Keywords: Cancer Diagnosis In Primary Care, early cancer detection, general practice, family medicine, cancer symptoms, referral guidelines, risk assessment, diagnostic delays
INTRODUCTION
Cancer remains a leading cause of mortality in developed nations, accounting for approximately a quarter of all deaths. Despite advancements in treatment, early diagnosis continues to be paramount in improving patient outcomes. A significant proportion of cancer cases are initially identified in primary care settings, often through symptom presentation, even for cancers with established screening programs. General Practitioners (GPs) are therefore at the forefront of cancer detection, navigating the complex landscape of differentiating malignant symptoms from benign conditions. This crucial role necessitates a deep understanding of cancer presentation in primary care and effective strategies for timely diagnosis and referral.
While the diagnosis of new cancer cases is relatively infrequent for individual GPs – an average UK GP might encounter one new cancer diagnosis per month – managing patients presenting with symptoms that could indicate cancer is a daily occurrence. This constant vigilance hones a GP’s skill in recognizing what is not cancer, building experience over time in identifying genuine malignancies. However, the current system often disproportionately focuses on diagnostic failures in confirmed cancer cases, overlooking the considerable expertise GPs demonstrate in effectively ruling out cancer in the vast majority of patients.
Over the past decade, advancements in diagnostic pathways, particularly the introduction of rapid investigation clinics offering specialist assessment within two weeks in the UK, theoretically should have streamlined cancer diagnosis. National referral guidelines exist to guide patient selection for these expedited pathways. Furthermore, access to primary care imaging has improved. Yet, the UK’s cancer mortality rates remain comparatively high, raising questions about the tangible impact of these improvements. Some argue that these advancements have yielded minimal improvement in overall cancer outcomes, with estimates suggesting that 7,500 to 10,000 lives are lost annually in the UK due to delayed diagnosis. However, this figure encompasses various aspects of diagnostic delay, as illustrated in Figure 1.
Figure 1. Intervals in Cancer Diagnostic Delays
Much of the existing research has focused on the total interval from symptom onset to treatment initiation (A–E in Figure 1). However, to effectively improve early cancer diagnosis in primary care, it is crucial to dissect this overall delay and pinpoint specific areas for intervention. This review will concentrate on the diagnostic process between point B (first presentation to primary care) and point D (referral for definitive investigation), using examples primarily from the four most prevalent cancers in the UK: breast, prostate, lung, and colorectal cancer, which have been the focus of extensive recent research.
In retrospect, patients can often recall the onset of their symptoms and their initial consultation with primary care. However, at the time of presentation, discerning malignancy from benign conditions can be challenging for clinicians. Interestingly, evidence suggests patients themselves are adept at filtering symptoms, often correctly attributing them to benign causes and not seeking medical attention. For instance, only a fraction of individuals experiencing rectal bleeding, a potential cancer symptom, actually consult their GP. This highlights a dual selection process: first, the patient’s decision to seek medical advice, and second, the GP’s decision to refer for specialist investigation. Across both stages, the likelihood of cancer is understandably higher in those who are selected for further investigation, indicating the effectiveness, albeit imperfect, of these selection processes. Notably, the relative increase in cancer risk following the patient’s decision to consult is more substantial than the risk increase after GP referral decisions. For example, the positive predictive value (PPV) of colorectal cancer associated with a single episode of rectal bleeding in the general population is around 0.1%. However, once reported to primary care, this risk escalates to approximately 2–3%, and further increases to 5–7% in patients referred to secondary care.
The Role of Primary Care as Gatekeeper
Primary care acts as a crucial ‘gatekeeper’ in cancer diagnosis, selecting patients who require specialist investigation or referral. Concerns have been raised that the UK’s relatively poorer cancer outcomes may stem from GPs having an overly high threshold for referral. This is juxtaposed with criticisms from specialist clinics regarding GP adherence to referral guidelines, which ironically are often acknowledged to have limited accuracy in identifying cancer. This creates a challenging situation for GPs, who must balance patient expectations with responsible resource allocation. This review aims to shed light on recent advancements in primary cancer diagnosis, particularly emphasizing risk-based approaches to referral, to aid GPs in this delicate balancing act.
Breast Cancer Diagnosis in Primary Care
Despite its significance and emotional weight, primary care-focused research to guide GPs in breast symptom referrals remains surprisingly limited. While extensive research exists on breast cancer screening and genetics, paradoxically, more studies examine GP compliance with referral guidelines than the evidence underpinning these guidelines. It’s crucial to remember that screening programs do not detect the majority of breast cancers; approximately three-quarters are diagnosed following symptom presentation in primary care.
Breast symptoms are a common presentation in primary care, accounting for nearly 3% of female consultations. Breast pain and palpable masses are the most frequently reported concerns, particularly among women aged 25–44. Nipple discharge or skin changes are less common. A breast lump carries a likelihood ratio (LR) of 15 for malignancy, whereas nipple complaints have an LR of 3.1, and breast pain an LR of 1.8. Overall, about 8% of women presenting to primary care with a breast lump are subsequently diagnosed with cancer. This risk is strongly age-dependent; breast cancer incidence is 15 times higher in women aged 45–64 compared to those under 25. Consequently, the cancer risk associated with a breast lump in women under 25 is approximately 0.5%. Nipple complaints have a cancer risk of less than 2%, and breast pain even lower at 0.9%. These figures, published after the development of NICE (National Institute for Health and Care Excellence) guidelines in the UK, support the concept of age-based thresholds for urgent breast lump referrals. The prominence given to nipple bleeding in referral guidelines might be questionable, given the low proportion of women with this symptom who have underlying cancer. Similarly, isolated breast pain is a very low-risk symptom.
Box 1. Key Metrics in Diagnostic Research
Likelihood Ratio (LR): This ratio quantifies how much more likely a symptom is to be present in patients with cancer compared to those without cancer. For example, if 20% of lung cancer patients report haemoptysis to their doctor, while only 1.5% of healthy individuals do, the positive likelihood ratio is 20/1.5 = 13.
Positive Predictive Value (PPV): This represents the probability that a patient actually has cancer when they present with a specific symptom. Using the same haemoptysis example, a person reporting haemoptysis to their GP has a 2.4% chance of it being due to lung cancer.
These metrics are interconnected. The PPV is influenced by both the LR of the symptom and the underlying incidence of the cancer in the population. A higher LR and a higher cancer incidence increase the PPV. Conversely, rarer cancers will generally have lower PPVs for individual symptoms. It’s important to consider the timeframe for symptom relevance. A cough experienced a year before lung cancer diagnosis is unlikely to be causally related.
In younger women, GPs heavily rely on clinical acumen to differentiate suspicious breast lumps. However, referral is often necessary for definitive reassurance. Younger women are at a higher risk of inflammatory breast cancer, a more aggressive subtype accounting for about 5% of all breast cancers. Inflammatory breast cancer presents with rapid onset of swelling, redness, and warmth of the breast, often mimicking infection, making differentiation challenging. ‘Peau d’orange’ skin changes are frequently observed. The underlying pathology involves lymphatic vessel blockage by cancer cells, contributing to a poorer prognosis.
Lung Cancer Diagnosis: Overcoming Nihilism and Improving Early Detection
Lung cancer diagnosis is often associated with a sense of nihilism, partly due to its historically poor prognosis and debates surrounding the benefits of early detection. This perspective can be shared by patients, sometimes stemming from feelings of guilt related to smoking, potentially delaying symptom presentation. By the time symptoms are reported, lung cancer is often advanced, limiting curative treatment options. Furthermore, lung cancer symptoms are frequently non-specific, making early identification challenging even when patients present promptly. These factors contribute to the low proportion of UK lung cancer patients eligible for surgical resection, and subsequently, lower survival rates.
While lung cancer screening with conventional sputum cytology and plain chest X-ray has proven ineffective, spiral CT screening shows promise in detecting smaller nodules. However, spiral CT screening is associated with a high false-positive rate, creating challenges in implementation. The Lung-SEARCH trial in the UK is investigating annual cytology augmented by computer-assisted image analysis in high-risk smokers with COPD. Large-scale spiral CT screening trials are ongoing, but early results from smaller studies have been less encouraging due to high false-positive rates.
Given the often lengthy period between symptom onset and presentation in lung cancer, efforts to enhance symptom awareness and encourage earlier consultation are crucial. Public health campaigns aimed at raising awareness of lung cancer symptoms, coupled with streamlined access to chest X-rays, have shown moderate success in shifting diagnoses towards earlier stages and increasing resectability rates. A campaign in Doncaster, UK, encouraging individuals with coughs to consult their GP, alongside relaxed chest X-ray criteria, resulted in a significant increase in stage I or II lung cancer diagnoses. Such initiatives, focusing on symptom awareness and accessible diagnostics, may offer more tangible progress in improving lung cancer outcomes than broad population screening.
Lung cancer diagnosis benefits from the relatively straightforward nature of chest X-ray, a readily available, cost-effective, and reasonably accurate initial investigation. While chest X-rays can be negative in a proportion of lung cancer cases, and some lesions may be missed, the false-negative rate is comparatively low for a primary care cancer test. False-positive chest X-rays are also relatively infrequent due to the size threshold for reliable nodule detection. Therefore, a low threshold for requesting chest X-rays in patients with potential lung cancer symptoms is justified. Expanding chest X-ray access for symptomatic individuals, particularly smokers over 60 presenting with cough, could expedite diagnosis in a significant proportion of lung cancer cases, potentially improving mortality. The number needed to X-ray to expedite one diagnosis by 9 months in this high-risk group is estimated to be 47, which may represent an efficient use of healthcare resources compared to population screening.
For patients presenting to primary care with symptoms suggestive of lung cancer, recent evidence informs investigation and referral decisions. Haemoptysis is a key symptom, carrying a significant risk of lung cancer, warranting a chest X-ray. While less common than other symptoms, haemoptysis has a higher PPV. Other symptoms like cough, dyspnea, weight loss, fatigue, and chest pain, though more prevalent in lung cancer, are also common in benign conditions, resulting in lower PPVs. However, in smokers and individuals over 70, the PPVs for these symptoms increase, potentially justifying chest X-ray investigation even at these lower risk levels, given the accessibility and low risk of the test. Despite the non-specific nature of many lung cancer symptoms, GP delays in referral remain a concern in some regions, highlighting the need for continued efforts to optimize diagnostic pathways and reduce delays.
Colorectal Cancer: Addressing Referral Guidance and Low-Risk Symptoms
Colorectal cancer diagnosis in primary care has been extensively researched, with systematic reviews and primary care-focused studies providing valuable insights. Rectal bleeding has been a particular focus, with studies consistently showing a significant cancer risk, especially in older individuals. Current UK referral guidelines for rectal bleeding, requiring a 6-week duration or co-occurrence with diarrhea for urgent referral, are arguably flawed. The inherent risk associated with rectal bleeding alone in individuals over 60, irrespective of accompanying symptoms, warrants investigation.
Referral guidelines for iron-deficiency anemia (IDA) also require re-evaluation. The current haemoglobin thresholds for urgent colorectal cancer investigation in the UK (10 g/dL for women and 11 g/dL for men) correspond to substantial colorectal cancer risks. Patients with even milder degrees of IDA, particularly in the absence of other clear causes, should be considered for colorectal cancer investigation, especially given that IDA is associated with poorer colorectal cancer prognosis.
Managing patients with low-risk but not negligible risk symptoms, such as abdominal pain, constipation, or diarrhea, poses a greater diagnostic challenge. These symptoms have lower PPVs for colorectal cancer. While colonoscopy remains the gold standard for investigation, it is more invasive and resource-intensive than chest X-ray for lung cancer. CT colonography, initially hoped to replace colonoscopy, has limitations in detecting smaller lesions. Emerging biomarkers, such as matrix metalloproteinase–9 and free DNA in rectal samples, hold future promise but require further research before clinical application in primary care. Scoring systems like CAPER and SELVA scores have been developed to risk-stratify patients with low-risk symptoms, but achieving high sensitivity without sacrificing specificity remains a challenge, potentially leading to over-investigation.
In the absence of readily available and highly accurate biomarkers or risk scores, GPs rely on subtle clinical cues and experience to identify colorectal cancer in patients with low-risk symptoms. The term “change in bowel habit,” while broadly defined as constipation or diarrhea, carries a nuanced clinical meaning for GPs, suggesting a higher suspicion of colorectal cancer. GPs often accurately identify patients with a “change in bowel habit” who warrant investigation, even when the symptoms themselves might seem non-specific. The subtle features that enable GPs to discern higher-risk individuals within the broader group of patients with low-risk symptoms remain poorly understood, but their existence is evident in clinical practice. This likely involves a subconscious Bayesian approach, integrating prior probability of cancer with subtle clinical findings to refine risk assessment.
Prostate Cancer: Symptom-Based Diagnosis and PSA in Primary Care
Similar to lung cancer, prostate cancer diagnosis has faced nihilistic perspectives, albeit from a different angle. The argument suggests that early lung cancer diagnosis is futile due to poor prognosis, while early prostate cancer diagnosis is unnecessary as many men will live long lives regardless. This nihilism is partly fueled by prostate cancer screening trials showing limited mortality benefit. While treatment of small, screen-detected prostate cancers may offer a marginal survival advantage, it comes at the cost of potential complications. However, the benefit of treating larger, symptomatic prostate cancers is less controversial. This distinction simplifies the primary care approach. Given the established benefit of treating symptomatic prostate cancer, investigating for prostate cancer in men presenting with lower urinary tract symptoms (LUTS) is appropriate. However, LUTS are non-specific, indicating prostate enlargement without differentiating benign prostatic hyperplasia (BPH) from malignancy. Common LUTS have PPVs for prostate cancer around 3%. Erectile dysfunction can also be an early symptom, carrying a similar risk. Therefore, the clinical management of men with LUTS should include digital rectal examination (DRE) and prostate-specific antigen (PSA) testing. PSA in this context is used diagnostically in symptomatic men, distinct from its use as a population screening tool. While some men with prostate cancer may have low PSA levels, these cancers often have a favorable prognosis, and missing them is of less clinical concern. Urologists sometimes use PSA velocity or free PSA levels to refine biopsy selection, but evidence supporting these measures in primary care is lacking.
Ovarian Cancer: Challenging the “Silent Killer” Myth
Recent primary care-based studies have illuminated ovarian cancer, challenging the long-held “silent killer” myth. These studies demonstrate that symptoms are common and are typically reported to GPs. Ovarian cancer is more accurately described as a “noisy killer,” but the symptoms are often non-specific, including fatigue, abdominal pain, and urinary frequency. Abdominal distension, despite being omitted from current guidelines, carries a relatively higher risk of ovarian cancer. A key challenge is that GPs may not readily consider ovarian cancer in the differential diagnosis for these non-specific symptoms, or underestimate its likelihood. Consequently, ovarian cancer is often perceived as a cancer frequently missed by GPs. However, evidence suggests that GPs are likely no worse (or better) at diagnosing ovarian cancer compared to other internal malignancies. Serum CA125 testing is available in primary care, but its sensitivity and overall value in this setting are debated. Transvaginal ultrasound is a more promising diagnostic tool, extensively evaluated in screening studies with good performance characteristics, although extrapolating screening data to symptomatic populations requires caution. Large-scale screening trials are underway, but definitive mortality data is still pending.
Other Cancers and the Role of Clinical Judgement
Research on primary care diagnosis of less common cancers is limited due to their rarity. Referral decisions for these cancers often rely heavily on GP experience and intuition. Existing limited evidence generally supports current clinical practice. For example, dysphagia carries a significant risk of esophageal cancer, and hematuria a risk of urinary tract malignancy, justifying investigation. GPs appear proficient in identifying malignant cervical lymphadenopathy and, importantly, in ruling out malignancy in most cases of lymph node enlargement. The risk of brain tumor in patients presenting with headache is very low, supporting the current approach of not routinely scanning patients with typical headaches.
CONCLUSION
Primary care is the cornerstone of cancer diagnosis, where initial suspicion arises and pre-symptomatic risk assessment takes place. Cancer as a diagnostic possibility is a near-daily consideration in general practice. The evidence base informing patient selection for referral, and equally importantly, for reassurance and non-investigation, is continuously expanding. While evidence sometimes challenges existing guidelines, GPs often appropriately rely on their clinical judgment. Most cancer patients receive effective primary care services. Future research should focus on strategies to identify and expedite diagnosis for the minority of patients who currently experience diagnostic delays, further optimizing cancer outcomes.
Acknowledgments
The author acknowledges Peter Wood for his critical review and helpful suggestions on an earlier draft.
Linked commentary
This article has an associated commentary: Neal RD. Cancer diagnosis — the role of urgent referral guidelines. Br J Gen Pract 2010; DOI: 10.3399/bjgp10X483427.
Funding
No specific funding was received for this work. William Hamilton is supported by a NIHR postdoctoral fellowship. The views expressed are those of the author and not necessarily those of the Department of Health.
Competing interests
William Hamilton has received funding for research studies from various sources, including NIHR School of Primary Care Research, the Department of Health, CRUK, and Colonix Ltd. None of these funding sources influenced the content of this review.
Discuss this article
Contribute to the discussion on this article at the Discussion Forum: http://www.rcgp.org.uk/bjgp-discuss
