Evaluation of a Machine-Learning Algorithm for Treatment Planning in Prostate Low-Dose-Rate Brachytherapy.

PURPOSE This work presents the application of a machine learning (ML) algorithm to automatically generate high-quality, prostate low-dose-rate (LDR) brachytherapy treatment plans. The ML algorithm can mimic characteristics of preoperative treatment plans deemed clinically acceptable by brachytherapists. The planning efficiency, dosimetry, and quality (as assessed by experts) of preoperative plans generated with an ML planning approach was retrospectively evaluated in this study. METHODS AND MATERIALS Preimplantation and postimplantation treatment plans were extracted from 100 high-quality LDR treatments and stored within a training database. The ML training algorithm matches similar features from a new LDR case to those within the training database to rapidly obtain an initial seed distribution; plans were then further fine-tuned using stochastic optimization. Preimplantation treatment plans generated by the ML algorithm were compared with brachytherapist (BT) treatment plans in terms of planning time (Wilcoxon rank sum, α = 0.05) and dosimetry (1-way analysis of variance, α = 0.05). Qualitative preimplantation plan quality was evaluated by expert LDR radiation oncologists using a Likert scale questionnaire. RESULTS The average planning time for the ML approach was 0.84 ± 0.57 minutes, compared with 17.88 ± 8.76 minutes for the expert planner (P=.020). Preimplantation plans were dosimetrically equivalent to the BT plans; the average prostate V150% was 4% lower for ML plans (P=.002), although the difference was not clinically significant. Respondents ranked the ML-generated plans as equivalent to expert BT treatment plans in terms of target coverage, normal tissue avoidance, implant confidence, and the need for plan modifications. Respondents had difficulty differentiating between plans generated by a human or those generated by the ML algorithm. CONCLUSIONS Prostate LDR preimplantation treatment plans that have equivalent quality to plans created by brachytherapists can be rapidly generated using ML. The adoption of ML in the brachytherapy workflow is expected to improve LDR treatment plan uniformity while reducing planning time and resources.

[1]  C. Catton,et al.  Brachytherapy improves biochemical failure-free survival in low- and intermediate-risk prostate cancer compared with conventionally fractionated external beam radiation therapy: a propensity score matched analysis. , 2015, International journal of radiation oncology, biology, physics.

[2]  J. Williamson,et al.  Update of AAPM Task Group No. 43 Report: A revised AAPM protocol for brachytherapy dose calculations. , 2003, Medical physics.

[3]  Alberto de la Zerda Lerner Inverse planning for low‐dose‐rate prostate brachytherapy by simulated annealing under fuzzy expert control , 2004 .

[4]  Ljubomir J. Buturovic,et al.  Cross-validation pitfalls when selecting and assessing regression and classification models , 2014, Journal of Cheminformatics.

[5]  Joshua B. Tenenbaum,et al.  Human-level concept learning through probabilistic program induction , 2015, Science.

[6]  J. Williamson,et al.  Update of AAPM Task Group No. 43 Report: A revised AAPM protocol for brachytherapy dose calculations. , 2004 .

[7]  Jean Pouliot,et al.  Class solution for inversely planned permanent prostate implants to mimic an experienced dosimetrist. , 2006, Medical physics.

[8]  Saeed Hassanpour,et al.  Artificial Intelligence in Medicine , 2015 .

[9]  Nawaid Usmani,et al.  Canadian prostate brachytherapy in 2012. , 2013, Canadian Urological Association journal = Journal de l'Association des urologues du Canada.

[10]  J. Lacy,et al.  Salvage Brachytherapy for Biochemically Recurrent Prostate Cancer following Primary Brachytherapy , 2016, Prostate cancer.

[11]  D. Jude Hemanth,et al.  Computational Intelligence Techniques for Pattern Recognition in Biomedical Image Processing Applications , 2013 .

[12]  Kuldip K. Paliwal,et al.  Feature extraction and dimensionality reduction algorithms and their applications in vowel recognition , 2003, Pattern Recognit..

[13]  Hiroyuki Takahashi,et al.  Salvage partial brachytherapy for prostate cancer recurrence after primary brachytherapy , 2014, International journal of urology : official journal of the Japanese Urological Association.

[14]  Jean Pouliot,et al.  Class solution for inversely planned permanent prostate implants to mimic an experienced dosimetrist. , 2006, Medical physics.

[15]  Purang Abolmaesumi,et al.  Automatic Prostate Brachytherapy Preplanning Using Joint Sparse Analysis , 2015, MICCAI.

[16]  K. Becker,et al.  Analysis of microarray data using Z score transformation. , 2003, The Journal of molecular diagnostics : JMD.

[17]  B. Carey,et al.  Correlation between pre- and postimplant dosimetry for iodine-125 seed implants for localized prostate cancer. , 2009, International journal of radiation oncology, biology, physics.

[18]  Z. Nugent,et al.  The dosimetric quality of brachytherapy implants in patients with small prostate volume depends on the experience of the brachytherapy team. , 2010, Brachytherapy.

[19]  Rich Caruana,et al.  Distributed tuning of machine learning algorithms using MapReduce Clusters , 2011, LDMTA '11.

[20]  A. Dicker,et al.  Salvage of suboptimal prostate seed implantation: Reimplantation of underdosed region of prostate base. , 2005, Brachytherapy.

[21]  C. Kirisits,et al.  Review of clinical brachytherapy uncertainties: Analysis guidelines of GEC-ESTRO and the AAPM☆ , 2014, Radiotherapy and oncology : journal of the European Society for Therapeutic Radiology and Oncology.

[22]  B. Prestidge,et al.  Evaluation of hybrid inverse planning and optimization (HIPO) algorithm for optimization in real‐time, high‐dose‐rate (HDR) brachytherapy for prostate , 2013, Journal of applied clinical medical physics.

[23]  Adam P Dicker,et al.  Comparative analysis of prostate‐specific antigen free survival outcomes for patients with low, intermediate and high risk prostate cancer treatment by radical therapy. Results from the Prostate Cancer Results Study Group , 2012, BJU international.

[24]  Ingrid Spadinger,et al.  Sci—Fri AM(2): Brachy—09: Using Bayesian Networks for Prostate Brachytherapy Inverse Planning , 2009 .

[25]  Impact of post-implant dosimetric parameters on the quality of life of patients treated with low-dose rate brachytherapy for localised prostate cancer: results of a single-institution study , 2015, Radiation oncology.

[26]  Aggelos K. Katsaggelos,et al.  Spectral approximation to point set similarity metric , 2013, 2013 IEEE International Conference on Multimedia and Expo Workshops (ICMEW).

[27]  J. Pouliot,et al.  Optimization of permanent 125I prostate implants using fast simulated annealing. , 1996, International journal of radiation oncology, biology, physics.

[28]  Measurement uncertainty analysis of low-dose-rate prostate seed brachytherapy: post-implant dosimetry , 2015, Australasian Physical & Engineering Sciences in Medicine.

[29]  M. Lahanas,et al.  A hybrid evolutionary algorithm for multi-objective anatomy-based dose optimization in high-dose-rate brachytherapy. , 2003, Physics in medicine and biology.

[30]  C. Lee Giles,et al.  Active learning for class imbalance problem , 2007, SIGIR.

[31]  Joseph Bucci,et al.  Decline in urinary retention incidence in 805 patients after prostate brachytherapy: the effect of learning curve? , 2006, International journal of radiation oncology, biology, physics.

[32]  Y. Yamada,et al.  Predicting biochemical tumor control after brachytherapy for clinically localized prostate cancer: The Memorial Sloan-Kettering Cancer Center experience. , 2012, Brachytherapy.

[33]  J. Camp,et al.  Shape analysis of the prostate: establishing imaging specifications for the design of a transurethral imaging device for prostate brachytherapy guidance. , 2014, Brachytherapy.

[34]  Jason Weston,et al.  Natural Language Processing (Almost) from Scratch , 2011, J. Mach. Learn. Res..