Download Detecting lung cancer relapse using self

Support Care Cancer
DOI 10.1007/s00520-013-1954-9
ORIGINAL ARTICLE
Detecting lung cancer relapse using self-evaluation forms
weekly filled at home: the sentinel follow-up
Fabrice Denis & Louise Viger & Alexandre Charron &
Eric Voog & Christophe Letellier
Received: 6 May 2013 / Accepted: 21 August 2013
# Springer-Verlag Berlin Heidelberg 2013
Abstract
Purpose We aimed to assess if patients' ratings of symptoms
can be used to provide an early indication of disease
recurrence or progression in lung cancer. We proposed a
simple self-evaluation form made of six clinical parameters
weekly scored by patients at home as a follow-up—here
named sentinel—to improve relapse detection. Its performances were compared to those of a routine imaging
follow-up.
Methods Patients with lung cancer were prospectively recruited
to weekly fill a form at home for self-assessing weight, fatigue,
pain, appetite, cough, and breathlessness during at least
4 months. Each patient reported weight and assessed the severity
of each symptom by grading it from 0 (no symptom) to 3 (major
symptom). A score was retrospectively designed for discriminating patients with relapse from those without. Accuracy of
relapse detection was then compared to values of the routine
planned imaging.
Results Forty-three patients were included in our center and
recruited for 16 weeks or more follow-up during which at least
one tumor imaging assessment was performed (CT scan or
PET-CT). Forty-one completed the form weekly. Sensitivity,
specificity, and positive and negative predictive values of
sentinel were high (86, 93, 86 % and 93 vs 79, 96, 92, and
90 % for routine imaging—p =ns) and well correlated with
relapse (pχ 2>0.001). Moreover, relapses were detectable
with sentinel on average 6 weeks earlier than the planned
imaging.
F. Denis : A. Charron : E. Voog
Jean Bernard Center/Victor Hugo Clinic, 9 rue Beauverger,
Le Mans, France
F. Denis (*) : L. Viger : C. Letellier
CORIA—University of Rouen, Av. de l’Université, BP 12,
76801 Saint-Etienne du Rouvray cedex, France
e-mail: [email protected]
Conclusion This study suggests that a personalized cancer
follow-up based on a weekly self-evaluation of six symptoms
is feasible and may be accurate for earlier detection of lung
cancer relapse, allowing integration in electronic devices for
real-time patient outcome follow-up.
Keywords Lung cancer . Follow-up . Supportive care .
Personalized medicine . Early relapse detection
Introduction
Lung cancer is the first cause of cancer-induced mortality in
the world and remains a poor prognosis [1, 2]. Although
pretreatment imaging evaluations are well-defined for the
diagnosis of lung carcinoma, very few studies were devoted
to the follow-up of patients after the completion of a therapy or
during maintenance therapy [3–5]. Moreover, there is no
personalized follow-up, and there is no evidence yet that an
expensive routine imaging-based follow-up may improve
prognosis of patients who could relapse. There is no randomized trial result comparing intensive follow-up using CT scan
and symptoms-only follow-up for relapse detection.
Nevertheless, a Brazilian retrospective study compared a strict
follow-up with frequent visits, imaging, and laboratory examinations in patients who underwent a complete resection of
non-small cell lung cancer to a follow-up with infrequent
visits that were scheduled mainly on the basis of patient's
symptoms evaluated during visits at the hospital [6]. No
difference in the survival outcome was noted, but symptom
follow-up was found to be more cost-effective than routine
imaging follow-up. Few relapses are indeed associated
with curative intent in this cancer, and at least 75 % of
them are symptomatic [7–9]. Some symptoms also appear
to have prognostic value in determining clinical course
and survival [10].
Support Care Cancer
In our institution, from July 30, 2012 to September 30, 2012, a
form designed for our “sentinel” follow-up was given to 43
consecutive symptomatic (or not) patients with surgical excision, complete response, or detectable but nonprogressive
lung carcinoma. Patients receiving chemotherapy were excluded except if they received bevacizumab maintenance.
An example of the form is shown in Fig. 1. Patient characteristics are reported in Table 1. Median age was 62 years (min,
42 years; max, 82 years). Fourteen patients had distant metastasis. Thirty-five patients received no treatment and eight were
under a maintenance treatment by bevacizumab for metastatic
disease during our study. Twenty-four patients had no detectable tumor by CT scan or PET-CT, and 19 had nonprogressive
detectable tumors. Thirty-seven patients had tumor initial
assessments by CT scan and six by PET-CT.
Patients were asked to report weight and five symptoms,
which were appetite loss, fatigue, pain, cough, and breathlessness, weekly. Words that were used for patient's self-assessment
were “fatigue” (same word as in French) to describe asthenia
and “breathlessness” (“essoufflement” in French) to describe
dyspnea. All terms were clearly explained by a physician
during the study entry visit. The five symptoms had to be
graded the same day as the weight was recorded. These six
items were reported in a form during at least 16 weeks. Each
patient self-assessed the severity of symptoms by grading them
from 0 (no symptom) to 3 (major symptoms). Weight loss was
graded by the physician as follows: grade 0 for no weight loss,
grade 1 for 1 kg loss since patient recruitment, grade 2 for 2 kg
loss, and grade 3 for 3 kg loss or more. Data collection was
prospectively performed.
Physical examination was performed during follow-up, every
16 weeks or earlier if needed (due to a patient call, for instance).
We also collected filled forms when planned PET-CT or CT scan
was performed during routine follow-up, on average 16 weeks
after the beginning of the study. PET-CT was systematically
used when CT scan was positive and biopsy/cytology was
performed when needed. PET-CT or other modalities was also
used in symptomatic patients without CT scan abnormalities. In
our institution, physical assessment and CT scan are repeated
every 16 weeks for all patients during maintenance therapy or
following combined radio-chemotherapy or following curative
surgery for 2 years, then it is repeated every 6 months for 3 years.
Due to the lack of a gold standard to follow-up patients with
lung cancer, we choose to combine all examinations to serve as a
reference. We then compared the efficiency of the common
follow-up in our hospital to sentinel.
Fig. 1 Example of a form filled by our patient. True negative case
with the sentinel follow-up. This patient had stage 4 NSCLC and
had a persistent complete response after chemotherapy. He received
bevacizumab maintenance during this study. Stable symptoms were
reported with “sentinel” during 20 weeks. No progression was observed with planned follow-up CT scan at 16 weeks. Self-graded
symptoms were associated with colors to improve visual assessment
by the physician
Up to now, one of the most common strategies for a routine
follow-up is to track the evolution of possible tumors by
performing regular clinical assessments with or without routine imaging at few months interval. In the case of lung cancer,
the most obvious symptoms are pain, breathlessness, coughs,
fatigue, and appetite [11]. For numerous cancers, one of the
most important parameters for the prognostic is the weight
loss [12, 13].
The objective of the present study was therefore to test
whether a personalized cancer follow-up based on a weekly
self-evaluation of six symptoms was reliable and helpful for
detecting cancer relapse. We thus designed a follow-up based
on patients' ratings of symptoms—here designated by “sentinel”—for early relapse detection. Its performances were then
compared to those of a routine planned follow-up imaging
based on CT scan or PET-CT.
Methods
Support Care Cancer
Table 1 Patient characteristics
Characteristics
Males/females
Pathology
Squamous cell
Adenocarcinoma
Small cell carcinoma
Large cell carcinoma
Initial Stage
I/II
IIIA
IIIB
IV
Detectable disease at the beginning of the study
Yes
No
Previous treatment
Surgery
Radiotherapy
Concomitant radio-chemotherapy
Chemotherapy
Disease status at study entry
Complete response after surgery
Partial or complete response after radio-chemotherapy
Stable disease after radio-chemotherapy
Partial or complete response after palliative chemotherapy
Stable disease after palliative chemotherapy
Treatment ongoing
None
Maintenance therapy
Planned imaging
CT scan
PET-CT
N =43
38/5
15
20
4
4
3
21
2
14
19
24
5
2
22
14
5/5
19/22
3/22
8/14
6/14
35
8
37
6
The dynamics underlying symptoms were retrospectively
investigated from the forms. The global score of the six clinical
parameters (the five symptoms plus the weight grade) was
computed for each week and for each patient (see Figs. 1 to 2,
for examples). The sentinel returns retrospectively a positive
relapse with the highest sensitivity (93 %) when at least one of
the three following conditions was verified:
1. Three kilogram weight loss during at least 2 weeks
2. Two symptoms simultaneously equal to grade 3 during at
least 1 week
3. A global score greater than or equal to 7 during at least
2 weeks
These conditions were tested for detecting relapse only
once the global score was less than 7 for 1 week. For instance,
a patient returned a global score greater than 6 during the first
3 weeks after inclusion and less than 7 at the fourth week: this
procedure was thus only applied starting from week 4.
This procedure was retained after a trial and error approach.
It allows detecting cancer relapse with the highest sensitivity,
specificity, and positive and negative predictive values.
Youden index and Q Yule coefficients were also computed
for all types of relapse (locoregional or distant failure). Results
were then compared to relapse detections obtained with the
routine imaging follow-up and with the complementary PETCT or pathological/cytological assessments if needed. Brain
MRI was only performed in the presence of neurologic symptoms and signs on physical examination. The Fisher exact test
was used to compare sensitivity, specificity, and positive and
negative predictive values between “sentinel” and routine
imaging follow-up. χ 2 test was used in 2×2 tables to assess
statistical association between the disease (relapsing or not)
and the “sentinel” or imaging results. All tests were two-sided
and “p” value was considered significant when less than 0.05.
Before inclusion, all patients signed written informed consent and “Informatic and Liberty National Commission” form
allowing data deletion according to patient's wish. The study
followed guidelines of our institutional ethical committee.
Results
The study included 43 consecutive patients. Forty-one were
compliant and filled exhaustively their form. Two patients
forgot or refused to fill their form. Patients were followed up
during at least 16 weeks (median time, 5 months), and all
presented at least one tumor imaging after reference imaging
assessment (which was performed at study initiation). Due to
relapse-associated acute symptoms, one patient visited the
oncologist earlier (4 weeks only after the beginning of the
study) and a CT scan was thus performed. Note that “sentinel”
detected the relapse because this patient reported grade 3
anorexia, grade 2 asthenia, and grade 3 pain caused by liver
metastasis. Weight loss was scored with grade 2: this patient
thus fulfilled condition II.
Fourteen patients presented a tumor relapse/progression as
determined by combining clinical imaging (CT scan and PETCT) and cytological/pathological examinations. All patients
who relapsed (n =14) were symptomatic and 12 of them were
positive according to “sentinel” (86 %). Two were positive for
condition II alone (2 symptoms graded to 3), 4 were positive
for condition III alone (global score ≥7 for at least 2 weeks), 1
was positive for both conditions I (3 kg of weight loss during
2 weeks) and II, 2 were positive for both conditions I and III, 2
were positive for conditions II and III, and 1 was positive for
conditions I, II, and III. Routine imaging (CT scan for 33
evaluable patients and PET-CT for the others) did not allow
detecting relapse in three patients. One patient had pleural
Support Care Cancer
Fig. 2 Example of a true positive
“sentinel” detection, which was a
false negative case with a routine
imaging follow-up: this patient
reported progressive weight loss
and asthenia. Planned CT scan
and PET-CT were negative.
Lumbar puncture led to discover
metastatic meningitis. Mild but
detectable symptoms thus
appeared 4 weeks before planned
imaging
infusion, which was stable as assessed by CT scans (at the
inclusion in this protocol and 16 weeks later), but cough
increased and pleural cytology became positive. The second
patient presented an increase of his pain but CT scan (the
routine imaging) did not show any tumor progression. Only
a complementary PET-CT and a biopsy allowed detecting a
cancer relapse. These two patients were detected using the
sentinel follow-up. A third patient had no relapse detected
with the routine imaging (CT scan and PET-CT were negative)
but was positive according to sentinel: he had a meningeal
carcinomatosis diagnosed using a brain MRI (Fig. 2).
Sentinel results were then compared to routing imaging
(CT scan or PET-CT) and/or cytology. We thus found retrospectively 12 true positive patients, 25 true negative patients,
2 false positive and 2 false negative patients (pχ 2 <0.001). For
all true positive “sentinel” patients, symptoms were detectable
on average 6 weeks before routine imaging (min, 1 week;
max, 12 weeks). The two false negative patients returned by
“sentinel” had symptoms. One had isolated superior cave
syndrome 16 weeks after the study initiation caused by mediastinal relapse. The second false negative case corresponds to
a patient presenting a grade 2 dyspnea and asthenia the week
of the planned routine imaging. Symptoms just started to be
self-assessed with increasing grades by the patient when the
routine imaging was scheduled. He had pericardial metastasis
and died 1 month later. This is one example for which the
routine imaging was planned at the occurrence of symptoms:
in such a case, sentinel would have detected this case 1 or
2 weeks later.
Sensitivity, specificity, and positive and negative predictive
values, Youden index, and Q Yule coefficient for evaluating
sentinel results versus those obtained with a routine planned
imaging are reported in Table 2. Sensitivity, specificity, and
positive and negative predictive values were 86, 93, 86, and
93 % for “sentinel” and 79, 96, 92, and 90 % for planned
imaging, respectively. There were no significant statistical
difference (p =ns) between “sentinel” and the routine imaging
follow-up for relapse detection. Patients identified as false
positive (false negative) by sentinel were not the same as the
patient identified as false positive (false negative) by routine
imaging.
We also calculated sensitivity, specificity, and positive and
negative predictive values, Youden index, and Q Yule coefficient of “sentinel” for evaluable patients with (n =18) or
without (n =23) detectable disease at the beginning of the
study. The relationship between the disease (progression/relapse or not) and the “sentinel” assessment remains significant
in both populations (pχ 2 <0.01 in patients without detectable
disease and pχ 2 <0.001 in patients with initial detectable
disease) as reported in Tables 3 and 4. Sentinel tended to be
more accurate than imaging in patients having detectable
disease than in patient without initial nondetectable disease
but the difference was nonsignificant.
Discussion
To the best of our knowledge, this study is the first to test a
personalized symptom-based follow-up in lung cancer. It suggests that weekly self-scored patient symptom dynamics may
lead to a follow-up with an excellent compliance and completeness and may be very helpful to detect cancer relapse or
tumor progression. A variety of assessment instruments have
been created to assess and follow up cancer symptoms.
However, there is no ideal instrument, and the wide variety
of instruments reflects the different settings for symptom
assessment as reported in a wide review by Kirkova and
colleagues [14]. However, a recent study showed that realtime data collection of patient with cancer is feasible and
allowed to monitor symptoms to assess toxicity and tumor
response [15]. Our study showed that relapse detection may be
performed using self-assessment of few symptoms which
Support Care Cancer
Table 2 Two by two tables and
results for evaluable patients
(n =41) with the sentinel and
the routine imaging (CT scan or
PET-CT) when compared with
a relapse/progression detection
using all modalities. Chisquared test indicates that there
is a significant association between the two forms of assessment and the disease (relapse or
no relapse)
Relapse
No relapse
12
2
11
3
2
25
1
26
Sentinel (%, CI 95)
Routine imaging (%, CI 95)
Sensitivity
Specificity
Positive predictive value
86 % (67–100)
93 % (83–100)
86 % (67–100)
79 % (57–100)
96 % (89–100)
92 % (77–100)
Negative predictive value
Youden Index
Q Yule coefficient
pχ 2
93 % (83–100)
0.78
0.97
<0.001
90 % (77–100)
0.75
0.98
<0.001
Sentinel positive
Sentinel negative
Imaging positive
Imaging negative
could be sent by the patients to the oncologist: relapse detection may thus be reliably observed almost in real time and, in
average, earlier than routine imaging. Self-evaluated symptoms were weight, fatigue, pain, appetite, cough, and breathlessness. Fatigue mainly results from a lack of sleep, and
weakness would have been more appropriate since it is more
related to the lack of energy induced by the tumor growth. Our
form to be filled up by the patients will thus be improved by
replacing “fatigue” by “weakness”. “Breathlessness” is commonly induced by a physical effort. “Shortness of breath” is
induced by a respiratory failure (as induced by a lung cancer,
for instance) and would be more rigorous. Nevertheless, all
our patients clearly defined what breathlessness is but may
have some difficulties to recognize what “shortness of breath”
is. This is why we choose to misuse “breathlessness.”
We thus used an original approach by prospectively using
self-evaluation forms weekly filled at home by the patients
having lung carcinoma and high risk of relapse. A score
Table 3 Two by two tables and
results for evaluable patients
without detectable disease
(n = 23) (with and without
maintenance therapy) at the beginning of the study. Chi-squared
test indicates that there is a significant association between the
two forms of assessment and the
disease (relapse or no relapse)
strongly associated with relapse was thus obtained. It can be
integrated in electronic devices such as patient smartphone.
In our population, we found that the sentinel score may be as
reliable as a routine planned imaging follow-up. Moreover, on
average, relapses detection (according to symptoms) were reported with “sentinel” 6 weeks before planned imaging thus
suggesting that sentinel would allow more efficient treatment
since it was initiated earlier than routine imaging follow-up. It
is indeed important to note that all symptomatic relapsing patients but one have waited oncologist planned visit and imaging
while symptoms began sometime many weeks earlier. This
suggests that an earlier adapted oncological or symptomatic
treatment could have been provided. The symptoms request
substantial needs of supportive care, and numerous patients do
not visit their physician, although symptoms may be severe when
associated with relapse [16]. Earlier alert by an electronic version
of sentinel would therefore help to prevent clinical deterioration
of patient before initiating treatment.
Relapse
No relapse
Sentinel positive
3
2
Sentinel negative
Imaging positive
Imaging negative
1
3
1
17
0
19
Sentinel (%, CI 95)
Routine imaging (%, CI 95)
75 % (32–100)
89 % (75–100)
60 % (17–100)
94 % (83–100)
0.64
0.92
<0.01
75 % (32–100)
100 %
100 %
95 % (85–100)
0.75
1
<0.001
Sensitivity
Specificity
Positive predictive value
Negative predictive value
Youden Index
Q Yule coefficient
pχ 2
Support Care Cancer
Table 4 Two by two tables and
results for evaluable patient with
detectable disease (n =18) with
and without maintenance therapy
at the beginning of the study. Chisquared test indicates that there is
a significant association between
the 2 forms of assessment and the
disease (relapse or no relapse)
Relapse
No relapse
9
1
8
1
0
8
2
7
Sentinel (%, CI 95)
Routine imaging (%, CI 95)
Sensitivity
Specificity
Positive predictive value
90 % (71–100)
100 %
100 %
80 % (55–100)
88 % (66–100)
89 % (69–100)
Negative predictive value
Youden Index
Q Yule coefficient
pχ 2
89 % (69–100)
0.9
1
<0.001
78 % (51–100)
0.68
0.93
<0.01
Sentinel positive
Sentinel negative
Imaging positive
Imaging negative
Sentinel may also be useful to limit false positive alerts and
may reduce the anxiety of patient due to slight symptoms because they would be aware that the physician is kept informed
continuously about their global status.
Symptomatic relapse (or progression) rate in our study
(100 % of relapse) is consistent with other published data. In
their study, Gilbert et al. showed that in 245 patients having
earlier stages (1 to 2B) who followed up after operation,
80 among 111 patients who relapsed were symptomatic
(72 %) and 67 % of relapses were diagnosed by a generalist
physician [8].
Our population was heterogeneous since it combined patients
with or without maintenance treatments. We choose to incorporate patients treated by maintenance bevacizumab due to the
good tolerance of this treatment and because it does not interact
with patient's relapse-induced symptoms. On the other hand,
“sentinel” was not used in patients having more “aggressive”
chemotherapy that would have induced symptoms perturbing
“sentinel” results. The size of the cohort is small, but χ 2 test and
Yule Q, coefficient that measure the link between the disease and
“sentinel”, results are quite convincing even in the subgroups of
patients with initial detectable or nondetectable tumors.
Lung cancer is a disease for which intensive follow-up
does not necessarily improve survival because there is often
no effective therapy against a recurrence. This expectation is
suggested by studies of follow-up in small cohorts of patients
curatively resected or irradiated non-small cell lung carcinoma
[3–6, 17]. Hence, patients with detectable but nonprogressive
tumors were incorporated in our study because no documented standard for the surveillance of these patients were set in
the NCCN, ACCP ASCO, or ESMO guidelines [18, 19]. We
thus think that a personalized follow-up based on weekly selfassessed symptoms should be developed for this population in
association with a routine imaging. A large randomized clinical trial conducted by the French Intergroup of Thoracic
Oncology (IFCT 0302) is in progress in France to compare,
in postoperative situations, a follow-up made of clinical and
thoracic X-ray planned every 6 months, and a more intensive
follow-up that adds CT scan and endoscopy to clinical examination every 6 months on survival. However, in this study,
planned visits to oncologists occur only every 6 months,
which thus provide only sparse data to detect tumor relapse.
Compare to such a follow-up, clinical data as defined in
sentinel could have been sent weekly to the oncologist by a
smartphone for instance and thus treated in real time by him.
This concept is already used in cardiology (by monitoring
blood pressure), diabetology (by monitoring glycemia), and
in nutrition care (by monitoring weight variation).
Due to the good compliance provided by our patients, sentinel could have improved the efficacy of a routine imaging by
avoiding many useless imaging. It could be useful too for
individualizing patient's follow-up thus avoiding anxiety producing unnecessary and expensive imaging for nonsymptomatic
and nonrelapsing patients. Reports of patients could appear
subjective or nonspecific because patients assessed their symptoms themselves, and benign disease may cause symptoms close
to those leading to conclude to a relapse. We only found two
false positive patients in our study. One had simultaneously
bronchitis and gastroenteritis, and the other had 3 weeks bronchitis. Subjectivity was limited by the patient weight, which is
an objective measure and strongly related with relapse when
associated with fluctuations of other symptoms. Moreover, our
conditions are mainly based on the dynamics (including duration) of symptoms and not only on the values of grades at a
given time. This real-time assessment of symptoms and their
underlying dynamics by patient may be useful to avoid underestimation of symptoms by physician or overestimation by
family caregivers as it appears to be in clinical practice [20].
Our study reveals a very strong agreement between patients'
ratings of symptoms and tumor dynamics. This approach
Support Care Cancer
occurs to be in agreement with the new theoretical analysis of a
chaotic cancer model describing the interactions between host
and tumor cells [21]. In this model, the best variable to observe
the cancer dynamics is sometimes the population size of host
cells and not necessarily the population size of tumor. At patient
body scale, this suggests that symptoms or weight loss, quite
related to the behavior of host cells, could be sometimes better
to observe tumor dynamics on host than observing the tumor
itself. This assumption is well supported by the present study
and others [12, 13].
Some studies suggest that symptom improvement could be
a prognostic factor for survival in cancer patients undergoing
palliative care. Sentinel could thus be used for that purpose
[22]. Moreover, as symptoms are prevalent and severe among
patients with cancer [23], clinical study seeking to evaluate the
impact of treatment on patients (thus including measurement
of symptoms) in real time are promising [15].
A prospective study is in progress to optimize our sentinel
follow-up and to test if sentinel improves the routine imaging
follow-up or not. This study is in progress and will also assess
long-term compliance and satisfaction of patients and physicians using an electronic version of “sentinel” to report symptoms and weight changes over time using their smartphone or
PC. Alerts are automatically transmitted to the oncologist. We
hopefully believe that sentinel and further real-time data collection for seeking tumor relapse open a new personalized
follow-up in lung cancer.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
Acknowledgments None
Conflicts of interest There are no financial disclosures, conflicts of
interest, for the authors, and no funding sources for the manuscript.
16.
17.
References
1. French National Institute report (2013) Survival of patients with
cancer 1989 – 2007
2. Howlader N, Noone AM, Krapcho M, Neyman N, Aminou R, et al.
(eds) (2012) SEER Cancer Statistics Review, 1975–2009 (Vintage 2009
Populations), National Cancer Institute. Bethesda, MD. Based on
November 2011 SEER data submission, posted to the SEER web site,
April 2012. Available: http://seer.cancer.gov/csr/1975_2009_pops09/.
Accessed 09 January 2012
3. Rubins J, Unger M, Colice GL (2007) Follow-up and surveillance of the
lung cancer patient following curative intent therapy, ACCP evidencebased clinical practice guideline (2nd edn.). Chest 132:355S–367S
4. Crino L, Weder W, van Meerbeeck J, Felip E (2010) On behalf of the
ESMO guidelines working group. Early stage and locally advanced
(non-metastatic) non-small-cell lung cancer: ESMO clinical practice
guidelines for diagnosis, treatment and follow-up. Ann Oncol
21(Suppl 5):v103–v115
5. Van Loon J, Grutters J, Wanders R et al (2009) Follow-up with
18FDG-PET-CT-CT after radical radiotherapy with or without chemotherapy allows the detection of potentially curable progressive
18.
19.
20.
21.
22.
23.
disease in non-small cell lung cancer patients: a prospective study.
Eur J Cancer 45:588–595
Younes RN, Gross JL, Deheinzelin D (1999) Follow-up in
lung cancer: how often and for what purpose? Chest 115(6):
1494–1499
Walsh GL, O'Connor M, Willis KM et al (1995) Is follow-up of lung
cancer patients after resection medically indicated and cost-effective?
Ann Thorac Surg 60(6):1563–1570
Gilbert S, Reid KR, Lam MY et al (2000) Who should follow
up lung cancer patients after operation? Ann Thorac Surg
69(6):1696–1700
Westeel V, Choma D, Clement F et al (2000) Relevance of an
intensive postoperative follow-up after surgery for non-small cell
lung cancer. Ann Thorac Surg 70(4):1185–1190
Walsh D, Rybicki L, Nelson KA, Donnelly S (2002) Symptoms
and prognosis in advanced cancer. Support Care Cancer 10:
385–388
Donnelly S, Walsh D, Rybicki L (1995) The symptoms of advanced
cancer: identification of clinical and research priorities by assessment
of prevalence and severity. J Palliat Care 11(1):27–32
Ross PJ, Ashely S, Norton A, Priest K, Waters JS, Eisen T, Smith IE,
O'Brien MER (2004) Do patients with weight loss have a worse
outcome when undergoing chemotherapy for lung cancers? Br J
Cancer 90:1905–1911
Khalid U, Spiro A, Baldwin C, Sharma B, McGough C,
Norman AR, Eisen T, O'Brien ME, Cunningham D, Andreyev
HJ (2007) Symptoms and weight loss in patients with gastrointestinal and lung cancer at presentation. Support Care Cancer
15(1):39–46
Kirkova J, Davis MP, Walsh D, Tiernan E, O'Leary N, LeGrand SB
et al (2006) Cancer symptom assessment instruments. J Clin Oncol
24(9):1459–1473
Pietanza MC, Basch EM, Lash A et al (2013) Harnessing technology to improve clinical trials: study of real-time informatics to
collect data, toxicities, image response assessments, and patientreported outcomes in a phase II clinical trial. J Clin Oncol 31(16):
2004–2009
Liao YC, Liao WY, Shun SC, Yu CJ, Yang PC, Lai YH (2011)
Symptoms, psychological distress, and supportive care needs in lung
cancer patients. Support Care Cancer 19(11):1743–1751
Yumuk PF, Mohammed N, Maat AP et al (2012) How do lung
cancer specialists follow their patients with stage II NSCLC
after definitive treatment ? a short report. Eur J Cancer 48(14):
2163–2165
Pfister DG, Johnson DH, Azzoli CG et al (2004) American society of
clinical oncology treatment of unresectable non-small-cell lung cancer guideline: update 2003. J Clin Oncol 22:330–353
NCCN Guidelines™ Version 1.2012 Non-Small Cell Lung Cancer
Oechsle K, Goerth K, Bokemeyer C, Mehnert A (2013) Symptom
burden in palliative care patients: perspectives of patients, their
family caregivers, and their attending physicians. Support Care
Cancer 21(7):1955–1962
Letellier C, Denis F, Aguirre LA (2013) What can be learned from a
chaotic cancer model ? J Theoretical Biology 322:7–16
Narducci F, Grande R, Mentuccia L, Trapasso T, Sperduti I, Magnolfi
E, Fariello AM, Gemma D, Gamucci T (2012) Symptom improvement as prognostic factor for survival in cancer patients undergoing palliative care: a pilot study. Support Care Cancer 20(6):1221–
1226
Reilly CM, Bruner DW, Mitchell SA, Minasian LM, Basch E, Dueck
AC, Cella D, Reeve BB (2013) A literature synthesis of symptom
prevalence and severity in persons receiving active cancer treatment.
Support Care Cancer 21:1525–1550