Radiography
Volume 18, Issue 1 , Pages 5-8, February 2012

Early diagnosis of cancer by imaging: The primary care perspective

Peninsula College of Medicine and Dentistry, Salmonpool Lane, Exeter EX2 4SG, UK

Received 21 July 2011; accepted 7 October 2011. published online 24 October 2011.

Article Outline

Abstract 

The United Kingdom (UK) performs relatively badly in cancer survival, much of which underperformance is attributed to delays in diagnosis. As a result, increased access to cancer diagnostics for general practitioners is proposed. Four cancer sites and their associated tests have been chosen: lung/chest X-ray, colorectal/colonoscopy, ovarian/trans-vaginal ultrasound and brain tumours/MRI scanning. This article elaborates on the need for these, which patients are likely to be chosen for investigation, and the effect this will have on radiology departments.

Keywords: Oncology, Radiology, Primary health care, Diagnosis

 

Back to Article Outline

Introduction 

It is well accepted that the UK has a relatively poor record in cancer outcomes when compared with other European countries. An estimated 10,000 lives are lost annually, with these ‘extra’ deaths apportioned to many different cancers.1 Much of this poor performance is attributed to delays in diagnosis. In simple terms, diagnostic delays can occur at four stages: patient late presentation, doctor non-recognition of the possibility of cancer, doctor recognition, but the patient is not investigated, and finally delays in the investigation process. The first two stages largely have to be addressed by education. This article concentrates on the last two of these four stages.

Once a General Practitioner recognises a patient may have cancer as a cause for their symptoms, it seems obvious that they should investigate. In fact it is far from obvious. The problem is that the risk of cancer has to be high enough to warrant investigation. An example may help: a man aged 60 has roughly a 0.25% risk of having colorectal cancer diagnosed in the next year – just by virtue of his age and sex.2 Yet you would not expect his GP to test him – though he would generally be advised to participate in the national screening programme. If this man complains of constipation to his doctor, the risk of cancer is now about 0.4%.3 All GPs know that constipation may be a symptom of colorectal cancer, yet once again this man would usually go without investigation. If he had diarrhoea, the risk is now 1%, but it is really only until he describes rectal bleeding, with a risk of 2–3%, that most GPs would refer for investigation.4 Even then, national guidance issued by NICE, argues that the rectal bleeding should be persistent or be accompanied by other symptoms before urgent referral is made.5 This gradient of risk is crucial – and probably underlies much of the UK’s poor cancer performance. Put crudely, GPs are discouraged from investigation of low-risk patients. Or at least they were until recently.

In recognition of the above, the UK Department of Health has brought in several initiatives. The most relevant to this article is ‘open-access’ to GP investigation for suspected cancer. This is being established for four cancers, though the long term aim is to expand the list of possible cancers, and of course, the matching list of imaging tests. The project goes ‘live’ in early 2012. The four cancer sites – and their matching tests – are lung (chest X-ray), colon (flexi-sigmoidoscopy/colonoscopy), ovary (ultrasound, ideally trans-vaginal) and brain (MRI). We will discuss the implications of this in the rest of the article, plus hint at what other cancer sites will enter this initiative subsequently.

Back to Article Outline

Lung cancer 

This is theoretically the easiest of the four cancer sites, in that chest X-rays (CXRs) have been available to GPs with open access for decades. The precise arrangements for the patient vary across radiology departments, with some providing true open access (the patient simply carries the GP request form), patient-booked appointments, department-booked appointments, and in some areas, separate services run by chest clinics which in the past offered open access for TB CXRs. The new initiative does not seek to change this, but does expect an increase in volume of testing.

GPs are being asked to lower their threshold for requesting a CXR. At the time of writing, the precise ‘rules’ for GPs to follow are not fully agreed, but it is likely that a ‘new’ cough in a smoker, or any cough persisting over three weeks without an obvious explanation and dyspnoea will be the main additional symptoms. Haemoptysis has always been seen as an indication for CXR. This recommendation is unchanged, though due recognition is given to the fact that less than half of lung cancer patients actually experience haemoptysis before diagnosis.6 This approach is being studied in a Cancer Research UK funded research project based in Wales, following successful piloting in Doncaster. A further initiative of giving GP paper charts outlining the specific risk of cancer with different symptom patterns, the risk assessment tool, or RAT, has been in place since early 2011 in several areas of England, with at least one early cancer being identified, despite being a non-smoker solely complaining of cough.

All these initiatives should lead to more requests for CXRs. Because CXRs have been around for so long, many of the problems associated with them have been recognised. The first one is the negative CXR in patients who actually do have a lung cancer. The percentage of false negatives has been reported both from secondary care, and (more germane to this article) primary care. Secondary care studies of patients with normal CXRs and lung cancer have examined patients complaining of either haemoptysis or dyspnoea. Cancer is diagnosed in 6–21% of patients who have haemoptysis and a normal X-ray, who are referred for further investigation.7, 8, 9, 10, 11 For dyspnoea and normal X-rays, the yield of cancers is under 1%.12, 13, 14 In primary care, nearly 25% of cancers have a negative CXR in the year before diagnosis.15 Some of these may be true negatives in that there was no lesion several months before diagnosis – lung cancer can grow extremely quickly, with the mean doubling time 126 days,16 but even so it does mean that a negative CXR cannot be wholly relied upon.

The second problem with CXRs and lung cancer is the indeterminate result. Many cancers present with a chest infection, and a CXR is requested to identify both infection and any underlying cancer. Frequently, the CXR shows shadowing which could simply be a chest infection, but the radiologist is unable to exclude a cancer. Lateral images sometimes help, but usually the patient has to be recalled for a repeat CXR in 4–6 weeks time. Radiology departments need good protocols here, covering whose responsibility it is to ensure a repeat CXR is taken (the GP’s? the department’s? the patient’s?) and how much is communicated to patient and GP.

A computerised tomography (CT) scan may also sometimes be useful in this scenario, and this brings similar questions – in particular is it better for the patient to proceed to immediate CT scan, or should the GP be asked to request one? There is no doubt that CT scanning reveals more. Several studies have demonstrated that low-dose helical CT of the lung detects more nodules and lung cancers, including early-stage cancers, than does chest radiography.17 The disadvantages of CT scanning are increased radiation and expense. However, radiation doses for CT scanning are falling and likely to fall further.18 In addition, one has to ask is CT scanning really that much more expensive than a plain CXR? This is not a simple question; one must take into account not only the cost of the procedure but also the efficacy of the test and the reduction in re-investigation. CT scanning has shown itself to be cost-effective and superior to CXR in the context of lung cancer screening. The recently published National Lung Screening trial has demonstrated a 20% reduction in mortality using low-dose CT compared to CXR: CT demonstrated three times more abnormalities and while there were more false positives, these were generally managed by ensuring stability on repeat CT scanning.17 In the medium term it is possible that GPs will be offered direct access to CT scanning. Certainly the use of direct access CT in other areas has been said to be effective and there is no reason to think this would not be the case with chest CT.19

Back to Article Outline

Colon cancer 

This cancer is difficult for GPs, but less so for radiology departments. GPs have the problem that the symptoms of colon cancer are extremely common in patients without colon cancer, making the actual chance of cancer in a symptomatic patient quite low.2, 3, 15, 20 This is compounded by the main test (colonoscopy) being uncomfortable and moderately expensive. Indeed the costs of diagnosis of colorectal cancer exceed the costs of primary treatment.21

As most patients today are investigated by colonoscopy or flexible sigmoidoscopy, radiology departments should be spared much of the increased workload generated by initiatives aimed at making GPs refer more patients. That said, in some areas, waiting times for colonoscopy are long (exacerbated in part by the national bowel screening programme) and barium enema waiting times are shorter. Barium enema is not as helpful as colonoscopy, in that it is less sensitive and cannot provide biopsy evidence. Abnormalities seen on barium enema generally require colonoscopy. CT colonography is another alternative to colonoscopy.22 CT has the advantage of being less invasive than colonoscopy and does not have blind spots (for example lesions can be missed behind folds on colonoscopy). A recent meta-analysis of 49 comparative studies of 11,551 patients recently concluded that CT colonography yields comparative sensitivity to colonoscopy, and suggested CT should be the initial colorectal screening test.23 That said, the next change in bowel screening is likely to be the introduction of a single flexi-sigmoidoscopy at the age of 50, so radiology departments may be spared for the time being.

Back to Article Outline

Ovarian cancer 

This is one of the most difficult cancers for GPs.24 In part, this is simply a function of the anatomy, in that the ovaries cannot be palpated, other than when they are considerably enlarged. Furthermore, few women are aware of the symptoms of ovarian cancer (contrast this to breast, where almost every woman understands the possibility of cancer with a breast lump). The symptoms are also frequently vague and non-specific, such as abdominal pain.25 There is also a very steep worsening in mortality with advancing stage in ovarian cancer, making it all the more important to identify this cancer at an early stage if at all possible. It is for these reasons that ovarian cancer was chosen for enhanced GP access to investigation.

There is considerable debate as to the value of performing a Ca125 blood test in primary care as a form of triage in patients who may have an ovarian cancer. This test was the basis of the main UK screening trial – which is yet to report its final results.26 NICE have recently reviewed the subject, and have recommended initial Ca125 testing by GPs, with women having a value over 35IU/ml being investigated further by ultrasonography. One problem is that some ovarian cancers do not have a raised Ca125. Despite – or perhaps because of – the debate over the place of Ca125 testing, open-access trans-vaginal ultrasonography (TVUS) is being offered from April 2012. Thus ultrasound departments need to prepare for an increase in test requests.

As with the issue of abnormal CXRs in suspected lung cancer, departments need to agree protocols for aftercare if a lesion is seen. This pathway needs to ensure that the requesting clinician (in this case, the GP) knows of the ‘abnormality’ – as they requested the test it is their responsibility for informing the patient. A separate matter is the responsibility for requesting any further investigation. The purist view is that this is again the requesting clinician’s responsibility. If delivery of the scan results to the GP is certain to be both speedy and secure (and if it isn’t – it should be!) then perhaps it is best to leave responsibility for further testing to general practice. An alternative, pragmatic, view is that the ultrasonographer is much more experienced than the GP in knowing the likely significance of the abnormality so knows better which investigative direction to follow. Both of these viewpoints have some merit: the crucial thing is that one and one only is chosen. The communication of negative results to the patient is easier to decide upon: this is clearly the responsibility of the GP. Finally, there is always a slight diagnostic danger when a cancer test is negative – it is all too easy to consider investigation complete, when there may be another important non-malignant explanation for the symptoms; for example a new diagnosis of inflammatory bowel disease is made in one patient in fifty investigated for suspected colon cancer.27 This is likely to be less of a problem with TVUS because of its small field of view. However a small numbers of alternative diagnoses will be visible and must be sought.

Back to Article Outline

Brain tumours 

The need for improved investigation of possible brain tumours is demonstrated by the high proportion presenting as emergencies – up to half, depending on the data source. Again this reflects the diagnostic difficulty for GPs as headache is the commonest symptom of a brain tumour presented to primary care. Yet headache is also one of the commonest symptoms not of a brain tumour presented to primary care! Indeed, the relative rarity of brain tumours means that the risk of a brain tumour in a patient describing a new onset of headache to their GP is one in a thousand.28 Therefore GPs ask about for additional symptoms before considering scanning – the so-called ‘headache-plus’ symptoms. Visual symptoms, dyscoordination, or personality change all raise suspicions. Additionally, public fear of brain tumours is high, and there is considerable pressure on GPs to offer brain scanning in patients with headache. Indeed, one of the main outcomes of a negative scan is a reduction in GP attendances.29

Open-access imaging of the brain is not new: indeed a recent article evaluated the service, offering CT scanning for patients with chronic daily headache in primary care.19 Sixty scans (1.4%) out of 4404 taken yielded a probable cause of the headache. A further 401 (9.1%) had incidental abnormalities, many of which would have required repeat scanning over time to ensure they were indeed unimportant. Of the sixty with abnormalities regarded as causative of the headache, four were meningiomas, two metastases, two pituitary tumours and two colloid cysts: four other lesions led to surgery.

CT brain scanning delivers quite a high radiation dose: one in 8100 women aged 40 will develop cancer from a single CT brain scan – with some forms of CT scanning posing a 1 in 80 risk of causing a cancer.30, 31 Therefore, a GP requesting a brain CT scan must be aware of the risks of the investigation and weigh this against the probability of disease. This is particularly true for young patients where the radiosensitivity of organs is much higher than in the older patient. In the young patient with headache, MRI would be more appropriate than CT because of its avoidance of radiation and its probability of demonstrating alternative causative pathologies. Thus, enhanced open access for GP brain scanning is intended to use MRI scans primarily.

Back to Article Outline

Conclusion 

Overall, offering better access to cancer investigations to GPs is likely to identify some cancers earlier in their natural history. This should translate into some of these being curable which would not previously have been so. These patients are the obvious ‘winners’ from the new initiatives. The cost will be borne by patients who were sent for testing and in whom the test was negative: they will have had some anxiety, though generally the negative test alleviates this. The cost is also borne by radiology departments and eventually by the taxpayer. All these factors will need to be evaluated over the course of the next few years.

Back to Article Outline

Conflict of interest statement 

The authors report no conflicts of interest. WH is a member of the DH committee, the Cancer Diagnostics Advisory Board, which is tasked with provision of enhanced GP access to cancer diagnostics.

Back to Article Outline

References 

  1. Richards MA. The size of the prize for earlier diagnosis of cancer in England. British Journal of Cancer. 2009;101(S2):S125–S129
  2. Hamilton W, Lancashire R, Sharp D, Peters TJ, Cheng KK, Marshall T. The risk of colorectal cancer with symptoms at different ages and between the sexes: a case-control study. BMC Medicine. 2009;7:17
  3. Hamilton W, Round A, Sharp D, Peters T. Clinical features of colorectal cancer before diagnosis: a population-based case-control study. British Journal of Cancer. 2005;93:399–405
  4. Ford AC, Veldhuyzen van Zanten SJO, Rodgers CC, Talley NJ, Vakil N, Moayyedi P. Diagnostic utility of alarm features for colorectal cancer: systematic review and meta-analysis. Gut. 2008;57(11):1545–1553
  5. NICE . Referral guidelines for suspected cancer. London: NICE; 2005;
  6. Hamilton W, Peters TJ, Round A, Sharp D. What are the clinical features of lung cancer before the diagnosis is made? A population based case-control study. Thorax. 2005;60:1059–1065
  7. Lederle FA, Nichol KL, Parenti CM. Bronchoscopy to evaluate hemoptysis in older men with nonsuspicious chest roentgenograms. Chest. 1989;95:1043–1047
  8. O’Neil KM, Lazarus AA. Hemoptysis indications for bronchoscopy. Archives of Internal Medicine. 1991;151:171–174
  9. Kaminski J. Frequency and causes of hemoptysis and role of bronchoscopy in patients with normal chest roentgenogram hospitalized in the Department of Physiopneumonology Silesian Medical University in the years 1961–1996. Pneumonologia i Alergologia Polska. 2001;69:663–668
  10. Hirshberg B, Biran I, Glazer M, Kramer MR. Hemoptysis: etiology, evaluation, and outcome in a tertiary referral hospital. Chest. 1997;112:440–444
  11. Colice GL. Detecting lung cancer as a cause of hemoptysis in patients with a normal chest radiograph: bronchoscopy vs CT. Chest. 1997;111:877–884
  12. Nielsen LS, Svanegaard J, Wiggers P, Egeblad H. The yield of a diagnostic hospital dyspnoea clinic for the primary health care section. Journal of Internal Medicine. 2001;250:422–428
  13. Pratter MR, Curley FJ, Dubois J, Irwin RS. Cause and evaluation of chronic dyspnea in a pulmonary disease clinic. Archives of Internal Medicine. 1989;149:2277–2282
  14. DePaso WJ, Winterbauer RH, Lusk JA, Dreis DF, Springmeyer SC. Chronic dyspnea unexplained by history, physical examination, chest roentgenogram, and spirometry. Analysis of a seven-year experience. Chest. 1991;100:1293–1299
  15. Stapley S, Sharp D, Hamilton W. Negative chest X-rays in primary care patients with lung cancer. British Journal of General Practice. 2006;56:570–573
  16. Chapman BE, Yankelevitz DF, Henschke CI, Gur D. Lung cancer screening: simulations of effects of imperfect detection on temporal dynamics. Radiology. 2005;234:582–590
  17. The National Lung Screening Trial Research Team . Reduced lung-cancer mortality with low-dose computed tomographic screening. The New England Journal of Medicine. 2011;365:395–409
  18. Naidich DP, Marshall CH, Gribbin C, Arams RS, McCauley DI. Low-dose CT of the lungs: preliminary observations. Radiology. 1990;175(3):729–731
  19. Simpson GC, Forbes K, Teasdale E, Tyagi A, Santosh C. Impact of GP direct-access computerised tomography for the investigation of chronic daily headache. British Journal of General Practice. 2010;60(581):897–901
  20. Hamilton W. The CAPER studies: five case-control studies aimed at identifying and quantifying the risk of cancer in symptomatic primary care patients. British Journal of Cancer. 2009;101:S80–S86
  21. York Health Economics Consortium . Bowel cancer services: costs and benefits. London: Department of Health; 2007;
  22. Johnson CD, Chen M-H, Toledano AY, Heiken JP, Dachman A, Kuo MD, et al. Accuracy of CT colonography for detection of large adenomas and cancers. The New England Journal of Medicine. 2008;359(12):1207–1217
  23. Pickhardt PJ, Hassan C, Halligan S, Marmo R. Colorectal cancer: CT colonography and colonoscopy for detection—systematic review and meta-analysis. Radiology. 2011;259(2):393–405
  24. Hamilton W, Menon U. Ovarian cancer. British Medical Journal. 2009;339:b4650
  25. Bankhead C, Kehoe S, Austoker J. Symptoms associated with diagnosis of ovarian cancer: a systematic review. BJOG: An International Journal of Obstetrics and Gynaecology. 2005;112:857–865
  26. Menon U, Gentry-Maharaj A, Hallett R, Ryan A, Burnell M, Sharma A, et al. Sensitivity and specificity of multimodal and ultrasound screening for ovarian cancer, and stage distribution of detected cancers: results of the prevalence screen of the UK Collaborative Trial of Ovarian Cancer Screening (UKCTOCS). Lancet Oncology. 2009;10:327–340
  27. Mahadavan L, Loktionov A, Daniels IR, Shore A, Cotter D, Llewelyn AH, et al. Exfoliated colonocyte DNA levels and clinical features in the diagnosis of colorectal cancer: a cohort study in patients referred for investigation. Colorectal Disease. 2011;
  28. Hamilton W, Kernick D. Clinical features of primary brain tumours: a case-control study using electronic primary care records. British Journal of General Practice. 2007;57:695–699
  29. Hamilton W. The price of diagnosis. British Journal of General Practice. 2008;58:837–838
  30. Smith-Bindman R, Lipson J, Marcus R, Kim K-P, Mahesh M, Gould R, et al. Radiation dose associated with common computed tomography examinations and the associated lifetime attributable risk of cancer. Archives of Internal Medicine. 2009;169:2078–2086
  31. Smith-Bindman R. Is computed tomography safe?. The New England Journal of Medicine. 2010;363:1–4

PII: S1078-8174(11)00095-2

doi:10.1016/j.radi.2011.10.002

Radiography
Volume 18, Issue 1 , Pages 5-8, February 2012

Article Tools