Remote Oncology Care: Review of Current Technology and Future Directions

Cancer patients frequently develop tumor and treatment-related complications, leading to diminished quality of life, shortened survival, and overutilization of emergency department and hospital services. Outpatient oncology treatment has potential to leave cancer patients unmonitored for long periods while at risk of clinical deterioration which has been exaggerated during the COVID19 pandemic. Visits to cancer clinics and hospitals risk exposing immunocompromised patients to infectious complications. Remote patient reported outcomes monitoring systems have been developed for use in cancer treatment, showing benefits in economic and survival outcomes. While advanced devices such as pulmonary artery pressure monitors and implantable loop recorders have proven benefits in cardiovascular care, similar options do not exist for oncology. Here we review the current literature around remote patient monitoring in cancer care and propose the use of reliable devices for capturing and reporting patient symptoms and physiology.


Introduction And Background
Cancer is a leading cause of morbidity and mortality in the United States, with over 1.8 million new cancer diagnoses and more than 600,000 cancer deaths estimated for 2020 [1]. Furthermore, cancer care significantly impacts the overall healthcare system through high rates of emergency department utilization, hospital admission, and costly treatments whose toxicities have potential to diminish quality of life [2][3][4].
Cancer patients frequently experience local and systemic symptoms causally related to their malignancy. Local symptoms are due to direct complications of the primary tumor or metastases and can include pain, neurological deficits, respiratory symptoms (cough, shortness of breath, hemoptysis) and obstruction (bowel, biliary, airway). Systemic complications include cachexia, paraneoplastic syndromes, electrolyte abnormalities, metabolic alterations, and hematologic changes. These symptoms are a frequent cause of hospitalization. In a series reported by Numico et al., 74% of hospitalizations of cancer patients were for conditions related to tumor involvement, with a minority for workup or treatment-related complications. The most common symptoms at the time of admission were dyspnea, pain, neurological (not specified by authors), fever, and gastrointestinal (vomiting, jaundice) [5]. Identifying and addressing cancer symptom burden prior to the need for admission is recognized as a critical need and has led to increased utilization of patient-reported outcome (PRO) tools discussed in detail below.
In addition to tumor-related morbidity, cancer patients suffer treatment-related complications. Given the myelosuppressive nature of radiation and chemotherapy, anemia, neutropenia, and thrombocytopenia with associated malaise, infection, and bleeding can occur. One of the most serious complications, neutropenic sepsis, is associated with up to 50% mortality rate, average hospital cost of over $49,000, and total expenditures of $1.1 billion per year in the US [6,7]. Less severe though far more prevalent toxicities such as nausea, vomiting, and diarrhea can have a large negative impact leading to volume depletion, metabolic and electrolyte imbalances and renal failure. Newer immunotherapies are associated with immunerelated adverse events with unpredictable timing and vagueness of symptoms, making identification and management especially challenging [2][3][4].
Modern technology can increase patients' connectivity to the healthcare system through mobile communications and remote physiologic monitoring [8]. Monitoring systems have been used most extensively for cardiovascular diseases such as congestive heart failure (CHF) and arrhythmia detection. For patients that require long-term monitoring for either of these, implantable devices like CardioMEMS™ and implantable loop recorders (ILR) have been shown to decrease rates of hospitalization, improve arrhythmia detection, and lower costs compared with usual care [9,10]. In oncology, while implantable devices are not available, studies have shown that monitoring patient-reported outcomes reduces visits to the emergency department, decreases follow-up costs and improves overall survival [11][12][13][14].
Since the outbreak of a novel SARS-CoV-2 in 2019 (Covid-19), the term "social distancing" has entered the common lexicon. However, social distancing has been a mainstay of oncologic care for decades as immunosuppressed cancer patients take precautions to minimize risk of infections. Patients are instructed to eliminate interactions with sick contacts and avoid large gatherings. Yet at the same time, patients attend frequent visits to outpatient cancer centers for clinical evaluations. For a neutropenic patient, each visit poses a risk for infectious exposure. The latter is further magnified since the onset of the Covid-19 pandemic. As a result, the Centers for Medicare and Medicaid services (CMS) has broadened its coverage of Telehealth services under the 1135 waiver authority and coronavirus preparedness and Response Supplementation Appropriation Act, which has significantly expanded the volume of patients receiving care via telemedicine [15]. Though this has reduced oncology patients' infectious exposure, it has created the dilemma for laboratory and vital sign monitoring for patients at risk for drug-induced toxicities. Tools that facilitate social distancing while maintaining connectivity to the healthcare system and providing objective data for ongoing management are urgently needed [16,17].

Monitoring systems for oncology
Representative prospective randomized studies of remote and/or electronic PRO in cancer care are summarized in Table 1. PRO-CTCAE™ (Patient-Reported Outcomes version of the Common Toxicity Criteria for Adverse Events) is a validated tool used to monitor and report toxicities related to cancer treatment in clinical trials. Basch et al. reported results from a prospective randomized study from Memorial Sloan Kettering Cancer Center evaluating the efficacy of an online symptom reporting tool [13,14]. A total of 766 patients with advanced solid tumors undergoing systemic cytotoxic chemotherapy were randomized to either an online symptom reporting platform or usual care. In the experimental arm, patients received weekly email prompts to report on 12 common treatment-related toxicities through a web-based portal called Symptom Tracking and Reporting (STAR). A severe or marked change in symptom reporting prompted an email alert to an oncology nurse; summary reports were made available to the treating physician at the time of clinical visits. Over the course of the study, the STAR platform was associated with an improvement in health-related quality of life scores (34% vs 18%, p < 0.001), with fewer visits to the emergency department (34% vs 41%, p = 0.02) compared to standard of care. Median overall survival was also 20% longer in the PRO arm (31.2 months vs 26.0 months, p = 0.04). The authors proposed that the mechanism of improved survival was related to early interventions including active remote symptom management, supportive medications, chemotherapy dose modifications, and referrals for specialty consultation that prevented downstream consequences. Additionally, the PRO group was able to tolerate continuation of chemotherapy for a longer duration (8.2 months vs 6.3 months, p = 0.002).   [12].

Patients
One of the earliest uses of patient-driven technology to monitor outcomes was the Cancer Care Monitor used by the West Clinic, Memphis, TN, in 2003 [27]. Prior to each scheduled oncology visit, patients would report their symptoms on computers and later digital tablets (Patient Care Monitoring) [28]. This validated survey was used to support a web and mobile-based application prospectively applied to postmenopausal women starting new anti-hormonal therapy to assess whether this tool would improve symptom burden and medication adherence [29]. Of significance, this application was able to increase medication adherence at eight weeks, when complimented with weekly reminders (100% vs 75%, p < 0.05).
Additional ongoing prospective randomized trials evaluating remote patient monitoring and patient reported outcomes in oncology are summarized in Table 2. In the United States, the Symptom Management Implementation of Patient Reported Outcomes in Oncology (SIMPRO) study aims to enroll 18,000 patients with thoracic, gastrointestinal, and gynecologic malignancies. Patients randomized to the intervention arm will enter symptom data into the mobile electronic symptom management system (eSyM). The primary outcome is 30-day emergency department treat and release rate as measured through medical record abstraction [30]. An additional ongoing US study, the THRIVE study, is a prospective, randomized trial evaluating a web-enabled application to improve adherence to hormonal therapy in women with breast cancer. The study will randomize 300 patients into three separate arms: 1. a group that receives weekly reminders through the application, 2. a group that receives weekly reminders and tailored feedback, and a usual care group. The primary endpoint is medication adherence as determined using an electronic pillbox [31].   [32]. The primary outcome is measurement of symptom burden, with secondary outcomes including quality of life, supportive care needs, anxiety, self-care self-efficacy, work limitations, cost effectiveness, and changes in clinical practice in response to PRO data. A similar study is being undertaken in Australia for hematologic malignancies utilizing the Patient Remote Intervention and Symptom Management System (PRISMS), a computer tablet-based software system that prompts patients to enter twice daily data regarding physical and emotional symptoms [33]. The PRISMS study aims to enroll 222 patients undergoing chemotherapy for chronic lymphocytic leukemia (CLL), Hodgkin's lymphoma, and non-Hodgkin's lymphoma (NHL) in two Australian hospitals. The primary outcome focuses on symptom burden due to nausea, mucositis, constipation, and fatigue.

Physiologic monitoring in oncology
Ubiquitous smartphones and miniaturization of sensor and communication technology has led to enormous data gathering enterprises in consumer and healthcare markets. Smartphone accelerometers, GPS tracking, and high-resolution cameras are critical to fitness and wellness applications but also as diagnostic tools such as remote skin cancer detection. These same devices when combined with subjective PRO feedback can be used to guide patient care or provide prognostic data.
Performance status (PS), commonly measured using the Karnofsky or Eastern Cooperative Oncology Group (ECOG) PS scale, is one of the strongest predictors of cancer survival outcomes and risk of treatment toxicity. Patients with a compromised PS have higher risks of morbidity and mortality, but current clinical PS assessments, despite their critical role in clinical trials, are limited by subjectivity and high rates of interobserver variability [39][40][41]. Many consumer and medical wearables contain three-axis accelerometers and three-axis gyroscopes, providing continuous readings of individual movement which holds significant promise as a means of overcoming the current limitations of physician identified performance status. Gresham et al. evaluated the feasibility of using activity data from FitBit Charge HR® wearable devices as a surrogate for PS in 37 patients with cancer, 92% of whom had stage IV disease [42]. They reported high correlations between average daily step counts and ECOG-PS (r = 0.63); each increase in daily step count by 1000 was associated with a significant decrease in adverse events, hospitalization, and hazard for death. Strong correlation was also reported between these activity metrics and PRO data. Subsequent systematic review of published studies evaluating the use of activity tracking in cancer care identified 41 trials including active cancer patients and those undergoing follow-up and survivorship care [43]. Most trials included breast cancer patients (65%) and focused on exercise (54%) or behavioral (29%) interventions. Twelve trials evaluated daily step counts, and the reported steps per day were slightly lower in patients on active cancer therapy (2885 to 8300 steps/day) compared to survivors (4660 to 11,000 steps per day). No studies included implantable monitoring devices.

Current technology limitations
While PRO platforms and wearable technology have potential to improve care, implementation in the clinic faces many challenges. Web-based PRO platforms rely on patient engagement, which can be increasingly problematic in an aging cancer population with less familiarity with technology. Furthermore, waiting for patients to report symptoms lacks a proactive preventative solution.
With regards to wearable technology, most studies demonstrate feasibility, however patient adherence remains a major limitation. Seventeen of the studies included in the Gresham systematic review reported adherence data, but adherence was defined differently in most studies [43]. The most commonly reported adherence metric was three consecutive days of activity tracking (range: 3-7 consecutive days), with a valid wear time as 5 to 10 hours per day. These definitions of adherence leave large gaps in time where no data is gathered or reported, limiting the ability to produce high fidelity analytics and accurate diagnostic tools. Additionally, for cancer patients facing several months of therapy, three to seven consecutive days of activity tracking does not represent reliable monitoring through the duration of treatment and follow-up.
Dreher et al. demonstrated similar adherence limitations in a study of FitBit® use in breast cancer patients, stating, "Adherence to wearing the Fitbit was low, with 16.9% of patients never syncing their device." For patients who did sync their devices, the median number of valid activity tracking days (defined as > 10 hours of use) during the 9-month duration of the study was only 44.5% (median = 39.6%, range 0% -100%) [44]. Even the much-publicized Apple Watch® atrial fibrillation study, a population consisting of Apple Watch® owners, showed poor adherence with only 21% of those who received an irregular pulse notification initiating the first indicated visit [45]. The studies utilizing wearable technology in oncology care demonstrate the importance of objective data gathering, but also highlight the limitations of systems that rely on patient adherence. Alternatively, the studies involving cardiac referenced in the introduction show significant benefits associated with the reliable and continuous monitoring from implantable sensor technology, including disease control, survival, and economic endpoints.

Future directions
Representative ongoing early phase or pilot studies evaluating emerging digital technology in cancer care are summarized in Table 3. Emerging research shows benefits in outcomes and costs of cancer care through use of remote monitoring technology especially electronic patient reported outcomes (ePRO). However, broad clinical adoption has been limited by a lack of commercially available oncology specific solutions, concerns about reimbursement, and limitations associated with low patient adherence.   A preferred solution would securely and reliably gather digital physiologic data without requiring patient activation, much like pulmonary artery pressure monitors and implantable loop recorders in cardiology. Additionally, such a system would contain mobile patient engagement and ePRO tools that could be tied to existing reimbursement framework facilitating rapid adoption. Powerful analytic tools including machine learning and artificial intelligence could then be applied to identify early signs of common complications and provide individual patient health and risk profiles, allowing oncologists to make more informed and personalized treatment recommendations. The use of an implantable device in oncology has already been explored, as oncologic cardiologists at MD Anderson Cancer Center have utilized implantable loop recorders and CardioMEMS® to monitor patients at high risk of complications from cardiotoxic systemic therapy, but data regarding safety and efficacy is lacking [46]. To our knowledge, Oncodisc, a San Francisco-based medical technology start-up, is the only company developing implantable monitoring technology and mobile ePRO solutions for oncology. The Oncodisc device, an intelligent implantable vascular access port, takes advantage of existing oncology workflow and a common minimally invasive procedure with established reimbursement.

Conclusions
The current Covid-19 pandemic has highlighted the need for reliable connected systems to facilitate home care while reducing hospitalizations and clinic visits, especially in immunocompromised cancer patients. Early studies show significant benefits to PRO-based systems, including lower costs and prolongation of survival. Further advances in sensor technology and mobile communications hold great promise for improving cancer outcomes while at the same time reducing costs. However, widespread adoption has been hampered by a lack of commercially available solutions. To that end, implantable physiologic sensor systems and associated data analytic tools, akin to those used in cardiovascular care, should be researched for oncology.