Clinical Breast Examination is an important tool in the screening and diagnosis of breast cancer. However, training healthcare providers in the confident use of the technique using patients can be problematic and time-consuming. This article gives insight into the design and development of biofidelic (life-like) simulation models which can be used in training to help detect cancer.
Clinical breast examination is one part of what’s called the ‘triple test’ for breast cancer. The triple test involves some form of imaging, such as mammography or MRI; biopsy as another arm; and Clinical Breast Examination (CBE). CBE is used by health professionals such as general practitioners, breast surgeons and breast-care nurses trained to recognise different types of abnormalities in the breast. It involves a visual examination, taking a medical history from the patient and physical examination. This includes palpation of the breast, that is, determining by touch which breast lumps are normal and which are suspicious.
The triple test is important because each of its pillars detect cancer in a different way. Imagine the different pillars as parts of a Venn diagram, all intersecting with each other. A proportion of cancers are only detected by mammography or imaging; a proportion are only detected by breast examination; and a proportion are only detected by biopsy. In the centre where the three circles overlap you have cancers known as triple positives.
Some cancers are detected using all three methods, but some are only detected with one of the methods and come up negative on the others. If you are one of the 10-20 per cent of women with breast cancer that have a negative to mammography your cancer won't be detected in the mass mammography screening programs, it’s detected in one of the two other ways.
Although CBE is considered by some to be a bit old-fashioned now because of the modern preference for tests, it's still very important. While mammography is extremely important—it is very specific and sensitive—people who live in remote areas or underdeveloped countries may not have any access to the kinds of imaging equipment used. Also, palpation sometimes detects ‘interval cancers’ — those that only become noticeable between imaging appointments — so CBE can act as a way you can quickly and easily check for the cancer between appointments of, for example, two years.
The fact is that any way that you can detect breast cancer is a good way and you shouldn't cut off any avenue when you're trying to do this because early detection saves lives. Breast cancer is still the most prevalent cancer in women, and the second most common cancer diagnosed worldwide.
Clinical breast examination can be taught using theory, lectures, videos, computer simulations/models, virtual reality simulators, and commonly what's known as a ‘standardised patient’ or as an ‘intimate examination associate’, a paid actor who agrees to have intimate examinations performed on them. Sometimes synthetic models of breasts are used.
Students can get allocated to a breast clinic for a rotation, but because there are so many rotations to choose from, some students may not have this experience. This means that some students graduate with nothing more than a training session or two from their pre-clinical years.
Much teaching now happens outside of major teaching hospitals and is different to how it used to be. There are sensitivities involved in having a large number of students examining someone who has just been diagnosed with breast cancer. Also, the gap between being diagnosed with breast cancer and going into surgery can be very brief which means patients with different types of breasts and different types of breast disease may not be on hand for the students to examine. Medical students would once have practised on each other but female medical students are not going to volunteer for that, just as the male students wouldn't volunteer for prostate examinations or digital rectal examinations.
Breast simulation models are used quite frequently in training but there can be pitfalls with them. Sometimes they have finger marks on them that indicate, without the student even touching the breast model, where the lumps are on the breast. Often the model is small and doesn't represent anyone with large breasts – an increasing proportion of the population. The model may not represent breast adiposity (fatty tissue) and there's often no normal nodularity (fibro-glandular tissue). When you actually do a breast examination, you'll feel these tissues inside the breast, and other normal anatomical structures such as the milk ducts and ribs. Often none of those are incorporated in a simulation model, just smooth silicone with a lump inserted, so the lump is quite easy to detect—there are no background structures to confuse, like ribs.
There is a range of what is ‘normal’ in breasts. Clinician and scientist William Goodson MD, a recognised leader in breast cancer, has said that in terms of the detection of breast cancer, there are two important elements. One is the softness (or hardness) of the breast, called the durity. The other is the nodularity. They were the two aspects that our research has focused on in terms of variation. We tested each individual element, like fat and how soft the fat was, the feeling of the skin, different bits of nodularity and different pathologies, such as fibroadenoma and cancer. Although there was a large range of softness and of nodularity we tried to focus on only using cases that Adelaide breast surgeon Dr Melissa Bochner said were important for teaching.
Various construction materials were used, mostly silicone, MRI scans, traditional casting and three-dimensional (3D) printing to build models with a lifelike look and feel, that is, they are biofidelic. The models were realistic in anthropometry (size and shape), feel (durity and nodularity) and appearance (skin feel and colouring). Visual biofidelity enhances the perception of feel. The anatomically correct layering of ribs was incorporated, soft adipose tissue, nodularity and additional signs of breast disease, both benign and pathological.
With the first model developed, the components were tested individually and once we were happy with them I went into surgery with Melissa. The patient, who had breast cancer, was having a mastectomy. The surgeons took the removed breast and gave it to me on a tray so I was able to feel it and poke it and build a comparable breast model right next to it. This model is right in the middle of the distribution for softness and is also one of the most commonly occurring breast softness type. That first model was relatively easy to develop. From there, getting the other variations and understanding what caused these variations was much more difficult.
Eventually six multilayered breasts were developed representing a range of normal human variation for durity, nodularity and adiposity. The models had much variation in them to do with either the thickness of the material or the way that they were layered. These models were given to a series of breast surgeons, who were asked to take them into the clinic and feel the breasts of the patients and feel the range of models and say whether they thought the breasts were similar or not for nodularity and softness. More than 80 per cent of the time the surgeons said that they were similar or that they couldn’t tell the difference. The breast models are broadly representative of people. However, the softest model is not as soft as the softest patient so potentially we could add another model as the very softest patient to improve the similarity at that end of the spectrum.
To our knowledge, these novel biofidelic models are the first models to incorporate normal human variability and be validated with real patients. They provide a standardised way of teaching health professionals normal from abnormal. The models are ready to go. However, they would need durability testing prior to manufacturing and some sort of manufacturing protocol if they were going to be mass produced. This would require additional funding.
This work into breast simulation models and medical manikins grew from a background in the fashion industry. I began working at the Flinders Medical Centre in Adelaide in 2008 including on a project that involved scanning and quantifying the size of women's breasts for a breast reduction study. Professor Harry Owen, an anaesthetist who is the Head of Medical Simulation, requested help in developing a breast examination model for the largebreasted person because examining people who have a larger BMI is quite a different proposition requiring different techniques. There weren’t any models that represented larger people with larger breasts. A grant was obtained and the research was published in 2011. That year it won the Ken Provins award for the best paper at the Human Factors Conference in Sydney (HFESA). The keynote speaker was Professor Richard Goossens, who is now my professor in Delft University of Technological Design, in the Netherlands. The university specialises in the crossover between medicine and design, including a section called Medisign.
A medical designer acts as a central hub to bring experts into the room, drawing relevant information from them and translating that into something that is a useful product fit for the purpose of whatever it needs to do. We have developed lactating manikins, a model with a pleural tap for students to practise the medical procedure involved in draining the fluid out of lungs, and smaller models such as one used for practising suturing.
One of the tools of the medical designer is anthropometry. Anthropometry is the science of making systematic measurements of the human body and is used in both fashion and medical design. Anthropo means human and metry is measure. Anthropometric measurements involve the size (e.g. height, weight), structure (e.g. neck circumference, shoulder and hip width, waist-to-hip ratio), and composition (e.g. lean body mass and percentage of body fat) of humans. It has been used historically, controversially, to try to link physical with racial and psychological traits, and is now used practically in industrial and clothing design, including for work, and ergonomics.
My involvement in anthropometry began when I conducted a survey in 2002 of the Australian population to see if we could find who was average. It turned out that there's no such thing as average, in fact everybody is a combination of small, medium and large. For example, you could be average for weight and short, or you could be the same weight but very tall and slim, and so forth. Some body measurements are related to each other. Your height is obviously going to be related quite strongly to things like your shoulder height and your hip height from the ground because leg length is associated with height. But circumferences which are related to weight are independent.
The World Engineering Anthropometry Resource (WEAR) of which I am a co-founder is a group of interested experts involved in the application of anthropometry data for design purposes. The members and partners are from around the globe. WEAR is a non-profit organisation registered in Europe.
Regardless of the technique of breast examination used, there is a need for standardised training to be introduced across Australia: there is currently no standardisation of methodology or performance evaluation for trainees to test whether they have the right skills or not. The current regime for testing students in South Australia for example, doesn’t actually test them for palpation skills in terms of the comprehension involved, that is, did they find a breast lump and was it normal or not?
Further unpublished research we have conducted is developing a test that could potentially be used for GP accreditation. The test gives very specific feedback and results such as: What lesions did they miss? Why did they miss the lesions? Then you could give remediation for that person on the part of the technique that they are missing.
Daisy Veitch, Melissa Bochner, Lilian Fellner, Christopher Leigh & Harry Owen. Design, development and validation of more realistic models for teaching breast examination. Design for Health. Vol. 2, 2018 - Issue 1, pp 40-57. Published online: 04 May 2018
1. Goodson WH. Clinical Breast Examination and Breast Self-Examination. In: Sauter ER, Daly MB, editors. Breast Cancer Risk Reduction and Early Detection, 2010.
2. Irwig L, Macaskill P, Houssami N. Evidence relevant to the investigation of breast symptoms: the triple test. The Breast. 2002;11:215-20.
3. World Cancer Research Fund. Worldwide data: World Cancer Research Fund; (Available from: http://www.wcrf.org/int/cancer-facts-figures/worldwide-data)
4. Goodson WH, Moore DH. Overall Clinical Breast Examination as a Factor in Delayed Diagnosis of Breast Cancer Archives of Surgery. 2002;137:1152-6.
5. Veitch D, Bochner M, Fellner L, Leigh C, Owen H. Design, development and validation of more realistic models for teaching breast examination. Design for Health. 2018.
6. Veitch D, Dawson R, Owen H, Leigh C. The Development of a Biofidelic Breast Cancer Large Size Female Patient Simulator using 1D and 3D Anthropometric Data. Ergonomics Australia. 2011;Special Edition.