Prevalence of palliative radiotherapy abstracts presented at the annual scientific meetings of the Canadian Association of Radiation Oncology: 2003–2021

Approximately half of all radiotherapy (RT) is delivered with palliative intent. Clinical research in palliative RT aims to manage symptoms, improve quality of life (QoL), evaluate supportive care, and determine optimal dose-fractionation schedules. Our aim was to describe the prevalence of palliative research at the Canadian Association of Radiation Oncology (CARO) Annual Scientific Meeting (ASM) over time and compare this analysis to previously published work which evaluated the years 1992–2002. Published abstracts (2003–2021) were independently reviewed by two authors who categorized each as curative-intent; palliative-intent; pertaining to both populations; or neither. Abstracts were considered palliative if they described incurable malignancy and interventions primarily for symptom control or QoL. Type of study, primary, site treated, and symptoms palliated were recorded. Descriptive and summary statistics were calculated including one-way ANOVA test for trend. Three hundred thirty-nine out of 4566 abstracts (7.4%, range 2.4–13.9% per year) were classified as palliative. 7.7% (26/339) described phase I–III trials. The main primary site was the lung (39/339) and the most common metastatic site was the bone (34.2%). QoL, symptom and toxicity outcomes were reported in 31.6% (107/339), 37.8% (128/339) and 17.7% (60/339), respectively. The most common symptom investigated was pain (38/339). The proportion of abstracts classified as curative, palliative or reporting toxicity endpoints demonstrated significant change over time (all p<0.0001). While proportion of palliative themed abstracts has increased with time, there remains a significant gap before equivalence with the prevalence of palliative RT in clinical practice is achieved.


Introduction
At least one-half of patients who die of cancer develop physical and psychosocial symptoms before death, many of which are not adequately controlled [1,2]. Cancer patients' symptom burden negatively affects both patients' and caregivers' quality of life (QoL) [3]. Palliative radiotherapy (PRT) is an effective modality to reduce multiple symptoms of advanced cancer, including pain secondary to bone metastases, neurologic symptoms related to brain metastases, bleeding caused by friable or fungating tumours, and airway or gastrointestinal tract obstruction [4]. PRT is used in emergency situations to preserve organ function and minimize the negative impact of cancer progression [5], and patients may receive multiple courses of PRT throughout their cancer journey [6].
Given that 30-50% of the total RT workload of a department is delivered with palliative intent [4,7,8], palliative care-related issues constitute a substantial portion of radiation oncology practice [9]. In one study of patients reviewed for PRT, one-fifth were seen at initial metastatic diagnosis, 17% were seen at the point of discontinuation of all further systemic therapy, and the remainder were somewhere in between [9].
Twenty years ago, ASTRO identified a need for Radiation Oncologists (ROs) to assume a greater role in supportive and palliative care, psychosocial support, coordination of multidisciplinary interventions, and end-of-life needs [10]. Similarly, the ESTRO core curriculum states that " [ROs] should be able to discuss the role of palliative care in… management…and also implement treatments to control symptoms or refer appropriately to other specialties" [11]. This has been echoed by the CARO statement on Radiation Oncology Scope of Practice in Canada, which outlines that 'Radiation Oncologists are required to be competent in…palliative and supportive care [and] symptom management' [personal communication, CARO]. All ROs should be comfortable providing supportive care, with referral to allied health providers or specialist palliative care if needs are complex or intractable [9,12]. Decisions regarding palliative treatment of symptoms can be complex, requiring cost-benefit decision-making in the context of a limited life expectancy [5]. The choice of an optimal PRT regimen and management of disease-and treatment-related symptoms should be evidence-based whenever possible [8].
Research is therefore essential to develop, refine and disseminate the roles of RT (and ROs) in supportive cancer care [8,13]. Oncology conferences provide an opportunity to improve the collective body of knowledge, discuss research developments, as well as provide continuing education [1,13,14]. The breadth of data presented provides a window into a field's research focus and current trajectory of inquiry [5]. Research is also important to increase consistency in working practices, and provide opportunities to apply positive outcomes from one setting to work in other settings [13]. However, despite the prevalence and importance of PRT delivery and RO involvement in routine supportive care, related research is infrequently reported at major conferences [1,8,[14][15][16] or in high-profile journals [5].
The Canadian Association of Radiation Oncology (CARO) Annual Scientific Meeting (ASM) provides opportunities to discuss new developments and disseminate knowledge to ROs and attendees from related disciplines. Our objective was to evaluate time-based trends in the number of palliative abstracts presented at the annual CARO ASM from 2003 to 2021, and compare this analysis to previously published work which evaluated the years 1992-2002.

Methods
Abstracts accepted for oral and poster presentations at the annual CARO ASM have been published in Clinical and Investigative Medicine (1992)(1993)(1994)(1995)(1996)(1997)(1998)(1999)(2000), International Journal of Radiation Oncology, Biology, Physics (2001), and Radiotherapy and Oncology (2002 to present). Published abstracts (2009-2021) were independently reviewed by two authors who categorized each as: curative-intent; palliative-intent; pertaining to both populations; or pertaining to neither population. Abstracts from the ASM between 2003 and 2008 which had been the subject of a previous study [16] were reclassified per current methodology for consistency. Additionally, we compared our results to previously published work which evaluated the years 1992-2002 with virtually identical methodology. Abstracts were considered palliative if they described incurable (locally advanced or metastatic) malignancy and interventions primarily for symptom control or QoL. Similar to Shi et al., outcomes of metastatic patients treated with curative intent, and abstracts describing definitive treatment of oligometastatic disease, were not considered palliative in nature [5].
Areas of patient care to which these abstracts related were also identified, including treatment toxicity. Type of study, primary site, type of intervention, site treated, symptoms palliated and regions of authorship were recorded. It was also noted whether research was performed through a dedicated PRT clinic [17]. Population distribution by province, and incidence and mortality rates by primary site and region were abstracted from Canadian Cancer Statistics 2021, the only source of national cancer information [18]. Canadian Cancer Statistics is a freely available publication produced by the Public Health Agency of Canada based on data provided by provincial and territorial cancer registries. Within this publication, actual cancer incidence data to 2017 and mortality data to 2019 were used (except for Quebec, where incidence data were available to 2010) to project 2021 rates, but do not take into account impact from COVID 19 [18]. For further information on methodology, the interested reader is directed to this report [18]. To determine number of practicing Radiation Oncologists by province, a search of the Royal College of Physicians of Canada directory was performed (1 August 2022) for fellows by address, minus retired.
Data were reported using descriptive statistics (mean, range and percentages). One-way ANOVA was performed for trend analysis and compared the mean proportion of palliative abstracts presented over time. All statistical analysis was performed using SPSS version 25 software (IBM SPSS Statistics for Windows, Version 25.0. Armonk, NY: IBM Corp, 2017).

Results
Over 19 years of the CARO ASM, a total of 4566 abstracts have been presented, ranging from 159 in 2003 to 384 in 2018 (average 240/year). 3022/4566 abstracts (66.6%) were classified as curative-intent. The proportion classified as curative ranged from 56.6% to 74.3%, demonstrating a significant trend over time (p<0.0001; Fig. 1). Of these 3022, 157 (5.2%) were aggregated multiple primary sites while in 184, the primary site was not stated, leaving 2681 which focused on a specific primary site. Figure 2 compares the projected national incidence of new cancer diagnoses versus proportion of curative abstracts by primary site. Of all subsites which could be compared, lung cancer was the sole primary site with parity, indicated by a non-significant p value (p=0.58).
Overall, 339/4566 (7.4%) were classified as palliative between the years 2003 and 2021 (Fig. 1). The proportion classified as palliative ranged from 2.4 to 13.9%. When including the additional 892 abstracts from 1992 to 2002 from Barnes et al. [14], of which 6.7% were classified as palliative, the trend over all years was again statistically significant (p<0.0001; Fig. 1).
Of the palliative abstracts, 7.7% (26/339) described phase I-III trials while 30.7% (104/339) were retrospective. Other common research types were prospective (18.6%), health services (18.0%) and survey (11.5%) studies. The most common metastatic sites described were bone (34.2%) and brain (13.9%), while multiple metastatic sites were grouped together in 16.5%. 139/339 (41.0%) aggregated multiple primary sites while in 86 the primary was not stated. Of the remaining, 114 abstracts focused on a specific primary site, including 39 investigating lung cancer and 22 with a genitourinary (GU) primary. Figure 3 compares projected national cancer mortality versus proportion of palliative abstracts by primary site. Of all subsites, lung, GU, breast, melanoma and H&N demonstrated significantly higher proportion of palliative abstracts versus expected proportion of cancer mortality, while gastrointestinal showed the reverse trend.
Overall survival was reported in almost one-third (30.7%) of palliative abstracts, while other common disease control outcomes, such as disease-free survival or local control, were described in 10% or fewer. QoL, symptom and toxicity outcomes were reported in 31.6% (107/339), 37.8% (128/339) and 17.7% (60/339) abstracts, respectively. While QoL and symptom outcomes did not demonstrate significant differences with time, toxicity endpoint reporting did (Fig. 4). The most common specific symptom investigated was pain. In terms of region of authorship, 37 abstracts were produced by authors from multiple provinces within Canada, 31 by authors collaborating from within Canada and outside Canada, and 11 solely by authors outside of Canada. Of the remaining abstracts authored by investigators from a single province, 57.7% (150/260) were reported by investigators from Ontario. The proportion of palliative abstracts by specific province of authorship, relative to population distribution, projected cancer mortality and proportion of boardcertified Radiation Oncologists is presented (Fig. 5). 15.9% (54/339) were co-authored by members of one or more rapid access PRT clinics.

Discussion
Despite the prevalence of patients with incurable disease seen in routine practice and statistical evidence of some improvement over time, palliative-themed research remains strikingly poorly represented at the CARO ASM. On average, 6.7% of all abstracts presented between 1992 and 2002 pertained to symptom control or palliative topics [14], compared to an average of 7.4% between 2003 and 2021. It is discouraging that the proportion of palliative abstracts presented at the national meeting has not closed the gap to proportion of real-world workload represented by PRT [14,16]. Potential explanations are listed in Table 1.
This finding is in fact quite similar to the first review performed of symptom control and palliative care-themed papers, by Sweeney et al., over twenty years ago [1]. They evaluated alternate years' abstracts presented at the American Society of Clinical Oncology annual congress between  1983 and 1999. Palliative abstracts accounted for 6.3%, with a significant decrease in the proportion accepted over the 9 years of 0.54% per 2-year interval. Interestingly, the authors were also able to discern acceptance rate, which was 37.8% (389/1030) for palliative abstracts versus 45.7% (6165/13,675) for all submissions (p<0.0001) [1]. Abstracts presented at the American Society for Radiation Oncology (ASTRO) Annual Meeting have also been reviewed [8,15]. Barnes et al. evaluated 3511 abstracts from the ASTRO ASM between 1993 and 2000. An average of 47 (1.3%, range 0.9-2.2% per year) pertained to symptom control or palliative care, with no significant change over the 8-year period [8]. A subsequent study assessed >11,000 abstracts accepted between 2001 and 2010 for palliativerelated endpoints, but has been reported itself only in abstract form [15]. The absolute number of abstracts categorized as "palliative care" increased from 6 in 2004 to 22 in 2010; as a proportion of total meeting abstracts, this ranged from a low of 0.5% in 2005 to a high of 1.3% in 2008. Up to 4.3% of all abstracts focused on symptom control or palliative care. Interestingly, 40% of palliative abstracts were submitted by Canadian authors [15].
Given these collective results, it is critical to increase efforts to engage in methodologically sound research activities, ideally facilitated by experienced researchers mentoring young investigators, to continue to expand the evidence base for supportive care and PRT [5,9,19]. This will help ensure equity of cost-effective treatments for cancer patients regardless of location or care setting [19]. Research supports innovation in practice including identification of new technologies and evaluation of the feasibility of their implementation [22].
To move the field forward most efficiently, however, research findings should be submitted for conference presentation [1]. Dissemination of findings externally at a high-visibility meeting is the most agile way for clinical practice to evolve. Presentation at congresses also encourages networking, collaboration, and mentoring, and provides opportunities for trainees to be exposed to a broad range of academic avenues [22]. Lo Presti et al. performed a systematic literature search of two databases for English papers which discussed the 'recognition of the role of palliative care in radiotherapy' plus at least one of ten additional issues [20]. 19 articles (2002-2019) were qualitatively summarized, the majority cross-sectional or observational. 7/19 identified a lack of 'research [and] palliative care topics in oral presentations or posters at radiotherapy meetings' as a barrier to comprehensive education of radiation oncology trainees [20].
Unfortunately, this trend of underrepresentation is also evident in fully published studies. Shi et al. reviewed 6386 full length original research articles published by the International Journal of Radiation Oncology, Biology, Physics (the "Red Journal") and 2406 manuscripts published by Radiotherapy and Oncology (the "Green Journal") between 2005 and 2014 [5]. The Red Journal publishes relatively more research from the USA, while the Green Journal's authors are more commonly from Europe and Canada. Articles were categorized as 'palliative care' (PC) if they pertained to any aspect of treatment of  [1,5,8,14,20,21] • Limited interest by ROs in this technology-driven specialty compared with curative treatments, with increased identification of oligometastases for which use of SBRT is often considered [5,8] • Lack of local departmental support [14] • Perception that ROs are not included as full members of the multidisciplinary care team in managing palliative patients [20] Patient population-related • Lower methodologic quality of this type of research, e.g. subjectivity of patientreported outcomes, descriptive studies [1,19] • Challenges in accrual and completion of study participation due to clinical decline/ death [5,14] • Traditionally, ROs did not participate in surveillance of palliative patients, handing over their care to the Medical Oncologist or Primary Care Practitioner [8] Conference organization-related • Perception that the CARO ASM is not the correct forum for this type of research [1] • Lack of reviewer expertise [1] • Lower acceptance rate [21] metastatic cancer, and as 'symptom control and palliative care' (SCPC) if they described patients with metastatic cancer for whom symptom palliation was the stated goal. Similar to our findings, the most common symptom palliated was pain and the most common disease site was bone metastases. Interestingly, the authors report that the Green Journal rate of publication of SCPC research was lower than the proportion of SCPC abstracts presented at the CARO ASM between 1992 and 2003, while the Red Journal's frequency of palliative publications essentially mirrored the extremely low proportion of SCPC abstracts presented at ASTRO prior to 2010 [15]. Given the absence of significant change in PC or SCPC themed articles with time in either journal, the authors concluded that study productivity in this arena has stagnated [5].
In the first CARO ASM review, 45% of palliative abstracts came from a dedicated PRT clinic [14]. These dedicated programs, often called "rapid access clinics", facilitate same day assessment, simulation and treatment start for symptomatic patients, typically by a multidisciplinary team with dedicated infrastructure [5,6,17]. Other aims include holistic assessment of unmet physical, practical, social, and emotional needs; supportive care education of trainees; and as a centralized point of access for clinical research studies [17]. It was therefore anticipated in 2004 that research submission by teams within rapid access clinics would increase concomitantly with the upsurge in popularity of these programs [14]. That was not borne out in our study, with 15.9% of abstracts produced by dedicated programs.
In terms of examining trends by primary site, 20% of abstracts grouped patients with different primary sites in the previous publication (1992)(1993)(1994)(1995)(1996)(1997)(1998)(1999)(2000)(2001)(2002); that number was just over 40% in our study. This aggregation restricts readers' ability to interpret findings and apply them to local patients or settings. Barnes et al. also previously identified that research is lacking in some disease sites where PRT is used commonly and is effective, such as primary brain tumours, head & neck and gynecological malignancies, a pattern which we also demonstrated.
It is not possible to discern whether the low rate of palliative-themed abstracts we report is related to a paucity of submissions for consideration, a high rejection rate, both, or other factors. It is possible that abstracts describing primarily symptom control or quality of life outcomes are being presented at different conferences, such as those with a particular Palliative or Supportive Care focus.
Efforts have been made by the CARO ASM organizers to increase the opportunity for presentation of palliative research with specific submission streams and an annual call for expert peer-reviewers. An award for the best supportive care-themed abstract has been presented since 2006. During and between meetings, the CARO Supportive Care Committee is active with educational sessions available to members nationally. This networking assists interested trainees in finding experienced mentors and promotes visibility of ongoing research.
Our results must be interpreted within the context of some limitations. As per Sweeney et al., we also found the decision taken about which category to apply to any particular abstract was not straightforward at times. To reduce bias, a single investigator initially reviewed and classified each, followed by independent review by a second author. Discordance was resolved by consensus. While we cannot exclude any remaining bias, justifications for the classification and the study investigators were consistent throughout, similar to Shi et al. [5]. In order to maximize consistency, we also re-assessed CARO ASM abstracts from years 2003 to 2008, which our group had previously evaluated [16]. We excluded patients with metastatic disease treated with radical intent, where alleviation of diseaserelated symptoms would have been a likely secondary goal [14]. We also excluded ablative-intent stereotactic radiosurgery or radiotherapy (SABR/SBRT) delivered for oligometastases, given the focus in that setting often pertains to local control, with correspondingly aggressive RT schedules. However, even if these were to have been included, resources devoted to investigating palliative topics would remain disproportionately lower than its representation in clinical practice.

Conclusions
While proportion of palliative themed abstracts has increased with time, there remains a significant gap before equivalence with the prevalence of palliative RT in clinical practice is achieved. Acknowledging the challenges of conducting studies in a population with limited life expectancy, research represents an essential way to advance comprehensive personalized patient care.
Author contributions All authors contributed to the study conception and design. Material preparation, data collection and analysis were performed by SS and AF. Statistical analysis was performed by SG. The first draft of the manuscript was written by AF and all authors commented on previous versions of the manuscript. All authors read and approved the final manuscript.

Declarations
Ethics approval Given the nature of this work, ethics approval was not required.