The use of genetic programming in the analysis of quantitative gene expression profiles for identification of nodal status in bladder cancer

BMC Cancer BioMed Central Research article Open Access The use of genetic programming in the analysis of quantitative gene expression profiles for identification of nodal status in bladder cancer Anirban P Mitra†1, Arpit A Almal†2, Ben George3, David W Fry2, Peter F Lenehan2, Vincenzo Pagliarulo4, Richard J Cote1, Ram H Datar*1 and William P Worzel2 Address: 1Department of Pathology, University of Southern California Keck School of Medicine, 2011 Zonal Avenue, HMR 312, Los Angeles CA 90033, USA, 2Genetics Squared Inc., 210 South 5th Avenue, Suite A, Ann Arbor MI 48104, USA, 3Department of Internal Medicine, Gundersen Lutheran Medical Center, 1900 South Avenue, La Crosse WI 54601, USA and 4Dipartimento Emergenza e Trapianti d'Organo, Sezione di Urologia, Università di Bari, Piazza G. Cesare 11, Bari 70124, Italy Email: Anirban P Mitra - ; Arpit A Almal - ; Ben George - ; David W Fry - ; Peter F Lenehan - ; Vincenzo Pagliarulo - ; Richard J Cote - ; Ram H Datar* - ; William P Worzel - * Corresponding author †Equal contributors Published: 16 June 2006 BMC Cancer 2006, 6:159 doi:10.1186/1471-2407-6-159 Received: 09 February 2006 Accepted: 16 June 2006 This article is available from: http://www.biomedcentral.com/1471-2407/6/159 © 2006 Mitra et al; licensee BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. Abstract Background: Previous studies on bladder cancer have shown nodal involvement to be an independent indicator of prognosis and survival. This study aimed at developing an objective method for detection of nodal metastasis from molecular profiles of primary urothelial carcinoma tissues. Methods: The study included primary bladder tumor tissues from 60 patients across different stages and 5 control tissues of normal urothelium. The entire cohort was divided into training and validation sets comprised of node positive and node negative subjects. Quantitative expression profiling was performed for a panel of 70 genes using standardized competitive RT-PCR and the expression values of the training set samples were run through an iterative machine learning process called genetic programming that employed an N-fold cross validation technique to generate classifier rules of limited complexity. These were then used in a voting algorithm to classify the validation set samples into those associated with or without nodal metastasis. Results: The generated classifier rules using 70 genes demonstrated 81% accuracy on the validation set when compared to the pathological nodal status. The rules showed a strong predilection for ICAM1, MAP2K6 and KDR resulting in gene expression motifs that cumulatively suggested a pattern ICAM1>MAP2K6>KDR for node positive cases. Additionally, the motifs showed CDK8 to be lower relative to ICAM1, and ANXA5 to be relatively high by itself in node positive tumors. Rules generated using only ICAM1, MAP2K6 and KDR were comparably robust, with a single representative rule producing an accuracy of 90% when used by itself on the validation set, suggesting a crucial role for these genes in nodal metastasis. Conclusion: Our study demonstrates the use of standardized quantitative gene expression values from primary bladder tumor tissues as inputs in a genetic programming system to generate classifier rules for determining the nodal status. Our method also suggests the involvement of ICAM1, MAP2K6, KDR, CDK8 and ANXA5 in unique mathematical combinations in the progression towards nodal positivity. Further studies are needed to identify more class-specific signatures and confirm the role of these genes in the evolution of nodal metastasis in bladder cancer. Page 1 of 16 (page number not for citation purposes) BMC Cancer 2006, 6:159 Background Cancer of the urinary bladder is the seventh most common cancer worldwide (3.2% of all cancers), with an estimated annual incidence of 330,000 new cases and to which 179,000 deaths are attributed each year [1,2]. In the USA, where more than 63,000 new cases of bladder cancer were expected in 2005, urothelial carcinoma (UC) is the most common histology (90%), followed by squamous cell carcinoma (6–8%), adenocarcinoma (2%), and a variety of other rare tumor types [3]. The standard TNM clinical stage classification system for bladder cancer recommended by the American Joint Committee on Cancer takes into account the depth of invasion of the bladder wall by the primary tumor (T), the presence and size of metastatic regional lymph nodes (N), and the presence or absence of distant metastases (M) [4]. Nodal involvement is considered to be an independent risk factor for recurrence and survival after cystectomy for organ-confined bladder cancer [5]. Consequently, extensive bilateral pelvic lymphadenectomy is now considered an integral part of the surgery, having been shown to significantly improve the prognosis of patients with muscularis propria-invasive bladder cancer [6,7]. Non-muscularis propria-invasive tumors (TNM Stages 0a, 0is, and I), confined to the bladder mucosa or subepithelial connective tissue (pTa, pTis, and pT1) without regional (N0) or distant (M0) metastases, are generally treated by transurethral resection of the tumor with fulguration, intravesical chemotherapy, and radiotherapy. Although cures are possible, up to 80% of these presumed "localized" tumors will eventually recur following initial resection, with up to 25% progressing to muscularis propria-invasive disease [8]. The confirmation of the existing true nodal status in a patient with bladder cancer thus assumes primary importance, along with the need to determine if the tumor has the molecular potential to metastasize to the lymph nodes later, provided undiagnosed micrometastasis has not occurred already. Molecular changes in bladder cancer have been shown to precede morphologic changes that can be identified visually [9]. Further, some tumors have specific molecular patterns that predispose them to be more morphologically aggressive, with a greater propensity to metastasize and recur, regardless of their clinical stage at diagnosis [10]. Hence, morphologic changes need to be complemented with molecular correlates for an accurate prediction of bladder tumor progression. The goal of this study was to create an objective and accurate tool for the identification of nodal status from primary tumor tissue. Since bladder cancer has a multifactorial etiology with a complex pathogenesis encompassing various pathways that involve more than a simple two directional (up/down) regulation of a few genes, we http://www.biomedcentral.com/1471-2407/6/159 felt that it was necessary to investigate a comprehensive panel of genes to define this complex disease. Utilizing bladder tissue biopsies from 60 primary UC subjec (...truncated)