Modern Principles of Early Diagnosis of Prostate Cancer

 

N.G. Kulchenko¹, A.A. Kostin², S.M.Chibisov¹, I.Z. Eremina¹, A.O. Tolkachev², S.P. Syatkin¹,

A.S. Skorik¹, G.I. Myandina¹, M.L. Blagonravov¹, V.I. Kuznetsov¹

¹Federal State Budgetary Institution of Higher Professional Education,

“People’s Friendship University of Russia”, Mikloukho-Macklay Street, 6, Moscow, 117198

²Federal State Budgetary Institution “National Medical Research Radiological Centre” of the Ministry of Health of the Russian Federation

 

ABSTRACT:

Among the main methods of prostate cancer diagnosing, such as finger rectal examination, prostate-specific antigen test, ultrasound examination of prostate, none of them possesses enough specificity and sensitivity. In our research, we used special statistic and mathematic processing (multifactor analysis and binary logic regression) of main prostate cancer symptoms, which allows to suspect prostate cancer with 97% accuracy and also to reduce the frequency of unnecessary biopsies of prostate.

 

 

 

INTRODUCTION:

Prostate cancer (PCa) takes the second place in the structure of oncological diseases in developed countries [2, 4]. Among the main methods of prostate cancer diagnosing, such as finger rectal examination, prostate-specific antigen test, ultrasound examination of prostate, none of them possesses enough specificity and sensitivity. [1, 3, 5].

 

The purpose of the research:

To improve the results of the early diagnosis of prostate cancer at the outpatient stage.

 

MATERIALS AND METHODS:

We conducted a retrospective analysis of prostate cancer risk factors among 122 men. All patients were divided into two groups: main - 59 men with confirmed prostate cancer (48.3%), and control - 63 patients with benign prostatic hyperplasia (BPH) (51.7%). Primarily all data of patients in both groups were subjected to a comparative analysis of great number risk factors of prostate cancer.

 

The material processing was carried out using the standard package: Statistics 6.0 (comparison of independent samples was carried out according to Student's test; qualitative characteristics were defined using contingency tables of Pearson, the probability of erroneous conclusions was defined by the level of significance p ≤ 0.05).

 

Next, we carried out statistical analysis:

1. Using the program "Common models of discriminant analysis, GDA», i.e. multivariate analysis;

2. Using binary logistic regression.

 

RESULTS AND DISCUSSION:

We compared the patients from both groups by the following factors: the duration of disease, the nature of the complaints (early weight loss, aversion to the food), anthropometric parameters (p> 0,05). Then, using multivariate analysis, we have identified more than 30 the most significant signs and risk factors of prostate cancer (age, height, weight, bad habits, complaints of the muscle weakness, testosterone level, hemoglobin level, the level of prostate-specific antigen (PSA), comorbidities (an inflammatory disease of the kidneys), blood group, Rh - factor, the nature of hair distribution, finger rectal examination, ultrasound examination of prostate, etc.). According to received data, we defined 5 the most informative signs, on which general practice doctors and urologists had to give attention, alertness and determinate the risk of developing prostate cancer at pre-hospital stage:

1. inflammatory diseases of the kidneys

2. inflammatory diseases of the bladder

3. the increase of blood PSA

4. blood group

5. the excessive weight of the patient

The comparison of PSA level in both groups of patients is shown in Table 1.         

Comparative characteristics of patients in both groups, according to ultrasound examination parameters (prostate volume), is shown in Figure 1.

 

 

Table 1. The comparison of PSA level in both groups of patients

 

Figure 1. Comparative characteristics of patients in both groups, according to ultrasound examination parameters (prostate volume).

 

Taking into account that among the signs of prostate cancer, included in the research, there may be a non-linear relationship, binary logistic regression was used to forecast the development of prostate cancer. With the help of statistical analysis - binary logistic regression, we determined the probability, that one or another case (the patient) on the basis of certain characteristics, came to the particular target group. The most informative risk factors for prostate cancer (symptoms), included in the logistic regression were the following: infectious diseases of the urinary tract, the PSA level, the level of creatinine, muscle weakness.

 

For the final prognosis of the prostate cancer risk probability in individual patient, we offer to use the following formula:

P = 1/1 + eΖ,

Where:

- e is a constant value, equal to 2.71;

- for the given patient from one of the analyzed groups classification (this is always the group with dependent variable, having a large code, in this case 1-PCa, and in adenoma group this code is accordingly 0);

- Z value must be calculated using the additional formula:

 

Ζ=66,0-21,64 х Пр1 -1,0 х Пр2-0,479 х Пр3 - 3,79 х Пр4,

 

Where:

1 -  infectious deseases

2  - PSA level

3- creatinine level

4- muscle weakness (these signs were calculated using binary logistic regression).

 

Using this mathematical calculation, having minimal results of clinical examination, one can assume the probability of prostate cancer in a patient with an accuracy approaching to 97.5%.

 

Having such an experience and widespread use of the method of multivariate analysis and logistic regression, the patient with PSA 4-9 ng/l could be avoided of "unnecessary biopsies".

 

Therefore, the use of discriminant analysis and logistic regression allowed to classify patients and to form a risk group of prostate cancer at the stage of screening, with sufficient accuracy.

 

CONCLUSIONS:

Binary logistic regression (mathematical treatment of clinical signs of the disease) in patients with suspected prostate cancer has a high specificity (97.5%). This method of statistical processing allows to select from a variety of symptoms the particular, to perform a biopsy of the prostate gland and thus, reduce the frequency of unnecessary biopsies and also to identify the groups of patients with a high or low risk of prostate cancer, that in turn reduces the cost of diagnosis and treatment of the patients.

 

REFERENCES:

1.        Hori S., Blanchet J.S., McLoughlin J. From prostate-specific antigen (PSA) to precursor PSA (proPSA) isoforms: a review of the emerging role of proPSAs in the detection and management of early prostate cancer // BJU Int.- 2013; 112(6): Pp. 17-28

2.        La Rochelle J., Amling C.L. Prostate cancer screening: what we have learned from the PLCO and ERSPC trials //Curr. Urol. Rep.  - 2010; 11 (3): Pp. 198-201.

3.        Mousavi S.R., Sadeghi-Naini A., Czarnota G.J., Samani A. Towards clinical prostate ultrasound elastography using full inversion approach // Med. Phys. - 2014; 41(3): Pp. 33-51.

4.        Oranusi C.K., Ugezu A.I., Nwofor A. Diagnosis of prostate cancer with needle biopsy: should all cases be biopsied before treatment? // Niger J.Clin.Pract.-2012; 15(1): Pp. 48-50.

5.        Tang P., Chen H., Uhlman M., Lin Y.R., Deng X.R., Wang B., Yang W.J., Xie K.J. A nomogram based on age, prostate-specific antigen level, prostate volume and digital rectal examination for predicting risk of prostate cancer // Asian J. Androl. - 2013; 14(1): Pp. 129–133.

 

 

Accepted on 27.01.2017           © RJPT All right reserved