276
In e na ional Jou nal o Ad ance and Applied Resea ch
www.ijaa .co.in
ISSN – 2347-7075
Impac Fac o – 8.141
Pee Re iewed
Bi-Mon hly
Vol. 6 No. 38
Sep embe - Oc obe - 2025
Pe o mance E alua ion Me ics o Machine Lea ning Classi ica ion
Techniques In a Oncological Disease P edic ion
Ms. P a iksha S. Cha an
Depa men o Compu e Science
D . D Y Pa il A s, Comme ce and Science College Aku di, Pune 44, INDIA
Co esponding Au ho –Ms. P a iksha S. Cha an
DOI - 10.5281/zenodo.14784834
Abs ac :
INDIA, Repo ed nea ly 13, 24, 413 new cance incidences in 2020. P os a e cance , an
oncological disease cases a e also signi ican ly inc eased wi h an inc easing mo ali y a e. Accu a e
p edic ion ools a e impo an in p os a e cance de ec ion. Machine lea ning p inciples a e bes sui ed
o de eloping mode n applica ions in heal hca e in o ma ics and biomedical Enginee ing and sciences.
The p oposed pape discussed ML based classi ie echniques o p edic ion o a ec ed p os a e cance
glands. Logis ic eg ession decision ee, k nea es neighbo and naï e Bayes classi ie a e implemen ed
o he p os a e cance pa ien da a. Compa a i e analysis wi h he help o classi ica ion me ics is
p esen ed. The p edic ion accu acy using logis ic eg ession is calcula ed as 91% which is be e han
he 73% o K nea es Neighbo classi ie echniques is 73%.
Keywo ds: Machine lea ning, Image Segmen a ion, Oncology, PBCRs, HBCRs, IARS, P os a e
cance , Managemen
In oduc ion:
An in oduc ion o machine lea ning
based classi ica ion algo i hms in oncological
diseases aims a building e icien and accu a e
p edic i e modelling ools. These ools will
no only help o p edic he isk le el o
se e i y o cance s bu also help in clinical and
adia ion oncology managemen [11] [14]. The
p edic i e ools designed using ML echniques
p o ide an op imal me hod in he apeu ic and
cance pa ien managemen .
New incidences o male p os a e
cance a e signi ican ly inc easing ac oss he
globe. I anks among he majo cance ypes
ound in men. The Popula ion Based Cance
Regis y (PBCRs) and Hospi al Based Cance
Regis y (HBCRs) which a e aimed o keep
eco ds o all cance incidences es ima ed ha
he p os a e cance incidences and dea hs a e
inc easing a a as e a e. The da a published
by In e na ional Agency on Cance esea ch
(IARC) shows ha 19292789 new cance
incidences o all ypes a e epo ed in 2020
[1][3]. This da a also shows ha new 1414259
p os a e cance incidences epo ed which a e
mo e han epo ed in 2018 [2] [3].
India epo ed a signi ican ise in he
numbe o cance incidences in all egions o
he coun y. A o al o 13324413 new cance
incidences o all ypes o cance s a e epo ed
wi h 851678 dea hs cases. Incidences o new
p os a e cance pa ien s also inc ease in majo
ci ies o India. The e a e 34540 new P os a e
cance pa ien s epo ed in 2020. These
pa ien s a e mo e han ha o bladde 21096,
Thy oid 20432, Gallbladde 19570 and 12642
Panc eas cance pa ien s [6]. P os a e cance is
a second leading cance disease in Pune,
IJAAR Vol. 6 No. 38 ISSN – 2347-7075
Ms. P a iksha S. Cha an
277
Delhi, Thi u anan hapu am and Kolka a and
hi d in ci ies like Mumbai and Bangalo e [5].
P os a e cance anks 12 h in all cance ypes
ound in India [6].
Fig. 1(a)
Fig. 1(b)
Fig.1 (a) (b) GLOBOCAN 2020 S a is ic by WHO’s IARC
Fig. 2 (a)
Fig. 2 (b)
Fig.2(a) (b). GLOBOCAN 2018 S a is ic by WHO’s IARC
P os a e Gland Ana omy And P os a e
Cance :
The ep oduc i e sys em o male has
seminal esicles, Bulbou e h al and P os a e
accesso y glands. These glands a e esponsible
o anspo a ion o semen and p oduc ion o
s uc u al p o ein and alkaline solu ion ha a e
needed in he o ma ion o spe ma opho es [7].
P os a e gland, a ib omuscula gland lies
in he on o he ec um a ea and su ounding
u e h a [7] [8]. P os a e gland ana omy
desc ibes his gland in di e en lobes and
h ee zones as shown in ig.3.
▪ La e al lobe
▪ Median lobe
▪ An e io lobe
▪ Pos e io lobe
▪ Cen al zone
▪ T ansi ion zone
▪ Pe iphe al zone
Fig.3. Sec ion o P os a e Gland Ana omy zone Rep esen a ion
IJAAR Vol. 6 No. 38 ISSN – 2347-7075
Ms. P a iksha S. Cha an
278
P os a e gland size changes wi h an
inc ease in men's age. The p os a i is condi ion
in he p os a e gland occu s by in ec ion.
When gland cell size changes ab up ly,
p os a e cance begins o o m. This cance
sp eads e y slowly and s a s causing
symp oms o yea s and yea s.
Digi al ec um examina ion [10],
p os a e speci ic an igen [9] Imaging es and
p os a e biopsy es s a e used o possible
cance symp oms examina ion and diagnosis
in male pa ien s.
T adi ional p os a e cance es ing
ools some imes become less accu a e and
digi al ec um analysis and biopsy may cause
in ec ion, bleeding, and pain due o ex e nal
ins umen use.
The new app oach has been
in oduced in cance diagnosis using ad anced
adiog aphic es s and machine lea ning
echniques. Wi h he incu sion o a i icial
in elligence echniques in spine ela ed
su ge y, a new app oach has been adop ed o
c ea e clinical analy ics ools and diagnos ic
ins umen s. ML and AI echnique makes
clinical p edic ions and decision-making asks
easie [12] [13]. These echniques also play an
impo an ole in ans o ming medical
imaging which inco po a es medical
adiology.
Newly designed diagnos ic imaging
sys ems a e u ilizing he capabili ies o AI and
ML echniques. Specially designed Algo i hms
help pe o mance uning o imaging sys ems
[15-18] and acili a e he quan i ica ion and
de ec ion o a ious possible clinical
scena ios.AI echniques no only show
imp essi e pe o mances in accu a ely
iden i ying he cance egion in p os a e gland
in male ep oduc i e sys em.
Machine Lea ning In P os a e Cance
De ec ion:
Resea che s wi h he help o he la es
machine lea ning echniques a e de eloping
ad anced me hods o de ec ion and analyzing
p os a e cance [9].
ML echniques can be bes sui ed wi h
imaging es s ha a e used o p os a e cance
diagnosis. Radiog aphy such as CT and MRI
a e he x- ay imaging es ing me hods ha
easily de ines all p os a e gland ana omical
de ails [11]. This me hod p o ides de ails
abou lesion loca ion, s uc u al issue changes
in gland and size o he p os a e gland.
A. The logis ic eg ession:
Logis ic Reg ession, a ML based
p edic i e analy ic algo i hm which uses
p obabili y s a is ics. Bina y and mul i linea
logis ic eg ession a e he wo ypes o he
eg ession used o da a classi ica ion.
Con e sion o da a samples in o i s
s a is ical p obabili ies, sigmoid unc ions a e
used in logis ic eg ession gi en by
( )
------ (1)
Following s eps a e equi ed o do
logis ic eg ession in classi ica ion p ocess
▪ Da a p ep ocessing
▪ Logis ic eg ession unc ion i ing
o aining da ase
▪ Tes esul p edic ion
▪ Tes esul isualiza ion
B. K Nea es Neighbo Classi ie
▪ Selec ion o no. o neighbo
▪ Euclidian dis ance calcula ion
be ween no. o neighbo s
▪ Coun ing no. o da a poin s o
e e y da a
▪ Assign new da a poin s
P oposed Wo k:
An aim o he p oposed pape is o
implemen he ou ML based classi ie s o
IJAAR Vol. 6 No. 38 ISSN – 2347-7075
Ms. P a iksha S. Cha an
279
p edic ion o he p os a e cance ype om he
gi en p os a e cance pa ien da a. The cance
da a is a CSV based ile ha has in o ma ion
abou p os a e cance esul s ha a e ei he
malignan o benign and geome ic
in o ma ion abou cance p os a e gland.
Logis ic eg ession and K- nea es
Neighbo (KNN) classi ie s a e implemen ed
o p edic he esul s based on he in o ma ion
p o ided in he gi en da a.
Following essen ial s eps pe o med
o he p edic ion p ocess using logis ic
eg ession and KNN Classi ie echniques
▪ Load he da ase
▪ Fea u e ex ac ion om he gi en da ase
▪ Da a collec ion o aining and es ing
pu pose
▪ T ain he model on he aining da a
▪ P edic ion on he basis o es ing da a
Obse a ions:
The alues o classi ica ion me ics
a e calcula ed by implemen ing he p oposed
classi ie s o he p os a e cance p edic ion.
The p os a e cance esul is ei he Benign o
Malignan . The benign p os a e cance is
ep esen ed by 0 and malignan by 0. The
alues o he me ics o each classi ie a e
gi en in espec i e ables.
Table 1: Classi ica ion Me ics o Logis ic Reg ession
Accu acy
P ecision
Recall
F1-Sco e
91%
+Ve
-Ve
Sensi i i y
Speci ici y
+Ve
-Ve
0.91
0.90
0.95
0.82
0.93
0.86
Fig 6 Con usion Ma ix o Logis ic Reg ession
TABLE 3 Classi ica ion Me ics o KNN Classi ie
Accu acy
P ecision
Recall
F1-Sco e
73%
+Ve
-Ve
Sensi i i
y
Speci ici y
+Ve
-Ve
0.84
0.57
0.73
0.73
0.78
0.64
IJAAR Vol. 6 No. 38 ISSN – 2347-7075
Ms. P a iksha S. Cha an
280
Fig.6. Con usion Ma ix o KNN Classi ie
Conclusion:
The compa a i e analysis be ween
logis ic eg ession and KNN based ML
classi ie s is p esen ed. The nume ical da a
ela ed o p os a e cance gland is used o his
compa a i e analysis. Logis ic eg ession
classi ie (LR) is accu a e o 91 % in
p edic ing he na u e o p os a e cance ei he
malignan o benign. The accu acy o KNN
is only 73% . The abili y o cap u e he all
posi i e samples called as T ue Posi i e Ra e
(TPR) o sensi i i y o gi en classi ie s
Logis ic eg ession and KNN a e 94% and
83% espec i ely. Thus he p os a e cance
p edic ion can be e icien ly pe o med by
using logis ic eg ession echnique.
Acknowledgmen :
I would like o exp ess my g a i ude o
D . Mohan Waman P incipal, Compu e
Science, D . D Y Pa il ACS College, Pune
INDIA o hei aluable guidance o
p o iding P os a e Medical Image Da abase.
Re e ences:
[1] Hyuna Sung, Jacques Fe lay, Rebecca L.
Siegel, Ma hieu La e sanne, Isabelle
Soe joma a am, Ahmedin Jemal; ―Global
Cance S a is ics 2020: GLOBOCAN
Es ima es o Incidence and Mo ali y
Wo ldwide o 36 Cance s in 185
Coun ies‖; CA: A Cance Jou nal o
Clinicians, Volume 71, Issue 3, 04
Feb ua y 2021
[2] F. B ay, J. Fe lay, I. Soe joma a am, R. L.
Siegel, L. A. To e, A. Jemal; ―Global
cance s a is ics 2018: GLOBOCAN
es ima es o incidence and mo ali y
wo ldwide o 36 cance s in 185
coun ies‖; CA: A Cance Jou nal o
Clinicians, Volume 70, Issue 4, 06 Ap il
2020
[3] In e na ional Agency o Resea ch on
Cance and Global Cance Obse a o y;
h ps://gco.ia c. /
[4] K ishnamoo hy Ha iha an, Venugopal
Padmanabha; ―Demog aphic and Disease
cha ac e is ics o p os a e cance in
India‖; Indian Jou nal o
u ology, Volume 32, Issue 2, Pages: 103-108,
2016
[5] P ameesha Pe e aa, Ayam Gup ac, Nidhi
Shuklab, Mam a Neh ac, Mukesh
Sha mad, Sneha Mish ad ,Vik am Singh
Chauhand, e al; ― Re isi ing P os a e
Cance in India‖; P ep in s , 26 No embe
2018
[6] P ashan Ma hu , K ishnan Sa hishkuma ,
Meesha Cha u edi, P iyanka Das, e al;
―Cance S a is ics, 2020: Repo F om
Na ional Cance Regis y P og amme,
India‖; Global Oncology, Ame ican
Socie y o Clinical Oncology, 16 July
2020
IJAAR Vol. 6 No. 38 ISSN – 2347-7075
Ms. P a iksha S. Cha an
281
[7] Omesh Singh, S ini asa Rao Bolla;
―Ana omy, Abdomen and Pel is,
P os a e‖; S a Pea ls Publishing, Janua y
2021
[8] Anil Bha sa , Sadhna Ve ma;
―Mul ipa ame ic MRI in P os a e
Cance ‖; BioMed Resea ch In e na ional,
ol. 2014, A icle ID 728539, 9 pages,
2014
[9] Michael Obana, Hen y O'Law ence;
―P os a e Cance Sc eening: PSA Tes
awa eness among adul males‖; J Heal h
Hum Se Adm. 2015
[10] Leen Naji, Ha kanwal Randhawa,
Zah a Sohani; ―Digi al Rec al
Examina ion o P os a e Cance
Sc eening in P ima y Ca e: A Sys ema ic
Re iew and Me a-Analysis‖; Analysis o
Family Medicine, 149-154, Ma ch
[11] Mohammed AliKazem; ―P edic i e
models in cance managemen : A guide
o clinicians‖; Else ie , The Su geon,
Volume 15, Issue 2, Pages 93-97, Ap il
2017
[12] Vincenzo Valen ini, Nicola Dinapoli,
And ea Damiani; ―The u u e o p edic i e
models in adia ion oncology: om
ex ensi e da a mining o eliable modeling
o he esul s‖; Fu u e Oncology, Ma ch
2013
[13] De ek J Van Boo en, Manish
Kuchakulla, Ragha Pai, Fabio S F ech,
Reshna Ramasahayam, P i ika Reddy,
Madhumi a Pa ma , Ranji h Ramasamy,
Himanshu A o a1; ―A Sys ema ic Re iew
o A i icial In elligence in P os a e
Cance ‖; Do ep ess Jou nal, Resea ch and
Repo s in U ology, 2021
[14] Ronan Thenaul , Ke in
Kaulanjan,Thomas Da de, Na halie Rioux
Lecle cq, Ka im Bensalah, Ma ie
Me mie , Zine-eddine Khene, Benoi
Pey onne , Shah okh Sha ia , Benjamin
P ade e , Romain Ma hieu; ―The
Applica ion o A i icial In elligence in
P os a e Cance Managemen —Wha
Imp o emen s Can Be Expec ed? A
Sys ema ic Re iew‖; Mdpi, Applied
Science, 2020
[15] Lidia Alcala Ma a , Juan An onio
Re ame o, Rajan T. Gup a, Robe o Ga cía
Figue as, An onio Luna; ―A i icial
In elligence–assis ed P os a e Cance
Diagnosis: Radiologic-Pa hologic
Co ela ion‖; RadioG aphics, Vo1ume 41,
Numbe 6, 1 Oc obe 2021
