Prostate cancer is the most common cancer in men, with around 1.1 million cases diagnosed and approximately 309,000 deaths in men worldwide in 2012 [
Carcinectomy and radiotherapy are the typical treatments for prostate cancer [
Pathologic staging prediction is very important because it provides physicians with optimal treatment and management strategies. For example, radical prostatectomy (RP), the surgical removal of the prostate gland, provides the best opportunity for cure when prostate cancer is localized and accurate prediction of the pathology stage can provide the most beneficial treatment approach [
Deep belief networks (DBN) are a deep learning technique and is an effective method for classification prediction [
In this paper, we propose a DBN-DS-based multiclassifier for pathologic stage prediction of prostate cancer. The proposed DBN-DS uses patient PSA level, Gleason score, and clinical T stage and three DBNs to predict the pathology stage by combining the predicted information from the classifier. The classifiers are created by learning data according to features. When output values are generated using each learned DBN classifier, the final predicted result is provided by stochastically calculating the predicted output from each DBN classifier using DS. This paper is organized as follows: Section
The study data comprised 6345 male patients extracted from the Korean Prostate Cancer Registry (KPCR) which is extended from Smart Prostate Cancer Data Base (SPCDB) at six tertiary medical centers in Korea [
A deep belief network (DBN) is a generative graphical model or a type of deep neural network composed of multiple layers of latent variables, with connections between the layers but not between the units within each layer. The DBN is composed of restricted Boltzmann machine (RBM) layers. The learning method in the DBN is done by configuring the visible layer and hidden layer 1 into a single RBM. The DBN is composed of multiple layers of RBMs [
In this study, we constructed a classifier for three input and two output variables to construct a multiclassifier, as shown in Figure
Multi DBN classifiers.
Dempster-Shafer (DS) is a mathematical theory that deals with the uncertainty and inaccuracy problems presented by Arthur Dempster and Glenn Shafer [
The DS expresses the degree of certainty as a section and sets mutually exclusive hypotheses such as probability. The set of objects is called the environment and is denoted by
Belief
The degree of trust depends on the reliability of the given evidence and on the overall environmental impact; the ratio of the degree is expressed by
The DS expresses the confidence measure for
In this study, three output data predicted from a multiclassifier were fused and calculated. The calculation process using DS shown in the figure as
As described above,
Next, the interval of the pass and fail of the evidential interval are summarized as
As described above, the evidential interval section is constructed for OCD and NOCD, and the higher probability value of OCD and NOCD was set as the final output value.
Uncertainty data processing is a critical issue in the data fusion process. The DS and the Bayesian methods were compared to deal with this uncertainty. Unlike Bayesian inference, DS can contribute different levels of information to each source. In addition, a popular approach to data fusion has been established; unlike the Bayesian method, reliability can be assigned to all subsets of a hypothetical group, making it possible to form distributions for all subsets [
The characteristics of the initial PSA variable in the OCD and NOCD groups are shown in Table
Summary of initial PSA by pathology stage (organ-confined or non-organ-confined disease) in 6345 patients with clinically localized prostate carcinoma.
Training set |
Validation set | |||
---|---|---|---|---|
OCD |
NOCD |
OCD |
NOCD | |
Initial PSA | ||||
Minimum | 4 | 4 | 4 | 4 |
Maximum | 160 | 440.60 | 81.13 | 164 |
Average | 9.535 (0.173) | 18.606 (0.622) | 9.377 (0.197) | 17.889 (0.653) |
The Gleason scores in the OCD and NOCD groups are shown in Table
Distribution of Gleason scores by pathology stage (organ-confined or non-organ-confined disease) in 6345 patients with clinically localized prostate carcinoma.
Gleason score | Training set |
Validation set | ||
---|---|---|---|---|
OCD (%) |
NOCD (%) |
OCD (%) |
NOCD (%) | |
3 | 3 (0.12) | 0 (0.00) | 0 (0.00) | 1 (0.11) |
4 | 5 (0.20) | 5 (0.33) | 6 (0.42) | 1 (0.11) |
5 | 6 (0.24) | 11 (0.73) | 8 (0.57) | 1 (0.11) |
6 | 1342 (54.16) | 378 (24.93) | 785 (55.52) | 235 (26.35) |
7 (3 + 4) | 565 (22.80) | 386 (25.46) | 306 (21.64) | 218 (24.44) |
7 (4 + 3) | 266 (10.73) | 277 (18.27) | 160 (11.32) | 159 (17.83) |
8 | 238 (9.60) | 326 (21.50) | 119 (6.42) | 174 (19.51) |
9 | 46 (1.88) | 147 (9.70) | 28 (1.98) | 95 (10.65) |
10 | 7 (0.28) | 31 (2.04) | 2 (0.14) | 8 (0.90) |
The clinical T stages in the OCD and NOCD groups are shown in Table
Distribution of clinical T stage by pathology stage (organ-confined disease and non-organ-confined disease) in 6345 patients with clinically localized prostate carcinoma.
Clinical T stage | Training set |
Validation set | ||
---|---|---|---|---|
OCD (%) |
NOCD (%) |
OCD (%) |
NOCD (%) | |
T1a | 9 (0.36) | 0 (0.00) | 3 (0.21) | 0 (0.00) |
T1b | 107 (4.32) | 49 (3.23) | 74 (5.23) | 18 (2.02) |
T1c | 988 (39.87) | 410 (27.04) | 556 (39.32) | 225 (25.22) |
T2a | 691 (27.89) | 380 (25.07) | 417 (29.49) | 241 (27.02) |
T2b | 278 (11.22) | 161 (10.62) | 151 (10.68) | 97 (10.87) |
T2c | 234 (9.44) | 224 (14.78) | 126 (8.91) | 127 (14.24) |
T3a | 150 (6.05) | 233 (15.37) | 66 (4.67) | 135 (15.13) |
T3b | 21 (0.85) | 104 (6.86) | 21 (1.49) | 49 (5.49) |
The proposed DBN and DS-based multiclassifier is shown in Figure
DBN-DS-based multiclassifier.
To evaluate the DBN-DS-based multiclassifier, the entire data set was divided into a 70% training set and a 30% testing set. The control groups included Decision Tree C4.5, naive Bayesian (NB), logistic regression (LR), back propagation (BP), support vector machine (SVM), random forest (RF), deep belief network, and Partin tables. The experiments compared the sensitivity, specificity, accuracy, and area under the curve (AUC) using confusion matrix [
Experimental results of all classification methods between the training and validation sets.
Training set | Validation set | |||||
---|---|---|---|---|---|---|
Sensitivity | Specificity | Accuracy | Sensitivity | Specificity | Accuracy | |
Partin table | 45.96% | 88.44% | 70.52% | 52.69% | 71.36% | 64.14% |
C4.5 | 64.46% | 91.32% | 80.46% | 56.61% | 85.22% | 74.15% |
NB | 64.46% | 93.30% | 81.64% | 58.86% | 93.78% | 80.27% |
LR | 60.65% | 92.16% | 79.42% | 57.29% | 85.64% | 74.67% |
BPN | 63.90% | 92.02% | 80.60% | 61.66% | 85.57% | 76.32% |
SVM | 52.13% | 89.21% | 74.35% | 52.13% | 84.87% | 72.20% |
RF | 57.37% | 86.43% | 74.86% | 56.73% | 70.93% | 65.44% |
DBN | 44.61 | 88.04 | 71.65% | 50.56% | 85.01% | 71.68% |
DBN-DS (proposed) | 65.13% | 94.29% | 82.60% | 61.77% | 93.56% | 81.27% |
In general, the results from a training set are better than those of a validation set because of differences in dataset volumes. Sensitivity was defined as the probability of correctly matching NOCD. Because NOCD has less data than OCD, it is difficult to match. The proposed method has a 61.77% improved performance compared to those of the other models. In other words, the probability of matching NOCD is very important because it is a prediction of the risk of the pathology stage. Specificity was defined as the probability of correctly matching OCD. NB had the highest specificity, with 93.78%, but its sensitivity was low. The proposed method showed 93.56% higher performance than those of the other models. The accuracy was defined as the probability of predicting both NOCD and OCD. The proposed model had the highest accuracy, at 81.27%. The AUCs are shown in Figure
ROC curve results of all classification methods using the validation set.
Results of a DBN-DS confusion matrix comparing the training and validation sets.
Variable | Training set | Validation set | |||||
---|---|---|---|---|---|---|---|
Sensitivity | Specificity | Accuracy | Sensitivity | Specificity | Accuracy | ||
DBN#1 | Initial PSA | 38.57% | 91.39% | 70.78% | 41.93% | 88.68% | 70.60% |
DBN#2 | Gleason score | 32.51% | 89.18% | 67.26% | 37.00% | 88.47% | 68.56% |
DBN#3 | Clinical T stage | 21.19% | 94.20% | 65.96% | 26.23% | 93.85% | 67.69% |
DBN#1, DBN#2 | Initial PSA, Gleason score | 41.48% | 93.71% | 73.50% | 41.48% | 93.00% | 73.07% |
DBN#1, DBN#3 | Initial PSA, Clinical T stage | 40.02% | 94.55% | 73.46% | 40.02% | 93.85% | 73.03% |
DBN#2, DBN#3 | Gleason score, Clinical T stage | 34.19% | 94.91% | 71.42% | 34.19% | 93.49% | 70.56% |
DBN#1, DBN#2, DBN#3 (proposed) | Initial PSA, Gleason score, Clinical T stage | 65.13% | 94.29% | 82.60% | 61.77% | 93.56% | 81.27% |
The ROC curve has the highest DBN-DS of 0.777. The error of all models was about 0.01, and the
Next, the DBN-DS was evaluated. The result of the confusion matrix for DBN-DS is shown in Table
Detailed ROC curve analysis results of all classification methods using the validation set.
AUC | 95% confidence interval | |||
---|---|---|---|---|
Lower bound | Upper bound | |||
Partin table | 0.620 ± 0.012 | 0.000 | 0.597 | 0.644 |
C4.5 | 0.709 ± 0.012 | 0.000 | 0.686 | 0.731 |
NB | 0.763 ± 0.011 | 0.000 | 0.741 | 0.785 |
LR | 0.715 ± 0.012 | 0.000 | 0.692 | 0.737 |
ANN | 0.736 ± 0.012 | 0.000 | 0.714 | 0.758 |
SVM | 0.685 ± 0.012 | 0.000 | 0.662 | 0.708 |
RF | 0.638 ± 0.012 | 0.000 | 0.615 | 0.662 |
DBN | 0.678 ± 0.012 | 0.000 | 0.655 | 0.701 |
DBN-DS | 0.777 ± 0.011 | 0.000 | 0.755 | 0.798 |
ROC curve results of DBN-DS using a validation set.
Detailed ROC curve result of DBN-DS using validation set.
AUC | 95% confidence interval | |||
---|---|---|---|---|
Lower bound | Upper bound | |||
DBN#1 | 0.653 ± 0.012 | 0.000 | 0.629 | 0.677 |
DBN#2 | 0.627 ± 0.012 | 0.000 | 0.603 | 0.651 |
DBN#3 | 0.600 ± 0.012 | 0.000 | 0.576 | 0.625 |
DBN#1, DBN#2 | 0.672 ± 0.012 | 0.000 | 0.649 | 0.696 |
DBN#1, DBN#3 | 0.669 ± 0.012 | 0.000 | 0.646 | 0.693 |
DBN#2, DBN#3 | 0.638 ± 0.012 | 0.000 | 0.614 | 0.663 |
DBN#1, DBN#2, DBN#3 (proposed) | 0.777 ± 0.011 | 0.000 | 0.755 | 0.798 |
Among the three variables, the initial PSA level had the highest prediction rate. The PSA level is closely related to pathologic stage and is the most important parameter in prostate cancer. Variables combined with PSA showed a high prediction rate. In other words, the reason for the high prediction rate was that the Gleason score and clinical T stage also affect the pathology. However, the combination of Gleason score and clinical T stage had a lower accuracy than that predicted by the initial PSA level alone. The two variables are uncertain because they are diagnosed according to the doctor’s experience. However, when combined with PSA level, the performance was much higher. In this study, we found that initial PSA was the most important predictor, and that the Gleason score and clinical T stage were also important predictors.
Prediction models for pathology staging of prostate cancer are based on clinical tests and can be used to predict the spread of cancer. It is possible to diagnose cancer more precisely at the postoperative, pathological stage and to determine the degree of metastasis of prostate cancer.
We proposed a DBN-DS-based multiclassifier approach to predict the pathologic stage of prostate cancer. The proposed method provides a predictive model to improve accuracy through deep learning and information fusion based on the relationship between data measured using clinical tests. The inputs include initial PSA level, Gleason scores, and clinical T stage variables. The output can be OCD or NOCD in pathological staging (pT). This approach was evaluated using an existing validated patient dataset that included 6345 patient records from the KPCR database, which collected data from six tertiary medical institutions.
The performance of the proposed DBN-DS was compared with that of the NB, LR, BPN, SVM, RF, DBN, and Partin tables. The results showed that the proposed DBN-DS had better sensitivity and accuracy than all other methods.
In a recent pathological staging methodology study, Cosma et al. [
Currently, the proposed DBN-DS method is implemented as a research tool. Once the clinical evaluation is completed, the proposed tool will be developed as an easy-to-use clinical decision support system that can be accessed by clinicians.
The authors declare that they have no conflicts of interest.
This work was supported by the National Research Foundation of Korea (NRF) grant funded by the Korean Government (NRF-2016R1A2B4015922).