Eur. Phys. J.

Plus (2018) 133: 92

DOI 10.1140/epjp/i2018-11950-y
THE EUROPEAN
PHYSICAL JOURNAL PLUS
Regular Article

Cancer treatment model with the Caputo-Fabrizio fractional

derivative

Mustafa Ali Dokuyucu1,a , Ercan Celik2,b , Hasan Bulut3,c , and Haci Mehmet Baskonus4,d
1
Department of Mathematics, Faculty of Art and Science, Agrı İbrahim Cecen Universtiy, Ağrı, Turkey
2
Department of Mathematics, Faculty of Science, Ataturk University, Erzurum, Turkey
3
Department of Mathematics, Faculty of Science, Fırat University, 23100, Elazığ, Turkey
4
Department of Computer Engineering, Faculty of Engineering, Munzur University, Tunceli, Turkey

Received: 17 January 2018 / Revised: 4 February 2018

Published online: 5 March 2018 – 
c Società Italiana di Fisica / Springer-Verlag 2018

Abstract. In this article, a model for cancer treatment is examined. The model is integrated into the
Caputo-Fabrizio fractional derivative first, to examine the existence of the solution. Then, the uniqueness
of the solution is investigated and we identified under which conditions the model provides a unique
solution.

1 Introduction

As is well known, cancer is one of the most important and most fatal diseases of the last century. A remedy has
been sought out in the last century for this malady with manifold varieties. Four different types of conventional cancer
treatments are commonly recognized today. Those treatment methods include surgery, chemotherapy, radiotherapy and
immunotherapy. In this article, we address the radiation therapy solely [1]. Radiotherapy is one of the most important
means to fight cancer. The goal is to kill the cancer cells by means of radiation. With this method, rapidly proliferating
cancer cells are targeted. However, the reproduction of the surrounding normal cells is adversely affected as well when
the cancer cells are annihilated. Therefore, several models have been developed previously providing concentrations
of the normal and cancerous cells [2]. Mathematical modeling is quite important to analyze real-world problems to
find better ways of understanding [3–10]. Furthermore, the modeling process provides important information on the
addressed process [11–13]. Some mathematical models have been developed previously for cancer treatment [2,14–
16]. Belostotski [2], provided a model representing the interaction between normal and cancerous cells. Radiation’s
entering into the body is of great importance and occurs in 4 different ways. By using the control theory, Belostotski
and Freedman [14] developed a model in which only the cancer cells were affected. Later, the authors revisited their
former model and developed a newer one, in which radiation affected normal cells only to a certain extent [1].
It is obvious that the information provided to the researchers about the process is as important as establishing
the model itself. In this article, we are going to apply the “cancer treatment by radiotherapy model” to the fractional
derivative of Caputo-Fabrizio first, in order to provide a better understanding and a solution of the process. Then,
we shall examine the existence of a rearranged model. We shall find the uniqueness of the solution under the initial
conditions.
This paper is organized as follows. In sect. 2, we give some fundamental definitions about fractional operators.
We present the cancer treatment model and integrate the model for the fractional operator. Then we try to find the
existence solution for the new model. Afterwards, we investigate the uniqueness solution using the fixed point theory
in sect. 3.

Focus Point on “Modelling Complex Real-World Problems with Fractal and New Trends of Fractional Differentiation” edited
by A. Atangana, Z. Hammouch, G. Mophou, K.M. Owolabi.
a
e-mail: [email protected]
b
e-mail: [email protected]
c
e-mail: [email protected] (corresponding author)
d
e-mail: [email protected]
Page 2 of 6 Eur. Phys. J. Plus (2018) 133: 92

2 Preliminaries
In this chapter some fundamental definitions [17–20] on the fractional derivative are given.

Definition 1. The fractional derivative defined by Caputo in 1967 is [19]:

 t
1 f (n) (r)
C ν
a Dt (f (t)) = dr, n − 1 < ν < n ∈ N.
Γ (n − ν) a (t − r)ν+1−n

Definition 2. The fractional integral defined by Riemann-Liouville is [20]:

 t
1
I ν f (t) = f (r)(t − r)ν−1 dr.
Γ (ν) a

Definition 3. The fractional derivative of order ν is defined by [17]

  
νM (ν) t
df (x) t−x
CF ν
a Dt (f (t)) = exp −ν dx.
1−ν a dx 1−ν

Here M (ν) is a normalization constant. Furthermore, M (0) = M (1) = 1. The definition is also written as follows:
 
νM (ν) t
t−x
CF ν
a Dt (f (t)) = [f (t) − f (x)] exp −ν dx.
1−ν a 1−ν

Remark 1. If η = 1−ν
ν ∈ (0, ∞), ν = 1
1+η =∈ [0, 1], then the above equation is supposed in the form
  
N (η) t
df (x) t−x
Dtη (f (t)) = exp − dx, N (0) = N (∞) = 1.
η a dx η

Furthermore,  
1 t−x
lim exp − = δ(x − t).
ν→0 ν ν
Note that, according to the above definition, the fractional integral of Caputo type of the function of order 0 < α < 1
is an average between function f and its integral of order one. This therefore imposes

2(1 − α) 2α
+ = 1.
(2 − α)M (α) (2 − α)M (α)

The above expression yields an explicit formula for

2
M (α) = , 0 ≤ α ≤ 1.
2−α
Because of the above, Nieto and Losada proposed that the new Caputo derivative of order 0 < α < 1 can be
reformulated as follows.

Definition 4. The fractional derivative of order ν is defined by [18]

  
1 t
 t−x
CF
Dν (f (t)) = f (x) exp −ν dx.
1−ν 0 1−ν

At this instant subsequent to the preface of the novel derivative, the connected anti-derivative turns out to be
imperative; the connected integral of the derivative was proposed by Nieto and Losada [18].

Definition 5. The Caputo-Fabrizio fractional definition rearranged by [18] is as follows:


2(1 − ν) 2ν t
CF ν
It (f (t)) = u(t) + u(s)ds, t ≥ 0.
2M (ν) − νM (ν) 2M (ν) − νM (ν) 0
Eur. Phys. J. Plus (2018) 133: 92 Page 3 of 6

3 Main results
It is well known that cancer is one of the most common diseases causing death in the last century. Emerging in various
parts of the human body, this disease becomes unresponsive to treatment when one intervenes late. There is a great
deal of research done for the treatment of cancer, which is the disease of the century. In this article, we are going
to examine the existence and uniqueness of the cancer treatment model. By developing models for the treatment of
cancer, we aim to contribute to the studies in this field.
It is assumed that healthy and cancer cells be located in the same area of the organism. Let ρ(t) denote the
concantration of healthy cells, and a(t) denote the concentration of cancer cells. Then the model is given by [1]
 
dρ(t) ρ
= α1 ρ 1 − − β1 ρa − D(t)ρ,
dt S1
 
da(t) a
= α2 a 1 − − β2 aρ − D(t)a, (1)
dt S2

where D(t) is the strategy of the radiotherapy. It is supposed that D(t) ≡ γ > 0, when t ∈ [nw, nw + L) (treatment
stage) and D(t) ≡ 0, when t ∈ [nw + L, (n + 1)w (no treatment stage), for all n = 0, 1, 2, . . ., where w is the radiation
treatment time. [1] The system (1) with the Caputo derivative is given as follows:
 
ρ
0 Dt (ρ(t)) = α1 ρ 1 − − β1 ρa − D(t)ρ,
CF ν
S1
 
a
0 Dt (a(t)) = α2 a 1 − − β2 aρ − D(t)a.
CF ν
(2)
S2

3.1 Existence of solutions

We will provide, in this section, the existence of the solutions for the cancer treatment model by radiotherapy. After
that, we also will present the uniqueness of the positive solutions.
Now, applying the fractional integral in eq. (2), we obtain the following:
   
2(1 − ν) ρ
ρ(t) − ρ0 (t) = α1 ρ 1 − − β1 ρa − D(t)ρ
2M (ν) − νM (ν) S1
 t   
2ν ρ
+ α1 ρ 1 − − β1 ρa − D(y)ρ dy,
2M (ν) − νM (ν) 0 S1
   
2(1 − ν) a
a(t) − a0 (t) = α2 a 1 − − β2 aρ − D(t)a
2M (ν) − νM (ν) S2
 t   
2ν a
+ α2 a 1 − − β2 aρ − D(y)a dy. (3)
2M (ν) − νM (ν) 0 S2

For simplicity, we choose our kernels as s(t, ρ(t)) and s(t, a(t)) as follows:
 
ρ
s(t, ρ(t)) = α1 ρ 1 − − β1 ρa − D(t)ρ,
S1
 
a
s(t, a(t)) = α2 a 1 − − β2 aρ − D(t)a.
S2

First we need to be able to identify an operator. We will then show that this operator is compact. So that the operator
T : H → H. Then we get
 t
2(1 − ν) 2ν
T ρ(t) = s(t, ρ(t)) + s(y, ρ(y))dy,
2M (ν) − νM (ν) 2M (ν) − νM (ν) 0
 t
2(1 − ν) 2ν
T a(t) = s(a, a(t)) + s(a, a(y))dy. (4)
2M (ν) − νM (ν) 2M (ν) − νM (ν) 0

Lemma 1. The mapping T : H → H is completely continuous.

Page 4 of 6 Eur. Phys. J. Plus (2018) 133: 92

Proof. Let M ⊂ H be bounded. There exists a constant l, m > 0, such that ρ < l and a < m. Let
L1 = max s(t, ρ(t)) and L2 = max s(t, a(t)).
0≤t≤1 0≤t≤1
0≤ρ≤l 0≤a≤m

∀ρ, a ∈ M , we have
  t 
 2 − 2ν 2ν 

|T ρ(t)| =  s(t, ρ(t)) + s(y, ρ(y))dy  ,
2M (ν) − νM (ν) 2M (ν) − νM (ν) 0
    t 
 2 − 2ν   2ν  
≤  
s(t, ρ(t)) +   s(y, ρ(y))dy ,
2M (ν) − νM (ν)  2M (ν) − νM (ν) 0
 
2 − 2ν 2ν
≤ + c1 |s(t, ρ(t))|,
2M (ν) − νM (ν) 2M (ν) − νM (ν)
 
2 − 2ν 2ν 
≤ + c1 L1 ,
2M (ν) − νM (ν) 2M (ν) − νM (ν)
2L1
T ρ ≤ [1 − ν + νc1 ] . (5)
2M (ν) − νM (ν)
Similarly,
  t 
 2 − 2ν 2ν 

|T a(t)| =  s(t, a(t)) + s(y, a(y))dy  ,
2M (ν) − νM (ν) 2M (ν) − νM (ν) 0
   t 
2 − 2ν   2ν  
≤  
s(t, a(t)) +   s(y, a(y))dy ,
2M (ν) − νM (ν)  2M (ν) − νM (ν) 0
 
2 − 2ν 2ν
≤ + c2 |s(t, a(t))|
2M (ν) − νM (ν) 2M (ν) − νM (ν)
 
2 − 2ν 2ν 
≤ + c2 L2
2M (ν) − νM (ν) 2M (ν) − νM (ν)
2L2
T a ≤ [1 − ν + νc2 ] . (6)
2M (ν) − νM (ν)
Hence T (M ) is bounded.
Now, in the following part, we will consider t1 < t2 and ρ(t), a(t) ∈ M and then, for a given > 0, if |t2 − t1 | < δ.
We have
 
 2 − 2ν 
T ρ(t2 ) − T ρ(t1 ) ≤  (s(t2 , ρ(t2 )) − s(t1 , ρ(t1 )))
2M (ν) − νM (ν)
  t2  t1 
 2ν 2ν 
+  s(y, ρ(y))dy − s(y, ρ(y))dy  ,
2M (ν) − νM (ν) 0 2M (ν) − νM (ν) 0
2 − 2ν 2ν
≤ |(s(t2 , ρ(t2 )) − s(t1 , ρ(t1 ))| + L1 |s(t2 , ρ(t2 )) − s(t1 , ρ(t1 ))|.
2M (ν) − νM (ν) 2M (ν) − νM (ν)
(7)
Now we will investigate the following:
   
 ρ(t2 ) − ρ(t1 ) 
|s(t2 , ρ(t2 )) − s(t1 , ρ(t1 ))| ≤ α1 (ρ(t2 ) − ρ(t1 )) 1 − − β1 (ρ(t2 ) − ρ(t1 )) a − γ (ρ(t2 ) − ρ(t1 ))
S1
≤ c3 |(ρ(t2 ) − ρ(t1 )| − c4 |(ρ(t2 ) − ρ(t1 )| − c5 |(ρ(t2 ) − ρ(t1 )|
≤ (c3 − c4 − c5 )|(ρ(t2 ) − ρ(t1 )|
≤ C|t2 − t1 |. (8)
Now putting eq. (8) and the integral part of eq. (7) in eq. (7), we get
2ν 2(1 − ν)
|T ρ(t2 ) − T ρ(t1 )| ≤ C|t2 − t1 | + L1 |t2 − t1 |, (9)
2M (ν) − νM (ν) 2M (ν) − νM (ν)

δ= 2(1−ν)
, (10)
2ν
2M (ν)−νM (ν) C + 2M (ν)−νM (ν) L1

such that |T ρ(t2 ) − T ρ(t1 )| ≤ .

Eur. Phys. J. Plus (2018) 133: 92 Page 5 of 6

When we can acquire the following for the function a with the same rules, we get

δ= 2(1−ν)
, (11)
2ν
2M (ν)−νM (ν) G + 2M (ν)−νM (ν) L2

such that |T a(t2 ) − T a(t1 )| ≤ are satisfied. Therefore T (M ) is equicontinuous, so that T (M ) is compact via the
Arzela-Ascoli theorem.
Theorem 1. Let N : [ρ1 , ρ2 ] × [0, ∞) → [0, ∞), then N (t, ·) is non-decreasing for each t in [ρ1 , ρ2 ]. There exist positive
constants, v1 and v2 , so that B(n)v1 ≤ S(t, v1 ), B(n)v2 ≥ N (t, v2 ), 0 ≤ v1 (t) ≤ v2 (t), ρ1 ≤ t ≤ ρ2 . Thus, the equation
has a positive solution.
Proof. We only need to consider the fixed point for the operator of T . We consider that T : H → H is completely
continuous. Let ρ1 ≤ ρ2 and a1 ≤ a2 . Here four variables are arbitrary:
  t
2 − 2ν  2ν
T ρ1 (t) =  s(t, ρ1 (t)) + s(y, ρ1 (y))dy
2M (ν) − νM (ν)  2M (ν) − νM (ν) 0
  t
2 − 2ν  2ν
≤ s(t, ρ2 (t)) + s(y, ρ2 (y))dy
2M (ν) − νM (ν)  2M (ν) − νM (ν) 0
≤ T ρ2 (x, t) (12)

and
  t
2 − 2ν  2ν
T a1 (t) =  s(t, a1 (t)) + s(y, a1 (y))dy
2M (ν) − νM (ν)  2M (ν) − νM (ν) 0
  t
2 − 2ν  2ν
≤ s(t, a2 (t)) + s(y, a2 (y))dy
2M (ν) − νM (ν)  2M (ν) − νM (ν) 0
≤ T a2 (x, t). (13)

Hence T is a non-decreasing operator, so that the operator T : v1 , v2  → v1 , v2  is compact and continuous via
lemma 1. In that case, H is a normal cone of T .

3.2 Uniqueness solutions

In the previous section, we proved, using the fixed point theorem, that the coupled cancer treatment model with the
Caputo-Fabrizio time fractional derivative has an existing solution. The goal of this section is to show the uniqueness
of solutions to the system (3) with the initial conditions. Let us assume, in addition, that we can find two special
coupled solutions (ρ1 , ρ2 ) and (a1 , a2 ). So the uniqueness of the solution is presented as follows:
   t 
 2 − 2ν  2ν 

|T ρ1 (t) − T ρ2 (t)| =   s(t, ρ1 (t) − s(t, ρ2 (t)) + (s(y, ρ1 (y)) − s(y, ρ2 (y))) dy 
2M (ν) − νM (ν)  2M (ν) − νM (ν) 0
 t 
2 − 2ν 2ν  
≤ |s(t, ρ1 (t) − s(t, ρ2 (t))| +  (s(y, ρ (y)) − s(y, ρ (y))) dy 
2M (ν) − νM (ν) 2M (ν) − νM (ν) 0  1 2 
2 − 2ν 2ν
≤ F1 |ρ1 (t) − ρ2 (t)| + F1 |ρ1 (t) − ρ2 (t)|. (14)
2M (ν) − νM (ν) 2M (ν) − νM (ν)

Thus we can write the above eq. (14) as


2 − 2ν 2ν
|T ρ1 (t) − T ρ2 (t)| ≤ F1 + F1 |ρ1 (t) − ρ2 (t)|.
2M (ν) − νM (ν) 2M (ν) − νM (ν)
Similarly,
   t 
 2 − 2ν  2ν 

|T a1 (t) − T a2 (t)| =   s (t, ρ1 (t)) − s(t, ρ2 (t)) + (s(y, ρ1 (y)) − s(y, ρ2 (y))) dy 
2M (ν) − νM (ν)  2M (ν) − νM (ν) 0
2 − 2ν 2ν
≤ F2 |a1 (t) − a2 (t)| + F2 |a1 (t) − a2 (t)|. (15)
2M (ν) − νM (ν) 2M (ν) − νM (ν)
Page 6 of 6 Eur. Phys. J. Plus (2018) 133: 92

Thus we can write the above eq. (15) as


2 − 2ν 2ν
T a1 (t) − T a2 (t) ≤ F2 + F2 |a1 (t) − a2 (t)|.
2M (ν) − νM (ν) 2M (ν) − νM (ν)

Therefore, if the following conditions hold:


2 − 2ν 2ν
F1 + F1 < 1
2M (ν) − νM (ν) 2M (ν) − νM (ν)

and 
2 − 2ν 2ν
F2 + F2 < 1,
2M (ν) − νM (ν) 2M (ν) − νM (ν)
mapping T is a contraction, we can say that the model has a unique positive solution using fixed point theorem.

4 Conclusion
We first integrated this article cancer treatment model with the new fractional derivative. After that, we found the
existence solution of the cancer treatment model. Finally, we analyzed how the model is the uniqueness positive
solution and under which conditions. We tried to help the researchers working on cancer education with this work.
When the results were examined, it was shown that the fractional derivative gives important information about the
process.

References
1. Z. Liu, C. Yang, Comput. Math. Methods Med 124, 1 (2014).
2. G. Belostotski, A control theory model for cancer treatment by radiotherapy, MS Thesis, University of Alberta (2004).
3. A. Atangana, I. Koca, Chaos, Solitons Fractals 89, 447 (2016).
4. A. Atangana, D. Baleanu, Therm. Sci. 20, 763 (2016).
5. A. Atangana, Neural Comput. Appl. 26, 1895 (2015).
6. A. Atangana, Neural Comput. Appl. 25, 1021 (2014).
7. A. Atangana, E.F.D. Goufo, BioMed. Res. Int. 2014, 7 (2014).
8. A. Atangana, B.S.T. Alkahtani, Entropy 17, 4439 (2015).
9. H.M. Baskonus, T. Mekkaoui, Z. Hammouch, H. Bulut, Entropy 17, 5771 (2015).
10. H.M. Baskonus, H. Bulut, AIP Conf. Proc. 1738, 290004 (2016).
11. D. Baleanu, J.A. Tenreiro Machado, C. Cattani, M.C. Baleanu, X.J. Yang, Abstr. Appl. Anal. 2014, 535048 (2014).
12. A.M. Yang, C. Cattani, C. Zhang, G.N. Xie, X.J. Yang, Adv. Mech. Eng. 6, 514639 (2014).
13. X.J. Yang, H.M. Srivastava, C. Cattani, Rom. Rep. Phys. 67, 752 (2015).
14. G. Belostotski, H.I. Freedman, Int. J. Appl. Mech. 25, 447 (2005).
15. H.I. Freedman, G. Belostotski, Differ. Equ. Dyn. Syst. 17, 115 (2009).
16. Z. Liu, S. Zhong, C. Yin, W. Chen, Appl. Math. Lett. 24, 1745 (2011).
17. M. Caputo, M. Fabrizio, Progr. Fract. Differ. Appl. 1, 73 (2015).
18. J. Losada, J.J. Nieto, Progr. Fract. Differ. Appl. 1, 87 (2015).
19. M. Caputo, Geophys. J. R. Astr. Soc. 13, 529 (1967).
20. I. Podlubny, Fractional Differential Equations, in Mathematics in Science and Engineering, Vol. 198 (Academic Press, San
Diego, USA, 1999).