Abstract

There have been an enormous number of publications on cancer research. These

integrated but unstructured cancer related articles are of great value for
cancer diagnosis, treatment and prevention. In this study, we use the basic
concepts underlying
text mining and semantic computing to discuss the current state of- the-art
text mining applications in cancer research. Using the abstract of literature
extracted from PubMed between 2008 and 2016, a total 925,648 articles are used
for subsequent text mining. Among the 925,648 articles, the top 5 most studied
cancer types were breast cancer (23.82%), lung cancer (10.54%), prostate cancer
(9.90%), rectal cancer (8.44%), and ovarian cancer (4.44%). The top 3 most
frequently occurred keywords in the abstracts of the 925,648 articles are
patients, cancer, and cell where each appear 1,445,688, 1,284,140, and 676,924
times, respectively. Analysis of the key concepts indicate that the most common
concepts are patients, cancer, cell and tumor. Our results suggest that while
the risk factors of cancer, treatment of cancer, and survival of cancer
patients were popular research topics, endof- life cancer care issues are less
studied. Further studies should explore these areas since they are as important
as treatment of the disease itself for many patients.

I. INTRODUCTION

Cancer is the leading cause of mortality and morbidity worldwide according to

the International Agency for Research on Cancer [1]. Cancer is one of the most
important study areas for biomedical researchers. It has been widely studied
for more than 100 years. The huge body and rapid growth of text on cancer
research provides a valuable resource [2]. As shown in Figure 1, there are
many publications on cancer research and the number of publications keeps
increasing every year. We searched PubMed with ‘cancer’’ in the abstract and
we retrieved more than 925,64 publications from 20011 to 2016. It is almost
impossible for people to read all of these publications and discover new
knowledge. The vast numbers of biomedical text provide a rich source of
knowledge for biomedical research. Text mining can help researchers to mine
information and knowledge from a mountain of text and it is now widely applied
in biomedical research. Many researchers have taken advantage of text mining
technology to discover novel knowledge to improve the development of biomedical
research especially those pertaining to malignant diseases, such as cancer.
Realizing the advantages of text mining will facilitate cancer research, by
helping to find new knowledge for cancer diagnostics, treatment, and
prevention.

Figure 1. Number of cancer papers in PubMed search with keyword “cancer”.

Text mining technology provides a solution for bridging the knowledge gap
between free-text and structured representation of related information in
cancer research.It has emerged as a potential solution for bridging the gap
between free-text and structured representation of cancer information. It
employs many computational technologies, such as machine learning, natural
language processing and deep learning, to find new exciting outcomes hidden
in the unstructured cancer related articles. There are many applications of
text mining on cancer-related articles, such as identifying malignant tumor
related Biomedical mentions, finding relationships among biomedical entities
such as protein-protein interaction, gene disease network, etc.), extracting
knowledge from text and generating hypotheses. Cancer is a complex disease,
which is related to a large number of genes. Biomedical research is interested
in mining from the cancer-related articles to study cancer diagnostics,
treatment and prevention. There have been a number of text mining applications
specifically focusing on extracting cancer related information; Spasić et
al.provide a comprehensive review of these. The review highlight a strong bias
towards symbolic techniques, i.e., the use of pattern matching for cancer-
related entity extraction to deliver good performance. Pletscher-Frankild et
al. present a system used text mining for extracting disease-gene associations
from biomedical abstracts. The system consists of a highly efficient
dictionary-based tagger for named entity recognition of genes and diseases,
and it combines with a scoring scheme that takes into account co-occurrences
both within and between sentences. Gonzalez et al. present an overview of the
fundamental methods for text and data mining, as well as recent advances and
emerging applications toward precision medicine. Baker et al. provide an
extensive Hallmarks of Cancer taxonomy and develop an automatic text mining
methodology of cancer-related articles from PubMed into the taxonomy. It offers
a great potential to organize and correctly classify cancer-related articles.
In this study, we explore the conceptual content of cancer research based on
the abstract of articles extracted from PubMed. Findings from this study may
be used to identify gaps in cancer research. In addition, we used Topic model
and GloVe for the major task in text mining so the broader scientific community
may be able to gain ideas and insights based on the existing cancer studies
for the development of their own primary studies.
 II. METHODS
A. Data Source
Data were retrieved and downloaded from PubMed, a website
(http://www.ncbi.nlm.nih.gov/pubmed/) that provides free access to biomedical
journal citations and abstracts mainly indexed by Medline [7]. The service is
administered by the National Center for Biotechnology Information of the United
States National Library of Medicine. Search terms including “(“neoplasms[MeSH
Terms] OR “neoplasms”[All] OR “cancer”[All]) AND (“2011/01: 2016/12/[DP]”)
were used in the search strategies. The publication date was limited to the
year between 2011 and 2016. The search was conducted on July 30, 2016 and a
total of 925,648 articles were retrieved.
B. Word Clouds
Word clouds are visualizations that display the words that frequently occur
in a text. These visualizations are particularly useful when one has no
preconceived idea of which concepts should occur in a text. In word clouds,
words that appear more frequently in a text are printed in larger fonts than
words that occur less often. The advantage of word clouds is that this
visualization is not biased by the use of a predefined set of concepts or an
ontology, but is driven by the raw content of the text. As such, they can
provide new ideas and insights on a particular concept and can function as a
starting point for more specific searches.
C. Topics Model
Previous text mining studies mainly rely on quantitative measures and suffer
from the lack of content analysis. To incorporate content analysis into cancer-
related articles, text mining techniques are applied. Topic-modeling
techniques are mostly adapted to identify the topics of a subject area while
analyzing that area more abundantly . The Latent Dirichlet Allocation (LDA)
is the best received topic-modeling technique. LDA is a populistic generative
model that assumes each document is mixture of latent topics, where each topic
is probability distribution over all words in the vocabulary. LDAis a three-
level hierarchical Bayesian model, in which each item of a collection is
modeled as a finite mixture over an underlying set of topics. It treats each
document as a mixture of topics, and each topic as a mixture of words. LDA
assumes the following generative process for each document W in a corpus
D:
1. Choose a multinomial distribution __ for document d, Dir(), where _ ∈ {1,…,}
and Dir() is a Dirichlet distribution with a symmetric parameter _ which
typically is spares (_ < 1)
2. Chose a multinomial distribution for topic t, Dir(), where _ ∈ {1,…,
} and is a Dirichlet distribution parameter
3. For each of the work positions i, j, where ∈ {1,…,} and _ ∈ {1,…,}
In this study, we use Gibbs sampling[11] to estimate LDA parameters. Gibbs
sampling is a Monte Carlo Markov-chain algorithm, powerful in statistical
inference and a method of generating a sample from a joint distribution when
only conditional distributions of each variable can be efficiently computed.
Gibbs sampling have widely for LSA parameters estimate[12].

D. Learning Word Embedding We download abstracts of 925,648 articles from

PubMed. Global Vectors (GloVe) for Word Representation models were trained on
cancer-related articles. The GloVe computes continuous vector representations
of words based on neural networks. It takes a corpus as input and produces a
vector space. It estimates continuous vector representations of words from a
large corpora. Each unique word in the corpus is represented as a vector and
words that share common contexts are positioned closely to each other. The
similarity of two words is determined by calculating the cosine similarity of
two vectors that represent the two words [13]. The intuitive idea behind the
GloVe model is to build a very big matrix, X, of co-occurrence words from a
corpus. Each cell of the matrix, Xij represents how many times has the row
word, i, appears in some context, j. By doing a simple normalization of the
values for each row of the matrix, we can obtain the probability distribution
of every context given a word. Once the probabilities have been calculated,
the relationship between the words can be calculated by doing the ratio between
the probabilities of the context given these two words.
 III. RESULTS

Of the 925,648 publications based on the data for the cancer publishing between
2011 and 2016. Table 1 lists the journals that publish from cancer-related
study between the years 2011 to 2016. The Top 5 journals with cancer related
research article are PLoS ONE, Oncotaget, Asia Pac. J. Cancer Preve., Tumour
Biol. and J. Clin. Incol., with 22543, 9390, 6616, 6364 and 5235 counts,
respectively. The 5-year impact factors of the journals ranged from 2.86 to
16.80, with a mean 9.83. The Top 5 most studied cancer types are breast cancer
(23.82), lung cancer (10.54), prostate cancer (9.90), rectal cancer (8.44) and
ovarian cancer (4.44) (Table 2).
Table1. Top 10 Journals that published cancer-related articles between
2011and 2016
Title (Five-Year Bioxbio Journal Impact* ) Frequency Normalized

PLoS ONE (2.86) 22543 0. 0247

Oncotarget (5.07) 9390 0. 2409
Asian Pac. J. Cancer Prev. (1.82) 6616 0. 1809
Tumour Biol. (3.33) 6364 0. 3056
J. Clin. Oncol. (16.80) 5235 0. 1874
Ann. Surg. Oncol. (3.91) 4683 0. 2255
Int. J. Cancer (6.77) 4252 0. 2851
Sci Rep (4.31) 3936 0. 0501
Oncogene (6.61) 3012 0. 2950
BMC Cancer (3.31) 2969 0. 2086

Five-year impact factors were based on

the 2016 journal citation reports
 Table2. Cancer type in the cancer-related articles between 2011 and 2016

(N= 925,648)

Cancer type Frequency (%)

Breast Cancer 220523 (23.82)

Lung Cancer 97569 (10.54)

Prostate Cancer 91662 (9.90)

Rectal Cancer 78156 (8.44)

Ovarian Cancer 41062 (4.44)

Cervical Cancer 34451 (3.72)

Pancreatic Cancer 31326 (3.38)

Colon Cancer 28631 (3.09)

Bladder Cancer 25645 (2.77)

Thyroid Cancer 15654 (1.69)

Skin Cancer 10800 (1.17)

liver Cancer 9868 (1.07)

A. Word Cloud
The term frequencies in the abstracts of 925,648 articles were visualized as
word clouds with a larger word size represent a higher frequency of appearance
among the articles, as shown in Figure 2. The Top 5 terms with the highest
frequencies both over different periods and over the entire period of 2011 to
2016 are patients, cancer, cell, tumor and study. The five terms are suppressed
in the display of the word cloud to allow a better visualization of the
remaining terms. The words ranked the sixth to the tenth in frequency are
expression, treatment, survival, associated and clinical.

B. Analysis of High word frequently association

Table 3 presents the results that the analysis of keyword association for the
top 15 most frequently occurred keywords in the cancer-related articles’
abstracts. The top 3 most frequently occurred keywords in the abstracts of the
925,648 articles were patients, cancer, and cell with 1,445,688, 1,284,140,
and 676,924 times, respectively. The remaining 12 keywords had a frequency of
occurrence ranged from 590685 times in “tumor” to 244349 times in
“significantly.” Each of the primary keywords was evaluated for their
correlations with other in the abstract (secondary keywords). Overall, the
correlation coefficients ranged from the highest at 0.52 for survival-overall
to the lowest at 0.10 for clinical-benefit and tumor-vitro.
Table3. Analysis high word frequently association of the corpus for the top 10 most frequently occurring word in the abstract

Keyword Frequency The secondary keyword with highest correlation with the keyword (correlation coefficient)
median(0.36), months(0.31), treated(0.27), rate(0.23), retrospectively(0.23), age(0.22), surgery(0.22),
patients 1445688
respectively(0.21), chemotherapy(0.21), retrospective(0.21)

prostate(0.24), lung(0.19), women (0.16), colorectal(0.16), association(0.15), gastric(0.15), colon(0.14),

cancer 1284140
screening(0.13), incidence(0.13), ovarian(0.13), increased(0.12), studies(0.12)

lines(0.49), proliferation(0.34), line(0.31), cycle(0.29), apoptosis(0.26), growth (0.25), human(0.24),

cell 676924 induced(0.21), inhibited (0.21), arrest(0.21), migration(0.21) inhibition (0.21 ), vitro(0.20), viability(0.20),
squamous(0.19), carcinoma(0.18), protein(0.18), stem(0.17), death(0.16), role (0.16),

tumors(0.29), growth(0.22), cells(0.17), microenvironment(0.17), mice(0.15), vivo(0.15), metastasis(0.14),

tumor 590685
suppressor(0.14), antitumor(0.14), progression(0.13), angiogenesis(0.13), xenograft(0.12), vitro(0.10)

study 537961 aim(0.23), significant(0.16), compared(0.15), age(0.15), purpose(0.15),

retrospective(0.15), prospective(0.15), groups(0.14), respectively(0.13), group(0.13),
expression 529424 protein(0.30), mRNA(0.30), gene(0.29), immunohistochemistry(0.27),

survival 356436 overall (0.52), year(0.34), prognostic(0.33), median(0.32), Kaplan Meier(0.30),

chemotherapy(0.19),proportional(0.16), time(0.13) associated 333354
multivariate (0.17), odds (0.17), factors (0.16), increased (0.15), regression (0.15), confidence (0.15),
polymorphisms (0.15), logistic (0.14), age (0.13), independently (0.13), case-control(0.13), SNP (0.13),

C. Topic Model
We employ topic modeling to explore the topics as a bunch of words in the
abstracts of cancer-related. The topic model analysis on the cancer-related
articles is presented in Table4. As
can be seen from the table, we find that a dominant word cancer and patient
are in these description texts, in addition there is meaningful difference
between these collections of word, from word about cancer study, patient
treatment, survival of patients, risk of patients, cancer type and cancer
cells study on genome in the topic. Topics 1 is cancer study to word about
cancer, patient, study and data. Topics 2 is treatment of patient about word
patients, cancer, surving and treatment. Topics 3 is about genomics and related
to cell biology. Topics 4 word about women, cancer and breast. This is not
surprising because breast cancer is a common cancer in women. Topics 5 includes
risk of patient and Topics 6 contains survival of patients. The topic modeling
process has identified groupings of terms that we can understand as human
readers.

D. Cancer-related articles on GloVe

We adopt GloVe to compute additional vector representations of word in cancer-
related articles. Table 5
shows that the top 15 most frequently occurred keywords and the most similar
words trained by GloVe on the cancer-related articles. The GloVe identifies
cancer type, such as breast, colorectal, lung, prostate, ovarian, gastric,
cervical bladder, pancreatic and carcinoma. and captures cancer-related
articles of attribute-related relatedness, such as patient similarities for
receiving, cases, treated, enrolled, diagnosed, surgery, undergoing, received,
advanced and primary. mRNA, overexpression, downregulation, upregulation,
protein, miRNA, levels, and furthermore, gene, correlated similar to
expressionand tumor similarities for metastasis, metastatic, invasion, growth,
size, tissue, malignant, lesion, cell, differentiation. The risk similar to
mortality, incidence, associated, occurrence, prevalence, odds, among,
predictors, adjusted, disease. We show that the text mining findings are
reliable, as per the PubMed scale, in that the word-word relationships inferred
from the literature-wide findings are similar.

Table 4. LDA with 6 topics

Topic 1 Topic 2 Topic 3
Cancer/Study patient/Treatment Cells/Genome
cancer 0.0205 patient 0.0248 cells 0.0198
patient 0.0178 cancer 0.0088 cancer 0.0163
tumor 0.0107 using 0.0056 expression 0.0135
study 0.0085 treatment 0.0054 tumor 0.0070
data 0.0052 study 0.0051 protein 0.0054
therapy 0.0043 clinical 0.0051 human 0.0054
activity 0.0041 tumor 0.0048 gene 0.0050
expression 0.0041 higher 0.0041 growth 0.0045
treatment 0.0037 significantly 0.0040 activity 0.0041
tumors 0.0036 cases 0.0040 study 0.0040
Topic 4 Topic 5 Topic 6
Breast Cancer Patient /Risk Patient /Survival
cancer 0.0233 patient 0.0346 patient 0.0360
patient 0.0097 cancer 0.0170 survival 0.0077
risk 0.0087 study 0.0071 treatment 0.0073
study 0.0064 risk 0.0068 tumor 0.0063
breast 0.0054 survival 0.0066 cancer 0.0056
studies 0.0050 treatment 0.0062 study 0.0052
women 0.0049 group 0.0053 group 0.0051
data 0.0047 years 0.0044 cases 0.0048
associated 0.0046 disease 0.0044 clinical 0.0043
using 0.0039 associated 0.0042 months 0.0042

Table 5. the top 15 most frequently occurred keywords and the most similar words according to GloVe
Keyword Top 15 most frequently occurred keywords most similar words

patients receiving, cases, treated, enrolled, diagnosed, surgery, undergoing, received, advanced, primary

cancer breast, colorectal, lung, prostate, ovarian, gastric, cervical bladder, pancreatic, carcinoma

cell proliferation, migration, lines, apoptosis, growth, human, viability, stem, vitro, differentiation

tumor metastasis, metastatic, invasion, growth, size, tissue, malignant, lesion, cell, differentiation
study conducted, present, retrospective, prospective, carried, examined, evaluated, investigate, aim,
review

expression mRNA, overexpression, downregulation, upregulation, protein, miRNA, levels, Furthermore, gene, correlated
treatment therapy, chemotherapy, adjuvant, regimens, therapies, neoadjuvant, combination, options, radiotherapy, effective.
survival OS (Overall survival), overall, disease-free, DFS (Disease-free survival), PFS (Progression-free survival), progression-free,

recurrence-free, disease-specific, outcome, CSS (Cause-specific Survival)

associated related, factors, risk, worse, significantly, correlated, increased, poor, significant, occurrence
clinical evaluation, outcome, regarding, prognosis, practice, data, review, implications, preclinical,
diagnostic

risk mortality, incidence, associated, occurrence, prevalence, odds, among, predictors, adjusted, disease

breast endometrial, prostate, ovarian, lung, bladder, colorectal, thyroid, melanoma, metastatic, cervical,
analysis multivariate, Multivariate, Cox, univariate, regression, univariate, logistic, Kaplan Meier, statistical, determined
disease failure, relapse, chronic, recurrence, malignancy, progression, metastases, diagnosis, risk, diabetes

significantly significantly, decreased, reduced, markedly, increased, whereas, higher, correlated, compared, lower
 IV. CONCLUSION

This study is the first to conduct computational text mining, semantic

computing, and visualization of articles indexed by PubMed on cancer research.
We use various approaches, including the topic model, Word Embedding, and word
cloud, to visualize the content of 925,648 articles in cancer-related articles
published between 2011 and 2016. We found that breast cancer, lung cancer,
prostate cancer, rectal cancer, ovarian cancer are the top 5 most studied
cancers based
on the cancer-related articles. In the word cloud visualization, we found that
survival of patients, the risk of patients, and cancer cell are frequently
mentioned from 2011to 2016. In high word-frequently association analysis. We
found the top 3 most frequently occurred keywords in the abstracts among the
925,648 articles are patients, cancer, and cell with 1,445,688, 1,284,140, and
676,924 times, respectively.
This study applies the LDA for text mining of a vast amount of cancer-related
articles. The results show that a dominant word cancer and patient care in
these description texts. In addition, there is a meaningful difference between
these collections of word, from word about cancer study, patient treatment,
survival of patients, risk of patients, cancer type and cancer cells study on
genome in the topic. We adopt GloVe to compute additional vector
representations of word in cancer- related articles. Using the Euclidean
distance between two word vectors provides an effective method for measuring
the linguistic or semantic similarity of the corresponding words.
The result suggests that cancer type such as breast, colorectal, lung,
prostate, ovarian, gastric, cervical bladder, pancreatic and carcinoma. A few
limitations of the present study should be mentioned. First, only articles
written in English and indexed by PubMed were included in the study. Second,
articles with no abstract could not be analyzed. Third, variations in the
length of abstract and inaccuracy in the content of the abstract might
potentially influence our results.

In this study of 925,648 published articles on secondary data analysis of the

cancer research, indexed by PubMed between 2011 and 2016, we found that the
top 5 most studied cancers are breast cancer, lung cancer, prostate cancer,
rectal cancer, ovarian cancer.

The conceptions of patients’ survival, cancer- related surgery, radiotherapy,

and chemotherapy are highly frequent from 2011 to 2016. Overall, there are
more articles focusing on the risk factors of cancer, treatment of cancer, and
survival of cancer patients, but relatively few articles on end-of-life cancer
care including hospice care, palliative care, and home palliative care. The
above studies showed that a topic model can accomplish the task of clustering
of cancer-related articles. Furthermore, each topic is interpreted as a
probability distribution over words. We believe that topic models are a
promising method with numerous applications to biomedical research. The GloVe
results show that the text mining findings are reliable, as per the PubMed
scale, in that the word-word relationships inferred from the literature-wide
findings are similar
 REFERENCES
[1] B. Stewart, C. W. L. CEDEX, France, 2014, World cancer
report 2014. 2014.
[2] F. Zhu, P. Patumcharoenpol, C. Zhang, Y. Yang, J. Chan, A.
Meechai, W. Vongsangnak, and B. Shen, “Biomedical text mining
and its applications in cancer research,” Journal of Biomedical
Informatics, vol. 46, no. 2, pp. 200–211, Apr. 2013.
[3] I. Spasić, J. Livsey, J. A. Keane, and G. Nenadić, “Text
mining of cancer-related information: Review of current status
and future directions,” International Journal of Medical
Informatics, vol. 83, no. 9, pp. 605–623, Sep. 2014.
[4] S. Pletscher-Frankild, A. Palleja, K. Tsafou, J. X. Binder,
and L. J. Jensen, “DISEASES: Text mining and data integration of
disease–gene associations,” Methods, vol. 74, pp. 83–89, Mar.
2015.
[5] G. H. Gonzalez, T. Tahsin, B. G. B. in, 2015, “Recent
advances and emerging applications in text and data mining for
biomedical discovery,” academic.oup.com.
[6] S. Baker, I. Ali, I. Silins, S. Pyysalo, Y. Guo, 2017,
“Cancer Hallmarks Analytics Tool (CHAT): a text mining approach
to organize and evaluate scientific literature on cancer,”
academic.oup.com.
[7] M. Castillo, “Is Your Journal Indexed in MEDLINE?,” American
Journal of Neuroradiology, vol. 32, no. 1, pp. 1–2, Jan. 2011.
[8] W. W. M. Fleuren and W. Alkema, “Application of text mining
in the biomedical domain,” Methods, vol. 74, pp. 97– 106, Mar.
2015.
[9] E. Yan, “Research dynamics, impact, and dissemination: A
topic-level analysis,” Journal of the Association for Science
and Technology, vol. 66, no. 11, pp. 2357– 2372, Nov. 2015.
[10] D. M. Blei, A. Y. Ng, and M. I. Jordan, “Latent Dirichlet
Allocation,” Journal of Machine Learning Research, vol. 3, no.
Jan, pp. 993–1022, 2003.
[11] T. L. Griffiths and M. Steyvers, “Finding scientific
topics,” presented at the Proceedings of the National Academy of
Sciences of the United States of America, 2004, vol. 101, no. S,
pp. 5228–5235.
[12] H. Jelodar, Y. Wang, C. Yuan, and X. Feng, “Latent
Dirichlet Allocation (LDA) and Topic modeling - models,
applications, a survey.,” arXiv, vol. cs.IR, 2017.
[13] J. Pennington, R. Socher, C. M. P. O. T. 2014, 2014,
“Glove: Global vectors for word representation,” aclweb.org.
 2019 Second IEEE International Conference on Robotic Computing

Application of Semantic Computing in

Cancer on Secondary Data Analysis

Charles C.N. Wang1,*, I-Seng Chang1, Phillip C.Y. Sheu2 and Jeffrey J. P. Tsai1
1
Department of Biomedical Informatics, Asia University, 500, Lioufeng Rd., Wufeng, Taichung 41354,
Taiwan
2
Department of Electrical Engineering and Computer Science, University of California, Irvine, 5200
Engineering Hall, Irvine, CA 92697,
*[email protected]

Presented by: Jan Dave E. Deocampo

MAT – General Science

IMRAD FORM
Reflection
People we spoke to whose cancer had been cured or was in remission, talked about the extent to which
they felt able to put the experience behind them and about what had helped them to overcome the
illness. While some said their cancer was always in the back of their mind, or that they were sometimes
reminded of it, others said they rarely thought about it and their life had moved on (see also ‘Facing the
future’). It was common for the illness to be referred to as merely 'an episode' in their life, a bad time, a
blip, or a page in the book of life, and it was important to live your life and not let it ruin the rest of your
life. Some said it was hard to believe now that it had happened to them and that time was a great
healer. People who were still attending hospital check-ups said they felt odd waiting among patients
who were still having cancer treatment.
I often say that there’s no single right way through cancer. What do I mean by that?
Some people aggressively treat their cancers with surgery and chemotherapy long past the time that
others would have switched to comfort measures.
Some people keep their cancer diagnosis a secret from nearly everyone while others make it a point to
tell strangers on the street.
Some people join support groups while others cringe at the thought of the idea.
Some people, when finished with treatment, try hard to not think about cancer ever again, while others
become engaged in cancer organizations and activism.
Some people want to hear only positive stories about cancer, while others want to hear the good, the
bad, and the ugly.
No single approach to cancer is right or wrong. What’s important is that you follow the approach that
helps you.
This is why I don’t like most books written about cancer. They tell you what you should do as if everyone
is the same. They’re written like cookbooks providing recipes.
We’re all different with varying personalities, living situations, and belief systems. In addition, our
cancers are all different. Your breast cancer is unlikely to be exactly like my breast cancer.
And remember that you can and will change your mind over time. A support group might not work for
you when you’re first diagnosed, but be open to the possibility when your treatment is ending.
Finally, no one really knows when they will choose to stop active treatment until they’re in that exact
situation. I can speculate what I would do, but I don’t know for sure. Neither do you.
As for the rest of us, we can support each person with cancer by listening with kindness and without
judgement.
To those people who do not have cancer please, do all the things to prevent and avoid it because cancer
is a “killing machine” you cannot cure it, well there are researches about cancer but we will not let the
time comes that we are going to experience cancer. There are lots of people who have cancer speaking
that they regret because they experienced that kind of illness. So as a citizen I eventhough I didn’t
experienced and hopefully will not triggered we to listen and apply what people with cancer suggest.
We should keep our immune system and eat healthy foods and avoid unhealthy llifestyle.