Introduction
In the year 2000, the Department of Health developed UK National Guidelines for referring suspected head and neck cancer cases. Termed the two-week-wait pathway, in England, this means suspected cancer patients should be seen by a specialist within 14 days of the primary care referral, to aid early detection. The target in England is 93 per cent.1 The National Institute for Health and Care Excellence (NICE) generated referral guidelines in 2005, which were revised in 2015 and last revised in 2021 without further new changes.2 Despite this, the cancer pick-up rate from two-week-wait referrals is not significantly different from non-urgent referrals.Reference McKie, Ahmad, Fellows, Meikle, Stafford and Thomson3,Reference Hobson, Malla, Sinha, Kay and Ramamurthy4 Since the introduction of the guidelines, cancer pick-up rates range from 6.2 per cent to 10 per cent.Reference McKie, Ahmad, Fellows, Meikle, Stafford and Thomson3,Reference Tikka, Pracy and Paleri5
There is a lack of published data exploring the pathway after the referral.Reference Douglas, Carswell and Montgomery6 There is also a paucity of research evaluating the effect of the revision of NICE referral guidelines in 2015 regarding cancer detection rates and compliance of referrals with the guidelines. This study explored these factors in diagnosed cancer patients from the two-week-wait out-patient otolaryngology clinic.
Cancer patients should receive their first definitive treatment within 31 days of the diagnosis,1 measured from the date when the patient is informed of the diagnosis and a management strategy is agreed, ideally after the multidisciplinary team (MDT) meeting. The patients should start first definitive treatment of their cancer within 62 days of their original referral. The target for these outcomes in England is 85 per cent. This means there are only 31 days for the specialist to diagnose and stage the cancer.1,2
Materials and methods
Analysis was based on existing, anonymised data, and therefore did not require ethics committee approval.
A retrospective audit was conducted for the period from January to June 2018 in a tertiary centre for head and neck surgery. Referral symptoms were analysed for all ENT two-week-wait referral patients, and sensitivity, specificity and positive predictive values were calculated for each symptom. In cases where cancer was diagnosed, the site of primary cancer, time taken to reach a diagnosis and first treatment, and staging and histology of the cancer were reviewed. Statistical analysis was conducted using Microsoft Excel® spreadsheet software for the cancer pick-up rate, compliance with the NICE referral criteria (Table 1) and compliance with the national cancer wait targets. The symptoms of ‘sore throat’ and ‘odynophagia’ were applied inconsistently and often overlapped, and so were merged under the refined variable ‘sore throat’.
Table 1. NICE 2015 referral criteria guidelines for two-week-wait referral to head and neck specialist clinic
NICE = National Institute for Health and Care Excellence
Clinical findings and outcomes were recorded anonymously. The data were collected from electronic clinic letters, patient information software (Integrated Clinical Environment (‘ICE’)) and InfoFlex digital health software (for access to MDT meeting information). This was an audit of clinical practice and formal ethical approval was not required.
Results
Demographics
A total of 1107 patients were referred through the two-week-wait pathway, with an average of 184 (standard deviation = 14) monthly referrals. The national target for specialist review within two weeks of referral was met 98.8 per cent of the time, with patients being seen within 6 days on average. History of smoking and alcohol were documented in 74 per cent of specialist letters. Performance status was rarely documented at all.
The male-to-female ratio of referral patients was 2:3. The youngest referral patient was 15 years old, with only 0.8 per cent of referrals from the 15–20 years bracket. Most referrals fell in the age groups 40–60 years and 60–80 years, accounting for 37.4 per cent of referrals each. However, of diagnosed cancer patients, 51 per cent were in the 60–80 years age group, with an average age of 60 years.
Of all referred patients, 66 (6 per cent) were diagnosed with cancer. Thirty-three per cent of the cancer patients were female. The prevalence rates of different head and neck cancers are illustrated in Figure 1. This pie chart shows that the commonest descending order of head and neck cancers were oropharyngeal squamous cell carcinoma (SCC), lymphoma and laryngeal SCC. Meanwhile, the highest primary site were similar for oropharynx and neck nodes (unknown primary and lymphoma). One per cent of head and neck cancer was metastatic disease in Virchow's node from a hepatobiliary primary. The histology of head and neck cancer was SCC in 68 per cent, lymphoma in 20 per cent and adenocarcinoma in 12 per cent. Half of the adenocarcinomas were metastatic lung cancer with neck nodes, while the other half were from the cervical oesophagus.
Fig. 1. Incidence of cancer in different head and neck regions. SCC = squamous cell carcinoma
Compliance with referral criteria
Fifty-two per cent of the referrals were consistent with the NICE 2015 criteria. In the 48 per cent of non-compliant referrals, 38 per cent of cases had isolated globus sensation misinterpreted as sore throat. The remaining referrals represented either inappropriate use of the service or misunderstanding of the episodic history of the patient. The frequencies of all ‘red flag’ symptoms are illustrated in Table 2 (including original 2005 criteria, which were removed in the 2015 revision), along with the sensitivity, specificity and positive predictive values in relation to cancer.
Table 2. Relationship between diagnosed cancer cases and presenting symptoms
NICE = National Institute for Health and Care Excellence; GI = gastrointestinal; N/A = not applicable; PPV = positive predictive value
In those patients diagnosed with cancer, neck lump was the most common presentation. This was followed by persistent hoarseness and sore throat. One patient diagnosed with laryngeal cancer presented with mixed globus sensation, and two other red flag symptoms of persistent unilateral sore throat and voice changes. Neck lump had the highest positive predictive value for cancer (17.4 per cent), followed by oral swelling (10.2 per cent).
The analysis showed remarkable positive predictive values for those criteria removed during the 2015 guideline revision; specifically, 9.7 per cent for otalgia and 4 per cent for sore throat. Oral bleeding – a symptom which is excluded from both the 2005 and 2015 NICE guidance – had a similar positive predictive value (8 per cent) to dysphagia in this study.
The cancer pick-up rate was equal (0.45 per cent) for each symptom of sore throat, dysphagia and oral swelling, while it was 0.34 per cent (two patients) for otalgia. The term ‘oral swelling’ described asymmetrical tonsillar enlargement or a visible oropharyngeal lesion.
Oropharyngeal cancer had the largest diversity in presentations (Figure 2), with the commonest being a cervical lymph node at level II (59 per cent).
Fig. 2. Oropharyngeal squamous carcinoma presentation symptoms.
Timing – compliance with national targets
The average time from tissue diagnosis to the MDT meeting was 14 days. The average time from MDT decision to the beginning of definitive treatment was 21 days. The MDT discussed 105 cancer cases during this six-month period for all head and neck subspecialties (including maxillofacial and thyroid). Sixty-three patients (60 per cent) were referred through the two-week-wait pathway, with 28 referrals (40 per cent) complying with the 62-day target. Forty-two patients (40 per cent) were referred through a routine pathway, with 13 referrals (30 per cent) compliant with the 62-day first definitive treatment target. The chi-square value was 0.86 and the p-value was 0.35 (non-significant at p < 0.05). The Fisher's exact test value was 0.45, and the p-value was not significant at p < 0.05.
From ENT two-week-wait referrals, there was 64 per cent and 52 per cent compliance with the 31-day and 62-day targets for cancer management, respectively. The mean time scales were 32 and 72 days respectively, while the medians were 27 and 67 days respectively. Of the 52 per cent of patients (n = 33) where the 62-day treatment target was missed, 55 per cent had a long wait between diagnosis and treatment, while 45 per cent had their lag during the investigations conducted to determine the diagnosis. The 104 days’ waiting target was breached in five patients (7.5 per cent).
Seventeen per cent of patients with diagnosed head and neck cancer (n = 11) died during this period. All mortality cases had advanced disease on presentation, except one with human papillomavirus positive, early oropharyngeal cancer, which progressed despite chemoradiation treatment.
Discussion
Epidemiology
Head and neck cancer is the 15th most common cause of cancer death in the UK, accounting for 2 per cent of all cancer deaths in 2017. However, over the last decade, head and neck cancer mortality rates have increased by around 14 per cent in the UK.1
Referral symptoms
Tikka et al. showed the commonest referral symptoms compliant with NICE guidelines were neck lump and hoarseness, with frequencies of 22.5 per cent and 20.2 per cent respectively, and positive predictive values of 17 per cent and 7.8 per cent respectively.Reference Tikka, Pracy and Paleri5 McKie et al. reported a slightly higher frequency for oral swelling (20 per cent) than for neck lump (19 per cent) and hoarseness (7 per cent); their study patients were from the maxillofacial surgery department, which one would expect to attract more referrals with oral cavity pathology than laryngeal pathology.Reference McKie, Ahmad, Fellows, Meikle, Stafford and Thomson3 Our study showed a similar positive predictive value for dysphagia (8 per cent) as in McKie and colleagues’ study (7.8 per cent), whereas Tikka and colleagues’ cohort study reported a much higher value (18.2 per cent).Reference McKie, Ahmad, Fellows, Meikle, Stafford and Thomson3,Reference Tikka, Pracy and Paleri5
Neck lump in our study had the highest positive predictive value for cancer (17.4 per cent), followed by oral swelling (10.2 per cent). A recent retrospective study conducted in London also found the highest positive predictive values for oral swelling and neck lump. The most frequent reasons for referral were persistent hoarseness, neck lump, and unexplained throat discomfort or pain;Reference Gao, Qin, Freeman, Oskooee and Hughes7 these were also the most common reasons for referral in our study, but with a different order of frequency.
In this study, patients were risk-stratified after their first visit; those who were neither discharged nor scheduled for biopsy were classified as medium risk.Reference Gao, Qin, Freeman, Oskooee and Hughes7 Of this group, 32 per cent were diagnosed with cancer, suggesting that they should remain on the two-week-wait pathway. The classification seemed more of academic value.
Tikka et al. showed similar figures for commonest presentations in a recent cohort study.Reference Tikka, Kavanagh, Lowit, Jiafeng, Burns and Nixon8 They emphasised the high predictive value of combining symptoms in the form of a risk calculator. They included unintentional weight loss and sore throat in an updated risk calculator.Reference Tikka, Kavanagh, Lowit, Jiafeng, Burns and Nixon8 Their findings are consistent with our study, though we have not evaluated weight loss, as it is a non-specific symptom associated with a wide range of diseases other than head and neck cancer.
Assessment of 2015 referral criteria update
Most primary care providers use a tick box form displaying the NICE 2015 criteria, for direct referral of suspicious cases to secondary care.Reference Duvvi, Thomas, Vijayanand and Reddy9 Several studies have shown the cancer pick-up rate to be similar for two-week-wait referrals and routine referrals,Reference McKie, Ahmad, Fellows, Meikle, Stafford and Thomson3,Reference Pracy10 which is supported by our finding of no significant difference between the two.
In the 2015 update, NICE removed the following criteria: persistent sore throat, unexplained tooth mobility not associated with periodontal disease for more than three weeks, unilateral unexplained pain in the head and neck area for more than four weeks, and unexplained persistent otalgia with normal otoscopy findings. Furthermore, a duration was not specified for the symptom of persistent hoarseness. Difficulty swallowing for three weeks was moved to the gastrointestinal two-week-wait referral criteria.2 Patients with ‘high’ dysphagia may have hypopharyngeal malignancy, which can be missed during flexible oesophagoscopy or can delay diagnosis if not initially referred to ENT.Reference Hobson, Malla, Sinha, Kay and Ramamurthy4
The positive predictive values reported in a cohort study were 5.9 per cent and 18.8 per cent for unexplained persistent sore throat and odynophagia respectively.Reference Tikka, Pracy and Paleri5 The authors advocate that the revised NICE 2015 referral criteria can miss 3 per cent of early head and neck cancer diagnoses. Our study demonstrated similar significant positive predictive values for sore throat (4 per cent) and otalgia (9.7 per cent). A recent study also agreed with the high positive predictive value for persistent sore throat (9.5 per cent).Reference Allam and Nijim11 The authors stated the need to refine the NICE criteria in the next update.Reference Allam and Nijim11 A recent study conducted in London showed a decrease in cancer detection rate, from 10 per cent to 3.76 per cent, with implementation of the 2015 NICE update; this could point to missing cases that are later diagnosed during routine referral.Reference Gao, Qin, Freeman, Oskooee and Hughes7 However, the authors did not compare cancer pick-up rate with routine referrals. They also recommended a review of the referral scheme in the current guidelines.Reference Gao, Qin, Freeman, Oskooee and Hughes7 Tikka and colleagues’ study from Scotland showed a similar failure of the NICE 2015 guidelines to improve the early detection rate in the UK.Reference Tikka, Kavanagh, Lowit, Jiafeng, Burns and Nixon8
Globus sensation as a sole symptom had no association with head and neck cancer.Reference Tikka, Pracy and Paleri5 When globus sensation co-existed with other symptoms such as otalgia or oral bleeding, the risk of head and neck cancer increased by 4 and 5.6 times respectively.Reference Tikka, Pracy and Paleri5 None of these symptoms on their own are an indication for specialist referral according to NICE guidelines;2 however, the combination would likely still benefit from referral. The same principle was witnessed in our study: dysphagia had a positive predictive value of 8 per cent, but when combined with sore throat this generated a positive predictive value of 12 per cent.
In light of the above evidence, we recommend a review of the NICE 2015 referral criteria for the addition of oral bleeding, and unexplained persistent unilateral sore throat or earache. The duration of persistent hoarseness also needs to be clarified.
• Head and neck cancer diagnosis from the two-week head and neck referral pathway is very low (6 per cent), only 1.5 times higher than routine referrals
• There is no significant difference between urgent and routine referrals in compliance with the 62-day treatment target for diagnosed cancer patients
• Neck lump is the most common reason for referral with the highest positive predicate value for cancer
• Oropharyngeal cancer is the commonest head and neck malignancy, which has variable presentations
• Investigation for metastatic head and neck cancer of an unknown primary or other multidisciplinary team involvement could delay target treatment
• The 2015 National Institute for Health and Care Excellence update would benefit from further review
Conclusion
The cancer pick-up rate from two-week-wait referrals is only 1.5 times higher than that from routine referrals. Neck lump is the most common reason for referral with the highest positive predicate value for cancer. Oropharyngeal cancer is the commonest head and neck malignancy, which has a wide range of presentations. The red flag symptoms given in the 2015 NICE update would benefit from further review for the inclusion of unilateral sore throat, unilateral otalgia and oral bleeding.
Data availability statement
The data used to support the findings of this study are available from the corresponding author upon request.
Competing interests
None declared
