Expert system classifier for adaptive radiation therapy in prostate cancer

A classifier-based expert system was developed to compare delivered and planned radiation therapy in prostate cancer patients. Its aim is to automatically identify patients that can benefit from an adaptive treatment strategy. The study predominantly addresses dosimetric uncertainties and critical issues caused by motion of hollow organs. 1200 MVCT images of 38 prostate adenocarcinoma cases were analyzed. An automatic daily re-contouring of structures (i.e. rectum, bladder and femoral heads), rigid/deformable registration and dose warping was carried out to simulate dose and volume variations during therapy. Support vector machine, K-means clustering algorithms and similarity index analysis were used to create an unsupervised predictive tool to detect incorrect setup and/or morphological changes as a consequence of inadequate patient preparation due to stochastic physiological changes, supporting clinical decision-making. After training on a dataset that was considered sufficiently dosimetrically stable, the system identified two equally sized macro clusters with distinctly different volumetric and dosimetric baseline properties and defined thresholds for these two clusters. Application to the test cohort resulted in 25% of the patients located outside the two macro clusters thresholds and which were therefore suspected to be dosimetrically unstable. In these patients, over the treatment course, mean volumetric changes of 30 and 40% for rectum and bladder were detected which possibly represents values justifying adjustment of patient preparation, frequent re-planning or a plan-of-the-day strategy. Based on our research, by combining daily IGRT images with rigid/deformable registration and dose warping, it is possible to apply a machine learning approach to the clinical setting obtaining useful information for a decision regarding an individualized adaptive strategy. Especially for treatments influenced by the movement of hollow organs, this could reduce inadequate treatments and possibly reduce toxicity, thereby increasing overall RT efficacy.

[1]  Yoshua Bengio,et al.  Exploring Strategies for Training Deep Neural Networks , 2009, J. Mach. Learn. Res..

[2]  J M Balter,et al.  A comparison of ventilatory prostate movement in four treatment positions. , 2000, International journal of radiation oncology, biology, physics.

[3]  K. Brock.  Results of a multi-institution deformable registration accuracy study (MIDRAS). , 2010, International journal of radiation oncology, biology, physics.

[4]  Yoshua Bengio,et al.  Understanding the difficulty of training deep feedforward neural networks , 2010, AISTATS.

[5]  F. Fazio,et al.  Evidence of limited motion of the prostate by carefully emptying the rectum as assessed by daily MVCT image guidance with helical tomotherapy. , 2008, International journal of radiation oncology, biology, physics.

[6]  Arjan Bel,et al.  Evaluation of delivered dose for a clinical daily adaptive plan selection strategy for bladder cancer radiotherapy. , 2015, Radiotherapy and oncology : journal of the European Society for Therapeutic Radiology and Oncology.

[7]  K. Brock,et al.  A magnetic resonance imaging study of prostate deformation relative to implanted gold fiducial markers. , 2007, International journal of radiation oncology, biology, physics.

[8]  K Lam,et al.  Measurement of prostate movement over the course of routine radiotherapy using implanted markers. , 1995, International journal of radiation oncology, biology, physics.

[9]  K. Otomo,et al.  Analysis of motion of the rectum during preoperative intensity modulated radiation therapy for rectal cancer using cone-beam computed tomography , 2015, Radiation oncology.

[10]  Pasi Fränti,et al.  Centroid index: Cluster level similarity measure , 2014, Pattern Recognit..

[11]  Stina Svensson,et al.  The ANACONDA algorithm for deformable image registration in radiotherapy. , 2014, Medical physics.

[12]  Bhudatt R Paliwal,et al.  The effect and stability of MVCT images on adaptive TomoTherapy , 2010, Journal of applied clinical medical physics.

[13]  Robert K. Colwell,et al.  Abundance‐Based Similarity Indices and Their Estimation When There Are Unseen Species in Samples , 2006, Biometrics.

[14]  Peter Ziegenhein,et al.  Fast CPU-based Monte Carlo simulation for radiotherapy dose calculation , 2015, Physics in medicine and biology.

[15]  Linyuan Lü,et al.  Empirical comparison of local structural similarity indices for collaborative-filtering-based recommender systems , 2010 .

[16]  A. Niemierko,et al.  Methodological issues in radiation dose-volume outcome analyses: summary of a joint AAPM/NIH workshop. , 2002, Medical physics.

[17]  M van Herk,et al.  Quantification of organ motion during conformal radiotherapy of the prostate by three dimensional image registration. , 1995, International journal of radiation oncology, biology, physics.

[18]  Lucila Ohno-Machado,et al.  A neural network-based similarity index for clustering DNA microarray data , 2003, Comput. Biol. Medicine.

[19]  Weiguo Lu,et al.  Evaluation of geometric changes of parotid glands during head and neck cancer radiotherapy using daily MVCT and automatic deformable registration. , 2008, Radiotherapy and Oncology.

[20]  Linyuan Lü,et al.  Similarity index based on local paths for link prediction of complex networks. , 2009, Physical review. E, Statistical, nonlinear, and soft matter physics.

[21]  Ke-Lin Du,et al.  Clustering: A neural network approach , 2010, Neural Networks.

[22]  G. Baldazzi,et al.  SU-E-J-96: Predictive Neural Network for Parotid Glands Deformation Using IGRT and Dose Warping Systems. , 2014, Medical physics.

[23]  C. Fiorino,et al.  Late fecal incontinence after high-dose radiotherapy for prostate cancer: better prediction using longitudinal definitions. , 2011, International journal of radiation oncology, biology, physics.

[24]  M. Zeng,et al.  Influence of tumor thrombus location on the outcome of external-beam radiation therapy in advanced hepatocellular carcinoma with macrovascular invasion. , 2012, International journal of radiation oncology, biology, physics.

[25]  C. Voyant,et al.  Dosimetric uncertainties related to the elasticity of bladder and rectal walls: adenocarcinoma of the prostate. , 2011, Cancer radiotherapie : journal de la Societe francaise de radiotherapie oncologique.

[26]  H. von der Maase,et al.  Impact of changes in bladder and rectal filling volume on organ motion and dose distribution of the bladder in radiotherapy for urinary bladder cancer. , 2004, International journal of radiation oncology, biology, physics.

[27]  M. Kattan,et al.  Is it time to tailor the prediction of radio-induced toxicity in prostate cancer patients? Building the first set of nomograms for late rectal syndrome. , 2012, International journal of radiation oncology, biology, physics.

[28]  Giuseppe Baldazzi,et al.  A support vector machine tool for adaptive tomotherapy treatments: Prediction of head and neck patients criticalities. , 2015, Physica medica : PM : an international journal devoted to the applications of physics to medicine and biology : official journal of the Italian Association of Biomedical Physics.

[29]  Yasumasa Nishimura,et al.  Volume and dosimetric changes and initial clinical experience of a two-step adaptive intensity modulated radiation therapy (IMRT) scheme for head and neck cancer. , 2013, Radiotherapy and oncology : journal of the European Society for Therapeutic Radiology and Oncology.

[30]  Ludvig Paul Muren,et al.  On the use of margins for geometrical uncertainties around the rectum in radiotherapy planning. , 2004, Radiotherapy and oncology : journal of the European Society for Therapeutic Radiology and Oncology.

[31]  Naoyuki Okamoto,et al.  Reduction of prostate motion by removal of gas in rectum during radiotherapy. , 2008, International journal of radiation oncology, biology, physics.

[32]  Robert K. Colwell,et al.  A new statistical approach for assessing similarity of species composition with incidence and abundance data , 2004 .

[33]  G. Rodrigues,et al.  Comparative analysis of image guidance in two institutions for prostate cancer patients. , 2014, Reports of practical oncology and radiotherapy : journal of Greatpoland Cancer Center in Poznan and Polish Society of Radiation Oncology.

[34]  T. Zhou,et al.  Effective and Efficient Similarity Index for Link Prediction of Complex Networks , 2009, 0905.3558.

[35]  Mehmet Engin,et al.  Early prostate cancer diagnosis by using artificial neural networks and support vector machines , 2009, Expert Syst. Appl..

[36]  Radhe Mohan,et al.  Quantification of volumetric and geometric changes occurring during fractionated radiotherapy for head-and-neck cancer using an integrated CT/linear accelerator system. , 2004, International journal of radiation oncology, biology, physics.

[37]  M. Adamczyk,et al.  Evaluation of combining bony anatomy and soft tissue position correction strategies for IMRT prostate cancer patients. , 2012, Reports of practical oncology and radiotherapy : journal of Greatpoland Cancer Center in Poznan and Polish Society of Radiation Oncology.

[38]  R. Boyce,et al.  Choosing the best similarity index when performing fuzzy set ordination on binary data , 2001 .

[39]  Michael I. Jordan,et al.  Learning with Mixtures of Trees , 2001, J. Mach. Learn. Res..

[40]  D. Nori,et al.  Influence of volumes of prostate, rectum, and bladder on treatment planning CT on interfraction prostate shifts during ultrasound image-guided IMRT. , 2009, Medical physics.

[41]  B. Paliwal,et al.  Feasibility study on effect and stability of adaptive radiotherapy on kilovoltage cone beam CT , 2011, Radiology and oncology.

[42]  S. Webb,et al.  Use of artificial neural networks to predict biological outcomes for patients receiving radical radiotherapy of the prostate. , 2004, Radiotherapy and oncology : journal of the European Society for Therapeutic Radiology and Oncology.

[43]  L. Muren,et al.  Organ motion, set-up variation and treatment margins in radical radiotherapy of urinary bladder cancer. , 2003, Radiotherapy and oncology : journal of the European Society for Therapeutic Radiology and Oncology.

[44]  Raj Bhatnagar,et al.  Similarity measures in formal concept analysis , 2011, Annals of Mathematics and Artificial Intelligence.

[45]  Ahmed N. Albatineh,et al.  Means and variances for a family of similarity indices used in cluster analysis , 2010 .

[46]  Akihiro Takemura,et al.  The use of positron emission tomography/computed tomography imaging in radiation therapy: a phantom study for setting internal target volume of biological target volume , 2015, Radiation oncology.

[47]  Igor Kononenko,et al.  Machine learning for medical diagnosis: history, state of the art and perspective , 2001, Artif. Intell. Medicine.

[48]  Wolfgang Birkfellner,et al.  High-performance GPU-based rendering for real-time, rigid 2D/3D-image registration and motion prediction in radiation oncology. , 2012, Zeitschrift fur medizinische Physik.

[49]  J Salk,et al.  Impact of the filling status of the bladder and rectum on their integral dose distribution and the movement of the uterus in the treatment planning of gynaecological cancer. , 1999, Radiotherapy and oncology : journal of the European Society for Therapeutic Radiology and Oncology.

[50]  Tom Vercauteren,et al.  Evaluation of deformable image coregistration in adaptive dose painting by numbers for head-and-neck cancer. , 2012, International journal of radiation oncology, biology, physics.