (AKU) has a primary health care (PHC) center established at Rehri Goth, which was utilized for clinical breast examination of women.

We used the modality of sequential screening; breast self-examination (BSE) followed by clinical breast examination. In the first step BSE was carried out and those who reported a lump or other danger signs including a change in size, appearance and skin of the breast, wrinkles or dimples in the skin of the breast, a lump in the breast, discharge from the nipple, pimples on the nipple, areola or the breast, nipple inversion or pain were then examined by a surgeon at PHC. Participants in whom the clinical breast examination was suggestive of breast cancer were taken for diagnostic mammography. The confirmed cases of breast cancer were referred to a nearby tertiary care hospital for treatment.

Prior to starting the screening program, efforts were put in by the lady health visitors of AKU and the local social activists to create awareness among the masses regarding the study. This helped to gain access to the participants. Five female university volunteer students were trained for field work. A five-day workshop was conducted by the principal investigator to train volunteers for educating the masses regarding breast cancer and the technique of breast self-examination.

The training of the trainers was followed by the implementation of the education and screening programme in the community. The field work was carried out from 19^th December, 2016 to 2^nd January, 2017. About thirty houses were included from each Para. Health education regarding breast cancer was given in all the households. A household was included for screening if there was at least one woman aged 40 years and above. If there was more than one eligible woman in the household, all of them were included. Written consent was taken prior to administering the questionnaire. The questionnaire included questions related to the age of menarche, age of menopause (if applicable), and the women’s parity. Participants were also asked about the history of any breast lump, pain or tenderness. The participants were then educated regarding the signs and symptoms and early detection of breast cancer, the technique of breast self-examination was demonstrated. The participants were also given brochures illustrating the danger signs of breast cancer and pictorial representation of breast self-examination.

At the last day visit to the community, the women were channelized to come to the AKU primary health care center on a specified date for clinical breast examination in case any danger sign for breast cancer was elicited. The clinical breast examination was performed by a lady surgeon at the AKU, PHC center.

A total of 526 houses of the 18 Paras were reached by the project team and breast cancer education was given. The houses were reached out on the basis of feasibility and 131 women aged >40 years were invited to participate. A total of 93 women participated in the study.

The mean age of women of the study was 53.6±13.2 years. All of them were married. The mean age of menarche was 13.4±1.4 years Majority of the women were multiparous with a mean of 5.7±2.7 children per woman. Only six participants were uniparous. About 77.41% of the women were pre-menopausal with the average age of menopause being 46.9±5.6years (Figure 1).

After the BSE of the 93 participants, 18 reported danger signs of breast cancer. Majority of the women, 17 reported a unilateral lump in the breast whereas one complained of nipple discharge (Table 1). Transport arrangements were made and the participants were brought to the AKU primary health care center. Only 9 out of 18 women turned up on the specified date and were examined by a lady surgeon. Informal interviews with the participants revealed that the rest of the women did not appear for clinical examination as “they were not allowed by the husbands to leave home”, “had to prepare meal for lunch”, “were afraid of finding out anything serious about their health”. The clinical examination of the nine participants revealed that two required a diagnostic mammography.

Our team members accompanied the two participants to ‘Dar-ul-sehat’, a tertiary care hospital for standard full field digital mammography. The mammograms were read by a consultant radiologist at Aga Khan University Hospital. One of the participants was diagnosed to have Paget’s disease of the breast whereas the other participant was labeled as invasive ductal carcinoma of the breast stage II. The detected Breast Cancer patient in our study was aged 45 years, pre-menopausal, had 5 children and her last pregnancy was 8-10 years back. Thus, one case of breast cancer was diagnosed after screening 93 women, giving a cancer detection rate of 1.04%. The diagnosed case was taken to a public sector tertiary care hospital providing free treatment, for further management.

We aimed to explore the usefulness of breast self-examination as a population-based screening tool in the context of Pakistan. Our study has revealed a detection rate of 1.04% for breast cancer by physical examination. Although clinical breast examination and screening mammography remain the mainstay for population based breast cancer screening, yet, in a limited resource setting like Pakistan, breast self- examination can be of worth in preventing and delaying morbidity and mortality associated with delayed identification of breast cancer [11]. A similar finding has been stated by the National Canadian Breast Screening Study which shows that there is no additional benefit in breast cancer mortality reduction by the use of screening mammography in addition to physical examination or usual care; rather there is 22% over-diagnosis of breast cancer with screening mammography [12]. However, some studies show additional benefit of screening mammography over clinical examination demonstrating the sensitivity, specificity and accuracy of mammography as 77.6%, 98.8% and 98.6% respectively; those of examination as 27.6%, 99.4% and 98.8% respectively [13]. In our study only, half of the high-risk women turned up for the clinical examination. There is a possibility that the detection rate could have been even higher had they been a part of the study till the end. Currently there is no population-based screening program for early detection of breast cancer in Pakistan. Based on our findings we conclude that physical examination can be used as a tool for screening of breast cancer in a low resource setting like Pakistan. Several studies in Pakistan have recommended Self-Breast Examination as a cost-effective and feasible strategy to substitute for more costly mammograms, in resource-limited settings and in areas where access to medical professionals may be inadequate or difficult [14,15].

In Pakistan, the efforts towards breast cancer screening and early detection at a population level are fragmented at best. Dedicated breast cancer care centers have been set up in Islamabad (2014) and Lahore (2017), which provides free of cost mass screening services to the local communities. Other public and private organizations have also provided sporadic screening services in local communities through mobile clinics or health camps. However, there is no surveillance mechanism developed as part of these initiatives, and no data is available with regards to screening results. We recommend that all efforts in this regard be unified, so that a scientific database can be created [16]. Efforts are being made towards development of nationwide hospital-based registries for cancer patients, primarily led by the Pakistan Atomic Energy Commission, but at the moment, paucity of data makes it difficult for generation of evidence-based decision making.

Some methodological and technical limitations need to be considered while interpreting the outcomes of this study. The volunteers educating the participants were university students. We learned from our experience that volunteers having some medical background such as medical or nursing students would be a better option for educating the masses. In this pilot project recruitment of participants was non-random and included women from a single community; therefore, the generalizability to the bigger population is not possible, however that was not the purpose of the study. Our study has revealed the importance of breast self-examination as a useful tool for population-based screening. Based on our findings, we have concluded that sequential screening is a feasible option for the detection of breast cancer in our community and that this study design can be applied to a larger cohort. The recommendation for prevention of breast cancer through screening in current practice in Pakistan is a grey area. Different clinicians have tailored guidelines according to local context, but there exists a gap for national level comprehensive policy. Based on our findings, we advocate that awareness needs to be created among the masses regarding breast cancer and the importance of breast self-examination. Unfortunately, we do not have a national breast cancer screening program. There is a dire need to combine efforts in a concerted and unified manner. Development of a national breast cancer screening policy followed by community-based programs will help us identify missed cases, and improve service delivery component to those in need.

We conclude that sequential screening is applicable to detect breast cancer in our setting. Large scale projects need to be implemented involving multiple communities for breast cancer detection using BSE and clinical examination to prevent or delay morbidity and mortality associated with breast cancer.

The authors would like to thank the team of Urban Health Program, AKU for their support and cooperation during the course of this study. We are also thankful to the volunteer students of Karachi University for their field work. We also acknowledge Prof. Dr. Zafar Fatmi and Dr. Ahmed Asaad Nafees for their valuable input in designing and implementing the study.

Written informed consent for publication was obtained from all the mothers of the participants.

The datasets analyzed for this paper are available from the corresponding author on reasonable request.

The authors declare that they have no competing interests.

This was not a funded project. All the expenses borne during the course of the study were taken care of by the principal investigator of the project.

MI came up with the conception and design of the study. She was involved in the rolling out of the project. She was a major contributor in writing the manuscript. MAK participated in writing the manuscript and critically revised the content. UR provided a supervisory role during the rolling out of the study. He also reviewed and revised the manuscript. SZ was the surgeon involved in carrying out the clinical breast examination in the outreach clinic.