One of the things conventional medicine prides itself on is its diagnoses. But much of the credit goes not to doctors, but to the technologists who have created increasingly sophisticated ways of peering into the human body. Ultrasound, computed tomography (CT), magnetic resonance imaging (MRI) and positron-emission tomography (PET) offer extraordinarily detailed views of every cubic centimetre of the body.
Nevertheless, pictures alone still cannot tell the whole story. So doctors have to fall back on the study of samples from the body - urine, blood and especially tissue.
Tissue sampling - or 'biopsy' ('view of the living' in Greek) - is mainly used to detect cancer in cancer-prone areas such as the skin, pancreas, lung, liver, kidney, breast, testicle, prostate, spleen, thyroid and cervix. Joints and parts of the mouth can be biopsied to confirm certain rheumatic diseases. In cases of severe anaemia or leukaemia, bone marrow can also be biopsied.
Biopsies are often guided by ultrasound or CT to make them safer and more accurate. There are four basic biopsy techniques:
* Aspiration - a thin needle (such as used for drawing blood) is inserted into the tissue to suck out thousands of cells
* Needle - a special cutting needle is used to remove a cone-shaped piece of tissue
* Excisional/incisional - a portion of tissue is removed by simple cutting, using either a scalpel or a tissue punch
* Endoscopic - a fibreoptic endoscope is inserted to reach hard-to-get-at parts of the body such as the gastrointestinal tract, bladder, womb, abdomen, joints and lungs. The sample is taken by tiny specialised forceps on the end of the endoscope cable.
How accurate is biopsy?
In theory, tissue sampling should be 100 per cent accurate. But this is not always the case. The fault can lie with both the laboratory procedure and failure to collect the right tissue, which can result in either not detecting a cancer (false negatives) or a diagnosis of cancer when none is present (false positives).
* Prostate biopsy. This is one of the most inaccurate tests of all, with up to a 25 per cent failure rate, mainly due to false negatives - despite the fact that the biopsy needle is guided by transrectal ultrasound (TRUS) imaging and that six separate samples are taken ('sextant biopsy', the standard procedure). So, doctors now advise that the number of samples be increased to eight or even 10, but this reduces the failure rate by only a few percentage points (Mol Urol, 2000; 4 : 93-7). Doctors at the Mayo Clinic, one of the world's top cancer centres, admit: 'Currently, biopsy does not provide adequate information for differentiating between clinically insignificant and life-threatening prostate cancer' (J Urol, 1995; 153: 1543-8).
* Liver biopsy. This has a better track record - '90 per cent to 95 per cent accuracy in experienced hands' - but is routinely lower than this (Acta Gastroenterol Belg, 2003; 66: 160-5). In practice, you have at least a one in 10 chances of a misdiagnosis.
* Breast biopsy. Each of the five main biopsy techniques has its own inaccuracies. Automated needle biopsy has up to a 40 per cent failure rate, while vacuum-assisted biopsy is better at 16 per cent (Eur J Cancer, 2003; 39: 1676-83). Excisional biopsy is better still with 2.6 per cent false negatives, but stereotactic, or large-core, biopsies are claimed to be over 99 per cent accurate. The downside of stereotactic needle biopsy is that, because only a tiny amount of tissue is removed, a negative result is no guarantee that no cancer is present. Multiple samples may have to be taken to achieve greater accuracy (Gynakol Geburtshilf Rundsch, 2002; 42: 201-11). Fine-needle aspiration is the most commonly used but, because of its inaccuracy, it is usually combined with ultrasound and mammography in a so-called 'triple test', which has an overall accuracy of more than 98 per cent (Can J Surg, 1996; 39: 302-11).
* Sentinel node biopsy. This technique is designed to detect breast cancer that has spread to the lymph nodes. At best, this has a one in 10 failure rate, but may be three times that in inexperienced hands (Arch Surg, 1999; 134: 764-7).
* Pancreas biopsy. This is only about 80 per cent accurate, mainly because it is difficult to distinguish cancer cells in the laboratory (Pathology [Phila], 1996; 4: 389-407).
* Lung biopsy. This can result in up to 30 per cent false negatives, and is one of the few biopsies that is less accurate than modern imaging techniques (Resp Care Clin North Am, 2003; 9: 51-76).
* Kidney biopsy. This has 4 per cent false-positive and 8 per cent false-negative rates (J Urol, 1986; 136: 1292-3).
* Testis biopsy. This has the lowest failure rate, with a record of over 99 per cent accuracy (J Urol, 1999; 162: 364-8).
What are the dangers of biopsy?
Although doctors tell patients that biopsies are 'relatively safe', this is relative to medical interventions in general, most of which are meant to be curative and may therefore be worth the risk.
But biopsies don't cure, they merely diagnose, so they should be as near as possible 100 per cent risk-free.
Yet, what doctors don't tell you is that, in fact, biopsies can be seriously hazardous to your health. Doctors refer these problems 'complications', a euphemism that can cover anything from mild pain to actual death.
* Prostate biopsy complications include blood in the urine, faeces and semen. Blood clots may form and prevent urination. Pain described as a 'dull ache' may persist for as much two weeks.
* Liver biopsy's main complication is bleeding from the point of needle entry, which occurs in about one in 100 cases; in some cases, blood transfusions or surgery are needed (Ann Intern Med, 1993; 118: 96-8). Other complications arise when the doctor mistakenly punctures nearby organs, such as the kidney, lung or colon. The gallbladder is particularly vulnerable and, if damaged, may leak bile into the abdomen, causing peritonitis. These risks taken together mean there is more than a one in 1000 chances of dying from a liver biopsy (BMJ [Clin Res Ed], 1988; 288: 1254-6).
* Lung biopsy is particularly hazardous, causing pneumothorax (air in the space between the lungs and the chest wall, leading to a collapse of the lungs) in about one in three cases.
* Colonoscopy can cause perforation or tears through the bowel wall, sometimes serious enough to require surgery. Bleeding may occur from the biopsy site, which may be severe enough to need surgery and blood transfusion. There may also be extreme abdominal pain and fever.
* Womb (endometrial) biopsy can cause infection in the uterus and fallopian tubes, or an actual tear in the womb lining.
The greatest complication of biopsy
The most serious complication is that biopsies can spread cancer, worsening the very disease they are meant to diagnose. Discovered 50 years ago, when a case of prostate cancer was found to have been 'seeded' (spread) by a biopsy (J Urol, 1953; 70: 937), doctors nevertheless continue to recommend biopsy for prostate cancer detection. Some advise radiotherapy at the biopsy site to prevent cancer seeding (Urology, 1976; 8: 513-5).
One large US military hospital has reported a 1 per cent mortality rate as a direct result of prostate cancer seeded by biopsy (J Urol, 1989; 142: 86-8), which means that the chances of dying from biopsy-seeded prostate cancer are about one in 100.
In breast biopsies, malignant cells are 'displaced' in about 30 per cent of cases (Am J Roentgenol, 1999; 173: 1303-13). The risk of developing cancer from the biopsy is estimated to be one in 15 chances (Acta Radiol Suppl, 2001; 42: 1-22).
Liver biopsy is worse, causing a 'significant risk' of cancer seeding in up to 16 per cent of cases - roughly one in six chances (Dis Colon Rectum, 2003; 46: 454-8). Pancreas, lung and kidney biopsies can all cause cancer seeding.