Sains Malaysiana 43(4)(2014): 567–573

A Hybrid Prognostic Model for Oral Cancer based on Clinicopathologic and Genomic Markers

(Model Hibrid untuk Prognosis Kanser Mulut berdasarkan kepada Penanda Klinikopathologi

dan Genomik)

SIOW-WEE CHANG¹²*, SAMEEM ABDUL KAREEM¹, AMIR FEISAL MERICAN ALJUNID MERICAN²& ROSNAH BINTI ZAIN³⁴

¹Department of Artificial Intelligence, Faculty of Computer Science and Information Technology

University of Malaya, 50603 Kuala Lumpur, Malaysia

²Bioinformatics Division, Institute of Biological Science, Faculty of Science, University of Malaya, 50603 Kuala Lumpur, Malaysia

³Department of Oral Pathology and Oral Medicine and Periodontology, Faculty of Dentistry, University of Malaya, 50603 Kuala Lumpur, Malaysia

⁴Oral Cancer Research and Coordinating Centre (OCRCC), Faculty of Dentistry, University of Malaya

50603 Kuala Lumpur, Malaysia

Diserahkan: 11 September 2012/Diterima: 22 Julai 2013

ABSTRACT

There are very few prognostic studies that combine both clinicopathologic and genomic data. Most of the studies use only clinicopathologic factors without taking into consideration the tumour biology and molecular information, while some studies use genomic markers or microarray information only without the clinicopathologic parameters. Thus, these studies may not be able to prognoses a patient effectively. Previous studies have shown that prognosis results are more accurate when using both clinicopathologic and genomic data. The objectives of this research were to apply hybrid artificial intelligent techniques in the prognosis of oral cancer based on the correlation of clinicopathologic and genomic markers and to prove that the prognosis is better with both markers. The proposed hybrid model consisting of two stages, where stage one with Relief F-GA feature selection method to find an optimal feature of subset and stage two with ANFIS classification to classify either the patients alive or dead after certain years of diagnosis. The proposed prognostic model was experimented on two groups of oral cancer dataset collected locally here in Malaysia, Group 1 with clinicopathologic markers only and Group 2 with both clinicopathologic and genomic markers. The results proved that the proposed model with optimum features selected is more accurate with the use of both clinicopathologic and genomic markers and outperformed the other methods of artificial neural network, support vector machine and logistic regression. This prognostic model is feasible to aid the clinicians in the decision support stage and to identify the high risk markers to better predict the survival rate for each oral cancer patient.

Keywords: ANFIS; clinicopathologic; genomic; oral cancer prognosis; Relief F-GA

ABSTRAK

Terdapat kurang kajian yang memaparkan penyelidikan prognostik yang menggabungkan kedua-dua klinikopatologi dan genomik. Kebanyakan kajian hanya menggunakan faktor klinikopatologi tanpa mengambil kira biologi tumor dan maklumat molekul, manakala beberapa kajian penyelidik yang lain menggunakan penanda genomik atau maklumat mikroarai sahaja tanpa menggunakan parameter klinikopatologi. Maka, kajian ini tidak dapat membuat prognosis pesakit dengan berkesan. Kajian terdahulu telah menunjukkan bahawa keputusan prognosis adalah lebih tepat dengan menggunakan kedua-dua klinikopatologi dan genomik. Tujuan utama kajian ini adalah untuk mengaplikasikan hibrid teknik kepintaran buatan dalam prognosis kanser mulut berdasarkan kepada korelasi penanda klinikopatologi dan genomik dan untuk membuktikan bahawa prognosis adalah lebih baik dengan kedua-dua penanda. Model hibrid yang dicadangkan terdiri daripada dua peringkat, dengan peringkat pertama terdiri daripada ReliefF-GA sebagai kaedah pemilihan untuk mencari ciri optimum subset dan peringkat dua dengan pengelasan ANFIS untuk mengelaskan sama ada pesakit hidup atau mati selepas beberapa tahun didiagnosis. Model ramalan prognostik yang dicadangkan telah diaplikasikan ke atas dua golongan dataset kanser mulut yang dikumpulkan di Malaysia, iaitu Kumpulan 1 dengan penanda klinikopatologi sahaja dan Kumpulan 2 dengan gabungan kedua-dua penanda klinikopatologi dan genomik. Keputusan yang didapati telah membuktikan bahawa model yang dicadangkan dengan ciri optimum yang dipilih adalah lebih tepat dengan kehadiran kedua-dua penanda klinikopatologi dan genomik dan mengatasi kaedah lain seperti rangkaian saraf buatan, mesin sokongan vektor dan regresi logistik. Model prognostik ini boleh dilaksanakan untuk memberi bantuan kepada pakar klinikal di peringkat membuat sokongan keputusan untuk mengenal pasti penanda risiko yang tinggi supaya dapat meramalkan kadar jangka hayat setiap pesakit kanser dengan lebih tepat.

Kata kunci: ANFIS; genomik; klinikopatologi; prognosis kanser mulut; Relief F-GA

RUJUKAN

Asakage, T., Yokose, T., Mukai, K., Tsugane, S., Tsubono, Y., Asai, M. & Ebihara, S. 1998. Tumor thickness predicts cervical metastasis in patients with stage I/II carcinoma of the tongue. Cancer 82: 1443-1448.

Catto, J.W.F., Abbod, M.F., Linkens, D.A. & Hamdy, F.C. 2006. Neuro-Fuzzy modeling: An accurate and interpretable method for predicting bladder cancer progression. The Journal of Urology 175: 474-479.

Dom, R.M., Abdul-Kareem, S., Abidin, B., Jallaludin, R.L.R., Cheong, S.C. & Zain, R.B. 2008. Oral cancer prediction model for Malaysian sample. Austral-Asian Journal of Cancer 7(4): 209-214.

Futschik, M.E., Sullivan, M., Reeve, A. & Kasabov, N. 2003. Prediction of clinical behaviour and treatment for cancers. Applied Bioinformatics 2(3 Suppl): S53-S58.

Gevaert, O., Smet, F.D., Timmerman, D., Moreau, D. & Moor, B.D. 2006. Predicting the prognosis of breast cancer by integrating clinical and microarray data with Bayesian networks. Bioinformatics 22(14): e184-e190.

Giacomarra, V., Tirelli, G., Papanikolla, L. & Bussani, R. 1999. Predictive factors of nodal metastases in oral cavity and oropharynx carcinomas. Laryngoscope 109: 795-799.

Jefferies, S. & Foulkes, W.D. 2001. Genetic mechanisms in squamous cell carcinoma of the head and neck. Oral Oncology 37: 115-126.

Kawazu, T., Kazuyuki, A., Yoshiura, K., Nakayama, E. & Kanda, S. 2003. Application of neural networks to the prediction of lymph node metastasis in oral cancer. Oral Radiology 2003(19): 137-142.

Kononenko, I. 1994. Estimating Attributes: Analysis and Extension of RELIEF. Paper read at ECML-94 Proceedings of the European conference on machine learning on Machine Learning.

Leite, I.C.G. & Koifman, S. 1998. Survival analysis in a sample of oral cancer patients at a reference hospital in Rio de Janeiro, Brazil. Oral Oncology 34(5): 347-352.

Li, H., Li, D., Zhang, C. & Nie, S. 2007. An application of machine learning in the criterion updating of diagnosis cancer. International Conference in Neural Networks and Brain 2005(1): 187-190.

Mehrotra, R. & Yadav, S. 2006. Oral squamous cell carcinoma: Etiology, pathogenesis and prognostic value of genomic alterations. Indian Journal of Cancer 43(2): 60-66.

Morton, R.P., Ferguson, C.M., Lambie, N.K. & Whitlock, R.M. 1994. Tumor thickness in early tongue cancer. Archives of Otolaryngology-Head & Neck Surgery 120: 717-720.

Muzio, L.L., Santarelli, A., Caltabiano, R., Rubini, C., Pieramici, T. & Trevisiol, L. 2005. p63 overexpression associates with poor prognosis in head and neck squamous cell carcinoma. Human Pathology 36: 187-194.

Oliveira, L.R., Ribeiro-Silve, A., Costa, J.P.O., Simoes, A.L., Di Matteo, M.A.S. & Zucoloto, S. 2008. Prognostic factors and survival analysis in a sample of oral squamous cell carcinoma patients. Oral Surgery, Oral Medicine, Oral Pathology, Oral Radiology, and Endodontology 106(5): 685-695.

Reichart, P.A. 2001. Identification of risk groups for oral precancer and cancer and preventive measures. Clinical Oral Investigations 5: 207-213.

Seker, H., Odetayo, M.O., Petrovic, D. & Naguib, R.N.G. 2003. A fuzzy logic based-method for prognostic decision making in breast and prostate cancers. IEEE Transactions on Information Technology in Biomedicine 7(2): 114-122.

Siow-Wee, C., Kareem, S.A., Kallarakkal, T.G., Merican, A.F., Abraham, M.T. & Zain, R.B. 2011. Feature selection methods for optimizing clinicopathologic input variables in oral cancer prognosis. Asia Pacific Journal of Cancer Prevention 12(10): 2659-2664.

Sun, Y., Goodison, S., Li, J., Liu, L. & Farmerie, W. 2007. Improved breast cancer prognosis through the combination of clinical and genetic markers. Bioinformatics 23(1): 30-37.

Thurfjell, N., Coates, P.J., Boldrup, L., Lindgren, B., Bäcklund, B., Uusitalo, T., Mahani, D., Dabelsteen, E., Dahlqvist, Å., Sjöström, B., Roos, G., Vojtesek, B., Nenutil, R. & Nylander, K. 2005. Function and importance of p63 in normal oral mucosa and squamous cell carcinoma of the head and neck. Current Research in Head and Neck Cancer. Molecular Pathways, Novel Therapeutic Targets, and Prognostic Factors, edited by Bier, H. Adv. Otorhinolaryngol. Basel, Karger 62: 49-57.

Walker, D.M., Boey, G. & McDonald, L.A. 2003. The pathology of oral cancer. Pathology 35(5): 376-383.

Williams, J.K., Carlson, G.W., Cohen, C., Derose, P.B., Hunter, S. & Jurkiewicz, M.J. 1994. Tumor angiogenesis as a prognostic factor in oral cavity tumors. The American Journal of Surgery 168: 373-380.

Zain, R.B. & Ghazali, N. 2001. A review of epidemiological studies of oral cancer and precancer in Malaysia. Annals of Dentistry University of Malaya 8: 50-56.

*Pengarang untuk surat-menyurat; email: siowwee@um.edu.my