Skip to main content Accessibility help

Patient classification of two-week wait referrals for suspected head and neck cancer: a machine learning approach

  • J W Moor (a1), V Paleri (a2) and J Edwards (a3)



Machine learning algorithms could potentially be used to classify patients referred on the two-week wait pathway for suspected head and neck cancer. Patients could be classified into ‘predicted cancer’ or ‘predicted non-cancer’ groups.


A variety of machine learning algorithms were assessed using the clinical data of 5082 patients. These patients had previously been referred via the two-week wait pathway for suspected head and neck cancer to two separate tertiary referral centres in the UK. Outcomes from machine learning classification were analysed in comparison to known clinical diagnoses.


Variational logistic regression was the most clinically useful technique of those chosen to perform the analysis and patient classification; the proportion of patients correctly classified as having ‘non-cancer’ was 25.8 per cent, with a false negative rate of 1 out of 1000.


Machine learning algorithms can accurately and effectively classify patients referred with suspected head and neck cancer symptoms.


Corresponding author

Author for correspondence: Mr James Moor, ENT Dept, Leeds General Infirmary, Leeds LS1 3EX, UK E-mail:


Hide All

Mr J W Moor takes responsibility for the integrity of the content of the paper



Hide All
1MacKay, DJC. Information Theory, Inference, and Learning Algorithms. Cambridge: Cambridge University Press, 2003
2Clifton, DA. Machine Learning for Healthcare Technologies. London: Institution of Technology and Engineering, 2016
3Oxford Cancer Intelligence Unit. Profile of Head and Neck Cancers in England: Incidence, Mortality and Survival. In: [5 October 2016]
4NHS England. Achieving World Class Cancer Outcomes: Taking the strategy forward. In: [5 October 2016]
5NHS England Interim Management and Support. Delivering Cancer Waiting Times: A Good Practical Guide. In: [5 October 2016]
6Drinnan, M, Paleri, V, Kumar, R, Mehanna, H. Efficacy of the two week wait referral system for head and neck cancer: a systematic review. Ann R Coll Surg Engl 2012;94:101–5
7Statistics vs. Machine Learning, fight! In: [17 November 2016]
8Breiman, L. Statistical modeling: the two cultures. Statistical Science 2001;16:199215
9Tikka, T, Pracy, P, Paleri, V. Refining the head and neck cancer referral guidelines: a two-centre analysis of 4715 referrals. Clin Otolaryngol 2016;41:6675
10Webb, AR, Copsey, AD. Statistical Pattern Recognition. Chichester: Wiley, 2011
11Scikit-learn. In: [17 November 2016]
12Seone, J, Takkouche, B, Varela-Centelles, P, Tomas, I, Seoane-Romero, JM. Impact of delay in diagnosis on survival to head and neck carcinomas: a systematic review with meta-analysis. Clin Otolaryngol 2012;37:99106
13Hippersley-Cox, J, Coupland, C, Robson, J, Brindle, P. Derivation, validation, and evaluation of a new QRISK model to estimate lifetime risk of cardiovascular disease: cohort study using QResearch database. BMJ 2010;341:c6624
14QCancer. In: [17 November 2016]
16Taylor, RA, Pare, JR, Venkatesh, AK, Mowafi, H, Melnick, ER, Fleischman, W et al. Prediction of in-hospital mortality in emergency department patients with sepsis: a local big data-driven, machine learning approach. Acad Emerg Med 2016;23:269–78
17Gultepe, E, Green, JP, Nguyen, H, Adams, J, Albertson, T, Tagkopoulos, I. From vital signs to clinical outcomes for patients with sepsis: a machine learning basis for a clinical decision support system. J Am Med Inform Assoc 2014;21:315–25
18Klann, JG, Anand, V, Downs, SM. Patient-tailored prioritization for a pediatric care decision support system through machine learning. J Am Med Inform Assoc 2013;20:e26774


Related content

Powered by UNSILO

Patient classification of two-week wait referrals for suspected head and neck cancer: a machine learning approach

  • J W Moor (a1), V Paleri (a2) and J Edwards (a3)


Altmetric attention score

Full text views

Total number of HTML views: 0
Total number of PDF views: 0 *
Loading metrics...

Abstract views

Total abstract views: 0 *
Loading metrics...

* Views captured on Cambridge Core between <date>. This data will be updated every 24 hours.

Usage data cannot currently be displayed.