Professional Documents
Culture Documents
Purpose
Purpose
original report
Multidimensional Telemonitoring of
Symptoms, Body Weight, Sleep, and
Circadian Activity: Relevance for
Chronomodulated Administration of
Irinotecan, Fluorouracil-Leucovorin, and
Oxaliplatin at Home—Results From a Pilot
Study
Purpose To assess the impact of chronomodulated irinotecan fluorouracil-leucovorin and oxalipla-
abstract
tin (chronoIFLO4) delivered at home on the daily life of patients with cancer in real time using a
home-based e-Health multifunction and multiuser platform. This involved multidimensional tele-
monitoring of circadian rest-activity rhythm (CircAct), sleep, patient-reported outcome measures,
Pasquale Innominato
and body weight changes (BWCs).
Sandra Komarzynski
Patients and Methods Patients received chronoIFLO4 fortnightly at home. Patients completed the
Abdoulaye Karaboué 19-item MD Anderson Symptom Inventory on an interactive electronic screen, weighed them-
Ayhan Ulusakarya selves on a dedicated scale, and continuously wore a wrist accelerometer for CircAct and sleep
Mohamed Bouchahda monitoring. Daily data were securely teletransmitted to a specific server accessible by the hospi-
tal team. The clinically relevant CircAct parameter dichotomy index I < O and sleep efficiency (SE)
Mazen Haydar
were calculated. The dynamic patterns over time of patient-reported outcome measures, BWC, I <
Rachel Bossevot- O, and SE informed the oncology team on tolerance in real time.
Desmaris
Results The platform was installed in the home of 11 patients (48 to 72 years of age; 45% men;
Magali Mocquery
27% with performance status = 0), who were instructed on its use on site. They received 26 cy-
Virginie Plessis cles and provided 5,891 data points of 8,736 expected (67.4%). The most severe MD Anderson
Francis Lévi Symptom Inventory scores were: interference with work (mean: 5.1 of 10) or general activity
(4.9), fatigue (4.9), distress (4.2), and appetite loss (3.6). Mean BWC was −0.9%, and mean SE
Author affiliations and remained > 82%. CircAct disruption (I < O ≤ 97.5%) was observed in four (15%) cycles before
support information (if chronoIFLO4 start and in five (19%) cycles at day 14.
applicable) appear at the
end of this article. Conclusion The patient-centered multidimensional telemonitoring solution implemented here was
Corresponding author: well accepted by patients receiving multidrug chemotherapy at home. Moreover, it demonstrated
Francis Lévi, MD, PhD, that chronoIFLO4 was a safe therapeutic option. Such integrated technology allows the design of
Cancer Chronothera- innovative management approaches, ultimately improving patients’ experience with chemothera-
py Team, Division of py, wellbeing, and outcomes.
Biomedical Sciences and
Cancer Research Centre, Clin Cancer Inform. © 2018 by American Society of Clinical Oncology
Office B-166, Warwick
Medical School, Warwick
University, Gibbet Hill
Rd, Coventry CV4 7AL,
United Kingdom; e-mail:
F.Levi@warwick.ac.uk.
© 2018 by American Society of Clinical Oncology ascopubs.org/journal/cci JCO™ Clinical Cancer Informatics 1
LV
80 ble-in-time infusion pump (Melodie; Domocare,
Montmirail, France). The treatment lasted a
70
total of 80 hours (from 2:00 a.m. on day 1 until
60 10:00 a.m. on day 4), and it was repeated every
2 weeks.
50
40
Data Monitoring
30
Visual displays of the rest-activity patterns and
20
quantification of estimated sleep parameters
10 were available on a secured central server,
accessible by the oncology nursing and medi-
0
0:00 12:00 0:00 12:00 0:00 12:00 0:00 12:00 0:00 cal staff. A first prospective alert algorithm was
Time of Day (clock hour) designed and implemented, on the basis of pre-
set threshold values for body weight loss and
MDASI cluster scores increment. The graphical
Fig 1. Administration the Internet. Patients were asked to use the plat- displays and warning alerts were checked daily
profile of the chrono- by the designated hospital cancer nurse special-
form for 30 days.4
modulated irinotecan
ists through a dedicated dashboard. In cases of
fluorouracil-leucovorin
and oxaliplatin (chronoI- lack of data transmission for > 24 hours, alert
FLO4) schedule, showing
Patients about high symptom severity, quick body weight
irinotecan (CPT, red), loss, or apparent deterioration of the rest-activ-
The patients were recruited at the chronother-
oxaliplatin (OHP, gold),
apy clinics in the medical oncology department ity rhythm, the oncology nurse would phone the
fluorouracil (FU, dark
blue), and folinic acid of Paul Brousse Hospital (Villejuif, France), patient and organize any appropriate interven-
(LV, light blue). within the inCASA cancer pilot.4 For the current tion required. The intervention spanned from
report, patients with advanced or metastatic col- telephone reassurance, a home visit by a techni-
orectal or pancreatic cancer with an indication cian or a nurse, a referral to the patient’s general
to be treated with triplet chemotherapy were practitioner or oncologist, to an emergency visit
offered to take part in the study. Availability of at the outpatient clinics or a hospital admission.
an Internet connection at home and signature
of an informed consent were mandatory to be
Data Analysis
eligible. The inCASA study was approved by the
local institutional review board and conducted MDASI scores were analyzed individually as they
according to the Declaration of Helsinki.35 Each were rated by the patients. Each item ranged
patient could be monitored for more than one from 0 (ie, symptom not present) to 10 (ie, as
course of fortnightly chronochemotherapy. bad as imaginable).34 Hence, higher values indi-
cated more severe symptoms. Daily body weight
change (BWC) was calculated with reference to
ChronoIFLO4
the first recorded value. Rest-activity recordings
In the therapeutic regimen chronoIFLO4, each were analyzed using the Action 4 and AW2 Soft-
drug was administered according to a time-stip- ware (Ambulatory Monitoring, Ardsley, NY) to
ulated semisinusoidal profile (Fig 1). Irinotecan compute sleep parameters on each individual
was delivered on day 1, from 2:00 a.m. until night, as well as the dichotomy index I < O.38 I
8:00 a.m., with peak delivery at 5:00 a.m. This < O was computed as the percentage of activity
was followed by a 3-day alternation of 12-hour epochs when in bed whose values were lower
infusions of oxaliplatin (25 mg/m2/d), from than the median level of activity when out of bed.
10:15 a.m. until 9:45 p.m., with peak delivery at Here, I < O was calculated over 72 hours, with
4:00 p.m., and fluorouracil (900 mg/m2/d) and 3-day moving windows, throughout the 14 days
folinic acid (400 mg/m2/d), both administered of duration of each chronotherapy course.4 This
for fluorouracil, 22 courses (84.6%) for oxal- efficiency remained between 81.9% and 90.8%
iplatin, and all courses for leucovorin. No reduc- along the 14-day course (Fig 3B). Nonetheless,
tion below 60% of the protocol dose was ever sleep efficiency was < 70% in approximately one
performed for any drug. Altogether, doses were tenth of the nights. Moreover, individual patterns
reduced for prior acute toxicity, as per routine indicated that poor sleep efficiency was overall
clinical practice. consistently present in specific cycles, with few
instances of single-night low values (Fig 3B). The
average total sleep time ranged from 8.6 hours
Toxicity to 9.7 hours, with a few individual instances of
No grade 5 toxicity occurred. No grade 3 to 4 extremely short or long sleep duration (Fig 3C).
toxicity was encountered, except for two uncom- The average timing of midsleep had a fairly nar-
plicated neutropenic events (Table 1). All the row range (1:02 a.m. to 5:30 a.m.) across the
other toxic events were mild or moderate, with nights assessed, with an overall median value at
grade 2 toxicity occurring in < 20% of patients. 3:31 a.m. (data not shown).
On average, the patients lost 0.3% to 2.4% of
their initial weight (mean, 0.9%), with transient
Multidimensional Data Pattern
grade 1 weight loss (ie, ≥ 5%) occurring in four
Overall, chronoIFLO4 induced minimal alter- instances (1.6% of data) for three individual
ations on the rest-activity circadian pattern, as patients (Fig 2). The three symptoms that were
illustrated by the display of the mean rest-activ- rated as the most severe ones by the patients
ity rhythm throughout the 14-day span of the 26 included fatigue, whose mean severity score
courses (Fig 2). Overall, activity was prominent was 4.9 (range, 3.4 to 6.1); distress, with mean
at daytime and rest at nighttime throughout the severity score of 4.2 (3.5 to 5.1); and appetite
whole chronochemotherapy course, although a loss, with mean severity score of 3.6 (range,
slight dampening was apparent during the sec- 2.4 to 4.8). Patients rated that their symptoms
ond week. Indeed, the dynamics of wrist actig- interfered at a higher level with their work (mean,
raphy–derived parameters, including circadian 5.1), their general activity (mean, 4.9), and their
dichotomy index I < O, total sleep duration, enjoyment of life (mean, 4.3; Fig 2). The mean
sleep efficiency, and midsleep clock hour (not day-to-day patterns of the symptoms previously
shown) revealed mild and transient disturbances shown as associated with circadian disruption48
along the 14-day span (Fig 3). Thus, mean I < O are depicted in Fig 3D. Furthermore, all average
values ranged from 96.3% to 98.5%, with indi- MDASI scores ranked below the midrange value
vidual patterns showing transitory spans with I of 5.5. Only approximately 10% of individual
< O values indicating circadian disruption (ie, ≤ daily item ratings were ≥ 7, with consistently >
97.5%39,40), most often followed by recovery in 20% of ratings being 0 across the fortnight of the
the last days of the cycle (Fig 3A). Average sleep cure (Fig 3E).
Accelerations/min
Continuous wrist actigraphy 125
0
Value
0
1
2
3
MDASI Items
4
5
6
7
Once daily 8
e-MDASI 9
10
5.0
0.0
(%)
Once daily scale –2.5
–5.0
–7.5
Day 1 2 3 4 5 6 7 8 9 10 11 12 13 14
chronoIFLO4
Fig 2. InCASA (Integrated Network for Completely Assisted Senior Citizen's Autonomy) platform outline, including the sensors at the patient’s home
(left) and the output at the hospital interface, with the average plots of the actual data derived over the 14-day span of the course (right). Chemotherapy
(chronomodulated irinotecan fluorouracil-leucovorin and oxaliplatin [chronoIFLO4]) was administered between day 1 and day 4. Top panel: wrist acti-
graph, providing continuous recordings, displayed as average accelerations per minute. Middle panel: touch-screen desktop, with an electronic version
of the MD Anderson Symptom Inventory (MDASI) questionnaire, completed daily, providing 19 items, whose average is displayed as a heat map. It
also collected data from the actigraph via infrared connection and data from the scale via Bluetooth connection and was linked via the Internet to the
bespoken secure server. Bottom panel: electronic scale, used once daily, providing average (blue line) and individual (gray dots) body weight change
(yellow-shaded box represents a ≥ 5% loss, corresponding to grade 1 toxicity, according to the National Cancer Institute Common Toxicity Criteria for
Adverse Events).
6
90
5
4
85
3
2
80
1
0 0
1 2 3 4 5 6 7 8 9 10 11 12 13 14
B 100 Day
chronolFLO4
90
80
70 E
SE (%)
60
MDASI score 0 1-2 3-6 7-10
50
100
Percentage of Instances
40
90
30
80
20
70
60
C 15 50
13 40
TST (hours)
30
11
20
9 10
7 0
5 1 2 3 4 5 6 7 8 9 10 11 12 13 14
following day. These findings, albeit obtained in daily collection.2,64 Most missing data here were
a limited patient sample, bear the potential for related either to technical issues or to the fixed
the development of timely proactive and per- nature of the platform involving a touch-screen
sonalized interventions. These can be attained desktop.4 Such limitations have now been
through coupling predictive algorithms of clini- addressed in a second-generation mobile plat-
cal deterioration or improvement with partici- form (eg, a wireless tablet) involving improved
patory and dynamic bespoken coaching for the technologies, biosensors, and dedicated patient
patient. Thus, such a model of care incorporat- education and services. The current system had
ing self-management could allegedly improve already been evaluated in terms of usability and
patients’ well-being, contributing to the assim- patients’ perception with the Whole Systems
ilation of cancer to a chronic illness.53 How- Demonstrator Service User Technology Accept-
ever, dedicated time-series forecasting analysis ability Questionnaire65 in a subset of end users,
combining multiparametric data are required, whose results displayed an overall satisfaction
together with the testing of an array of potential rate of 84%.4
behavioral interventions, all integrated in patient
and health care professional platforms.4,18,21,54-56 A second-generation platform, on the basis of
This approach has the potential to provide ben- the blueprint of the solution here tested, ought to
efit not only to patients with cancer but also incorporate the monitoring and temporal analysis
patients with other chronic conditions, as pro- of additional biorhythms, including, for instance,
posed in primary care and in particular employ- skin temperature and heart rate variability,66-69
ment conditions.18,57-59 which can further increase the accuracy of the
predictive model and of appropriate response
Moreover, the combination of dose-dense contin-
through broadening the spectrum of physiologic
uous monitoring of physiologic biorhythms and
functions under remote surveillance. It should
repeated assessments of PROMs allows time-se-
also include a measure of the complex construct
ries correlative and predictive analyses, which
of health-related quality of life alongside symp-
can help better understand the physio-patho-
tom assessment.70
genesis of patients’ complaints. We showed
here an example providing further evidence
Altogether, the groundwork for patient-centered,
for an association between circadian rhythms
home-based cancer care involves the design of
and sleep function and systemic symptoms, in
effective and well-tolerated ambulatory chemo-
agreement with existing literature.32,48,49,60-63
therapy regimens, alongside optimized integra-
We acknowledge that a limitation of our study tive telemonitoring. In this study, the dynamics
was the approximately 30% rate of missing of multiple assessments concurred to estimate a
data. This figure was, notwithstanding, similar to fairly limited impact of chronoIFLO4 on patients’
that recently reported for electronic PROMs in physiology and daily life, with usual satisfactory
large studies and better than that observed with recovery between courses.
In conclusion, this pilot study confirmed the of integrated patient-centered solutions, accu-
feasibility and the clinical and scientific rele- rately telemonitoring relevant physiology and
vance of a dense, multidimensional near–real- capturing subjective assessments, thus provid-
time telemonitoring, even in patients receiving ing innovative insight to improve patients’ experi-
an aggressive anticancer regimen. In addition, ence with chemotherapy, safety, and outcomes.
chronoIFLO4 displayed an adequate safety pro-
DOI: https://doi.org/10.1200/CCI.17.00125
file, despite being administered at home. Alto- Published online on ascopubs.org/journal/cci on February
gether, these findings support the development 22, 2018.
Affiliations
Pasquale Innominato, North Wales Cancer Centre, Betsi Cadwaladr University Health Board, Denbighshire; Pasquale
Innominato, Sandra Komarzynski, and Francis Lévi, Warwick Medical School, Coventry, United Kingdom; Pasquale
Innominato, Sandra Komarzynski, Ayhan Ulusakarya, Mohamed Bouchahda, and Francis Lévi, Institut National de la Santé et
de la Recherche Médicale, Unit 935; Ayhan Ulusakarya, Mohamed Bouchahda, Mazen Haydar, Rachel Bossevot-Desmaris,
Magali Mocquery, Virginie Plessis, and Francis Lévi, Assistance Publique–Hôpitaux de Paris, Paul Brousse Hospital,
Villejuif; and Abdoulaye Karaboué, AK-SCIENCE, Research and Therapeutic Innovation, Vitry-sur-Seine, France.
Support
Supported by the European Commission through the Information and Communication Technologies Policy Support
Programme project inCASA (Integrated Network for Completely Assisted Senior Citizen's Autonomy) Contract No. CIP
250505, Framework Programme for Research and Technological Development (FP) 7 and the Coordinating Action on
Systems Medicine Contract No. 305033 CaSyM, FP7 SP1 HEALTH; the Conseil Régional d’Ile de France and Banque
Publique d’Investissement of France, through the Fonds Unique Interministériel 12 Contract No. PiCADo; and the
Institut de Recherche en Santé Publique Contract No. CLOCK-DOM1.
Prior Presentation
Presented in part at the European Society of Medical Oncology annual meeting, Madrid, Spain, September 8-12, 2017.
REFERENCES
1. Basch E, Deal AM, Kris MG, et al: Symptom monitoring with patient-reported outcomes during
routine cancer treatment: A randomized controlled trial. J Clin Oncol 34:557-565, 2016
2. Wysham NG, Wolf SP, Samsa G, et al: Integration of electronic patient-reported outcomes into
routine cancer care: An analysis of factors affecting data completeness. JCO Clin Cancer Inform
doi:10.1200/CCI.16.00043
3. Basch E, Dueck AC, Rogak LJ, et al: Feasibility assessment of patient reporting of symptomatic
adverse events in multicenter cancer clinical trials. JAMA Oncol 3:1043-1050, 2017
4. Innominato PF, Komarzynski S, Mohammad-Djafari A, et al: Clinical relevance of the first
domomedicine platform securing multidrug chronotherapy delivery in metastatic cancer patients
at home: The inCASA European Project. J Med Internet Res 18:e305, 2016